IPCS INCHEM Home


    INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY



    ENVIRONMENTAL HEALTH CRITERIA 160





    ULTRAVIOLET RADIATION



    This report contains the collective views of an international group of
    experts and does not necessarily represent the decisions or the stated
    policy of the United Nations Environment Programme, the International
    Labour Organisation, or the World Health Organization.

    An Authoritative Scientific Review of Environmental and Health
    Effects of UV, with Reference to Global Ozone Layer Depletion


    Published under the joint sponsorship of
    the United Nations Environment Programme,
    the International Labour Organisation,
    and the World Health Organization

    World Health Orgnization
    Geneva, 1994

          The International Programme on Chemical Safety (IPCS) is a joint
    venture of the United Nations Environment Programme, the International
    Labour Organisation, and the World Health Organization.  The main
    objective of the IPCS is to carry out and disseminate evaluations of
    the effects of chemicals on human health and the quality of the
    environment.  Supporting activities include the development of
    epidemiological, experimental laboratory, and risk-assessment methods
    that could produce internationally comparable results, and the
    development of manpower in the field of toxicology.  Other activities
    carried out by the IPCS include the development of know-how for coping
    with chemical accidents, coordination of laboratory testing and
    epidemiological studies, and promotion of research on the mechanisms
    of the biological action of chemicals.

    WHO Library Cataloguing in Publication Data

    Ultraviolet radiation

    (Environmental health criteria: 160)

    1. Ultraviolet rays           2. Radiation effects
    3. Environmental exposure     I.Series

          ISBN 92 4 157160 8         (NLM Classification: WD 605)
          ISSN 0250-863X

          The World Health Organization welcomes requests for permission
    to reproduce or translate its publications, in part or in full.
    Applications and enquiries should be addressed to the Office of
    Publications, World Health Organization, Geneva, Switzerland,
    Which will be glad to provide the latest information on any changes
    made to the text, plans for new editions, and reprints and
    translations already available.

    (c) World Health Organization 1994

          Publications of the World Health Organization enjoy copyright
    protection in accordance with the provisions of Protocol 2 of the
    Universal Copyright Convention.  All rights reserved.

          The designations employed and the presentation of the material in
    this publication do not imply the impression of any opinion whatsoever
    on the part of the Secretariat of the World Health Organization
    concerning the legal status of every country, territory, city, or area
    or of its authorities, or concerning the delimitation of its frontiers
    or boundaries.

          The mention of specific companies or of certain manufacturers'
    products does not imply that they are endorsed or recommended by the
    World Health Organization in preference to others of a similar nature
    that are not mentioned.  Errors and omissions excepted, the names of
    proprietary products are distinguished by initial capital letters.

    CONTENTS

         PREFACE

    1. SUMMARY AND CONCLUSIONS

         1.1. Physical characteristics
         1.2. Action spectrum and minimum erythemal dose
         1.3. Cellular and molecular studies
         1.4. Animal studies
         1.5. Health effects on humans
                1.5.1. Skin
                1.5.2. Immune system
                1.5.3. Eye
         1.6. Environment
         1.7. Guidelines on exposure limits and protective measures

    2. PHYSICAL CHARACTERISTICS

         2.1. Electromagnetic spectrum
         2.2. Radiometric quantities and units
         2.3. UV production
                2.3.1. Thermal emitters
                2.3.2. Electrical gaseous discharges
                2.3.3. Stimulated emission
         2.4. Summary

    3. UV SOURCES

         3.1. The sun
                3.1.1. Factors affecting solar UV levels
                3.1.2. Ozone depletion effects
                3.1.3. Trends in UV 19
                3.1.4. Theoretical models
                3.1.5. UV monitoring
                3.1.6. Analysis of UV data
                3.1.7. Conclusions
         3.2. Artificial sources
                3.2.1. Incandescent sources
                3.2.2. Gaseous discharge sources
                3.2.3. Gas welding
                3.2.4. Arc welding
                3.2.5. Lasers
                3.2.6. Sunbeds

    4. HUMAN EXPOSURE

         4.1. Sunlight
         4.2. Skin exposure geometry
         4.3. Ocular exposure geometry
         4.4. Workplace

                4.4.1. Outdoor work
                4.4.2. Indoor work
                4.4.3. Research
                4.4.4. Commerce
                4.4.5. Medicine and dentistry
         4.5. Elective exposure

    5. DOSIMETRIC CONCEPTS

         5.1. Interaction of UV with matter
         5.2. Biological weighting factors and spectrally
                weighted quantities
         5.3. Measurement techniques
                5.3.1. Detectors
                5.3.2. Radiometers
                5.3.3. Spectroradiometers
                5.3.4. Personal dosimetry
         5.4. Calibration

    6. CELLULAR AND MOLECULAR STUDIES

         6.1. Introduction
         6.2. Interactions with biomolecules
                6.2.1. Cellular chromophores
                6.2.2. Cellular targets
         6.3. Action spectra
         6.4. Biomolecular damage
                6.4.1. Nucleic acids
                6.4.2. Membranes
                6.4.3. Proteins
         6.5. Cellular defenses
                6.5.1. DNA
                6.5.2. Human excision repair disorders
                6.5.3. Antioxidant pathways
                6.5.4. Summary
         6.6. Cellular consequences of damage
                6.6.1. Membrane disruption
                6.6.2. Activation of genes
                6.6.3. Cell death
                6.6.4. Mutation, chromosomal damage and transformation
         6.7. Conclusions

    7. ANIMAL STUDIES

         7.1. Skin carcinogenesis
                7.1.1. Domestic animals
                7.1.2. Experimental animals
                7.1.3. Interactions between radiations
                        of different wavelengths
                7.1.4. Dose response
                7.1.5. Effect of pattern of exposure

                7.1.6. Action spectrum
                7.1.7. Interaction between UV and chemicals
                7.1.8. Mechanisms of UV carcinogenesis
                7.1.9. Conclusions
         7.2. Immune responses
                7.2.1. Immune function assays
                7.2.2. Susceptibility to rumours
                7.2.3. Susceptibility to infectious disease
                7.2.4. Susceptibility to immunologically-mediated
                        disease
                7.2.5. Conclusions
         7.3. Ocular studies
                7.3.1. Introduction
                7.3.2. General effects
                7.3.3. Caractogenesis
                7.3.4. Retinal effects

    8. HUMAN STUDIES: THE SKIN

         8.1. Characteristics
                8.1.1. Structure and optical properties
                8.1.2. Skin types
         8.2. Beneficial effects
                8.2.1. Vitamin D3
                8.2.2. Skin adaptation
                8.2.3. Other benefits
         8.3. Acute effects
                8.3.1. Erythema and sunburn
                8.3.2. Skin pigmentation and tanning
                8.3.3. Photosensitization
         8.4. Chronic effects on the skin other than cancer
         8.5. Cancer
                8.5.1. Nonmelanocytic skin cancer
                8.5.2. Cutaneous melanoma
                8.5.3. Cancer of the lip
                8.5.4. Ocular cancers
                8.5.5. Other cancers
                8.5.6. Action spectrum
                8.5.7. Dose response
                8.5.8. Effects of pattern of exposure
                8.5.9. Interactions between UV and other agents
                8.5.10. Mechanisms of UV carcinogenesis
         8.6. Conclusions

    9. HUMAN STUDIES: IMMUNE FUNCTION

         9.1.   Immune function assays
         9.2. Susceptibility to tumours, infectious and autoimmune
                diseases
         9.3. Conclusions

    10. HUMAN STUDIES: THE EYE

         10.1. Introduction
         10.2. The eye
         10.3. Study design
         10.4. Diseases of the external eye
                10.4.1. Photokeratitis and photoconjunctivitis
                10.4.2. Climatic droplet keratopathy
                10.4.3. Pinguecula
                10.4.4. Pterygium
                10.4.5. Hyperkeratosis, carcinoma-in-situ, and squamous
                        cell carcinoma of the conjunctiva
         10.5. Diseases of the lens
                10.5.1. Cataract
                10.5.2. Exfoliation syndrome
                10.5.3. Anterior lens capsule
         10.6. Diseases of the choroid and retina
                10.6.1. Uveal melanoma
                10.6.2. Age-related macular degeneration
         10.7. Conclusions

    11. EFFECTS ON PLANT AND AQUATIC ECOSYSTEMS

         11.1. Introduction
         11.2. Effects on terrestrial plants
                11.2.1. UV penetration into the leaf
                11.2.2. Changes in growth
                11.2.3. Effects on plant function
                11.2.4. Species competition
                11.2.5. Plant diseases
                11.2.6. UV-protection systems
         11.3. Effects on aquatic ecosystems
                11.3.1. Effects on phytoplankton
                11.3.2. UV increase and primary biomass production
         11.4. Conclusions

    12. HEALTH HAZARD ASSESSMENT

         12.1. Introduction
         12.2. Elective exposures
                12.2.1. Medical exposures
                12.2.2. Phototherapy of seasonal effective disorder
                        (SED)
                12.2.3. Sunbeds
                12.2.4. Sunbathing
         12.3. Adventitious exposures
                12.3.1. Outdoor exposures
                12.3.2. Artificial sources

    13. INTERNATIONAL GUIDELINES ON EXPOSURE STANDARDS

    14. PROTECTIVE MEASURES

         14.1. Introduction
         14.2. Education
         14.3. Protection factors
         14.4. Clothing
         14.5. Sunscreens
         14.6. Tanning devices
         14.7. Occupational protection
         14.8. Protection in medicine and dentistry
         14.9. Nutrition
         14.10. Additional protective agents
         14.11. Eye protection

    15. FUTURE RESEARCH

         15.1. Introduction
         15.2. INTERSUN
         15.3. Solar and personal UV monitoring
                15.3.1. Solar monitoring
                15.3.2. Personal monitoring
         15.4. Terrestrial plants
         15.5. Aquatic ecosystems
         15.6. Human health
                15.6.1. Skin
                15.6.2. Immune system
                15.6.3. Eye
         15.7. Laboratory studies
         15.8. Education
         15.9. Administration

    16. PREVIOUS EVALUATIONS BY INTERNATIONAL BODIES

         16.1. United Nations Environment Programme
                16.1.1. Ozone
                16.1.2. Human health
         16.2. International Agency for Research on Cancer
         16.3. World Health Organization
         16.4. International Commission on Non-Ionising
                Radiation Protection

    REFERENCES
    

    PREFACE

         The World Health Organization (WHO), in collaboration with the
    United Nations Environment Programme (UNEP) and the International
    Non-Ionizing Radiation Committee (INIRC) of the International
    Radiation Protection Association (IRPA), published the first
    Environmental Health Criteria (EHC) monograph on Ultraviolet
    Radiation in 1979. At the United Nations Conference on the
    Environment and Development (UNCED) in 1992 it was declared under
    Agenda 21 that there should be activities on the effects of
    ultraviolet radiation. Specifically:

    "(i) Undertake, as a matter of urgency, research on the effects on
    human health of the increasing ultraviolet radiation reaching the
    earth's surface as a consequence of depletion of the stratospheric
    ozone layer;

    (ii) On the basis of the outcome of this research, consider taking
    appropriate remedial measures to mitigate the above-mentioned
    effects on human beings".

         Within the United Nations mandate, and that of the 1993 WHO
    Global Strategy for Health and Environment, this monograph has been
    drafted to provide the essential authoritative review on which
    future research programmes in UV can progress.

         IRPA initiated activities in NIR by forming a Working Group on
    Non-Ionizing Radiation in 1974. This Working Group later became the
    INIRC at the IRPA meeting held in Paris in 1977. In May 1992 the
    INIRC was chartered as an independent scientific commission called
    the International Commission on Non-Ionizing Radiation Protection
    (ICNIRP). ICNIRP continues the work of the IRPA/INIRC by reviewing
    the scientific literature on NIR and making assessments of health
    risks of human exposure to such radiation. Using the Environmental
    Health Criteria monographs, developed in conjunction with WHO,
    ICNIRP recommends guidelines on exposure limits, drafts codes of
    practice, and works in conjunction with other international
    organizations to promote safety and standardization in the
    non-ionizing radiation fields.

         A UNEP/WHO/ICNIRP Task Group to review the final draft of the
    updated Environmental Health Criteria on Ultraviolet (UV) Radiation
    met in Geneva from 8-11 December, 1993. Dr W. Kreisel, Executive
    Director, World Health Organization opened the meeting on behalf of
    WHO. Dr H. Gopalan and Mr R. Matthes welcomed the participants on
    behalf of UNEP and ICNIRP respectively.

         The first revision of this publication was compiled by members
    of the IRPA/INIRC and more recently by members of ICNIRP. Chapters
    were prepared by Drs B. Armstrong, J-P. Cesarini, L. Court, P.
    Dolan, G. Johnson, A. Kricker, A. McKinlay, M. Repacholi, M.

    Selgrade, D. Sliney and R. Tyrrell. These chapters together with the
    report of an informal consultation held in Geneva, in August 1993 on
    "The Effects of Solar UV Radiation on the Eye", the International
    Agency for Research on Cancer (IARC) report of the meeting on
    "Health, Solar UV Radiation and Environmental Change " held in Lyon
    in October 1992, and the meeting on Immunotoxicology held in
    Bilthoven in July 1993, were used by Drs M. Repacholi and D. Sliney
    to compile the review draft in September 1993. Comments were
    received from a broad cross-section of specialists in the UV fields
    and their reviews were gratefully received. These comments were then
    incorporated for Task Group review. Final scientific editing of the
    text was completed by Drs J-P. Cesarini, A. McKinlay, M. Repacholi
    and D. Sliney. Sincere thanks to Christine Cornish and Nancy Smith
    for their assistance in the preparation of this text. An editorial
    group consisting of Drs M. Repacholi, H. Gopalan and T. Kjellström
    coordinated the preparation of this monograph.

         This monograph comprises a review of the data on the effects of
    exposure on biological systems pertinent to the evaluation of human
    health risks. Its purpose is to give an overview of the known
    biological effects of UV, identify gaps in knowledge and provide
    direction for further research. This monograph will assist health
    authorities, regulatory and similar agencies to provide guidance on
    health risks from exposure to UV and limits for occupational and
    general public exposure.

         Earlier reports were not necessarily included, as they were
    reviewed in the 1979 monograph. Every effort has been made to
    distinguish clearly between established biological effects and those
    that have been reported as preliminary or isolated results, or as
    hypotheses proposed to explain observed results. The conclusions are
    based on established knowledge of interactions of UV with biological
    systems.

         Subjects reviewed include: the physical characteristics of UV;
    measurement techniques; applications and sources of exposure;
    mechanisms of interaction; biological effects; guidance on
    protective measures; and recommendations on exposure limits.

         This monograph will also serve as a scientific database for the
    planned WHO/UNEP/IARC/ICNIRP International Research Programme on
    Health, Solar UV Radiation and Environmental Change (INTERSUN). The
    general objectives of INTERSUN are to:

    (i) evaluate the quantitative relationship between solar UV at the
    surface of the earth and human health effects, develop reliable
    predictions of the health consequences of changes in UV, provide
    baseline estimates of the incidence of health effects of UV in
    representative populations around the world, and develop practical
    ways of monitoring change in these effects over time in relation to
    environmental and behavioural change;

    (ii) provide essential input into the development of environmental
    and public health policies and actions in relation to depletion of
    stratospheric ozone; and

    (iii) provide a framework for monitoring the effects of solar UV and
    the impact of prevention programmes.

         Health agencies and regulatory authorities are encouraged to
    set up and develop programmes that ensure effective protection
    against the health effects of UV. It is hoped that this criteria
    monograph will provide useful information for such endeavours.

    TASK GROUP MEETING ON THE REVISION OF ENVIRONMENTAL HEALTH CRITERIA
    ON UV RADIATION

    Members

    Dr J.P. Cesarini         LRTPH, Fondation Rothschild, Paris, France

    Dr L. Court              Centre de Recherches du Service de Santé
                             des Armées, La Tronche Cédex, France

    Dr A. McKinlay           National Radiological Protection Board,
                             Didcot, United Kingdom

    Dr D. Sliney             US Army Environmental Hygiene Agency,
                             Aberdeen Proving Ground, MD, USA

    Dr M.J. Selgrade         Immunotoxicology Branch, USEPA, Research
                             Triangle Park, NC, USA

    Dr G. Johnson            Institute of Ophthalmology, University of
                             London, London, United Kingdom

    Dr B. Armstrong          International Agency for Research on
                             Cancer, Lyon, France

    Dr H. Van Loveren        National Institute of Public Health &
                             Environmental Protection, Bilthoven,
                             Holland

    Dr A. Koppikar           Office of Health & Environmental
                             Assessment, USEPA, Washington DC, USA

    Dr R.M. Tyrrell          Swiss Institute for Cancer Research,
                             Epalinges/Lausanne, Switzerland

    Dr J.C. van der Leun     Academisch Ziekenhuis Utrecht, CX Utrecht,
                             Holland

    Dr I.A. Badr             King Khaled Eye Specialist Hospital, King
                             Saud University, Riyadh, Saudi Arabia

    Dr G. Kulandaivelu       School of Biological Sciences, Madurai
                             Kamaraj University, Madurai, India

    Dr E. Turowski           Institute for Water, Soil and Air Hygiene,
                             Berlin, Germany

    Observers

    Dr B. Weatherhead        Surface Radiation Research Branch, Boulder,
                             Colorado, USA

    Dr H. Gopalan            United Nations Environment Programme
                             (UNEP), Nairobi, Kenya

    Dr A. Kulmula            World Meteorological Organization (WMO),
                             Geneva, Switzerland

    Dr A. Kricker            International Agency for Research on
                             Cancer, Lyon, France

    Dr U. Feister            Deutscher Wetterdienst, Potsdam, Germany

    Dr W.H. Weihe            President, International Society of
                             Biometeorology (ISB), Brannenburg, Germany

    Dr R. Matthes            Bundesamt fur Strahlenschutz, Ingolsädter,
                             Oberschleissheim, Germany

    Dr C. Remé               Professor of Ophthalmology,
                             Universitätsspit, Zurich, Switzerland

    WHO Secretariat

    Dr W. Kreisel            Executive Director, Health and Environment
                             Mr G. Ozolins Manager, Prevention of
                             Environmental Pollution
    Dr T. Kjellström         Prevention of Environmental Pollution
    Dr M. Repacholi          Prevention of Environmental Pollution
                             (Secretary)
    Dr B. Thylefors          Manager, Programme for the Prevention of
                             Blindness
    Dr P.-H. Lambert         Chief, Microbiology and Immunology Support
                             Services
    Dr A.-D. Négrel          Programme for the Prevention of Blindness
                             Dr V. Koroltchouk Cancer and Palliative
                             Care

    NOTE TO READERS OF THE CRITERIA MONOGRAPH

         Every effort has been made to present information in the
    criteria monographs as accurately as possible without unduly
    delaying their publication. In the interest of all users of the
    Environmental Health Criteria monographs, readers are kindly
    requested to communicate any errors that may have occurred to the
    Director of the International Programme on Chemical Safety, World
    Health Organization, Geneva, Switzerland, in order that they may be
    included in corrigenda, which will appear in subsequent volumes.

    1.  SUMMARY AND CONCLUSIONS

         This monograph is concerned with the effects of ultraviolet
    (UV) radiation exposure on human health and the environment. Such a
    review of the scientific literature is considered timely in view of
    the consequences of increased levels of UV at the surface of the
    earth resulting from depletion of stratospheric ozone.

    1.1  Physical Characteristics

         Exposure to UV occurs from both natural and artificial sources.
    The sun is the principal source of exposure for most people. Solar
    UV undergoes significant absorption by the atmosphere. With
    depletion of the stratospheric ozone people and the environment will
    be exposed to higher intensities of UV. The consequences of this
    added UV exposure are considered so serious that it was a major
    topic for discussion at the World Environment Conference, held in
    Rio de Janeiro in 1992. In Agenda 21, adopted by the Conference, it
    was specifically recommended to "undertake, as a matter of urgency,
    research on the effects on human health of the increasing
    ultraviolet radiation reaching the earth's surface as the
    consequence of depletion of the stratospheric ozone layer." It is
    this issue that underscores the current need to better understand
    the potential health and environmental risks of UV exposure.

         UV is one of the non-ionizing radiations in the electromagnetic
    spectrum and lies within the range of wavelengths 100 nm (which
    corresponds to a photon energy of approximately 12 eV) to 400 nm.
    The short wavelength limit of the UV region is often taken as the
    boundary between the ionizing radiation spectrum (wavelengths < 100
    nm) and the non-ionizing radiation spectrum. UV can be classified
    into UVA (315 - 400 nm), UVB (280 - 315 nm) and UVC (100 - 280 nm)
    regions, although other conventions for UVA, UVB and UVC wavelengths
    bands are in use.

         Most artificial sources of UV, except for lasers, emit a
    spectral continuum of UV containing characteristic peaks, troughs
    and lines. These sources include various lamps used in medicine,
    industry, commerce, research and the home.

         UV-induced biological effects depend on the wavelengths of the
    radiation. It is necessary for a proper determination of hazard to
    have spectral emission data. These consist of spectral irradiance (W
    m-2 nm-1) measurements or calculations of emissions from the
    source. The total irradiance (W m-2) is obtained by summing over
    all wavelengths emitted. The biological or hazard weighted
    irradiance (W m-2 effective) is determined by multiplying the
    spectral irradiance at each wavelength by the biological or hazard
    weighting factor (which quantifies the relative efficacy at each
    wavelength for causing the effect) and summing over all wavelengths.
    Such factors are obtained from action spectra.

    1.2  Action Spectrum and Minimum Erythemal Dose

         An action spectrum is a graph of the reciprocal of the radiant
    exposure required to produce the given effect at each wavelength.
    All the data in such curves are normalized to the datum at the most
    efficacious wavelength(s). By summing the biologically effective
    irradiance over the exposure period, the biologically effective
    radiant exposure (J m-2 effective) can be calculated. For UV
    induced erythema, the action spectrum adopted by the International
    Commission on Non-Ionizing Radiation Protection (ICNIRP),
    International Commission on Illumination (CIE), the International
    Electrotechnical Commission (IEC) and various national bodies, is a
    composite curve obtained by statistical analysis of many research
    results on the minimum radiant exposure of UV at different
    wavelengths necessary to just cause erythema.

         The most commonly used quantity for describing the erythemal
    potential of an exposure to UV is the number of minimum erythemal
    doses (MEDs) represented by the exposure. An MED is the radiant
    exposure of UV that produces a just noticeable erythema on a
    previously unexposed skin. It corresponds to a radiant exposure of
    monochromatic radiation at the maximum spectral efficacy for
    erythema (around 300 nm) of approximately 150 to 2000 J m-2
    effective, depending on skin type. In this text 200 - 300 J m-2
    effective is used as the value of 1 MED for comparative safety
    purposes for white skin.

    1.3  Cellular and Molecular Studies

         To produce any change, UV must be absorbed by the biomolecule.
    This involves absorption of a single photon by the molecule and the
    production of an excited state in which one electron of the
    absorbing molecule is raised to a higher energy level. The primary
    products caused by UV exposure are generally reactive species or
    free radicals which form extremely quickly but which can produce
    effects that can last for hours, days or even years. DNA is the most
    critical target for damage by UVB and UVC radiations. While a
    considerable amount of knowledge is available concerning the
    interaction of UV with nucleic acids, controversy exists as to which
    lesion constitutes the most important type of pre-mutagenic damage.

         Cell death, chromosome changes, mutation and morphological
    transformations are observed after UV exposure of procaryotic and
    eucaryotic cells. Many different genes and several viruses
    (including HIV) are activated by UV exposure. The genes activated by
    UVB and UVC are different from those activated by UVA. Studies of
    DNA repair defective disorders have clearly established a link
    between UV induced DNA damage in skin and various types of cancer.

    1.4  Animal Studies

         Solar UV has been shown to produce cancers in domestic and food
    animals. In experimental animals UV causes predominantly squamous
    cell carcinomas. UVB is most effective at producing SCCs, although
    they are produced by UVA but at much higher intensities, similar to
    the levels needed for erythema and tanning. The effectiveness of UVC
    is unknown except at one wavelength (254 nm). At this wavelength the
    effectiveness is less than UVB.

         Melanomas are much less common and only two animal models have
    been found for induction of melanoma by UV alone. An initial action
    spectrum determined for a type of hybrid fish indicates a peak in
    the UVB range but also shows a high level of effectiveness in the
    UVA. Basal cell carcinomas are rare in animals.

         Exposure to suberythemal doses of UV have been shown to
    exacerbate a variety of infections in rodent models. UV affects
    infections both at the site of exposure and at distant sites. Recent
    work indicates that systemic infections without skin involvement may
    be affected. Enhanced susceptibility appears to result from T-helper
    cell activity. The mechanisms associated with this suppression
    appear to be the same as those identified with suppression to
    contact and delayed type hypersensitivity responses. Suppression of
    these immune responses appears to be mediated by release of soluble
    mediators from UVB exposed skin which alters the antigen
    presentation by Langerhans and other cells so that they fail to
    activate TH 1 cells. The resulting immune suppression is antigen
    specific, can occur regardless of whether or not antigen is applied
    at the site of exposure, and is relatively long lasting. UV exposure
    also prevents the development of protection immunity to a variety of
    infections in mice and rats.

         Many studies in experimental animals have demonstrated that UV
    exposure can cause both acute and delayed effects such as cataract,
    photokeratitis, damage to the corneal epithelium and various retinal
    effects. Studies of photochemical retinal injury in aphakic monkeys
    have shown that the retina is six times more vulnerable to
    photochemical damage from UV than the visible wavelengths.

    1.5  Health Effects on Humans

    1.5.1  Skin

         Acute effects on the skin consist of solar erythema, "sunburn",
    which, if severe enough, may result in blistering and destruction of
    the surface of the skin with secondary infection and systemic
    effects, similar to those resulting from a first or second degree
    heat burn. Although UVC is very efficiently absorbed by nucleic
    acids, the overlying dead layers of skin absorb the radiation to
    such a degree that there is only mild erythema and, usually, no late

    sequelae, even after repeated exposures. Much less is known about
    the biological effects of UVA. However, doses of UVA, which alone
    may not show any biological effect, can, in the presence of certain
    environmental, consumer and medicinal chemical agents, result in
    injury to tissues (phototoxicity, photoallergy, enhancement of
    photocarcinogenesis).

         Chronic skin changes due to UV consist of skin cancer (both
    melanoma and non-melanocytic), benign abnormalities of melanocytes
    (freckles, melanocytic naevi and solar or senile lentigines), and a
    range of other chronic injuries resulting from UV exposure to
    keratinocytes, blood vessels and fibrous tissue, often described as
    "photoaging" (solar elastosis). The much increased rates of skin
    cancer in patients with xeroderma pigmentosum, who have a deficiency
    in the capacity to repair UV-induced DNA damage, suggest that direct
    UV damage of the DNA may be a step in the cause of these cancers.
    This suggestion has also been supported by the observation of UV
    specific mutations of the p53 tumour suppressor gene in a proportion
    of patients with non-melanocytic skin cancer. Oxidative and immune
    suppressant effects may also contribute to the capacity of UV to
    cause skin cancers.

         Cancer of the lip is much more common in fair than dark skin
    populations and is associated with outdoor work. However possible
    confounding with tobacco and alcohol use has not been adequately
    controlled in any study.

         The worldwide incidence of malignant melanoma has continued to
    increase. Strong epidemiological evidence exists that sun exposure
    causes cutaneous melanoma and non-melanocytic skin cancer. Their
    incidence is less in darker than light skin groups living in the
    same geographical area. Risk of skin cancer decreases with
    increasing pigmentation. The anatomical site most seen for squamous
    cell carcinoma (SCC) is the head and neck, areas most exposed to the
    sun. Incidence of both melanoma and non-melanocytic skin cancer are
    increased in areas of high ambient solar UV radiation. Melanoma is
    strongly related to frequency of recreational exposure to the sun
    and to history of sunburns.

         There is suggestive evidence that exposure to sunlamps may
    increase the risk of melanoma, but the studies conducted so far have
    not consistently controlled confounding factors.

    1.5.2  Immune system

         A number of studies suggest that UV exposures at environmental
    levels suppress immune responses in both rodents and man. In rodents
    this immune suppression results in enhanced susceptibility to
    certain infectious diseases with skin involvement and some systemic
    infections. Mechanisms associated with UV-induced immunosuppression
    and host defence mechanisms which provide for protection against

    infectious agents, are similar in rodents and man. It is therefore
    reasonable to assume that exposure to UV may enhance the risk of
    infection and decrease the effectiveness of vaccines in humans.
    However additional research is necessary to substantiate this.

    1.5.3  Eye

         The acute effects of UV on the eyes consist of the development
    of photokeratitis and photoconjunctivitis, which are unpleasant but
    usually reversible and easily prevented by appropriate eyewear.
    Chronic effects on the eye consist of the development of pterygium
    and squamous cell cancer of the conjunctiva and cataracts. A review
    of the studies suggests that there is sufficient evidence to link
    acute ocular exposure to photokeratitis but our knowledge of the
    effects of chronic exposure is less certain. While there is
    sufficient evidence that cortical and posterior subcapsular
    cataracts (PSC) can be caused by UVB in laboratory animals, there is
    limited evidence to link cortical and PSC cataracts in humans to
    chronic ocular exposure to UVB. Insufficient information is
    available to separate out the other factors contributing to cataract
    formation, or to state the proportion of cataracts which can be
    attributed to UVB exposure. There is also limited evidence to link
    the development of climatic droplet keratopathy and pterygium, but
    insufficient evidence to link uveal melanoma with UV exposure.

    1.6  Environment

         Increased levels of UV due to ozone layer depletion may have
    serious consequences for living organisms. A 10% reduction in ozone
    could lead to as much as a 15-20% increase in UV exposure depending
    on the biological process being considered. While the impact on
    human health, crop production, fisheries etc. is largely unknown,
    adverse effects of increased exposure to UVB have been reported on
    plant growth, photosynthesis and disease resistance. Further, the
    impact of increased UV levels on aquatic ecosystems (the major
    contributor to the earth's biomass) may be substantial.
    Phytoplankton, at the base of the aquatic food chain, serves as food
    for larvae of fish and shrimp. These in turn are consumed by fish,
    which subsequently provide an essential food source for many human
    beings and other animals. A significant reduction in phytoplankton
    from increased UVB exposure will directly affect the human and
    animal marine food source.

    1.7  Guidelines on Exposure Limits and Protective Measures

         Guidance on exposure limits for UV are described in chapter 13.
    International guidelines define exposure limits (ELs) below which it
    is expected that nearly all people may be repeatedly exposed without
    adverse effects. The ELs are intended to be used to evaluate
    potentially hazardous exposures from, for example, solar radiation,
    arcs, gas and vapour discharges, fluorescent lamps and incandescent

    sources. The ELs are generally below levels which are often used for
    the UV exposure of patients required as part of medical treatment
    and below levels associated with sunbed exposure. ELs are not
    intended to apply to exposure of pathologically photosensitive
    individuals, to people concomitantly exposed to photosensitising
    agents or to neonates.

         Finally this monograph describes protection and control
    measures such as the containment of UV sources, and methods for
    personal protection including the use of sunscreen preparations,
    clothing, eye and skin protection, and behavioural modifications.

         While topical application of sunscreen is a preferred method of
    absorbing UVB, some preparations do not absorb the longer wavelength
    UVA effectively. Moreover, some have been found to contain
    ingredients that are mutagenic in sunlight. There is still much
    research necessary before the impact on health of increased levels
    of UVA will be known. In the meantime people using sunscreens should
    use only those with the highest sun protection factor (SPF) and be
    aware that they are for their protection from the sun and not for
    tanning purposes. Use of wide brimmed hats, protective clothing and
    UV absorbing eye glasses is still the best personal protection
    against the adverse effects of UV exposure.

         With increasing levels of solar UV resulting from depletion of
    the ozone layer, and the continuing rise in the level of melanoma
    worldwide, people should become more aware of their UV exposure and
    take appropriate precautions. These precautions include staying out
    of the sun during the period around noon (the period when the UV
    levels are highest), or wearing UV protective clothing, hats and sun
    glasses. Broad spectrum (UVB and UVA protective) sunscreens should
    be used when other means of protection are not feasible. These
    sunscreens should be used to reduce exposure rather than lengthen
    the period of exposure to the sun. Protection of young children is
    particularly important for the prevention of long-term consequences
    of UV exposure. In general behavioural patterns must change to
    protect against increasing solar UV levels.

    2.  PHYSICAL CHARACTERISTICS

    2.1  Electromagnetic Spectrum

         Oscillations of electromagnetic fields can cause energy to be
    transported in the form of electromagnetic radiation. Examples of
    this type of radiative transport of energy include radiofrequency
    waves, light and x rays. Ultraviolet (UV) radiation is one form of
    electromagnetic energy in the optical region of the electromagnetic
    spectrum. All electromagnetic radiation is characterized by
    frequency f and wavelength lambda. These two quantities are linked
    through the relationship:

                                 f = c/lamda

    where c is the speed of light (3 108 m s-1). The energy of a
    single photon is determined by the wavelength of the photon as
    described by the relationship

         (photon energy) = hf = hc/lamda    where h = 1.24 eV nm

         Non-ionizing radiation (NIR) is the term generally applied to
    all forms of electromagnetic radiation whose primary mode of
    interaction with matter is other than by producing ionization. NIR
    refers to electromagnetic fields and radiation with wavelengths
    exceeding 100 nm, which is equivalent to quantum (photon) energies
    below 12.4 eV, the minimum energy needed to break the weakest
    macromolecular bonds. The non-ionizing spectrum encompasses all
    fields of radiation from UV to DC fields.

         For purposes of health protection, the optical portion of NIR
    can be subdivided into several wavelength ranges, as shown in table
    2.1. The nomenclature was standardized by the International
    Commission on Illumination (CIE) however, some scientists use a
    modification of this system by shifting the 315 nm break point
    between UVA and UVB to 320 nm; the reader should always check the
    definition given in each publication. For the purposes of this
    document, the CIE convention is followed.

    Table 2.1  Optical Radiation Spectral Bands
                                                                       

         Ultraviolet radiation              100 - 400 nm
         UVA                                315 - 400 nm
         UVB                                280 - 315 nm
         UVC                                100 - 280 nm
         Visible radiation (light)          400 - 760 nm
         Infrared radiation (IR)            760 - 106 nm = 1 mm
                                                                       
         UV of wavelengths less than 180 nm has no direct biological
    effect on humans since it is effectively absorbed in a few

    centimetres of air. For this reason, the spectral region below 180
    nm is frequently referred to as the  vacuum ultraviolet region.

    2.2  Radiometric Quantities and Units

         Radiometric quantities are absolute physical quantities used to
    describe the characteristics of a source or radiation field. For UV,
    ten generic radiometric terms are summarized in table 2.2. Each of
    these quantities can be defined for a certain wavelength or
    frequency range, or can be integrated over the whole spectrum of a
    given source. Since UV is normally absorbed over a surface, with
    very limited penetration depth, the most common quantities used to
    describe exposure dose and dose rate to UV are: radiant exposure
    (incident energy divided by the receptor surface area) and
    irradiance (incident power divided by the receptor surface area)
    respectively. Radiant exitance is the power per area of the emitted
    radiation  at the emitting surface. The receptor surface is most
    often considered as a flat plane. However, for some biological and
    chemical purposes, the radiation incident on a cylindrical or
    spherical surface may also be considered.

         For the purpose of radiation protection, physical quantities
    are needed to describe sources and fields of radiation as well as
    the interaction of this optical radiation with matter. These are
    described in IRPA/INIRC(1985).

    2.3  UV Production

         Sources of electromagnetic radiation can be categorized in
    several different ways; for example they can be grouped according to
    the type of material or the type of equipment that produces the
    radiation. At the submicroscopic level, the manner in which the
    radiation originates can be described in terms of nuclear,
    electronic or molecular transitions between energy states or by the
    acceleration of charged particles. Common sources of UV emission
    involve energy transitions between electronic states of molecules in
    materials.

    2.3.1  Thermal emitters

         When the temperature of a material increases, electrons in the
    molecules are raised to higher energy states, a variety of energy
    transitions take place and photons are emitted. The higher the
    temperature the greater is the fraction of these photons at higher
    energies. Matter at temperatures above 2500 K emits a significant
    number of photons in the UV spectral range. The emission spectra of
    such incandescent sources are characteristically smooth - so-called
    "continuum", possibly with superimposed spectral emission lines.


    
    Table 2.2  Radiometric terminology for UV:
               Useful radiometric quantities and units
                                                                                                           
    Quantity             Symbol     Defining equation                           Unit and abbreviation
                                                                                                           

    Radiant energy       Qe         Qe = Integral Operator Phie dt              joule (J) = 1 watt second

                                         dQe
    Radiant energy       We         We =                                        joule per cubic metre
    density                              dV                                     (J m-3)

                                           dQe
    Radiant flux         Phie , P   Phie =                                      watt (W)
    (radiant power)                         dt

                                         dPhie
    Radiant exitance     Me         Me =                                        watt per square metre
                                           dA                                   (W m-2)

                                       = Integral Operator Le cosTheta dOmega

    Irradiance or                        dPhie
    Radiant Flux         Ee         Ee =                                        watt per square metre
    Density                                dA                                   (W m-2)

                                         dPhie
    Radiant intensity    Ie         Ie =                                        watt per steradian (W sr-1)
                                         dOmega

                                             d 2Phie
    Radiance1            Le         Le =                                        watt per steradian per
                                         dOmega dA cosTheta                     square metre (W sr-1 m-2)

    Table 2.2 (contd).
                                                                                                           
    Quantity             Symbol     Defining equation                           Unit and abbreviation
                                                                                                           

    Radiant exposure                He = dQe                                    joule per square metre
    (dose in             He                                                     (J m-2)
    photobiology)                         dA

                                         P
    Radiant efficiency2  ne         ne =                                        unitless
    (of a source)                        Pi

    Optical density3     De         De = - log10(gammae)                        unitless
                                                                                                           

    NOTE:  All terms in this table are radiometric terms and should not be confused with photometric terms.
    The symbol A represents surface area, Omega is solid angle, Theta is the incident (zenith) angle,
    V represents volume, t is time. The units may also refer to narrow spectral bands in which the term is
    preceded by the word  spectral and the unit is then per wavelength interval and the symbol has a
    subscript lamda. For example, spectral irradiance Egamma has units of W m-2 nm-1 or W cm-2 nm-1. While
    the metre is the preferred unit of length, the centimetre is still commonly used for many of the above
    terms and the nm or µm are most commonly used to express wavelength.

    1.  At the source L =     dI      and at a receptor L =       dE       
                          dA cosTheta                       dOmega cosTheta

    2.  Pi is electrical input power in watts

    3.  This formula applies only to situations where the radiation is not scattered, but only absorbed.
        In this case gamma represents the fraction of transmitted energy.


    
         The spectral emission of a heated material is governed by basic
    physical laws. The Stefan-Boltzmann Law expresses the total
    radiation emitted by a heated material as a function of its
    temperature. The total radiant power emitted by a theoretically
    perfect radiator, a so-called black body, is proportional to the
    fourth power of its temperature (in kelvin, K). Any practical
    thermal emitter in equilibrium emits less radiant power than its
    equivalent black body at the same temperature and unlike a black
    body the amount depends on the nature of its surface. The spectral
    distribution of black body radiation is described by the Planck
    Radiation Law. Wien's Displacement Law is illustrated in figure 2.1,
    and describes mathematically the spectral distribution of emission
    from a black body and the wavelength of the maximum emission shift
    to shorter wavelengths as temperature increases. When a material is
    heated so that incandescence occurs, it first appears red and, with
    increasing temperature, progresses to white or blue.

    2.3.2  Electrical gaseous discharges

         Radiation may be produced when an electric current is passed
    through a gas or vapour. Atoms may be ionized if sufficient energy
    is transferred from a moving electron. Alternatively the moving
    electron may not impart sufficient energy for ionization, but
    instead may impart energy to raise the electrons of the gas to an
    excited (higher) energy level. When they return to a lower energy
    level, or their ground state, radiation of one or more
    characteristic wavelengths is emitted. The wavelengths of emission
    are determined by the type of gas or vapour present in the discharge
    and appear as spectral emission lines. The width of the lines and
    the amount of radiation in the interval between them depends on the
    pressure of the discharge. At low pressure, fine lines with little
    or no continuum are produced. As the pressure of the discharge is
    raised the lines broaden and their relative magnitudes alter. The
    magnitude of the continuum increases. The electrical gaseous
    discharge is the basis of operation of many UV emitting lamps.

    2.3.3  Stimulated emission

         Radiation can be produced by the specific electronic transition
    process of stimulated emission. This depends on the ability of the
    radiating medium to undergo "population inversion", i.e., achieving
    a condition where there are more atoms or molecules in a higher
    energy excited state than in a lower one. Once population inversion
    occurs an avalanche of photons can be generated by stimulated
    emission. Initially, spontaneously emitted photons stimulate other
    excited atoms to emit photons of the same energy in phase with one
    another. This is the basis of operation of the laser.

    FIGURE 2.1

    2.4  Summary

         UV is a part of the non-ionizing region of the electromagnetic
    spectrum. Its position in the spectrum can be characterised by
    wavelength and it is quantified using radiometric quantities and
    units. Production of UV is possible by different mechanisms that are
    all based on atomic or molecular excitation. Excitation in a medium
    can be achieved by thermal, electrical or optical energy absorption.
    Emission of radiation is caused either by de-excitation or by
    stimulated synchronised processes.

    3.  UV SOURCES

    3.1  The Sun

         The sun is the main source of UV. The broad spectrum and
    intensity of UV from the sun are due to the high temperature at its
    surface and its size. The intensity of solar UV reaching the earth's
    atmosphere would probably be lethal to most living organisms on the
    earth's surface without the shielding afforded by the atmosphere.
    Solar UV undergoes absorption and scattering as it passes through
    the earth's atmosphere with absorption by molecular oxygen and ozone
    being the most important processes. The ozone layer prevents almost
    all UV of wavelengths lamda < 290 nm and a substantial fraction (in
    excess of 90% of the total energy) from 290 - 315 nm from reaching
    the earth's surface. Thus the terrestrial environment is exposed to
    UV between 290 nm and 400 nm. The spectrum, both before passage
    through the atmosphere and at sea level, is shown in figure 3.1

         The solar radiation reaching the top of the earth's atmosphere
    is affected by the solar output and the earth-sun distance.
    Variations in solar output are much smaller than the variations
    caused by atmospheric attenuation factors. The irradiance incident
    on the top of the earth's atmosphere is summarised in table 3.1, but
    will vary with the exact distance of the earth from the sun at a
    particular time. The extreme values associated with this variation
    are approximately ± 3.3% above and below the annual mean and occur
    in January and July respectively.

    Table 3.1  Spectral distribution of solar radiation prior to
    attenuation by the earth's atmosphere (Frederick et al., 1989).
                                                                       
    Wavelength band     Irradiance (W m-2)            Percent of total

    UVC                          6.4                         0.5
    UVB                         21.1                         1.5
    UVA                         85.7                         6.3
    Total UV                   113.2                         8.3
    Visible and IR            1254                          91.7
                                                                       

         The atmosphere has a profound effect on the irradiance which
    reaches the surface of the earth. In January (in the northern
    hemisphere) or July (in the southern hemisphere) when the solar
    elevation is low, direct UV travels a longer path through the
    atmosphere and a large amount of scattering occurs. In addition,
    much of the resultant scattered UV propagates downwards to the
    earth's surface at angles to the horizontal that are larger than the
    solar elevation, hence travelling a shorter and less absorptive
    path. This results in large ratios of scattered to direct UV. During
    the summer the ratio of diffuse to direct UV is smaller.

    FIGURE 3.1

    3.1.1  Factors affecting solar UV levels

         The total solar UV reaching the earth's surface, termed global
    UV can be divided into two components: direct and diffuse. Global UV
    reaching a horizontal surface is the quantity most often measured.
    For biological entities such as people and trees, UV hitting
    cylindrical or spherical surfaces may be more important.

         The amount and spectral distribution of solar UV irradiance
    reaching the earth's surface depends on a number of factors,
    including

    (a)  wavelength of the UV
    (b)  solar zenith angle, which depends on latitude, date of the year
         and time of day
    (c)  solar source spectrum incident at the top of the atmosphere (d)
         ozone column thickness and vertical distribution
    (e)  molecular absorption and scattering (including localized
         gaseous pollutants)
    (f)  aerosol absorption and scattering (including anthropogenic
    (g)  absorption, scattering and reflection by clouds
    (h)  reflectance characteristics (albedo) of the ground
    (i)  shadowing by surrounding objects
    (j)  altitude above sea level

         The presence of cloud cover, air pollution, haze, or even
    scattered clouds, plays a significant role in attenuating UV. UVB
    and UVA irradiances are reduced due to scattering by water droplets
    and/or ice crystals in the clouds. Clouds can block a significant
    portion of the UV which would have otherwise reached the surface.

         Cloud cover and type are highly variable. The transmission of
    UV radiation through clouds depends on cloud height, type and
    optical density. The resultant effect on UV transmission is
    difficult to assess particularly in the case of partial cloudiness.
    The effect of cloudiness on the solar irradiation of a horizontal
    plane can be approximated by

              F = 1 - 0.056 C

    where C is the total cloud index in tenths of sky covered from 0 to
    10, 10 being complete sky cover. Thus for complete cloud cover, the
    transmitted UV irradiance would decrease by 72% and for half cloud
    cover by 44%. In extreme cases cloud cover can decrease UV
    irradiance by over 90%. Estimates of the average reduction of UVB
    due to clouds (relative to cloudless skies) based on satellite
    measurements of backscattered solar UV are 30% at 60 degrees
    latitude, 10% at 20 degrees latitude and 20% at the equator.

    3.1.2  Ozone depletion effects

         Over 90% of the total atmospheric ozone resides in the
    stratosphere (the upper atmosphere). The total ozone column is
    important for filtering solar UV. Only UVB is affected by changes in
    the ozone column. UVC is almost completely absorbed by ozone and
    oxygen in the atmosphere; even with severe ozone reduction UVC would
    still be effectively absorbed by the remaining oxygen.

         As stratospheric ozone levels decrease, the resulting higher
    levels of solar UVB could increase the production of reactive OH
    radical molecules, potentially increasing the chemical reactivity of
    the troposphere. In polluted areas with sufficient concentrations of
    oxides of nitrogen (NOx) (above 0.5 ppb by volume) and hydrocarbon
    compounds this enhanced reactivity is calculated to result in
    greater levels of tropospheric ozone and other potentially harmful
    oxidized products, such as hydrogen peroxide and acids.

         The amount of ozone in the free troposphere is increasing for
    example over central europe and other parts of the globe. The effect
    of these trends on UV radiation needs to be studied further (WMO,
    1993). For high sun angles, tropospheric ozone is a more effective
    absorber of UV radiation than stratospheric ozone because of the
    increased path length of scattered radiation in the lower atmosphere
    (Brühl & Crutzen, 1989).

         The total ozone column is not uniform but varies with latitude
    and time of year. At the same latitude, away from the equator and
    tropics, the total ozone column tends to be greater in spring than
    in autumn. Thus, even though the sun angles are the same on March 21
    and September 21, the differences in total ozone column result in
    more UVB in early autumn than in early spring. If cloud cover and
    atmospheric pollution are not taken into account, in the tropics,
    the relatively constant ozone column and the similar solar angles
    throughout the year, result in little variation in solar UVB with
    season.

         Changes in total column ozone were observed by both ground
    based and satellite based instruments from the WMO Global Ozone
    Observing System. The measurements show pronounced ozone depletion
    over Antarctica in months 9-12 (local early spring) (UNEP-WMO,
    1989). Increases in biologically effective UVB were observed during
    this same time period (NSF, 1993). Over representation of UVA in the
    common broad band meter action spectrum, relative to most biological
    action spectra, produces less pronounced relative and absolute
    changes in assessed UVB. Therefore under cloud free conditions
    common broad band meters are less sensitive in assessing the ground
    level UVB effects of ozone depletion than detectors with a steeper
    action spectrum response such as for DNA damage or for erythema.

    3.1.3  Trends in UV

         There have been no significant changes in ozone at the equator.
    Total column ozone over the Northern mid-latitudes has decreased by
    several percent over the past two decades. Efforts to detect changes
    in UV over long time periods have failed due to changes in column
    ozone.

         Besides the problems indicated in the interpretation and
    comparison of measurement data from different sites and sources,
    there are also insufficient direct solar terrestrial UVB
    measurements for constructing a global climatology or trend
    assessment due to some of the following (Driscoll 1992):

    (a)  problems in establishing appropriate instrumentation: either
         highly sophisticated and reliable spectral instruments or broad
         band instruments which fairly represent the sensitivities of
         different biological and chemical targets (with different
         wavelength dependencies and sensitivity to different
         orientations),

    (b)  difficulty of maintaining accurate field instrument
         calibrations over many years, and

    (c)  practical limitations in establishing a global monitoring
         network especially with the potential for disturbance from
         locally polluted areas.

         With insufficient spectral measurement data collected over long
    periods, data from the more extensively used broad-band measurement
    systems have been scrutinised. UV data from eight stations in the US
    showed decreases in UVB between 0.5% and 1.1% per year during the
    time period of 1974-1985 (Scotto  et al., 1988). However, this
    result does not agree with theoretical predictions and may be due to
    problems in the long-term calibration of meters or local pollution
    in mainly urban or semi-urban sites (Munkata, 1993; Smith & Ryan
    1993). In Russia, a 12% decrease of UVB was observed in Moscow
    between 1968 and 1983 with a concurrent 15% increase in turbidity
    and a 13% increase in cloudiness. At the Jungfraujoch observing
    station in the Swiss alps (3.6 km above sea level), increases of
    0.7±0.2% per year in UVB were observed under clear sky conditions
    between 1981 and 1989. A comparison of summer spectral data weighted
    with the CIE erythemal action spectrum using the same
    instrumentation at Lauder, New Zealand (45°S) and Neuherberg,
    Germany (48°N) showed weighted UV irradiances were 1.6 times larger
    in New Zealand due to decreased ozone column thickness (24% lower
    than in Germany). Spectral filter measurements at 39o N between 1976
    though 1990 showed large increases in monthly maximum values which
    were not statistically significant (Correll  et al., 1992).

    3.1.4  Theoretical models

         To overcome some of the deficiencies in interpreting the
    results from solar UV measurements, theoretical models have been
    used to predict UV levels. However, the theoretical determination of
    the spectral distribution of global UV including the effects of
    cloud cover and ground reflection is extremely complex. There are
    three categories of models to predict UV transmission. Empirical and
    semi-empirical models are useful for assessing daily or annual
    erythemal doses as a function of latitude, solar zenith angle, cloud
    cover and ground albedo. These models have been used for a long time
    at a variety of locations. Two of the most common examples are the
    models of Green  et al. (1974) and Diffey (1977)

         Two stream models simulate the radiative transfer through the
    atmosphere. They have been well verified with field data from
    Antarctica as well as mid-latitudes (e.g. Frederick  et al., 1989).

         Multistream models allow for the angular distribution of UV
    transmission. This is useful for more detailed dosimetry (Stamnes
     et al., 1988).

         The maximum erythemally effective UVB irradiances, calculated
    by the Diffey model, are shown in table 3.2 for the Northern
    Hemisphere, as a function of latitude and time of year at sea level
    (Driscoll 1992). There are strong seasonal and latitudinal
    variations in UVB. Under cloudless skies, the UVB is more intense in
    summer and at all times of year is greater at lower latitudes. The
    solar elevation angle determines the length of the path of the sun's
    rays as they penetrate the atmosphere. When the sun is low in the
    sky, the path through the atmosphere is longer and the filtering
    action of the air is therefore increased. When the sun is directly
    overhead, the sun's rays have the shortest path through the
    atmosphere. Approximately 50% of the daily UV is received during the
    middle four hours around noon when the sun is high in the sky
    (Sliney, 1987).

         The number of minimum erythemal doses (MEDs) in a 3 hour
    exposure period around noon for fair skin is shown as a function of
    latitude and time of year in table 3.3 for the Northern Hemisphere.
    For example, at 55°N (Newcastle), a calculated value of 6 MED would
    be received for a sensitive skin type in a 3 hour exposure around
    noon on a clear day in July (1 MED per half hour exposure).

         Comparing the values in table 3.2 with the results of
    calculations using the model developed by Frederick  et al. (1989)
    for clear sky conditions at noon, the latter are typically 50%
    higher. This may be due to differences in the biological action
    spectra used or in the spectral irradiance data calculated. The R-B
    meter response characteristic used by Frederick  et al. (1989) does

    not follow the CIE reference erythema action spectrum used in the
    model developed by Diffey (1977) at wavelengths greater than 300 nm.

    3.1.5  UV monitoring

         Recent public and scientific concern about ozone depletion and
    increased UV have lead to the establishment of many UV monitoring
    centres in the last few years. Five years ago less than fifty UV
    monitoring stations were operating around the world. Today more than
    250 monitoring centres are underway for a variety of reasons.
    Governmental agencies, scientific institutions, universities and
    private groups have begun to monitor UV. The World Meteorological
    Organization (WMO) has established a global network called Global
    Atmosphere Watch (GAB). It presently has eight observatory stations
    that make continuous spectral and broad band UV measurements. The
    Global Environment Facility is supporting the creation of 10-15
    additional stations in developing countries. Various national and
    multi-national agencies are also operating and establishing UV
    monitoring networks.

    Table 3.2  Calculated clear sky noontime erythemally effective UVB
    irradiances (mW m-2 CIE erythemally weighted) on a horizontal
    surface as a function of latitude (°) and time of year for the sea
    level in the Northern Hemisphere using typical ozone values
    (Driscoll, 1992).
                                                                       

          Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sep  Oct  Nov  Dec

     0    229  251  260  229  210  207  214  233  278  248  226  221
     5    212  232  242  235  214  213  220  241  258  227  207  206
    10    203  220  229  248  226  225  232  255  244  215  198  175
    15    155  199  201  227  221  226  234  240  223  198  155  135
    20    127  183  191  215  234  239  247  227  212  184  127  111
    25     95  134  165  185  203  227  219  200  194  143  101   84
    30     79  110  157  176  192  215  208  190  186  117   85   56
    35     42   80  114  146  160  188  182  166  148   93   50   31
    40     28   61   93  139  152  178  172  158  121   73   33   21
    45     16   33   66  105  126  158  153  128   93   42   21   13
    50     12   22   54   85  119  150  145  104   77   28   15    8
    55      6   14   33   65   92  130  114   81   48   18    8    4
    60      4   10   21   53   75  105   93   67   31   12    5    2
    65      0    6   14   35   59   85   76   42   21    7    2    0
    70      0    3   10   23   48   70   62   29   15    4    0    0
    75      0    0    6   15   30   53   39   20    9    0    0    0
    80      0    0    3   11   20   33   26   14    5    0    0    0
    85      0    0    2    7   14   23   18    9    2    0    0    0
    90      0    0    0    4   10   16   13    5    0    0    0    0
                                                                       

    Table 3.3  Number of MEDs in a 3 h exposure period for a sensitive
    skin type (1 MED = 200 J m-2 effective) for the erythemally
    effective UVB irradiances given in table 3.2 as a function of
    latitude (°) and time of year for the Northern Hemisphere (Driscoll
    1992).
                                                                       

          Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sep  Oct  Nov  Dec

     0     12   14   14   12   11   11   12   13   15   13   12   12
     5     11   13   13   13   12   12   12   13   14   12   11   11
    10     11   12   12   13   12   12   13   14   13   12   11    9
    15      8   11   11   12   12   12   13   13   12   11    8    7
    20      7   10   10   12   13   13   13   12   11   10    7    6
    25      5    7    9   10   11   12   12   11   10    8    5    5
    30      4    6    8   10   10   12   11   10   10    6    5    3
    35      2    4    6    8    9   10   10    9    8    5    3    2
    40      2    3    5    8    8   10    9    9    7    4    2    1
    45      1    2    4    6    7    9    8    7    5    2    1    1
    50      1    1    3    5    6    8    8    6    4    2    1    0
    55      0    1    2    4    5    7    6    4    3    1    0    0
    60      0    1    1    3    4    6    5    4    2    1    0    0
    65      0    0    1    2    3    5    4    2    1    0    0    0
    70      0    0    1    1    3    4    3    2    1    0    0    0
    75      0    0    0    1    2    3    2    1    0    0    0    0
    80      0    0    0    1    1    2    1    1    0    0    0    0
    85      0    0    0    0    1    1    1    0    0    0    0    0
    90      0    0    0    0    1    1    1    0    0    0    0    0
                                                                       

         The reasons for monitoring UV are generally divided into the
    following four areas:

         (i)  to provide information to the public on UV levels and
              variations,

        (ii)  establishing a basic UV climatology,

       (iii)  studying cause and effects of UV transmission, and

        (iv)  detecting long term variability.

         While these reasons are not mutually exclusive they can often
    dictate the type of care taken with instrument maintenance. The
    quality and care needed for trend detection requires a great deal
    more effort than for UV measurements to provide public information.

         There are many difficulties in making accurate field
    measurements of ambient UV. Care must be taken that no part of the
    instrument is shaded from either direct or indirect sun; nearby
    surfaces must not change during the time period of monitoring; and

    routine maintenance such as snow and debris removal must be
    conducted systematically to avoid trend bias. The value of the
    measurements for trend detection rely on diligent maintenance of the
    site including co-located measurements. The data must have spectral,
    temporal and angular resolution appropriate for validation of model
    calculations and for evaluation of biological effects (spectral
    resolution better than <1 nm). Broad band instruments are useful
    for establishing climatology and operations where spectral
    instruments are not practical.

         Frequent and reliable calibration of the instrumentation is
    necessary. UV monitoring for the detection of trends is difficult
    and failure to detect a trend may be due to lack of awareness of
    changes in instrument behaviour and correction of it.
    Intercomparison of different types of UV monitoring instruments have
    shown that individual instrument characteristics can cause
    substantial differences in measurements. Further intercomparisons
    are necessary to understand and reduce the uncertainties of
    measurements.

    3.1.6  Analysis of UV data

         Atmospheric UV data are difficult to analyze. To date all UV
    data have been analyzed separately by each investigator and in a
    different manner. This variation in analysis makes the
    interpretation and comparison of results difficult. Unfortunately,
    the data are normally not easily obtained and re-analyzed for
    comparative studies.

         UV data pose difficult problems because they are not generally
    independent and normally distributed, therefore robust time series
    techniques should be used to analyze the data correctly. Standard
    techniques, depending on how they are applied, can greatly over or
    under estimate the confidence in the observed trends, as well as
    supply spurious trends when used on UV data. Bishop (1992) has
    suggested that using proper statistical techniques, at least ten
    years of data will be necessary to correctly detect a trend of 5%
    per decade. This estimate was made based on analysis of a single
    station's data independently. However, analysing groups of data may
    allow trend detection in less time.

    3.1.7  Conclusions

         The earth's atmosphere has a profound effect on the UV
    irradiance reaching the earth's surface. It absorbs a large fraction
    of the incident UV and changes part to diffuse radiation. Variations
    in UV intensity depend on solar zenith angle, atmospheric ozone,
    cloudiness, aerosol load and other factors.

         One of the most important tools to determine climatological
    values and long-term changes in solar UV radiation is the monitoring

    of solar UV. Atmospheric parameters which modulate UV need to be
    observed in parallel to help explain the changes observed in the
    measurements. The existing data base is much too scarce to derive
    climatological values and trends on global or regional scales.
    Efforts are being made to establish national and international
    networks for measurements of solar UV radiation.

         Models that simulate typical values and predictions of changes
    in UV radiation for different atmospheric conditions are also used.
    The results from these models need to be compared with measurements.

    3.2  Artificial Sources

         Artificial sources of UV are commonplace. There are few
    artificial sources that result in human exposure to UV greater than
    that from the sun. However, exceptions are those used for medical
    therapy and diagnosis, cosmetic tanning. Industrial sources are
    generally effectively enclosed, but accidental exposure may occur.
    There are very few non-laser sources of optical radiation that emit
    UV solely.

         Any unfiltered optical source, whose emissions are due to the
    heating of a material e.g. a filament lamp, that emits significant
    quantities of UV will also emit visible and infrared radiations. In
    the case of high temperature tungsten halogen lamps biologically
    significant amounts of shorter wavelength UVB are also emitted.
    Essentially the same holds true for high intensity (gaseous)
    discharge (HID) lamps. Some incandescent and HID lamps have
    sufficient intrinsic filtration in the glass envelope of the lamp.
    However, additional filtration, afforded by incorporation of the
    lamp in a suitably filtered luminaire, may be necessary.

         Most man-made sources of UV can be grouped together in the
    categories shown below. The spectrum of the UV emitted varies from
    one source to another

     Incandescent sources
         tungsten lamps

     Gas discharges
         mercury lamps (low-, medium- and high-pressure)
         mercury lamps with metal halides
         xenon lamps
         hydrogen and deuterium lamps
         flash tubes

     Electric discharges
         welding arcs
         carbon arcs

     Fluorescent lamps
         fluorescent lighting tubes
         fluorescent sunlamps (UVB emitters)
         fluorescent UVA tubes

     Lasers
         excimer laser
         dye laser
         gas laser

    3.2.1  Incandescent sources

         When a material is heated a large number of energy transitions
    occur within its molecules and optical photons are emitted. An
    ideally efficient emitter (radiator) is termed a black body
    radiator. The total radiant power and its spectral distribution
    depend only on the temperature of the black body. The spectral
    emissions in terms of spectral radiant exitance of a black body
    radiator for different temperatures are illustrated in figure 3.2
    and tabulated in table 3.4 (McKinlay  et al., 1988). The wavelength
    corresponding to the peak of the spectral emission of a black body
    radiator varies inversely as its temperature and as the temperature
    increases an increasing amount of UV is emitted. Incandescent
    sources whose temperatures are greater than about 2900 K emit
    significant amounts of UV with respect to possible effects on human
    health. The optical emission of the sun corresponds approximately to
    that of a black body radiator at a temperature of around 6000 K.

    FIGURE 3.2

         By integrating over all wavelengths the area under an absolute
    plot of spectral radiant exitance against wavelength, the total
    radiant exitance M can be determined by the formula:

              M = sigmaT4

    where sigma = Stefan-Boltzmann constant (5.67 x 10-3 W m-2 K-4)
    and       T = absolute temperature (K).

    Table 3.4  Approximate radiant exitances of black body sources at
    different temperatures

                                                                      
    Temperature         UVB/C                UVA                UVA
    K              200-315 nm         315-340 nm         340-400 nm
                      [W m-2]            [W m-2]            [W m-2]
                                                                      

    1500           8.6 x 10-5         5.6 x 10-4         2.7 x 10-2
    2000           2.4 x 10-1         7.9 x 10-1               14.2
    2500                   29                 62                634
    3000                  760               1147               8095
    3500                 8050               9220            5 x 104
    4000            4.8 x 104          4.4 x 104            2 x 105
    5000            6.3 x 105            4 x 105          1.4 x 106
    6000            3.6 x 106          1.7 x 106            5 x 106
    10000           1.4 x 108          3.2 x 107          6.9 x 107
                                                                      

         In practice no material emits radiation with a black body
    spectrum. However, tungsten at high temperatures (such as used for
    the filaments of ordinary and tungsten halogen incandescent lamps)
    and molten metals approximate to theoretical black bodies. Such
    sources are termed grey bodies. The ratio of the actual spectral
    radiant exitance to the theoretical black body spectral radiant
    exitance is a measure of how closely the grey body approximates in
    its spectral emission at a particular wavelength to a black body and
    is termed the spectral emissivity (epsilonlamda). Similarly, by
    summation over all wavelengths, the total radiant exitance M of a
    material is given by

              M = epsilon sigma T4

    where epsilon is the emissivity.

         The emissivity depends on the material and on its surface
    structure, and is a function of wavelength. For example, for
    tungsten at 3000 K (typical of a tungsten halogen lamp filament),
    the emissivity varies from an average of about 0.45 in the UVA
    region to about 0.2 in the IRB ( 1.4 - 3.0 µm) region.

         The incandescent lamp is the oldest type of electric lamp still
    in common use. The optical radiation from incandescent lamps results
    from the heating of tungsten filaments which, with the exception of
    lamps for some photographic and graphic arts applications, are not
    heated to temperatures in excess of about 3000 K. Typical gas-filled
    incandescent filament lamps operate between 2700 and 3000 K with
    electrical input powers up to 500 W and the peaks of their spectral
    emissions are in the infra-red (IRA) region. The emissions of UV by
    tungsten filament lamps are generally negligible with respect to
    human health.

         In applications where greater power (up to 5 kW) is required,
    tungsten-halogen (often called quartz-halogen) lamps are often used.
    These are quartz envelope tungsten filament lamps to which a halogen
    vapour (usually iodine) is added. The presence of the vapour enables
    operation at an increased gas pressure compared with conventional
    incandescent lamps, and evaporation from the filament is minimized.
    These features result in improved luminous efficacy and longer life.
    In order to operate efficiently, the temperature of the wall of the
    bulb must be maintained at not less than 260 °C. The majority of the
    bulbs of tungsten-halogen lamps are made from silica (quartz) whose
    thermal properties are most suitable. The combination of filament
    temperatures which are likely to be in the range 2900 to 3450 K and
    quartz bulbs results in a significantly higher level of emission of
    potentially harmful UV compared with ordinary tungsten filament
    lamps. The incorporation of suitable secondary filtration to reduce
    UV emissions to an acceptable level is an important feature of the
    design of any illumination system using tungsten halogen lamps
    (McKinlay  et al., 1989).

    3.2.2  Gaseous discharge sources

         The electrical excitation of a gas or vapour is a typical
    mechanism for generating optical radiation. An electric current is
    passed through a gas or a mixture of gases ionized to produce
    electrons and positive ions. Emissions are the result of electronic
    transitions in the atoms of a material from low to high energy
    states (absorption and excitation) followed by transitions from the
    high to low energy states (de-excitation and emission). This process
    is often combined with the process of luminescence, whereby the
    characteristic (line) photon emissions of the gas are absorbed by a
    luminescent material (luminophor or phosphor) which in turn emits
    optical radiation, typically as a continuum over a range of longer
    wavelengths. The 253.7 nm UV emission from a low pressure mercury
    vapour discharge is used as a source of excitation in low pressure

    fluorescent lamps. By raising the pressure of the discharge to a few
    atmospheres the emission lines increasingly broaden effectively
    forming a continuum. In some cases the 253.7 nm line emission will
    be self-absorbed by the vapour of the discharge.

         Low pressure discharge lamps

          Germicidal lamps

         The low pressure mercury-discharge lamp is often used for the
    purpose of germicide and disinfection. Such lamps are very efficient
    emitters of UV. Approximately 50% of the electrical power is
    converted to UV of which up to 95% is emitted at a wavelength of
    253.7 nm (see figure 3.3). Germicidal type lamps are available in a
    range of sizes, shapes and powers. Small, low-wattage (5-10 W)
    germicidal lamps are often used as fluorescence-inducing lamps for
    the purposes of, for example, chromatographic analysis and the
    fluorescence identification and authentication of documents. The
    quartz envelopes of some lamps in this category transmit 185 nm
    wavelength radiation characteristic of mercury vapour.

          Fluorescent lamps

         The most common application of the low-pressure discharge is
    fluorescent lamps. These operate by means of a discharge between two
    electrodes through a mixture of mercury vapour and a rare gas,
    usually argon. Light is produced by conversion of 253.7 nm mercury
    emission to longer wavelength radiations by means of a phosphor
    coating on the inside of the glass envelope of the lamp. Lamps are
    available with many different phosphors and envelopes to produce a
    wide range of spectral emissions covering the visible (light), UVA
    and UVB regions. While the continuum emissions of fluorescent lamps
    are characteristic of the phosphors the narrow peak, spectral
    emissions are dominated by the characteristic line emission spectrum
    of the low-pressure mercury vapour discharge.

          General lighting fluorescent lamps

         These lamps are available in a range of physical sizes, powers
    and phosphors. The range of phosphors includes a large selection of
    "near white" and "special colour" lamps. In relation to other light
    sources the fluorescent lamp is particularly efficient, with about
    20% of the input energy resulting in useful light. Detailed spectral
    analysis of the UV emissions of different general lighting
    fluorescent lamps have shown that in general UVB and UVC emissions
    are extremely low due to the marked attenuation of wavelengths <
    320 nm afforded by the glass envelope.

    FIGURE 3.3

         Frequently in office and industrial environments where
    fluorescent lamps are used, the luminaire assembly incorporates a
    diffuser or controller. Three materials are commonly used in the
    construction of diffusers; opal acrylic, opal styrene and opal
    polycarbonate. Controllers are luminaire covers that are configured
    with small prisms or lenses. Two commonly used materials are clear
    acrylic and clear styrene. Some luminaires incorporate opal
    (diffusing) sides and a clear figured [controller] base. The use of
    diffusers and controllers results in the absorption and reflection
    of the radiation emitted by the associated lamp. The UV-attenuating
    properties of different diffusers is demonstrated by the measurement
    data in table 3.5.

         During the past few years the further development and improved
    design of general lighting fluorescent lamps have been evident in
    the production of compact fluorescent lamps. These lamps are
    essentially low power small diameter fluorescent tubes folded in a
    compact form. They are most readily available commercially with cool
    white phosphors but other phosphors are also available. For a given
    illuminance their spectral emissions of UV are essentially no
    different from those of full size tubular fluorescent lamps as shown
    in table 3.6.
        Table 3.5  Measurements of UV irradiance from various diffusers/controllers
    with a white fluorescent lamp as the source. Irradiance totals in each
    waveband in mW m-2 (percentage in parentheses), (McKinlay et al., 1988).
                                                                               
    Diffuser Type          UVA            UVB                 UVACGIH**
                           mW m-2         mW m-2              [mW m-2]effective
                                                                               

    Bare lamp              22.32 (100d)   3.45 (100)          59 x 10-3 (100)
    Clear acrylic+         16.35 (73)     2.91 (84)           48 x 10-3 (81)
    Clear styrene+         2.87 (13)      0 (0)               0 (0)
    Opal styrene*          0.92 (4)       3 x 10-3 (< 0.1)    0.02 x 10-3 (<0.1)
    Opal polycarbonate*    0.20 (< 1)     12 x 10-3 (< 1)     0.09 x 10-3 (<1)
                                                                               

     + Surface figured with small prisms
     * Reeded surface
    ** ACGIH occupational hazard weighted irradiance: 1 maximum permissible
    exposure for an 8 h working day is equivalent to 10-3 [W m-2]effective ..
    
    Table 3.6  Measurements of UV from compact fluorescent lamps,
    normalised to an illuminance of 500 lux (Whillock et al., 1990)
                                                                      
    Lamp type                          UVA            UVB
                                       mW m-2         µW m-2
                                                                      

    Luma (7 W) LC7                     47 x 103       0
    Luma (7 W) LC7 with diffuser       197            0
    Osram (11 W) Dulux EL              38 x 103       0.1
    Philips (9 W) SL9                  37 x 103       0
    Sylvania (13 W) Lynx CFD           43 x 103       30.81
    Thorn (16 W) 2D                    54 x 103       2.48
    Tungsram (16 W) Globulux           663            0
                                                                     

          "Special" applications fluorescent lamps

         Apart from a number of colour-rendering fluorescent lamps,
    which are essentially variations of general lighting fluorescent
    lamps, a number of special applications fluorescent lamps have been
    developed.

         A common example of a UVB emitting fluorescent lamp is the FS
    type lamp with spectral emission shown in figure 3.4. Such lamps
    were previously used for cosmetic tanning and are now often used as
    a source of UVB in biological experiments. The blacklight lamp uses
    a nickel/cobalt-oxide (Woods glass) envelope that is almost entirely
    opaque to light. The phosphor chosen for this type of lamp emits
    around 370 nm in the UVA. Such lamps are used for a number of
    commercial, scientific and industrial fluorescence purposes as well
    as for display and entertainment. Three types of fluorescent lamps
    that emit UVA in printing and copying merit mention viz; lamps
    suitable for diazo printing with a principal emission around 360 nm;
    those intended for modern fast diazo printing with a main emission
    around 420 nm; and those used for photocopying with predominantly
    green emissions.

         A medical treatment for hyperbilirubinaemia in neonates
    (neonatal jaundice) consists of irradiating the newborn child with
    phototherapy lamps emitting in the wavelength range approximately
    400 to 470 nm. Some lamps used for this purpose have emission
    spectra that extend into the UV region. Fluorescent lamps in
    printing and copying include lamps for diazo printing with a
    principal emission around 360 nm.

         The development of a range of phosphors with enhanced UVA
    emissions has led to the widespread use of fluorescent lamps in
    sunbeds, solaria and PUVA (Psoralen + UVA) treatment cabinets.

         High pressure discharge lamps

         The designation "high pressure discharge" lamps is taken in
    this monograph to include the families of lamps often called high
    intensity discharge (HID), such as mercury vapour or metal halide
    lamps, where an electric arc is not created. At still higher
    pressures, arcs may be produced, e.g., the xenon or mercury compact
    (short-arc) lamps.

          Mercury and metal halide lamps

         High-pressure mercury vapour lamps are widely used for lighting
    in commerce, streets, displays, floodlighting and a large number of
    printing, curing and other industrial applications. The spectral
    emissions of the discharge are in the blue, green and yellow regions
    of the spectrum and a large amount of UV is also generated. The
    general construction of high-pressure mercury lamps is a fused
    silica (quartz) discharge tube containing the mercury/argon vapour
    discharge mounted inside an outer envelope of soda-lime or
    borosilicate glass.

         The outer glass envelope effectively absorbs most residual UV.
    Consequently the quantity of potentially harmful UV emitted by such
    lamps depends critically on the integrity of this envelope. In the
    USA, but apparently not in Europe, it is a legally enforceable
    manufacturing requirement that breakage of the outer envelope must
    either cause the lamp to fail to operate, in which case the lamps
    are described as "self-extinguishing" and are marked with the letter
    "T", or if not "self-extinguishing" they should be marked with the
    letter "R". In the latter case a warning notice must be included
    with the packaging of the lamp (FDA., 1988). Data from measurements
    made at 2 m from a mercury HID lamp with the outer envelope removed
    illustrate the importance of this aspect of safety design, as shown
    in table 3.7.

    FIGURE 3.4


    
    Table 3.7  UV emissions from HID mercury vapour general lighting (USA) lamps; effective irradiance
    in mW m-2effective, (ACGIH), (Piltingsrud et al., 1978).
                                                                                                                                        
    Test condition                      A                           B                           C                           D

    Lamp type                With outer   Without        With outer   Without        With outer   Without        With outer   Without
                             bulb         outer bulb     bulb         outer bulb     bulb         outer bulb     bulb         outer bulb
                                                                                                                                        

    General Electric         2.5          110*           < 0.1        160            12.5         3640           0.2          0.2
    H400 A33-1
    Clear

    Westinghouse             < 0.1        1040           < 0.1        190            0.2          3680           < 0.1        0.3
    H33 GL 400/DX
    White

    General Electric         < 0.2        75*            < 0.1        180            0.5          3900*          < 0.1        < 0.1*
    H400D x 33-1
    White

    General Electric         < 0.3        122*           < 0.1        6*             1            510*           < 0.1        < 0.1*
    MV400/BUH
    [Metal halide]
                                                                                                                                        

    Test Conditions:
    A: Lamp mounted vertically - measurements at 2 m on mid-line axis of lamp.
    B: Lamp mounted horizontally - measurements at 2 m on central axis of lamp.
    C: Lamp mounted horizontally in reflector shield with no face plate - measurements at 2 m on central axis of lamp.
    D: Lamp mounted horizontally in reflector shield with glass face fitted - measurements 
       at 2 m on central axis of lamp.

    * Lamps did not operate at normal intensity.


    
         The family of metal halide lamps encompasses a number of
    different types of high pressure mercury lamps whose discharges all
    contain additives. The additives are most typically metal halides
    chosen to produce either a strongly coloured emission (usually a
    single halide), to produce a more broadly spectrally uniform
    emission (multi-halide) or to enhance the UV (most often UVA)
    emission. Compared with ordinary high pressure mercury lamps the
    luminous efficacies of metal halide lamps are high. They are used
    for a range of industrial and commercial applications that include
    photochemical processing, graphic and photographic illumination,
    studio lighting, reprography and are also used for UVA cosmetic
    tanning equipment, for some medical applications and for solar
    radiation simulation. The UV irradiances of some metal halide lamps
    used in filtered industrial applications requiring an activating
    range of wavelengths between 320 and 440 nm, e.g. lithographic
    platemaking and printed circuit photo-resist etching, are shown in
    table 3.8 (McKinlay  et al., 1988).

    Table 3.8  UV irradiances measured at 1 m from typical graphics arts
    metal halide mercury lamps (McKinlay et al., 1988)
                                                                      
    Lamp type      Power          UVC            UVB            UVA
                   [W]            [W m-2]        [W m-2]        [W m-2]
                                                                      

    HPA 400        400            0.5            3.2            9.0
    HPA 1000       930            2.3            9.0            23.0
    HPA 2000       1750           4.5            19.0           48.0
                                                                      

         The emission of a device incorporating iron additive halide
    lamps for phototherapy is illustrated in figure 3.5. The importance
    of incorporating filtration in such devices in order to remove
    unwanted components of the UV spectrum is illustrated.

    FIGURE 3.5

         Xenon, compact and linear arcs

         Where an optical source of very high radiance is required and
    of small size a very high pressure arc lamp may be used. These have
    a filling gas of mercury vapour, mercury vapour plus xenon gas or
    xenon gas. Metal halide types are also available. Two physical types
    are commonly used; the compact (short) arc and the linear arc. The
    spectral emission of xenon lamps, which at wavelengths shorter than
    infra-red, closely matches that of a black-body radiator at about
    6000 K. This enables their use in photography and as solar radiation
    simulators.

         The spectral emission of xenon lamps, which at wavelengths
    shorter than infrared closely matches that of a black body radiator
    at about 6000 K, enables their use as solar radiation simulators.
    Their emission spectrum is continuous from the UV through to the IR
    regions. Large amounts of UVA, UVB and UVC are emitted by unfiltered
    lamps to the extent that they can present a significant health
    hazard if incorrectly used. The luminance of compact xenon arcs may
    approach that of the sun and in some lamps with greater than 10 kW
    rating the luminance may exceed that of the sun. They therefore
    present a potentially severe retinal hazard if viewed.

    3.2.3  Gas welding

         Oil, coal and gas flames normally operate at temperatures below
    about 2000 K and consequently emit virtually no UV. Oxyacetylene and
    oxyhydrogen flames burn at much higher temperatures and emit UV
    mostly in the UVA region.

    3.2.4  Arc welding

         By comparison with gas flame processes, the emissions of UV
    from arc welding are very high (see figure 3.6) and many data on the
    optical radiation emissions associated with a variety of electric
    arc welding processes have been published (e.g., Sliney & Wolbarsht,
    1980).

    3.2.5  Lasers

         All lasers have three basic components (see figure 3.7): (a) a
    laser (active) medium; (b) an energy source (pumping system) and;
    (c) a resonant optical cavity. A pumping system is necessary to
    provide energy to electrons to raise them to excited states and
    achieve population inversion. Optical pumping, using an intense
    source of light such as a xenon flashtube; electron-collision
    pumping, using an electrical discharge and; chemical pumping, using
    the energy released from making and breaking chemical bonds, are all
    used for this purpose. A resonant optical cavity is formed by
    mirrors placed at each end of the laser medium. The construction is
    such that the beam passes through the laser medium several times and

    the number of emitted photons is amplified during each transit. One
    of the mirrors is chosen to be partially transmitting thus enabling
    part of the beam to be emitted from the cavity (figure 3.7).
    Examples of lasers that emit UV are presented in table 3.9.

    Table 3.9  UV laser emissions and characteristics,
               (McKinlay, 1992)
                                                                      
    Type        Name                             Spectral emissions
                                                                      

    Excimer     Argon fluoride [(ArF)            193 nm
                Krypton fluoride (KrF)           248 nm
                Xenon chloride (XeCl)            308 nm
                Xenon bromide (XeBr)             282 nm
                Xenon fluoride (XeF)             351 nm
                Krypton chloride (KrCl)          222 nm

    Dye         Excimer-, Nitrogen-,             300 nm-
                Flash lamp pumped                340 nm-

    Gas         Nitrogen (N2)                    337 nm
                Helium cadmium (He-Cd)           325 nm

    Gas ion     Argon (Ar+0                      333 nm-
                                                 357 nm
                                                 363 nm
                Krypton (Kr+)                    337.5 nm
                                                 350.7 nm
                                                 356.4 nm
                                                                      

         A recent generation of UV lasers are the excimer lasers. The
    active media used are rare gas halides such as ArF, KrCl, XeBr or
    XeF. The molecules of those gases have only a short lifetime in the
    ground state, however they are very stable in the excited state. As
    a buffer medium, helium is often used. Population inversion is
    easily achievable because the ground state is very unstable.

    3.2.6  Sunbeds

         Sunbeds are broadly used for cosmetic tanning purposes. The
    expression "sunbed" includes tanning equipment consisting of a UV
    emitting lamp or a number of such lamps incorporated in a bed,
    canopy or panel, or any combination thereof. There are four distinct
    types of lamps in use, each with different UV emission
    characteristics. Those are UVA, low-pressure fluorescent tubes; UVA,
    filtered high-intensity discharge lamps; UVB, low-pressure
    fluorescent tubes; UVB, filtered high-intensity discharge lamps.

    FIGURE 3.6

    FIGURE 3.7

         The emission characteristics and the health risks associated
    with the use of each type of lamp are different. The last two lamp
    types are associated with high levels of UVB and are now little
    used. They have been almost universally replaced by the
    predominantly UVA emitting lamp types.

    4.  HUMAN EXPOSURE

    4.1  Sunlight

         Outdoors, exposure to UV constantly changes during the day.
    People are largely unaware of the degree of these changes. At noon,
    when the sun is overhead, the level of UV at a wavelength of 300 nm
    is ten times greater than at either three hours before (9 am) or
    three hours after noon (3 pm). An untanned person with fair skin may
    receive a mild sunburn in as little as 25 minutes at noon (depending
    on the time of year and the latitude) but would have to lie in the
    sun for at least two hours to receive the same dose after 3 pm. The
    global biologically effective UV falling on a horizontal surface
    occurs primarily during the midday hours, about 50% during the four
    hours centred on noon-time zenith (Sliney, 1987).

         Scattering of sunlight by air molecules (due to Raleigh
    scattering) favours UV and blue light (hence the blue sky). For
    longer pathlengths through the atmosphere when the sun is low in the
    sky, more UV and sunlight is scattered. The sun which is white at
    noonday becomes yellow and then orange as less UV and blue light are
    present in the direct rays. When the sun is overhead staring at the
    sun for 90 seconds would cause solar retinitis. A few hours later,
    with the sun much lower in the sky, it would take many minutes to
    reach a hazardous retinal dose, and it is virtually impossible to
    cause any eye damage at sunset. Thus, the geometry of exposure as
    well as the spectrum (hue) plays a major role in determining the
    hazards from direct exposure from the sun (Sliney, 1983, 1986).

         Estimation of an individual's lifetime UV exposure requires
    knowledge of the ambient solar UV levels, history of outdoor
    exposure and the relative exposures at the different anatomical
    sites. Studies on the anatomical distribution of solar UV have been
    reported (Diffey et al 1979; Rosenthal  et al., 1985; Holman  et
     al., 1983; Gies et al 1992a; Roy  et al., 1988). The relative
    doses at various body sites have been determined using UV sensitive
    polysulphone film on rotating manikins and headforms. It was found
    that even though the relative doses to the face and eyes are higher
    in winter, due to the lower solar elevation, the absolute doses are
    higher in summer. The presence of a brimmed hat reduced the face
    exposure by a factor of at least two and the eye exposure was
    reduced by a factor of 4 to 5 (Diffey  et al., 1979, Roy et al
    1988).

         Polysulphone film badges have also been used to quantify the
    solar UV exposure received by different subjects and results
    compared to those calculated from personal diaries and measured
    ambient solar UVB (Gies  et al., 1992a). In general, when UV
    exposure activities took place under close supervision, good
    correlations were obtained between the polysulphone badges and the
    ambient/diaries approach. Results from a recent study (Roy and Gies,

    1993), of indoor, outdoor and retired workers indicated that
    exposures to badge locations of up to 30% of ambient were recorded.
    Through studies of this type, knowledge is gained on the amount and
    pattern of exposure from routine activities and this can then be
    applied in the design of educational campaigns to modify outdoor
    behaviour and reduce UV exposure. The use of polysulphone or CR39
    plastics (Wong  et al., 1989; Sydenham  et al., 1991) as
    contact-lens dosimeters have been proposed, but lack of sensitivity
    and personal comfort has discouraged their use in field studies.

    4.2  Skin Exposure Geometry

         UV incident on human skin can follow one of three courses, it
    can undergo absorption, reflection, or scattering. Thus, the actual
    radiant exposure received by the various layers of the skin will be
    lower than the incident exposure. Reflection not only occurs at the
    surface of the stratum corneum, but at all interfaces changing in
    refractive index. Scattering occurs because of the different
    structural elements, such as hair follicles and sebaceous glands,
    and also by cellular components, such as mitochondria and ribosomes.
    The remaining UV can penetrate into deeper skin layers.

         UV penetrates into the dermis exposing a variety of cells and
    structures, depending in part on the thickness of the human stratum
    corneum and epidermis. The depth of penetration is wavelength
    dependent the longer the wavelength the deeper the penetration
    (Bruls  et al., 1984). From figure 4.1 it is seen that the same
    incident exposure of UVA or UVB radiation will result in a higher
    actual exposure of UVA than UVB at a given depth. For example, if
    the incident UV exposure was 100 kJ m-2 then 50 kJ m-2 of 365 nm
    radiation would be present at a depth of 30 µm and only 19 kJ m-2
    at a depth of 70 µm. For 313 nm radiation only 33 kJ m-2 would be
    present at 30 µm and 9.5 kJ m-2 at 70 µm. While less than 1% of
    the UVC wavelengths can barely penetrate the epidermal layer (Bruls
     et al., 1984), about 1% of the incident UVA dose can penetrate
    into the subcutaneous tissue (Parrish  et al., 1978). The
    distribution and size of melanin particles also plays an important
    role in protecting epidermal cells. Melanin particles have a
    distribution within the stratum corneum and epidermal cells
    depending upon skin type. In dark skin types (5 and 6) these
    particles are positioned within cells to provide optimum optical
    protection for the cell nuclei and in adequate size in the stratum
    corneum (Kollias  et al., 1991).

    FIGURE 4.1

    4.3  Ocular Exposure Geometry

         People seldom look directly at the sun when it is overhead and
    very hazardous to view. It is not very hazardous to view when the
    sun is low in the sky and falls within the normal field-of-view.
    When the sun is more than 10° above the horizon, the natural
    tendency is to partially close the eyelids or squint (called squint
    reflex), thus shielding the retina from direct exposure. These
    factors reduce the exposure to the cornea to a maximum of about 5%
    of that falling upon the exposed top of head (Sliney, 1986). If the
    squint and other behavioural factors are not considered, the dose to
    the eyelid would be approximately 20% of the dose falling on a
    horizontal surface.

         Although the cornea is more sensitive to UV injury than the
    skin, people seldom experiences a corneal burn when in sunlight.
    Using the action spectrum for human photokeratitis and
    mathematically weighting this with the midday solar spectrum, the
    time to achieve the threshold for photokeratitis is about 100
    seconds (Sliney, 1987, Rosenthal  et al., 1988). Again, the
    geometry of exposure precludes photokeratitis except when ground
    reflectance exceeds approximately 10%. When the sun is overhead and
    UV exposure is most severe, the brow ridge and upper lid shield the
    cornea, and if the eye is turned away from the sun, the more intense
    scattered UV from overhead strikes the cornea at a grazing angle of
    incidence where most is reflected and little is absorbed (Sliney,
    1983). Only when the incident UV rays are parallel to the pupillary
    axis are most rays (approx. 98%) absorbed.

         When looking at snow, UV is reflected directly into the eye;
    hence, the traditional eye protector of the Inuit or Eskimo, the
    slit, in whalebone or in a seal-skin mask, provided geometrical
    rather than spectral protection against UV exposure (Hedblom, 1961).
    The lack of protection above and to the sides of sunglasses is a
    serious shortcoming. However, to obtain a quantitative idea of this
    component of exposure to the eye, measurements were made using a
    simulated ocular geometry in sunlight (Sliney, 1986). The human eye
    received 10 to 25 % of the UV dose when wearing glasses with lenses
    opaque to UV compared to no lens in the glasses. Therefore, unless
    goggles with side-shields are used, UV transmission factors in
    lenses much less than 2-5 % do not provide the eye protection
    suggested by the transmission factor (Sliney, 1986).

         The strong dependence of reflectance with angle of incidence is
    termed Fresnel's Law of Reflection. This law not only explains the
    survival of the cornea exposed overhead to UV, but also the glare
    experienced over water. When the sun is overhead, water reflects
    about 2% of the UV upward (Sliney, 1986). When low in the sky, much
    of the sunlight is reflected while the UV and blue light are
    filtered by the atmosphere. Nevertheless, the strong reflections

    from water at these low sun angles create discomfort glare and UV
    exposure of the cornea is further reduced because of the squint.

         If dark lenses are placed over the eyes, the natural aversion
    to bright light, which leads to the squint reflex (that greatly
    lowers retinal UV or exposure to the eye), would be disabled. This
    may appear to be an unusual way to consider the comfort that shaded
    lenses provide. However, poor sunglasses may actually lead to a
    higher UV exposure (Sliney, 1983).

         However, quantifying the protective value of the upper and
    lower eye lids when they close to squint is difficult. In terms of
    UV exposure, at least a twenty-fold reduction is likely. For
    shielding the retina from the direct image of the sun, the upper lid
    probably provides a protection factor exceeding a thousand (Sliney -
    personal communication). If a brimmed hat is worn, the direct image
    of the sun on the retina is rare and overhead UV exposure is
    virtually eliminated. However, while using a hat the lid opens
    further and ground reflection of UV could become important.

         On an overcast day, the lids open wider, and although the UVB
    irradiance is reduced by cloud cover, the actual UVB dose rate to
    the eye from atmospheric scattering near the horizon may be reduced
    by a factor of only two (Sliney, 1983). Hence, on a cloudy day the
    eye may receive a greater UVB dose than on a bright sunny day.
    However, a heavy overcast may attenuate the UVB sufficiently, that
    this observation may not be true. As sunglasses are not typically
    worn on an overcast day, one could argue that the concern over
    sunglasses increasing total ocular exposure is unimportant. However,
    sunglasses should have sufficient UV filtration so that ocular
    exposure does not actually increase when they are worn on a sunny
    day.

         Eye and head movements can further reduce UV exposure. Most
    humans in bright sunlight squint or avoid looking into the sun
    sector of the sky. These behavioural and physiological factors are
    not taken into account by simple UV measurements. There is an
    obvious need to determine accurately the corneal and lenticular
    exposure to ambient UV. Indeed, the results of previous
    epidemiological studies of cataract may be questioned because of
    inadequate dosimetry. Some epidemiological studies may have reached
    incorrect conclusions regarding risk from UV or sunlight exposure by
    assigning inaccurate exposure levels to population groups, because
    they assumed that overhead UV exposure accurately predicts corneal
    exposure.

    4.4  Workplace

    4.4.1  Outdoor work

         Although humans have adapted and acclimatized to solar UV it
    nevertheless represents the most hazardous source of optical
    radiation likely to be encountered by the average person and with a
    few exceptions by the worker. People who work outdoors will be
    subjected to involuntary UV exposure. The highest exposure will
    occur during the two hours period either side of noon. Workers must
    be made aware of this and take appropriate precautions as discussed
    in chapter 13.

         In a study of various work and recreational situations
    (Challoner  et al., 1976; Diffey  et al., 1982), it was found that
    outdoor workers had the highest exposure, receiving approximately
    10% of the ambient level, similar to that received during sailing
    and sightseeing. Higher exposures were found for skiing (20%) and
    the largest during sunbathing on a beach (80%). Office workers
    received about 3% of the total ambient radiation with about half
    that figure from weekend exposure (Leach  et al., 1978). The actual
    exposures vary depending on the time of day and year, duration and
    frequency of exposure.

    4.4.2  Indoor work

         Examples of indoor workplace exposures to UV are given below.

          Photoprocesses

         Many industrial processes involve a photochemical component.
    The large-scale nature of these processes often necessitates the use
    of high-power (several kilowatts) lamps such as high pressure metal
    halide lamps which emit significant amounts of UV (Diffey, 1990a).
    The principal industrial applications of photopolymerisation include
    the curing of protective coatings and inks and photoresists for
    printed circuit boards. The curing of printing inks by exposure to
    UV takes only a fraction of a second. UV drying units can be
    installed between printing stations on a multicolour line, so that
    each colour is dried before the next is applied. Another major use
    of UV curing has been for metal decorating in the packaging industry
    (Phillips, 1983). UVA is also used to inspect printed circuit boards
    and integrated circuits in the electronics industry.

          Sterilization

         UV with wavelengths in the range 260-265 nm is the most
    effective for sterilization and disinfection since it corresponds to
    a maximum in the DNA absorption spectrum. Low-pressure mercury
    discharge tubes are often used as the UV source as more than 90% of
    the radiated energy lies in the 254 nm line, figure 3.2. These lamps

    are often referred to as "germicidal lamps", "bactericidal lamps" or
    simply "UVC lamps" (Diffey, 1990a).

         UVC radiation has been used to disinfect sewage effluent,
    drinking-water, water for the cosmetics industry and swimming pools.
    Germicidal lamps are sometimes used inside microbiological safety
    cabinets to inactivate airborne and surface microorganisms (Diffey,
    1990a). The combination of UV and ozone has a very powerful
    oxidizing action and can reduce the organic content of water to
    extremely low levels (Phillips, 1983).

          Welding

         Welding equipment falls into two broad categories: gas welding
    and electric arc welding. Only the latter process produces
    significant levels of UV, the quality and quantity of which depend
    primarily on the arc current, shielding gas and metals being welded
    (Sliney & Wolbarsht, 1980). The levels of UV irradiance around
    electric arc welding equipment are high; effective irradiance at 1 m
    at an arc current of 400 A ranged from 1 to 50 W m-2 and the
    unweighted UVA irradiance ranged from 3 to 70 W m-2, depending on
    the type of welding and the metal being welded (Mariutti & Matzeu,
    1987; Cox, 1987). It is not surprising therefore that most welders
    at some time or another experience "arc eye" or "welder's flash"
    (photokeratitis) and skin erythema. The effective irradiance at 0.3
    m from many types of electric welding arcs operating at 150 A is
    such that the maximum permissible exposure time for an 8-h working
    period on unprotected eyes and skin varies from a few tenths of a
    second to about 10 s, depending on the type of welding process and
    the material used (Cox, 1987).

    4.4.3  Research

         Sources of UV are used by most experimental scientists engaged
    in photobiology and photochemistry and in molecular biology. These
    applications, in which the effect of UV irradiation on biological
    and chemical species is of primary interest to the researcher, can
    be differentiated from UV fluorescence by absorption techniques
    where the effect is of secondary importance (Diffey, 1990a).

    4.4.4  Commerce

         Sunlamps emitting UV have been used for tanning, particularly
    in northern Europe and North America. Prior to the mid-1970s, the
    source of UV in sunlamps was usually an unfiltered mercury arc lamp
    which emitted a broad spectrum of radiation, including large
    quantities of UVB and UVC. There are four distinctly different types
    of ultraviolet tanning lamps in use, each with different UV emission
    characteristics viz.; low pressure UVA fluorescent lamps; filtered
    high intensity discharge UVA lamps; low pressure UVB fluorescent
    lamps and; filtered high intensity discharge UVB lamps (IRPA/INIRC

    1991). The emission characteristics and the health risks associated
    with each type of lamp are different (Diffey and McKinlay 1983, Gies
    et al 1986, Diffey 1987). Tanning systems incorporating lamps that
    emit predominantly UVB are now little used and have been almost
    universally replaced by the low pressure UVA fluorescent and
    filtered high intensity discharge UVA types. Tanning requires
    deliberate exposure of the skin to UV; however, the eyes must be
    protected. Similarly, staff working in tanning salons must ensure
    their exposure is kept to a minimum.

         Many contaminants of food products can be detected by UV
    fluorescence techniques. For example, the bacterium Pseudomonas
    aeruginosa, which causes rot in eggs, meat and fish, can be detected
    by its yellow-green fluorescence under UVA irradiation. One of the
    longest established uses of UVA fluorescence in public health is to
    demonstrate contamination with rodent urine, which is highly
    fluorescent.

         Many flying insects are attracted by UVA radiation,
    particularly in the region around 350 nm. This phenomenon is the
    principle of electronic insect traps, in which a UVA fluorescent
    lamp is mounted in a unit containing a high-voltage grid. The
    insect, attracted by the UVA lamp, flies into the unit and is
    electrocuted in the air gap between the high-voltage grid and a
    grounded metal screen. Such units are commonly found in areas where
    food is prepared and sold to the public (Diffey, 1990a).

         UVA blacklight lamps are sometimes used in discotheques to
    induce fluorescence in the skin and clothing of dancers. The levels
    of UVA emitted are usually low (< 10 W m-2) (Diffey, 1990a).

         Signatures, banknotes and other documents can be authenticated
    by exposing them to UVA, under which they fluoresce. UVA exposure of
    the user is normally to hands and irradiance is low (< 10 W m-2)
    (Diffey, 1990a).

    4.4.5  Medicine and dentistry

         In both medical and dental applications of UV, the patient is
    deliberately exposed to either treat or diagnose a disease or
    disorder. While non-target areas of the patient must be covered,
    care must also be taken to ensure staff are fully protected from UV
    exposure (see chapter 13).

         The diagnostic uses of UV are confined largely to fluorescing
    of the skin and teeth. UV exposure is limited to small areas (< 15
    cm in diameter) and the UVA radiation dose per examination is
    probably no more than 5 x 104J m-2. Diagnostic techniques are
    limited to the use of fluorescence to identify the presence of
    various fungal and bacterial infectious agents on the skin or in
    wounds. The source of UV most commonly used for this purpose is the

    Woods lamp which emits predominantly UVA radiation (365 nm).
    Exposure of the patient's skin to potentially
    carcinogenically-effective UV is insignificant (Diffey, 1990a).

         UV phototherapy is a well established method for the treatment
    of a number of skin conditions and in particular for psoriasis
    (Green  et al., 1992). Two general types of UVB emitting lamps are
    used in phototherapy, high intensity discharge (HID) mercury vapour
    and mercury/metal halide vapour lamps and low pressure mercury
    vapour fluorescent lamps. The spectral emissions of these different
    types of lamps vary greatly and this, with other factors such as the
    number and power of the lamps, the lamp to skin distance and the
    individual sensitivity of the patient, determines the treatment
    time. A treatment dose is chosen to cause erythema in a particular
    patient. As treatment progresses and the skin becomes increasingly
    acclimatised, individual treatment doses are increased accordingly.
    Treatments may be given several times per week. Phototherapy has
    also been used for the treatment of other skin conditions including
    severe itching, acne, eczema, polymorphic light eruption (PLE),
    pityriasis rosea and urticaria, and for renal failure (Green  et
     al., 1992).

         Seasonal affective disorder (SAD) is frequently treated by
    exposure to sources of high illumination. Most treatment regimes
    have employed sources incorporating so-called "full spectrum
    lighting" whose emissions contain small amounts of UVA and UVB
    radiations (Terman  et al., 1990).

         PUVA

         UV photochemotherapy is UV phototherapy used in conjunction
    with the oral or topical application of a chemical
    (photosensitising) agent to the patient. The most widely used
    treatment is PUVA which involves the use of the photosensitising
    agent 8-methoxypsoralen (8-MOP) in conjunction with UVA irradiation
    for the treatment of psoriasis. The irradiation sources used for
    such treatment incorporate either low pressure mercury vapour UVA
    fluorescent lamps, or HID mercury vapour or mercury/metal halide
    vapour lamps with an added filter to effectively attenuate the UVB.
    The UVA irradiance on the skin from sources incorporating
    fluorescent lamps is generally of the order of 60 W m-2 and for
    the HID mercury systems, around 250 W m-2. Treatments consist of a
    starting dose of between 500 and 40,000 J m-2, depending on skin
    type, followed by further incremental doses two to three times
    weekly. Total treatment radiant exposures of between 106 and 2.5
    106 J m-2 have been reported for successful results. However,
    studies have shown large uncertainties in the dosimetry associated
    with PUVA treatment and point to equally large consequential
    uncertainties about the actual treatment doses delivered (Diffey et
    al 1980).

          Hyperbilirubinaemia

         Phototherapy is sometimes used in the treatment of neonatal
    jaundice or hyperbilirubinaemia. The preferred method of treatment
    is to irradiate the baby for several hours a day for up to one week
    with visible light, particularly blue light. The lamps used for
    phototherapy, although intended to emit only visible light, may also
    have a UV component (Gies & Roy, 1990; Sliney & Wolbarsht, 1980).

          Dentistry

         Irradiation of the oral cavity with a Woods lamp can produce
    fluorescence under certain conditions. This has been used in the
    diagnosis of various dental disorders, such as early dental
    cavities, the incorporation of tetracycline into bone and teeth,
    dental plaque and calculus (Hefferren  et al., 1971).

         Pits and fissures in teeth have been treated using an adhesive
    resin polymerized with UVA. The resin is applied with a fine brush
    to the surfaces to be treated and hardened by exposure to UVA
    radiation at a minimal irradiance of 100 W m-2 for about 30 s
    (Eriksen, 1987; Diffey, 1990a).

    4.5  Elective Exposure

         By comparison to occupational exposure to UV, where control
    measures are generally instituted to protect the eye and skin,
    elective exposures can be much greater. Elective exposure results
    from outdoor recreational activity, from intentional exposure to
    sun-tanning equipment and to sunlight at the beach and elsewhere.
    Actual dose estimates of elective exposure experienced by those who
    attempt to maintain a tan may exceed 100 MED per year (Challoner  et
     al., 1976; Diffey, 1993a; Diffey  et al., 1982).

    5.  DOSIMETRIC CONCEPTS

    5.1  Interaction of UV with Matter

         Attenuation of UV occurs due to absorption and scattering.
    Reflection, refraction and diffraction are phenomena related to
    boundaries between media. All of these interactions may change the
    direction, intensity and the wavelength of UV. The quantities
    describing interactions of UV with matter are dependent on the
    wavelength.

         The total energy absorbed in a material is influenced by
    reflections from its surfaces. The attenuation coefficient describes
    the attenuation of UV within tissue and is the sum of the absorption
    and scattering coefficients. Penetration depth in a tissue is
    inversely proportional to the exponential attenuation coefficient.

         Absorption requires transfer of radiative energy to matter.
    Apart from possible photomechanical effects, UV is absorbed as a
    result of electron transitions at the atomic and molecular level.
    Such molecular absorption can lead to photochemical reactions.

         In most molecules the ground state or the singlet state (S0)
    consists of two paired electrons. On absorption of radiant energy
    one of the electrons can make a transition to an excited state
    (S1), provided that the incident photon energy corresponds to an
    existing level in the absorbing molecule. As shown in figure 5.1,
    the molecule can release this absorbed energy from the excited state
    by (a) transition directly back to the ground state, S0, or (b)
    transition to the generally long-lived excited triplet state, T1,
    and then discharging the remaining energy to return to S0. In
    either case, transition back to S0 may be radiative (a photon
    emitted) or non-radiative (energy dispersed by vibrational-
    rotational relaxation). The radiative S1 -> S0 transition is
    called "fluorescence" while radiative T1 ->S0 is called
    "phosphorescence". In addition, molecules in the S1 or T1 states
    can return to the ground state either by forming photoproducts or by
    transferring the energy to an acceptor molecule. The effectiveness
    of the photochemical process can be amplified by "photosensitizers"
    or impeded by "quenchers".

    FIGURE 5.1

    5.2  Biological Weighting Factors and Spectrally Weighted Quantities

         Both radiant exposure H and irradiance E are quantities
    integrated over the total spectrum of interest. To describe the
    irradiance or radiant exposure in a very narrow spectral
    (wavelength) interval (Delta lambda), the quantities of  spectral
     irradiance Elambda and  spectral radiant exposure Hlambda are
    employed. These have the units of: W m-2 nm and J m-2 nm,
    respectively. If ElambdaDelta lambda is the irradiance in a narrow
    interval Delta lambda around the wavelength lambda, then the
    integrated irradiance E over the wavelength interval lamda1 to
    lambda2 can be written as:

                   lambda2
               E = Sigma Elambda Delta lambda
                   lambda1

         Elambda and Hlambda are spectral functions. Other
    spectroradiometric quantities exist with analogous definitions,
    e.g., spectral radiant power, spectral radiant energy, etc, as shown
    in table 2.2. The spectroradiometric quantities are important in
    photobiology, and are critical in any discussion of the biological
    effects of UV. The biological effects of UV are strongly wavelength
    dependent. As a measure of these effects, "effective" or
    "biologically active" quantities have been introduced (CIE, 1987).
    The effective irradiance Eeff is defined as:

              Eeff = Sigma Elambda Slambda Delta lambda

    and similarly, the effective radiant exposure is

              Heff = Sigma Hlambda Slambda Delta lambda

    where Slambda is called the relative spectral effectiveness
    function or  action spectrum.

         The  action spectrum gives the relative biological response of
    a tissue to irradiation at different wavelengths, and ideally will
    correspond to the absorption spectrum of critical absorbing
    molecules, or "chromophores." In biological systems, however, the
    action spectrum function is modulated by the shielding (i.e.
    thickness) of overlying tissue (e.g., stratum corneum of the skin),
    energy transfer, and the action of sensitizers and quenchers. The
    action spectrum is therefore specific for a certain effect arising
    in a certain tissue layer. For example, the action spectrum for
    erythema (skin) and photokeratitis (cornea) differ.

         Based on a statistical analysis of the results of minimum
    erythema dose studies carried out over the past 20 years or so, and
    including data by Parrish  et al. (1982) on the erythemal efficacy
    of UVA, the CIE has promulgated a reference action curve (McKinlay

    and Diffey, 1987), as shown in figure 5.2. This function consists of
    three straight lines when plotted on a semi-logarithmic scale, and
    although individual action spectra would not have the two inflection
    points, the function can be readily expressed by three mathematical
    functions. The function has been adopted internationally by the CIE
    and the IEC and is being used by national organizations and
    authorities for the determination of the erythemal potential of an
    exposure to a given source of UV. A slightly different action
    spectrum, which cannot be so readily expressed mathematically, has
    been recommended by IRPA (1991) for risk assessment in occupational
    health (see figure 5.2).

         The erythemal response of the skin to UV is usually inferred
    from the minimal erythema dose (MED). This value is determined by
    exposing adjacent areas of skin to increasing doses of UV, and
    recording the lowest dose to achieve erythema at a specified time,
    usually 24 h after irradiation.

         The visual detection of erythema is subjective and is affected
    by unrelated factors such as viewing geometry, intensity and
    spectral composition of ambient illumination, colour of unexposed
    surrounding skin, (Chamberlain and Chamberlain 1980, Diffey and
    Robson, 1992), and the experience and visual acuity of the observer.
    The difficulty in judging accurately a minimal erythema response is
    reflected by the varying definitions proposed for this value: these
    range from the dose required to achieve a just perceptible erythema
    (Everett  et al., 1965, Kelfkens and van der Leun, 1989); to that
    dose which will just produce a uniform redness with sharp borders
    (Wucherpfenning, 1931).

         The MED will vary according to the wavelength range over which
    the effective UV is summed and for radiation protection purposes is
    generally taken to lie in the range 200 to 300 J m-2 effective.

    5.3  Measurement Techniques

         There are three distinct types of measurement systems employed
    in the detection of UV: radiometers, spectroradiometers and
    dosimeters. Radiometers and spectroradiometers are direct-reading
    instruments that use electro-optical (physical) detectors to convert
    the incident radiation into an electrical signal. Radiometers
    measure all incident radiant power over a wide spectral range;
    whereas, spectroradiometers measure the radiant power distribution
    over a wide spectral range. Either by electronic means or by
    computer control, radiometers or spectroradiometers may be modified
    and calibrated to operate as "dosimeters" by time-integration of the
    output signal from the detector.

    FIGURE 5.2

         However, by dosimeters, one usually means devices that by
    nature respond directly to incident dose, i.e. radiant exposure.
    Dosimeters may be further optically modified and calibrated to
    respond according to an action spectrum, thereby serving as a
    direct-reading instrument for dose to a particular organ.

    5.3.1  Detectors

         The term detector normally refers to an electro-optical device
    which converts an optical signal (e.g UV or light) into an
    electronic signal which can be recorded. An important characteristic
    of any detector is the responsivity, defined as the quotient of the
    detector output (e.g amperes, A) and the radiant power incident upon
    the detector (e.g. watts, W). Thus units for responsivity may vary.
    The unit for irradiance-responsivity is often ampere-per-watt-per-
    square-metre (A W-1 m-2). The spectral responsivity is the
    responsivity as a function of wavelength. The term detectivity is
    used to compare the detection capability (the smallest quantity of
    radiation that can be detected) of different types of detectors, the
    higher the detectivity the more sensitive the detector. Thermal
    detectors, such as thermopiles and pyroelectric detectors, have a
    much lower detectivity (less sensitive) than photodiodes, phototubes
    and photomultipliers.

         The most common types of detectors for UV are semiconductor
    (junction) photodiodes, vacuum photodiodes (phototubes) and
    photomultipliers. Junction photodiodes are usually silicon (Si)
    photodiodes that may be enhanced to improve their UV responsivity
    and with a spectral responsivity between 190 and 1100 nm.
    Gallium-arsenide-phosphide (GaAsP) photodiodes have a spectral
    responsivity between 190 and 670 nm, and gallium-phosphide (GaP)
    photodiodes have a spectral responsivity between 190 and 520 nm.

         Chemical detectors, such as photographic film emulsions or
    polymer films of polysulphone or CR-39 resins, respond to incident
    radiant exposure (J m-2). Their responsivity is generally strongly
    wavelength dependant, and attempts are made to simulate a
    photobiological action spectrum directly.

         The choice of an optimum detector for a specific instrument or
    measurement situation depends upon the requirements for ease of data
    collection, portability, electrical power requirements, size and
    accuracy. Each detector type has advantages and disadvantages.
    Important parameters to consider for instrument requirements are:
    spectral responsivity, noise-equivalent-power (NEP), linearity,
    time/frequency response, stability over time, environmental
    operating conditions, maintenance, ease of operation, the
    requirements upon additional electronics, and cost.

         Biological detectors are also used. For example, biofilms using
    dried spores of  bacillus subtilis, immobilized on transparent

    polyester plastic sheets. After irradiation the biofilm is incubated
    in a growth medium and the proteins synthesized after spore
    germination, stained and evaluated by photometry. The biologically
    effective dose is calculated using a calibration curve. The UV
    response of this biofilm is additive and follows the reciprocity law
    in the normal range of fluence rates investigated. The response is
    independent of temperature (-20°C to 70°C) and humidity. The biofilm
    can be stored for up to 9 months at room temperature without
    significant influence on the viability of spores. These detectors
    have been used in Antarctica and in space to measure the biological
    consequences of ozone variations (Quintern  et al., 1992).

    5.3.2  Radiometers

         A radiometer is a detection system that measures incident
    radiation. A UV-radiometer usually measures irradiance in watts per
    square metre (W m-2). The basic layout of a radiometer is shown in
    figure 5.3.

         The diffuser shown in figure 5.3 may not always be present.
    When used, its purpose is to make the angular responsivity of the
    radiometer proportional to the cosine of the angle of incidence
    (measured from the normal to the diffuser surface). This arrangement
    is often called a  cosine-corrected radiometer. The diffuser may be
    made from a flat or slightly curved piece of ground quartz or
    teflon, or it may be an integrating sphere. The purpose of the
    optical filter is to limit the spectral responsivity to a certain
    band, having a lower and upper wavelength cutoff. The band-width can
    vary upward from about 5 nm, but is often several tens or even
    hundreds of nanometres wide. Hence, radiometers are called
    broad-band meters as opposed to spectroradiometers. Ideally the
    spectral responsivity is constant within the band and zero outside,
    but in practice this is not possible. If the detector is sensitive
    to radiation outside the passband of the filter, which is normally
    the case, there will always be a non-zero responsivity outside the
    band. Signals which are produced by radiation from outside the band
    are called "out-of-band signals," "out-of-band leakage," or just
    "stray light." This limitation of radiometers can often be a serious
    problem, and is particularly troublesome in UV radiometers based on
    Si photodiode detectors, because the spectral responsivities of Si
    photodiodes extend to 1100 nm in the infrared. If one attempts to
    use such radiometers to measure small quantities of in-band
    radiation in the presence of large quantities of out-of-band
    radiation, the results will be prone to large errors. This is
    frequently encountered when attempting to measure the very small
    component of biologically active UV present in a light source
    spectrum.

    FIGURE 5.3

         For many purposes it is desirable to have a UV radiometer which
    has a spectral responsivity equal to or closely resembling a certain
    action spectrum. If this is achieved, the radiometer signal is
    directly proportional to the "effective" or "biologically active"
    irradiance, because the radiometer will spectrally "weight" the
    different wavelengths according to the action spectrum Slambda. A
    well known example of such a radiometer is a photometer (lux-meter
    or luminance-meter) which has a spectral responsiveness that closely
    matches the photopic (visual) response of the human eye. Radiometers
    are commercially available which have spectral responsivities that
    match, for example, the UV erythema action spectrum (McKinlay and
    Diffey, 1987) and UV hazard action spectrum adopted by the
    IRPA/INIRC (1991) and ACGIH.

    5.3.3  Spectroradiometers

         A  spectroradiometer is a radiometer that is capable of
    measuring spectral radiometric quantities directly, such as spectral
    irradiance or spectral radiance. The major difference between the
    layout of a spectroradiometer and the layout of a radiometer is the
    waveband selecting device which in a radiometer is usually a
    broadband filter (see figure 5.3), whereas in a spectroradiometer it
    is a monochromator or a spectrograph. Radiation entering a
    monochromator or a spectrograph is dispersed by a grating or a prism
    and only a small band of radiation is passed to the detector, the
    so-called bandwidth. The spectroradiometer bandwidth can be selected
    according to the application, but it is typically 1 - 5 nm. The
    waveband passed to the detector can be changed manually or
    automatically by rotating the grating or prism; the instrument is
    scanned over the spectral range of interest. This type of instrument
    is called a scanning spectroradiometer.

         In a spectrograph, a portion of the spectrum is incident upon a
    photographic film, or a linear photodiode array which can be read
    diode by diode. This can occur very rapidly and a spectrum displayed
    almost instantly. This type of spectroradiometer is therefore useful
    for studies where time resolution is important. Another advantage is
    that there are no moving parts compared to a scanning
    spectroradiometer. However, spectrographs have disadvantages over
    scanning systems, such as spectral resolution and detectivity.

         Spectroradiometers are more complex to operate and maintain
    than radiometers; they are considerably more expensive; and there
    are many pitfalls for the inexperienced user. Spectroradiometers
    generally employ the same kind of input optics as radiometers.

         In an ideal spectroradiometer, the monochromator passes a small
    band of wavelengths to the detector, and passes no radiation outside
    this band. In practice the out-of-band radiation (leakage or
    stray-light) that is passed to the detector is of the order of 0.1
    percent of the amount of in-band radiation depending upon the

    quality and size of the monochromator. In certain measurement
    situations this may give rise to errors (just as in radiometers):
    the signal caused by out-of-band radiation may be of the same order
    of magnitude as the signal caused by in-band radiation. For example,
    if one attempts to measure the solar UV spectrum in the 290-310 nm
    region, this error occurs because the solar spectrum decreases about
    five orders of magnitude from 310 nm to 290 nm. To overcome this
    error, a double monochromator may be used. A double monochromator is
    essentially two identical monochromators coupled in tandem, the
    output of the first becomes the input of the second monochromator.
    The stray light of a double monochromator is typically 0.01 percent.

    5.3.4  Personal dosimetry

         For personal monitoring of UV doses, radiometers are too bulky.
    A thin-film polymer (eg polysulphone) dosimeter overcomes this
    problem. It is a thin (0.04 mm) clear plastic film that may be worn
    as a small badge comparable to those used to monitor ionizing
    radiation. This allows monitoring of UVB doses on mobile subjects.
    Polysulphone changes its absorbance (or transmittance) when
    irradiated with UV (mainly by UVB). The received dose (radiant
    exposure) is determined by measuring the change in absorbance of the
    film before and after exposure (Diffey, 1989a). CR-39 plastic resins
    have also been explored as UVB dosimeters; however, the low
    sensitivity requires longer exposure periods (Wong  et al., 1989).

    5.4  Calibration

         Improper or inadequate calibration of UV radiometers,
    spectroradiometers and dosimeters is a serious and common source of
    error. It is important to maintain a good calibration record for the
    instruments, but only experience on instrument stability will
    determine how often calibration is needed. The error caused by
    calibration provides the minimum uncertainty that can be obtained in
    measurement situations. Other sources of uncertainty include
    geometry and spectrum of source emission, detector-source geometry
    (angular errors), environmental influences and time factors.

         Radiometers can be calibrated by using a source of known
    irradiance. This may either be a line source, such as a mercury lamp
    or a laser, or by a broadband source, such as a tungsten halogen
    lamp. It must be realized that the irradiance-responsivity will
    depend upon the source used for calibration. For example, a
    radiometer which has been calibrated against a line source will give
    erratic readings if used to measure a broadband source. In practice
    it is advisable to have radiometers calibrated against a source
    which emits a spectrum similar to those of the sources to be
    measured.

         Complete calibration of radiometers and spectral radiometers
    include analysis of the cosine response of the instrument, the

    azimuthal response and the temperature sensitivity of the
    instrument. Recent work on instrument calibration and
    intercomparison has shown that instruments can vary greatly in these
    quantities even when the instruments agree on a simple spectral
    calibration. In the field, instruments will measure different
    quantities depending on the temperature and angle of incident
    radiation.

         Spectroradiometers are calibrated against standard lamps of
    known spectral irradiance (or radiance). Such lamps can be obtained
    from the standards institutes in various countries. Intercomparison
    of lamp calibration from institute to institute can vary by as much
    as eight percent in the UV region, while better agreement is common
    in the visible region of standard lamps. This discrepancy should be
    noted when comparing results based on instruments calibrated from
    different lamp standards.

         A tungsten halogen lamp is used as the standard lamp for
    wavelengths between 250 and 2500 nm, whereas a deuterium lamp is
    used in the region between 180 and 300-400 nm. It is good practice
    to operate the standard lamp only when calibrating a
    working-standard lamp and then use the working-standard for routine
    calibration of a spectroradiometer. For very accurate work it is
    recommended to maintain three calibrated lamps in order to find out
    whether a change in response/calibration, at least of a
    spectroradiometer, is caused by changes in the standard lamp used
    for calibration or by changes in the spectroradiometer.

         Dosimeters are calibrated in the same way as radiometers except
    that exposure time is an integral part of the calibration process.
    To obtain a reliable calibration for radiometers and dosimeters, it
    is advisable to calibrate them against a source which emits a
    spectrum similar to the one that is to be measured.

    6.  CELLULAR AND MOLECULAR STUDIES

    6.1  Introduction

         This chapter provides an overview of the evidence for cellular
    and molecular effects of UV exposure on biological systems. Since UV
    exposure has been associated with skin and eye cancers in humans,
    emphasis will be given to the process of carcinogenesis. It is known
    that UV exposure results in photochemical modification of the
    genetic material (DNA), but most of this damage is accurately and
    efficiently repaired by the cell. However if the amount of damage is
    too great some of the alterations to the DNA may remain as permanent
    mutations. It has been proposed that if unrepaired damage occurs to
    regulatory genes this may be involved in the process of
    carcinogenesis. In this context mutations to and activation of genes
    may be important.

         Other responses likely to result from UV exposure of cells
    include increased cellular proliferation, which could have a tumour
    promoting effect on genetically altered cells, as well as changes in
    components of the immune system present in the skin. There is
    evidence to suggest that UV exposure could elicit an
    immunosuppressive effect which may compromise the body's ability to
    identify and destroy tumour cells of the skin.

         This review includes summary descriptions of UV action and
    repair of damage to biomolecules, particularly in DNA, cellular
    chromophores and other target molecules, as well as damage to the
    cell membrane and proteins. Consequences of damage to the cell, its
    membranes and activation of genes are also reviewed.

    6.2  Interactions with Biomolecules

         UV must be absorbed to produce a chemical change. At solar UV
    intensities normally encountered, the first step in a photochemical
    reaction is the absorption of a single photon by a molecule and the
    production of an excited state in which one electron of the
    absorbing molecule is raised to a higher energy level. Such
    radiative transition can only occur efficiently when the photon
    energy of the radiation is close to the energy difference of the
    atom in the initial and final state (energy level). The
    photochemistry that may then occur will therefore depend upon the
    molecular structure and the wavelength of UV as well as the specific
    reaction conditions. The primary products generated by UV absorption
    are generally reactive species in a metastable excited state or free
    radicals both of which form extremely fast. Dark chemical reactions
    then occur often within microseconds but they may last for hours, as
    is the case for the lipid peroxidation chain reaction. Finally these
    relatively rapid processes are translated into photobiological
    responses which may occur within seconds but can take years or even
    decades to be manifested.

    6.2.1  Cellular chromophores

         Since chromophores (see chapter 5 for definition) are
    characterized by the wavelengths at which they absorb, the nature of
    the critical chromophores will change as a function of wavelength
    throughout the UV range. The peak absorption of DNA is dictated by
    its component nucleic acids and occurs at around 260 nm. There is a
    sharp drop in absorption through the UVB range and absorption is
    undetectable by conventional means at wavelengths longer than 320
    nm. Using special detectors Sutherland and Griffin, (1981) have
    measured DNA absorption at wavelengths as long as 360 nm. Although
    overall protein absorption peaks in the UVC range, the aromatic
    amino acids such as tryptophan (lambda max = 280 nm at pH7) and
    tyrosine (lambda max = 275 nm at pH7) exhibit absorptions that
    extend into the UVA range so that direct damage to proteins can
    occur at much longer wavelengths than direct damage to DNA.

         Several cellular components such as quinones, flavins, steroids
    and porphyrins are important UVA chromophores. Porphyrins, which
    exhibit an absorption with a peak around 405 nm, have been
    implicated in the lethal action of UVA and near-visible light in
    certain bacteria. Mutants in the bacterium  Escherichia coli (E
    coli) which are deficient in the synthesis of L-amino levulinic
    acid, the first step in heme synthesis, are resistant to UVA
    radiation (Tuveson and Sammartano, 1986) strongly suggesting that
    porphyrin intermediates can be phototoxic. Porphyrin intermediates
    evidently also arise during heme synthesis in humans. Indeed,
    supplementation of human cells with amino levulinate (ALA) bypasses
    the synthase step and leads to accumulation of protoporphyrin IX
    (PPIX), the immediate precursor to heme and a strong
    photosensitiser. ALA appears to be preferentially taken up by skin
    cancer cells and the selective photosensitisation of such cells by
    PPIX is the basis of a new type of phototherapy based on endogenous
    sensitisers (Kennedy  et al., 1990). Iron chelators can enhance the
    sensitising effect by preventing the insertion of iron into the PPIX
    macrocycle thus preventing the formation of the relatively
    non-photoactive heme product. Accumulation of iron-free porphyrins
    is the basis of the acute photosensitivity of skin in patients with
    a variety of porphyrins. However, it is not clear to what extent
    PPIX, for example, leads to UVA-mediated cytoxicity in normal human
    skin cells.

         Aside from DNA and proteins, the main chromophores in human
    skin are urocanic acid and melanins (for review see Anderson &
    Parrish, 1981). Trans-urocanic acid (4-amidazoleacrylic acid) is the
    deamination product of histidine generated by histidase and because
    of its broad absorption in the UV region has been considered as
    contributing to a modest extent to the natural sunscreen properties
    of skin. However, most notable is that the trans form can be
    photoisomerised to the cis form and it has been proposed that this
    conversion is a key factor in UV-induced immunosuppression (DeFabo

    and Noonan, 1983). However, considerable controversy now surrounds
    this original suggestion (eg. see Gibbs 1993).

         Melanins are the major UV absorbing chromophores in skin,
    exhibiting an extremely broad spectrum of absorption over the UVB,
    UVA and visible ranges. Melanins are complex polymeric proteins that
    are produced by melanocytes and transferred to keratinocytes.
    Although often considered to be neutral density filters, this is not
    strictly correct since melanins usually degrade upon UV exposure.
    There is some evidence that melanin may function as a
    photosensitizer of DNA damage.

         In addition to iron-free porphyrins which generate singlet
    oxygen upon exposure to UVA radiation, other small molecules also
    have the potential to generate active oxygen intermediates upon UVA
    exposure (Tyrell, 1992). For example, the photochemical degradation
    of tryptophan by wavelengths which include the more energetic
    portion of the UVA spectrum is able to generate hydrogen peroxide
    and N-formyl kynurenin (McCormick  et al., 1976). Although the
    level of hydrogen peroxide generated  in vivo by such a pathway
    would appear to be in the low micromolar range it could nevertheless
    be crucial to biological processes since iron complexes (such as
    citrate) that are present in the cytoplasm will react with hydrogen
    peroxide to generate the highly reactive hydroxyl radical in a
    superoxide driven Fenton reaction (see Gutteridge, 1985; Imlay  et
     al., 1988). Since the reaction is driven by the continual
    reduction of ferric iron to the ferrous state by superoxide anions,
    a cellular source of superoxide anions is also required. In this
    context it should be noted that both hydrogen peroxide and hydroxyl
    radical are generated by UVA irradiation of NADH and NADPH
    (Czochralska  et al., 1984; Cunningham  et al., 1985). However, it
    is not at all clear whether this is really the key source of
    superoxide anions or whether the main source is as a consequence of
    normal cellular metabolism.

    6.2.2  Cellular targets

         The chromophores for UV effects are not necessarily the
    critical targets which mediate the effects. The most important
    cellular target for UV is considered to be DNA since the crucial
    genetic material exists in unique and very low copy numbers in
    cells. Radiation in the UVB range is absorbed by DNA and leads to
    photochemical damage, so that DNA would certainly appear to be the
    primary chromophore and site of damage for most of the biological
    effects of short wavelength UV. DNA damage induced by UVB radiation
    is the key factor leading to sunlight-induced mutations in
    cancer-related genes and therefore in initiating the carcinogenic
    process. At longer wavelengths, targets may change. For example, the
    destruction of mitochondria may be a key factor in the breakdown of
    cellular integrity following certain types of photosensitisation.
    Membrane damage clearly takes on added significance when the UV

    radiation employed (e.g.from sunlight) includes a strong component
    of longer wavelengths. Breakdown of membranes can lead to aberrant
    signal transduction as well as dramatic alterations in transport
    pathways. Leakage of essential components or an influx of
    extracellular molecules such as calcium can have severe cytostatic
    and even lethal consequences for the cell and will clearly have an
    influence on overall tissue/organ function.

    6.3  Action Spectra

         An action spectrum is a measure of the relative effectiveness
    of different wavelengths, within the spectral region of study, to
    produce a given response (see chapter 5 for definition). Clearly the
    number of photochemical and photobiological endpoints that can be
    measured is as large as the number of effects themselves and before
    undertaking action spectroscopy for a given end-point, there should
    exist a good  a priori reason for the study. Numerous types of UV
    induced DNA damage have now been recognized that include stand
    breaks (single and double), cyclobutane-type pyrimidine dimers, 6-4
    pyo photoproducts and the corresponding Dewar isomer, thymine
    glycols, 8-hydroxy guanine, and many more. In addition, DNA-protein
    cross-links are produced during UV exposure. Larger scale genetic
    alterations include chromosome breakage, sister chromatid exchanges
    and chromatid aberrations. Although partial UV action spectra are
    now available for many of these lesions, the most studied have been
    the different types of pyrimidine dimers. Since the indirect
    oxidising component of radiation damage increases with increasing
    wavelength, there is a dramatic shift in the type of lesion induced
    as the wavelength increases. Pyrimidine dimers are characteristic of
    the direct absorption that occurs at shorter UV wavelengths whereas
    strand breaks and 8-hydroxy guanine type lesions become increasingly
    important at longer wavelengths. At a higher level of complexity,
    action spectra for cell death, mutation,  in vitro transformation,
    growth delays, cell permeability, etc, may also be measured.

         One goal of determining action spectra has been to correlate
    end-points with a specific type of initial damage. However, such
    evaluations are complex partly because the absorbing chromophores
    and crucial lesions will often change as a function of wavelength.
    At the whole organism level, action spectra may be determined for
    effects on entire organs, for example, various markers of leaf
    damage in plants, erythema induction in skin and even tumour
    induction.

         Studies in cultured cells may be of value in predicting
    responses in whole organs such as skin but various parameters must
    be evaluated. In particular, the penetration of UV to the critical
    chromophores as a function of wavelength must be taken into account
    by considering the transmission through overlying tissue. As an
    example, we may wish to calculate the relative cytotoxic action of
    the different wavelengths in sunlight to cells at the basal layer of

    the epidermis. In order to do this, an action spectrum must be
    available for the cytotoxic action of individual wavelengths on
    cultured epidermal keratinocytes. In addition, data is needed for
    transmission through human skin to the target cells since this will
    change as a function of wavelength. Finally if the effect of a
    particular UV source is required (such as terrestrial sunlight under
    defined conditions) then this must be determined by
    spectroradiometry or from available data. The relative biological
    effectiveness of individual wavelengths of sunlight in killing cells
    at the basal level of the epidermis may then be predicted by
    convoluting these three spectra. In practice, interactions
    (synergistic, additive or antagonistic) exist between different UV
    wavelength regions and these must also be taken into account in the
    evaluation of the biological effects of broad spectrum sources.
    Although action spectra may be of value for predictive evaluations,
    the primary aim of studies with cultured cells is often to determine
    chromophores. For this purpose, action spectra must be corrected in
    order to express results as action per photon before comparison with
    absorption spectra of critical biological molecules. Thus, DNA is
    known to be an important chromophore in the UVB region. The results
    are far more difficult to interpret in the longer wavelength UVA
    region but at least for simpler organisms such as bacteria,
    porphyrins have often been implicated. Action spectra, particularly
    in plants, are often modified by protective absorbing molecules such
    as carotenoids.

         In response to questions posed by the threat to the ozone
    layer, fairly detailed action spectra have now been determined for
    squamous cell carcinoma in hairless mice and transmission
    corrections made to estimate the spectrum in humans. DNA is clearly
    the primary chromophore in the UVB region but a significant second
    peak occurs in the long UVA wavelength region.

    6.4  Biomolecular Damage

         UV radiation can damage many cellular targets including the
    nucleic acids, proteins and lipids. For the non-solar UVC
    wavelengths, DNA is clearly the most important target and many
    photochemical changes can occur as a result of direct absorption.
    The genotoxic action of solar UVB radiation is also of critical
    importance, although the spectrum of DNA damage begins to change as
    oxidative events become more important. At longer UVA wavelengths,
    indirect effects mediated by active oxygen intermediates are common
    and except for events directly related to DNA modification (e.g.
    mutation), it is difficult to discern the crucial targets.

    6.4.1  Nucleic acids

         Cyclobutane type pyrimidine dimers

         Cyclobutane type pyrimidine dimers were the first type of
    UV-induced base damage to be identified (Beukers & Berends, 1960)
    and being the most frequent lesion induced by either UVC or UVB
    radiation, they have been the most studied.

         Cyclobutane-type pyrimidine dimer formation arises from the
    production of reactive excited states (normally the forbidden but
    long-lived triplet state) following absorption of UV radiation. The
    action spectrum for dimer formation as shown in figure 6.1, closely
    resembles that for the extinction coefficients of the appropriate
    monomers, cytosine(C) or thymine(T) for wavelengths as long as 313
    nm (Ellison & Childs, 1981) so that the mechanism of formation is
    probably similar. Although pyrimidine dimer formation (e.g. thymine
    lozenge thymine or T lozenge T) has been measured in the UVA range,
    six orders of magnitude more energy is required at 365 nm as
    compared to 254 nm (Tyrrell, 1973) and the mechanism of formation is
    unclear. Most of the original observations concerning dimer
    induction were made in isolated DNA, bacteriophage systems or in  E
     coli but essential results have since been confirmed in mammalian
    cells including human skin fibroblasts. Action spectra determined in
    human and mouse skin are sharply attenuated at shorter wavelengths,
    but are otherwise basically similar to those obtained  in vitro.
    The action spectra for cyclobutane pyrimidine dimer formation in
    naked DNA, cell cultures and epidermal DNA is given in figures 6.2
    and 6.3. Based on studies with  E coli, the ratio of T lozenge T, C
    lozenge T to C lozenge C changes appreciably with wavelength. For
    example, the ratio of T lozenge T to C lozenge T dimers is 0.63:1 at
    313 nm (Ellison & Childs, 1981) but increases to 6:1 at 365 nm
    (Tyrrell, 1973).

         Although repair of DNA damage is dealt with separately below,
    it is worth noting at this point that an extremely specific
    light-dependant repair process, photoreactivation, exists for repair
    of pyrimidine dimers in situ. Repair proceeds via photolyase which
    has been shown to be present almost ubiquitously throughout the
    animal world from  E coli to non-placental mammals. Recent evidence
    suggests that photoreactivation may be less important in human cells
    (Li  et al., 1993). Nevertheless, the photoenzymatic splitting of
    pyrimidine dimers has provided a powerful technique since it has
    provided a way of linking a specific type of DNA damage (the
    pyrimidine dimer) with defined biological effects (e.g. cell death,
    tumour formation).

    FIGURE 6.1

    FIGURE 6.2

    FIGURE 6.3

         Thy (6-4) pyo photoproducts

         A second type of pyrimidine dimer can be formed by UV (Varghese
    & Wang, 1967) which was originally termed a pyrimidine adduct and is
    now more commonly referred to as the thy (6-4) pyo photoproduct.
    More recently a pyrimidine nucleoside-cytidine lesion was recognized
    in highly reiterated sequences of human DNA (Lippke  et al., 1981)
    which is almost certainly the precursor of the thy (6-4) pyo
    photoproduct (Brash & Haseltine, 1982; Franklin,  et al., 1982).
    The (6-4) photoproduct is formed with much greater frequency between
    cytosines located 5' of adjacent pyrimidines. Action spectra that
    compare induction of cyclobutane-type pyrimidine dimers are all
    similar between 265 nm and 302 nm (Patrick, 1977; Chan  et al.,
    1986; Matsunaga,  et al., 1991; Rosenstein & Mitchell, 1987) as
    shown in figure 6.4. However, at longer wavelengths not all the
    spectra agree. At least for human skin fibroblasts the relative
    level of (6-4) photoproduct induction is only half that for
    cyclobutane-type dimers. The reason for this appears to be that the
    (6-4) photoproducts are converted to a Dewar pyrimidine isomer by UV
    radiation at wavelengths peaking in the UVB range. This has been
    confirmed in studies by Mitchell & Rosenstein (1987) using
    radio-immuno assays for the (6-4) photoproduct and its Dewar isomer.

         Although (6-4) photoproducts form with 5-10 fold lower
    efficiency than cyclobutane type dimers, they may be formed with
    equal efficiency at certain sites (Kraemer  et al., 1988).
    Furthermore, in the UVB region, a significant level of Dewar
    pyrimidine isomers will be formed and the precise ratio and level of
    these three major types of base damage will depend on the
    wavelength(s) of radiation employed, as well as irradiation
    conditions.

         Other types of dimeric base damage which include purines have
    been isolated from DNA heavily irradiated with UVC. These include
    the thymine-adenine dimer (Bose & Davies, 1984) and adenine
    dehydrodimer (Gasparro & Fresco, 1986). Such photoproducts occur at
    fairly low frequency (approximately 1 percent of that of
    pyrimidine-type dimers. The relative levels and significance of
    these photoproducts in the UVB or longer wavelength regions has not
    yet been determined.

    FIGURE 6.4

         Monobasic DNA damage

         Damage to a single base is a relatively low frequency event in
    UV damaged DNA. Simple hydrates of cytosine and thymine can
    certainly be formed (Fisher & Johns, 1976) but these low frequency
    events are unreliable and therefore extremely difficult to study
    from a biological viewpoint. They may be related to the low
    frequency cytosine photoproduct identified by sequencing techniques
    by Gallagher  et al. (1989). Sequencing techniques in combination
    with endonuclease V treatment have also led to the detection of a
    class of rare and unidentified purine or purine-pyrimidine sites
    after broad spectrum UV irradiation (Gallagher & Duker, 1986).

         A group of lesions induced by both UV and ionizing radiation
    are ring-saturated thymines of the 5, 6- dihyroxydihydrothymine type
    (thymine glycols). Although original measurements indicated that
    they were induced almost as frequently as pyrimidine dimers at 313
    nm as against 21:1 at 254 nm (Cerutti & Netrawali, 1979), it has
    been concluded from more recent data using sequencing analysis
    (Mitchell  et al., 1993) that they do not occur at a significant
    rate in UVC or UVB irradiated DNA.

         Increasingly sophisticated chemical methods are now becoming
    available to measure oxidative DNA damage (Cadet  et al., Methods
    in Enzymology, in press). Since the proportion of this type of
    damage will increase with increasing wavelength, such techniques
    will soon be applied to consolidate the picture of UV-induced based
    damage throughout the entire solar UV spectrum. Furthermore, an
    interesting enzyme has been isolated (Boiteux  et al., 1987) termed
    the Fapy glycosylase because of its ability to excise the
    ring-opened form of N7-methylguanine from DNA. The enzyme will also
    recognize DNA products generated in photosensitisation reactions
    involving singlet oxygen (Epe  et al., in press) and
    8-hydroxyguanine may be a major component. Since the longer UV
    wavelength in sunlight clearly generate biologically relevant levels
    of singlet oxygen and other active intermediates (Tyrrell, 1991;
    Basu-Modak & Tyrrell, 1993), this type of analysis may also be
    usefully applied to the solar UV spectrum. Certainly, near-visible
    light alone is able to generate Fapy glycosylase-sensitive lesions
    in the DNA of cultured mammalian cells and these, at least in part,
    are probably 8-hydroxyguanine (Epe  et al., 1993).

         DNA strand breaks

         Few DNA strand breaks are induced by UVC radiation but they
    constitute an increasing proportion of the total lesions as

    wavelength is increased. For example, in  E. coli the ratio of DNA
    strand breaks to pyrimidine dimers is 1:44 at 313 nm (Miguel &
    Tyrrell, 1983) whereas at 365 nm one strand break is formed for
    every two pyrimidine dimers (Tyrrell  et al., 1974). Action spectra
    have now been determined for the induction of DNA single strand
    breaks in human skin cells which show that breaks occur throughout
    the UVA and into the near-visible range (Peak  et al., 1987). Since
    break measurements involve alkaline denaturation, 10-20 percent of
    the so-called breaks are due to the fragility of chemical bonds at
    apurinic and apyrimidinic sites. It is important to note that unlike
    most of the common forms of base damage, induction of strand breaks
    is strongly dependent on oxygen throughout the UVB and UVA ranges
    (Tyrrell, et al., 1974; Peak & Peak, 1982; Miguel and Tyrrell, 1983)
    and at the longest wavelengths may involve generation of singlet
    oxygen (Peak  et al., 1987). Figure 6.5 gives the action spectra
    for single strand breaks and figure 6.6 for double strand breaks in
    cultured cells.

         DNA-protein cross links

         Photochemically induced DNA-protein cross links, mostly
    involving cysteine are clearly formed  in vitro (Smith, 1976;
    Shetlar, 1980) and appear to be formed  in vivo, particularly in
    the UVA range (Bradley,  et al., 1979; Peak & Peak, 1991). Despite
    the obvious importance of this type of damage, there is little known
    at the molecular level concerning the nature of the damage formed
     in vivo. Action spectra for DNA-protein cross links in cultured
    cells is given in figure 6.7 (from Peak  et al., 1985a, b).

         Ribonucleic acid

         Messenger RNA is readily susceptible to modification by UV
    radiation. However, given the fairly rapid turnover of most
    molecules of this type (half-lives generally of the order of minutes
    and hours) and the capacity for  de novo synthesis in the absence
    of DNA damage, mRNA is not generally considered a critical target of
    radiation damage in mammalian cells. Bacterial transfer RNA is often
    extremely photosensitive because of the presence of an unusual
    nucleoside 4-thiouridine and this leads to some fascinating
    photobiological phenomena of ecological significance (Jagger, 1985).
    However, such findings are of little significance to eukaryotic
    cells.

    FIGURE 6.5

    FIGURE 6.6

    FIGURE 6.7

    6.4.2  Membranes

         UV-induced changes in membrane permeability and membrane
    transport systems would be expected to have fairly dramatic
    consequences for human skin and eyes. Unsaturated fatty acids are
    readily oxidised to hydroperoxides. Several reports have now shown
    that UV radiation can peroxidise membrane lipids (Desai,  et al.,
    1964; Roshchupkin  et al., 1975; Putvinsky  et al., 1979; Azizova,
     et al., 1980). Using liposomal models (Mandal & Chatterjee, 1980;
    Bose  et al., 1989), it has been shown that UVA radiation causes a
    dose-dependent increase in lipid peroxidation as measured by various
    techniques and that this can be largely inhibited by membrane
    antioxidants such as butylated hydroxytoluene. In  E coli,
    sensitivity to UVA correlates with the levels of unsaturated fat in
    membranes (Klamen & Tuveson, 1982; Chamberlain & Moss, 1987). An
    agent which enhances singlet oxygen lifetime, deuterium oxide,
    enhances the level of membrane damage, sensitivity to UVA and lipid
    peroxidation. There are now reports that both UVB and UVA (Morlière
     et al., 1991; Punnonen  et al., 1991; Vile  et al., 1994)
    radiation can cause lipid peroxidation at biologically relevant
    fluences in the membranes of human fibroblasts and keratinocytes.

         Cell leakage experiments have been used to assess membrane
    damage in yeast (Ito and Ito, 1983) and a similar technique has now
    been used to show UVA-mediated enhancement of membrane leakage in
    human skin fibroblasts (Gaboriau  et al., 1993; Vile  et al.,
    1994). UVA also causes changes in membrane fluidity as assessed
    using a lipophilic fluorescent probe. UVA-induced alteration of
    membrane transport systems received considerable attention in
    prokaryotes (see Jagger, 1985) but little information is available
    in mammalian cells. A recent study has demonstrated that UVA
    radiation inhibits both receptor-mediated (low density lipoprotein)
    and nonspecific (sucrose) uptake of exogenous molecules
    (Djavuheri-Mergny  et al., 1993). These findings may be related to
    other studies in human fibroblasts which have shown that broad
    spectrum UV sources cause cytoskeletal damage as manifested by
    dose-dependent microtubule disassembly (Zamansky & Chou, 1987).

    6.4.3  Proteins

         As the wavelength is increased through the UVB and UVA regions
    damage to proteins becomes increasingly important because of the
    absorption properties of the aromatic amino acids relative to
    nucleic acids. In addition, many proteins that include the
    antioxidant enzymes catalases and peroxidases contain heme groups
    thus making them UVA chromophores and potentially photosensitisers.
    Indeed, catalase is sensitive to sunlight (Mitchell & Anderson,
    1965), probably as a result of UVA absorption (Kramer & Ames, 1987).
    In bacteria there is evidence that endogenous catalase and possibly
    alkyl hydroperoxide reductase are actually photosensitisers
    (Eisenstark & Perrot, 1987; Kramer & Ames, 1987) but no similar

    experiments to have been carried out in mammalian cells. Repair
    enzymes are also sensitive to UVA radiation and there is evidence
    that UV-induced repair disruption plays a role in cell death and
    mutagenesis (Haynes, 1966; Webb, 1977; Tyrrell, 1982). Considerable
    attention has been given recently to metal ion catalysed oxidation
    of protein since this is clearly a physiologically relevant process
    (for review see Stadtman, 1990). Numerous proteins have been shown
    to be modified by a free radical generating, system modelling those
    that cells are exposed to during normal metabolism or exogenous
    insult. It appears that active oxygen/free radical species are
    generated at specific metal binding sites on proteins and that this
    leads to reactions with amino acid residues at specific steric
    locations.

    6.5  Cellular Defences

    6.5.1  DNA

         Even one or two pyrimidine dimers in the entire genome of a
    bacterium such as  E coli may be lethal if DNA repair processes are
    defective. However, DNA repair processes are not only crucial for
    cell survival but also for the maintenance of genetic stability
    since DNA damage is continuously generated as a by-product of
    metabolism even in the absence of exogenous insults. Because of the
    availability of mutants, most of the original studies concerning
    removal of DNA damage employed bacteria and bacteriophage. The first
    process to be identified was photoreactivation later known as
    photoenzymatic repair. In this process, a DNA photolyase forms a
    specific complex with a pyrimidine dimer which can be split in the
    presence of UVA and/or visible light to leave the original
    pyrimidines in the DNA. The process has been widely used to
    correlate a specific type of DNA damage (the pyrimidine dimer) with
    a given biological effect (from cell death in bacteria to cancer in
    fish), although recent evidence for an enzyme which recognizes (6-4)
    photoproducts may cast doubt on the validity of this approach.
    Photoenzymatic repair is entirely error free (i.e. it does not lead
    to mutations) and occurs in a wide variety of single and
    multiple-celled organisms. Although photoenzymatic repair and
    associated photolyases occur in non-placental mammals such as the
    marsupials, evidence for photoreactivation in placental mammals,
    including humans, has been fragmentary and controversial. Recent
    evidence suggesting that photoreactivation of UV damage may not be a
    significant process in human skin (Li  et al., 1993) could be of
    crucial importance when considering potential interactions between
    wavelengths.

         Evidence for two broad categories of DNA repair, originally
    denoted as dark repair processes, were revealed by bacterial studies
    (see Friedberg, 1984 for review). Post-replicational repair, as the
    name implies, can occur only after DNA synthesis and includes many
    of the complex pathways of genetic recombination. Humans with the

    genetic disease Xeroderma pigmentosum (XP) of the variant
    complementation group are deficient in this type of repair. However,
    the SOS repair process in bacteria which involves the inducible
    activation of a post-replication error-prone (mutagenic) repair
    pathway, has no equivalent in humans. On the other hand, excision
    repair is the most widespread of DNA repair processes and occurs
    from bacteria to man. In excision repair, damaged or incorrect bases
    are excised from the genome and replaced by the correct nucleotides.
    The first step in the process involves recognition of the damage and
    incision at or close to the damaged site. This can occur by two
    distinct mechanisms. Damaged bases may be recognised by a series of
    specific glycosylases. In this case, incision occurs by a two-step
    reaction which involves the sequential activities of the DNA
    glycosylase and an apyrimidinic/apurinic endonuclease. However, a
    second mechanism of incision involves the direct action of a
    damage-specific DNA incising activity without the need for
    glycosylase action. At least in  E Coli, incisions are made at
    either side of the damaged base so that exonucleolytic removal of
    bases is unnecessary in this case. Repolymerization of the gap
    created by endonuclease/exonuclease activity is an essential step
    and can take place using the undamaged strand as a template. The
    action of DNA ligase restores the original integrity of the
    double-stranded DNA.

         DNA excision repair in humans is also an extremely complex
    process as evidenced by the considerable genetic heterogeneity in XP
    which is a rare but extensively studied disease whose cellular basis
    involves a defect in DNA repair. Genomic clones of several human DNA
    repair genes have now been derived by transfecting DNA from repair
    proficient human cells into a series of UV-sensitive mutant rodent
    cell lines (Westerveld  et al., 1984). A considerable amount of
    information is now available concerning the molecular genetics of
    eukaryotic DNA excision repair (Hoeijmakers & Bootsma, 1990). The
    repair genes isolated to date often have DNA binding and/or nucleic
    acid helicase domains. Most recently, a repair gene product has been
    identified as one of the components of a transcription factor
    complex involved in the transcription of polymerase II genes
    (Schaeffer  et al., 1993). This is currently an extremely active
    area of research and further genes are expected to be isolated.

          In vitro repair systems are now available (Wood  et al.,
    1988). In addition, animal models are now being developed using
    homologous recombination techniques and genetic manipulation of
    embryonic stem cells.

    6.5.2  Human excision repair disorders

         Several human diseases involve defects in DNA repair. The most
    studied example is the genetic disorder XP which actually involves
    several types of excision repair defect as well as the variant form
    which is defective in post-replication repair. These cell lines are

    crucial to basic studies of DNA repair in humans. Cell lines derived
    from the excision-defective individuals are many times more
    sensitive to inactivation and mutation by UVC and UVB radiation
    (Arlett  et al., 1992). This appears to be directly correlated with
    the fact that individuals with this disease are extremely prone to
    tumours of the skin, eye and lips (Kraemer  et al., 1987). Indeed,
    it is the study of cancer incidence in these individuals which
    provides the strongest evidence that the induction of photoproducts
    in skin by UV is the first event in a sequence which eventually
    leads to basal cell carcinoma, squamous cell carcinoma and melanomas
    in man.

         The above conclusion is now complicated by studies with cells
    from patients with the rare disease, Trichothiodystrophy. Although,
    cells from most of the patients studied so far have a marked defect
    in DNA excision repair, the patients do not show elevated evidence
    of skin tumours. They are, nevertheless, sun-sensitive in terms of
    their erythemal response (Bridges 1990). A partial explanation may
    be that XP patients but not Trichothiodystrophy patients have
    defects in both DNA repair and oxidative metabolism and both these
    processes are involved in carcinogenesis (Vuillaume  et al., 1992).

         Fibroblasts from patients with Cockayne's syndrome are
    specifically defective in excision of dimers from DNA undergoing
    active transcription (Mayne  et al., 1988) and are sensitive to
    killing and mutation by UVC radiation. (Arlett & Harcourt 1983).
    However, these individuals are not exceptionally cancer-prone
    (Barret  et al., 1991).

    6.5.3  Antioxidant pathways

         Small antioxidant molecules

         Recent reviews have evaluated the nature and importance of
    small antioxidant molecules in human blood plasma (Stocker & Frei,
    1991), the eye (Spector, 1991) and human skin (Fuchs & Packer 1991;
    Tyrrell 1991).

         Glutathione is a major constituent of lens epithelial cells
    (Rathbun 1989) and these cells have a high capacity for maintaining
    glutathione in the reduced state because of extremely active hexose
    monophosphate shunt activity. This compound may be involved in a
    number of antioxidant reactions that are relevant to UV-mediated
    oxidative stress including detoxification of hydrogen peroxide (as a
    cofactor of GSH peroxidase), detoxification of free radicals,
    reduction of protein disulphides, and competition with protein
    thiols for oxidising species. The role of ascorbate as an effective
    antioxidant in lens is less clear because of the potential of
    ascorbate to be involved in the generation of deleterious components
    (Spector, 1991). Vitamin E is also a potentially useful antioxidant

    but there is no general agreement as to the usefulness of vitamin E
    therapy for cataract.

         Free radical intermediates have been implicated in UV- induced
    carcinogenesis in investigations originally stimulated by the
    isolation of the putative carcinogen cholesterol epoxide from human
    skin (Black 1987). Various components present in skin are potent
    antioxidants including ascorbate, uric acid, carotenoids and
    sulphydrils. Carotenoids have been shown to inhibit UV-induced
    epidermal damage and tumour formation in mouse models (Mathews-Roth
    & Krinsky, 1987). In cell culture models using human skin cells, it
    has been clearly shown that glutathione depletion leads to a large
    sensitization to UVA (334 nm, 365 nm) and near-visible (405 nm)
    wavelengths as well as to radiation in the UVB (302 nm, 313 nm)
    (Tyrrell & Pidoux, 1986,1988). There is a direct correlation between
    the levels of sensitisation and cellular glutathione content.
    Additional evidence that glutathione is a photoprotective agent in
    skin cells is derived from experiments which have demonstrated that
    glutathione levels in both dermis and epidermis are depleted by UVA
    treatment (Connor & Wheeler, 1987).

         Water-soluble antioxidants in plasma include glucose, pyruvate,
    uric acid, ascorbic acid, bilirubin and glutathione. Lipid soluble
    anti-oxidants include alpha-tocopherol, ubiquinol-10, lycopene,
    ß-carotene, lutein, zeaxanthin and alpha-carotene. Since, the long
    wavelengths in sunlight can penetrate through tissue and into blood
    at the longer wavelengths, these defenses may be critical under
    certain circumstances.

         Antioxidant proteins including enzymes

         The major classes of antioxidant enzymes characterized to date
    in eukaryotic cells are superoxide dismutase (which converts
    superoxide anion to hydrogen peroxide), catalase (which destroys
    hydrogen peroxide) and glutathione peroxide and associated enzymes
    (which in addition to metabolizing hydrogen peroxide can also reduce
    hydroperoxides such as those that result from lipid peroxidation).
    Both glutathione peroxide and catalase are present in the lens,
    although the latter is present at low levels and concentrated in the
    epithelial cells (Bhuyan & Bhuyan 1983). Spector (1991) claims that
    these enzymes are present at sufficient concentrations to handle the
    normal levels of hydrogen peroxide generated in the lens. This
    author suggests that the thioredoxin/thioredoxin-reductase system
    may also be involved in the defense of lens against oxidative
    stress. This system can quench free radicals and also reduce some
    protein disulphides.

         All the major antioxidant enzymes are present in skin but their
    role in protecting cells against oxidative damage generated by UV
    radiation has not been elucidated. Acute UV exposures lead to
    changes in glutathione reductase and catalase activity in mouse skin

    but insignificant changes in superoxide dismutase and glutathione
    peroxidase (Fuchs  et al., 1989). Consistent with original studies
    in bacteria, neither endogenous catalase nor superoxide dismutase
    play a major role in protecting cells against the lethal effects of
    UVA irradiation (Tyrrell & Pidoux 1989). Iron plays a critical role
    in oxidative reactions as a catalyst in the Fenton reaction so that
    cellular levels of free iron need to be kept low. The intracellular
    storage protein, ferritin may therefore play a critical role in
    cellular antioxidant defense. UVA radiation (and other oxidant
    stress) leads to high levels of expression of the heme oxygenase 1
    gene (HO1) (Keyse & Tyrrell 1989) which in turn leads to the
    catabolism of heme and release of free iron. The increased ferritin
    levels that result appear to be directly responsible for a
    UVA-mediated adaptive response involving the protection of human
    fibroblast membranes against subsequent UVA radiation damage (Vile
     et al., 1994).

    6.5.4  Summary

         Repair of UV-induced DNA damage is crucial in removing
    potentially mutagenic damage from cells although errors in repair
    can themselves lead to mutations. The repair capacity of human skin
    cells therefore directly relates to the probability of initiation of
    the carcinogenesis process and eventually tumour formation. At
    longer UV wavelengths, an increasing component of oxidative damage
    to DNA, membranes and proteins influences the biological effects.
    Both endogenous and exogenous photosensitisers normally generate
    active oxygen intermediates. Cellular antioxidant defense mechanisms
    are therefore crucial for the prevention or removal of the damage
    caused by the oxidising component of UV radiation.

    6.6  Cellular Consequences of Damage

    6.6.1  Membrane disruption

         Lipid and protein damage by UV associated with cytoskeletal
    damage may lead to severe disruption of plasma membrane functions
    including a breakdown in the permeability barrier and interruptions
    in active transport functions. Crucial signalling molecules such as
    cytokines may be inappropriately released leading to aberrant cell
    to cell communication and toxins present in the external cellular
    environment may gain free access to the interior of the cell.
    Critical ion pumps may be damaged, thereby influencing a wide range
    of processes that rely on ion homeostasis. Breakdown of internal
    lipid membranes in eukaryotic cells will also be highly disruptive
    leading to many pathological consequences including mitochondrial
    damage, leakage of proteases from disrupted lysosomes and breakdown
    of the nuclear membrane permeability barrier. Details in this
    important area of UV effects are sparse but it is clear that
    consequences of UV damage to membrane components need to be better

    understood and incorporated into current models for cellular and
    organ damage.

    6.6.2  Activation of genes

         UV inducible defense pathways in bacteria and human cells

         In bacteria, UVC radiation induces a large set of genes under
    the regulatory control of the  rec A gene which lead to enhanced
    DNA repair, mutagenesis, prophage induction and inhibition of cell
    division (for review see Walker, 1987). Many groups have sought a
    similar SOS response in eukaryotic cells. Reactivation of several
    different UVC damaged viruses (human cytomegalovirus or Herpes
    Simplex virus) has been observed upon infection into UVC treated
    cultured human skin fibroblasts. The effects are not generally large
    (Dion & Hammelin, 1987; Abrahams  et al., 1984). In one study, a
    parallel phenomenon has been observed using split doses of UV in
    arrested cultures of human fibroblasts (Tyrrell, 1984). Cell
    populations irradiated with low doses of UV developed enhanced
    resistance to a second UV challenge dose with a maximum response
    occurring between 2-4 days. The level of reactivation was much
    higher in a repair deficient (XP variant cell line). As for viral
    reactivation, protein synthesis is necessary. No mutation studies
    have been carried out under similar conditions. Furthermore, these
    phenomena have not been investigated in the solar UVB region
    although it is likely that the response will be similar. Although
    these adaptation responses appear to reflect enhanced gene
    expression, there are few clues as to the molecular mechanism
    underlying these observations.

         Bacteria also respond strongly to oxidative stress by either
    the oxyR pathway (Christman  et al., 1985) or the soxRS pathway
    (Greenberg  et al., 1990) which is induced by superoxide-
    generating redox cycling agents such as menadione and paraquat. The
    oxyR pathway is induced by agents such as hydrogen peroxide and the
    regulatory protein encoded by this gene controls expression of at
    least 9 proteins including the antioxidant enzymes, catalase,
    glutathione reductase and alkyl hydroperoxide reductase. This
    pathway is emphasised here because it also appears to be involved in
    protection of bacterial cells against the cytotoxic action of UVA
    radiation (Kramer & Ames 1987). However, although several eukaryotic
    genes are induced by both UVA radiation and hydrogen peroxide (see
    below), none of them involves such a clear-cut direct activation of
    antioxidant enzymes.

         Gene activation in mammalian cells by UVC and UVB radiations

         A large number of genes have been shown to be activated by UVC
    radiation (listed in Table 1 of a review by Keyse 1993). However, in

    many cases the doses required are very high, so that by the
    criterion of colony-forming ability, the majority of the population
    is dead. Although it is a reasonable assumption that the same genes
    will be induced by UVB radiation, this has only been shown in
    certain cases. Damage to DNA may be a critical intermediate in
    triggering the response since induction of certain genes has been
    shown to occur at much lower doses in mutant cell strains lacking
    DNA repair. (Stein  et al., 1989, Miskin & Ben-Ishai, 1981).
    Furthermore, crude action spectra for induction of several genes by
    UV radiation correspond to the action spectra for DNA damage
    induction (Stein  et al., 1989). However, at least for the
    induction of the  fos gene (see action spectrum in figure 6.8),
    events at the membrane also appear to be involved (Devary  et al.,
    1992). There is still a controversy as to whether or not the initial
    signal occurs in the nucleus and is then transduced to the membrane
    or whether the crucial initiating events occur at the membrane
    itself. UVC activation of the HIV-1 promoter also appears to involve
    membrane events (Devary  et al., 1993).

         Metallothioneins are also induced by both UVC and UVB radiation
    (Angel  et al., 1986). However, in general, the relationship
    between UVC/UVB activation of genes and cellular defense against
    short wavelength UV radiations remains obscure. The synthesis of a
    constitutive damage specific DNA-binding protein has been shown to
    be stimulated by UV radiation in both monkey and human cells
    (Hirschfield  et al., 1990). Although the binding activity was
    shown to be absent in certain strains of XP (group E) (Chu & Chang
    1988), subsequent studies showed that not all strains from this
    group lacked the damage specific protein (Kataoka & Fujiwara 1991).

         Gene activation in mammalian cells by UVA radiation

         The eukaryotic genes induced by UVA radiation are, for the most
    part, distinct from those induced by the shorter wavelengths. This
    appears to be largely due to the fact that UVA radiation acts
    indirectly by generating active oxygen intermediates rather than
    being directly absorbed by biomolecules. Although UVA radiation can
    activate antioxidant enzymes such as catalase in prokaryotic cells,
    there is no evidence that UVA radiation can activate any of the
    common antioxidant pathways in mammalian cells. In contrast to the
    studies with UVB/UVC radiations, UVA radiation can activate genes at
    doses which kill only a small percentage of the total population in
    cultured cell models. This is crucial since it indicates that gene
    activation by UVA radiation may actually relate to events in living
    tissue and could reasonably be expected to be involved in protective
    pathways.

    FIGURE 6.8

         For simplicity, we can distinguish three groups of genes that
    are activated in cultured mammalian cells by UVA radiation. These
    may involve secreted molecules in cell to cell interactions:
    collagenase (Scharffetter  et al., 1991) and intracellular adhesion
    molecule 1; proteins involved in early cellular events and signal
    transduction (protein kinase C, phospholipase A2, and a phosphatase,
    see Keyse, 1993) or catabolic enzymes (currently only represented by
    heme oxygenase 1). To date only heme oxygenase 1 has been implicated
    in a cellular defense pathway in which human fibroblasts adapt to
    oxidative damage to membranes (Vile  et al., 1994).

         UV activation of viruses

         UV radiation has long been known to activate Herpes Simplex
    virus. Experimental evidence now indicates that UVC and UVB
    radiation can activate the promoter of the human immune deficiency
    virus (see figure 6.9) as well as complete HIV (Valerie  et al.,
    1988). This activation may contribute to the faster development of
    AIDS in seropositive individuals. Current evidence in both cell
    systems and transgenic animal models have led to the conclusion that
    UVA radiation cannot activate the virus (see Beer & Smudzka, 1991).
    However UVA at solar radiation levels can clearly activate binding
    of the NFKB transcription factors to the corresponding DNA binding
    site which is present in several promoters including the long
    terminal repeat of the HIV promoter (Tyrrell, R.M Personal
    communication). This issue clearly merits further study in view of
    the importance to human health.

    6.6.3  Cell death

         Cells irradiated with UV may show changes in permeability,
    inhibition of macromolecular synthesis, loss of ability to divide
    and total metabolic disruption eventually leading to cellular
    disruption. A discussion of how to define and measure cell death is
    beyond the scope of this overview, but the most commonly accepted
    parameter employed is the loss of the ability of a cell to divide
    and form colonies. Many classical studies of UV effects were based
    on these parameters and survival curves and their interpretation
    have been extensively discussed (Jagger, 1985). Inactivation rates
    can be derived from survival curves and used to construct action
    spectra for inactivation of cultured organisms ranging from viruses
    to human cells. Such survival curves are particularly useful for
    defining sensitive and resistant cell populations, often an
    indication of genetic defects in repair processes. Action spectra
    for inactivation of cell populations generally follow DNA absorption
    at wavelengths as long as 320 nm. However cells are killed more
    efficiently than predicted by DNA damage at longer wavelengths as a
    result of multiple effects of the longer wavelength including
    disruption of DNA repair and the increasing importance of targets
    such as the membrane.

    FIGURE 6.9

         At solar UV levels that lead to erythema and acute skin burn in
    human beings, extensive cell death may occur. This may be
    particularly important to the eventual appearance of melanomas which
    have been linked to severe sunburn early on in life in
    epidemiological studies. It should also be noted that action spectra
    data taken together with solar spectroradiometric measurements and
    the known transmission of human skin have led to the conclusion that
    the UVA component of sunlight is a major factor in the cytotoxic
    action of sunlight at the basal layer of the epidermis (Tyrrell and
    Pidoux 1987).

    6.6.4  Mutation, chromosomal damage and transformation

         Studies in microorganisms have unambiguously shown that damage
    induced by UV throughout the whole range is mutagenic (Webb, 1977).
    Until the present decade, UVB and UVA studies in mammalian cells
    were much less conclusive because most of them employed broad
    spectrum lamps and mutagenic effects could be partially or entirely
    attributed to wavelengths at the short end of the emitted spectrum.
    Nevertheless, a large body of work now supports the conclusion that
    both UVC and UVB radiations are mutagenic to cultured human cells.
    Studies using predominently UVA sources or monochromatic radiations
    in three different cultured human cell systems have reported
    positive (Enninga  et al., 1986), ambiguous (Jones  et al., 1987)
    or negative (Tyrrell 1984) results. These results are clearly a
    function of the system employed and merely underline the fact that
    although pre-mutagenic lesions are induced by wavelengths longer
    than 320 nm in human cells, the efficiency of mutation induction is
    several orders of magnitude lower than that induced by the short
    wavelengths. It should be stressed at this point that UVA radiation
    is clearly carcinogenic in animal models (De Gruijl  et al., 1993).

         Most DNA photoproducts are likely to be pre-mutagenic lesions
    because of the potential for error occurring during repair. Recent
    attention has focused on the relative mutagenicity of the
    cyclobutane type pyrimidine dimers compared with the second most
    common DNA lesion formed by UVB and UVC radiation, the (6-4) pyo
    photoproduct (reviewed by Mitchell  et al., 1993). Most of the
    conclusions are based on the specific type of sequence changes
    (transitions, tandem transitions) that occur in a variety of test
    systems. In certain cases, clear evidence has emerged that the minor
    (6-4) pyo photoproduct may be highly mutagenic relative to the
    pyrimidine dimer (Leclerc  et al., 1991). Much additional study
    will be required to resolve the mutagenic potential of each
    photoproduct, especially considering the rapid expansion in numbers
    of additional photoproducts that are now being recognised.

         Except for UVC radiation, most of the studies of the
    chromosomal effects of UV radiation have been carried out with broad
    spectrum sources. The data has been extensively tabulated in a

    recent evaluation of UV effects (IARC 1992). A similar comment
    applies to the studies on morphological transformation.

    6.7  Conclusions

         DNA is the most critical chromophore and target for damage by
    UVB and UVC radiation. The fraction of oxidative type damage
    involving other chromophores and additional targets increases with
    increasing wavelength. The determination of accurate action spectra
    in cultured cells and animal models is critical to obtaining clues
    as to the nature of these chromophores and for predictive evaluation
    in the many cases in which human data is lacking.

         A considerable amount of knowledge is available concerning the
    interaction of UV radiation with nucleic acids. Controversy still
    exists as to which type of lesion constitutes the most important
    type of pre-mutagenic damage, although (6-4) pyo photoproducts and
    cyclobutane type pyrimidine dimers may both be relevant. Damage to
    membranes and other organelles is also being given increasing
    attention since non-DNA events may also be involved in UV mediated
    biological effects.

         Studies of DNA repair defective disorders in humans have
    clearly established a link between UV induced DNA damage in skin and
    various types of cancer. A clear understanding of endogenous defense
    pathways including antioxidant defense is essential for
    understanding the origins of UV-related human disease and to the
    elaboration of adequate protective measures.

         Cell death, chromosome changes, mutation and morphological
    transformations are observed after irradiation of prokaryotic and
    eukaryotic cells with UV. Many different genes and several viruses
    (including HIV) are activated by UV radiation. However, the genes
    activated by UVC and UVB are different from those activated by UVA
    radiation.

    7.  ANIMAL STUDIES

    7.1  Skin Carcinogenesis

    7.1.1  Domestic animals

         Skin tumours have been reported in some domestic and food
    animals including cats, dogs, cows, sheep and goats (Dorn  et al.,
    1971; Emmett, 1973; Madewell  et al., 1981). That the tumours
    observed often develop in sparsely haired, light-coloured skin
    suggests that sunlight was involved. Cancers of the external
    membranes of the eye are also observed, particularly in cattle, and
    are thought to be related to sun exposure (Russell  et al., 1956).

    7.1.2  Experimental animals

         The experimental induction of skin cancers in mice following
    exposure to a mercury arc lamp was first reported by Findlay (1928).
    Since then, carcinogenicity of UV has been investigated in many
    experiments, mostly in mice, less often in rats, and infrequently in
    other species (Blum et al, 1959; Urbach  et al., 1974; Kripke &
    Sass, 1978; WHO, 1979; van der Leun, 1984; Epstein, 1985).
    Spontaneous skin tumours are rare in rodents but have been
    consistently observed following experimental UV exposure with a
    clear dose-response relationship in well-conducted studies. A recent
    review concluded that there was sufficient evidence that UV caused
    skin cancer in experimental animals (IARC, 1992).

          Broad Spectrum UV

         The carcinogenicity of sunlight was tested in two studies in
    mice and rats (Roffo, 1934; 1939). In the first study of 600 rats,
    165 (70%) of 235 which survived the acute heat load of exposure to
    sunlight for 5 hours a day developed tumours of the skin (mainly
    squamous cell carcinomas on the ears) or conjunctiva (spindle cell
    sarcomas). No tumours developed when sunlight was filtered through
    glass. Similar results were obtained in a second study in 2 000 rats
    and mice.

         "Solar-simulated radiation" has been studied in a number of
    experiments (Forbes  et al., 1982; Staberg  et al., 1983; Young
     et al., 1990; Menzies  et al., 1991). In the study by Forbes  et
     al. (1982), 1 000 hairless mice were exposed for up to 80 weeks to
    radiation from a xenon arc lamp passed through various filters to
    simulate sunlight. More than 90% of the mice developed tumours,
    particularly squamous cell carcinomas.

          UVB

         Many studies have been conducted with sources emitting mainly
    UVB radiation. A few of the more informative studies in different
    species are summarised below.

          Mouse: Freeman (1975) exposed one ear of each haired albino
    mouse in groups of 30, three times a week, to one of four
    wavelengths of narrow band UVB (290, 300, 310 and 320 nm), produced
    by a high-intensity diffraction grating monochromator with a
    half-power band-width of 5 nm, at doses proportional to the MED of
    each wavelength for untanned human skin. Tumours were produced in
    about 50% of surviving animals at each wavelength except 290 nm
    where no tumours were produced although all animals survived. Most
    tumours were squamous cell carcinomas (SCC).

         De Gruijl  et al. (1983) exposed six groups of 22 to 44 (total
    199) male and female Skh-hr 1 hairless albino mice to daily doses of
    from 57 to 1 900 J m-2 of mainly UVB from Westinghouse FS40TL12
    sunlamps. Most of the mice developed skin tumours, mainly squamous
    cell carcinomas, even though the highest daily dose was
    sub-erythemal. A power relationship (linear on log-log scales) was
    observed between the daily dose and the time required for 50% of the
    animals to develop tumours (figure 7.1). SCC developed in 71% of
    mice in the lowest dose group and two skin tumours of different
    types were observed late in the lives of 24 control mice (received
    no UVB radiation).

          Other Species: Stenbäck (1975a) exposed groups of shaven NMR
    rats, Syrian golden hamsters and guinea pigs to mainly UVB from
    Westinghouse FS40T12 sunlamps for 60 weeks at weekly doses of 5.4 to
    10.8 J m-2. Tumours developed in 16 of 40 rats (mainly papillomas
    on the ears), 14 of 40 hamsters (mainly papillomas of skin not on
    the ears) and 2 of 25 guinea pigs.

         A number of experiments were carried out in which groups of a
    South American opossum,  Monodelphis domestica, were exposed to
    mainly UVB from Westinghouse FS40 sunlamps.  M domestica is unusual
    in showing photoreactivation of cyclobutylpyrimidine dimers (Ley,
    1985). Regular exposures of 250 mJ m-2 produced melanomas in 5 of
    13 surviving animals by 100 weeks from commencement of the
    experiment. No melanomas were seen in a much larger colony not
    exposed to artificial UV (Ley  et al., 1989). In other experiments,
    nonmelanocytic skin tumours (mainly fibrosarcomas and squamous cell
    carcinomas) and fibrosarcomas of the corneal stroma were produced
    (Ley  et al., 1987, 1989). The latter were delayed in appearance
    and reduced in number when UV exposure was followed immediately by
    photoreactivating light.

    FIGURE 7.1

         A group of 460 hybrid fish of two strains developed by crossing
    platyfish ( Xiphophorus maculatus) and swordtails ( Xiphophorus
     helleri) were exposed for 5 to 20 days to mainly UVB from
    Westinghouse FS40 sunlamps filtered through acetate sheets
    transmitting > 290 nm (150 or 300 mJ m-2) and >340 nm (850 and
    1700 J m-2). Between 19% and 40% of exposed fish developed
    melanocytic tumours compared with 12% and 2% in control fish (Setlow
     et al., 1989).

         In an attempt to confirm a suspected association between
    sunlight and cancer of the eye in cattle, four Hereford cattle were
    exposed to mainly UVB from Westinghouse FS40 sunlamps. Visible
    tumours developed in the eyes of three of them and one was confirmed
    histopathologically as a preneoplastic growth (Kopecky  et al.,
    1979).

          UVC

         Experiments have been carried out in which animals have been
    exposed to low-pressure mercury discharge germicidal lamps which
    emit most of their radiation at 254 nm with weaker spectral lines in
    the UVB, UVA and visible spectra (IARC, 1992). Examples are given
    below.

         A group of 40 mice was irradiated with germicidal lamps at
    weekly doses of 3 x 104 J m-2. By 52 weeks, 97% of mice had
    developed skin tumours, the majority of which were squamous cell
    carcinomas (Lill, 1983). In another study, groups of 24 hairless
    albino mice were irradiated daily at 230, 1460 or 7000 J m-2. The
    prevalence of tumour-bearing mice (with mainly squamous cell
    carcinomas) increased with time to over 75% in each exposure group
    (Sterenborg  et al., 1988). A comparison with other studies showing
    response to exposure to UVB led the authors to conclude that the
    small amounts of UVB emitted by the germicidal lamps could not
    explain the response observed. Interaction between UVC and UVB could
    not be excluded. Yields of keratoacanthoma like skin tumours,
    proportional to dose, were observed in small groups of rats exposed
    to radiation from Westinghouse G36T6L germicidal lamps (Strickland
     et al., 1979).

          UVA

         A number of studies have been conducted in which skin tumours
    have been produced in hairless mice by UVA under conditions in which
    the exclusion of UVB from the exposure was adequately documented

    (IARC, 1992). For example, groups of 24 albino hairless mice were
    exposed to UVA from Philips TL40W/09 fluorescent tubes filtered
    through 10mm of glass, highly absorbent for UVB, for 12 h/day, seven
    days a week, for one year. The daily dose was 220 kJ m-2. Skin
    tumours appeared in all animals and histopathological examination of
    the larger lesions showed the majority to be SCC. On the basis of
    the known dose-response to UVB, it was estimated that 100,000 times
    more UVB than was residually present in the exposure would have been
    necessary to produce the incidence of tumours observed (van Weelden
     et al., 1986, 1988). Similar findings were obtained in a study in
    which UVA at > 340 nm was produced by passing radiation from
    Philips HPA 400 W lamps through liquid filters (Sterenborg and van
    der Leun, 1990).

    7.1.3  Interactions between radiations of different wavelengths

         Simple photoaddition is commonly assumed for the interaction of
    different wavelengths of UV in producing cancer. That is, exposure
    at each wavelength contributes to the effective dose in an additive
    way. Several studies, however, suggest that the true position is
    more complex and the subject has been reviewed in detail by van der
    Leun (1987, 1992).

         For exposures to UV at different wavelength ranges administered
    simultaneously, or in close temporal proximity, both reductions and
    increases in the carcinogenic effect have been reported by
    comparison with what would have been expected on the basis of simple
    addition. Following detailed review, these effects were described as
    "nonexistent, unproven or small" (IARC, 1992) and this conclusion is
    consistent with the results of a recent study (Berg  et al., 1993).
    Such interactions play only a small role in the evaluation of risks
    of UV (Health Council of the Netherlands, 1986).

         There is a well-established protective effect of visible light
    against UV carcinogenesis (and other effects) in  M. domestica
    (Ley, 1993) and lower animals (see, for example, Setlow  et al.,
    1993) which possesses the photoreactivating enzyme, photolyase.
    Whether or not humans possess a DNA photolyase or show
    photoreactivation is controversial (Ley, 1993, Li  et al., 1993).

         Several experiments, with somewhat conflicting results, have
    been carried out in which exposures to UV to one wavelength range
    have been separated in time from exposures to another (Forbes  et
     al., 1978; Staberg  et al., 1983; Bech-Thomsen  et al., 1988a,
    b; Slaper, 1987). On review of these studies, it was concluded that

    the combined effects tend to be slightly less than what would be
    expected from simple photoaddition (IARC, 1992).

    7.1.4  Dose-response

         The accurate and quantitative description of the relationship
    between UV and the occurrence of skin tumours, usually SCC, has been
    based on experiments in which mice have been exposed regularly,
    usually daily, to UV from standard sources. In most relevant
    experiments, a UV dose, usually much lower than that of the outdoors
    environment, was delivered daily or several times per week until
    skin tumours developed. The UVB dose which induces tumours in mice
    is lower than for acute reactions such as erythema or oedema and up
    to 33 times lower in one experiment which produced an abundance of
    skin cancers (De Gruijl  et al., 1983). The higher the dose given,
    the less time it takes for tumours to appear. In most experiments,
    the time taken for 50% of mice to develop tumours has ranged between
    a few months and one year but can be brought down to as low as 18
    days (IARC, 1992; Willis  et al., 1981).

         Quantitative dose-effect relationships have been derived for
    mice exposed regularly (usually daily) to UV. The median time to
    first tumour (tm) has been used as the measure of effect.
    Dose-effect relationships of the following form have been proposed
    (IARC, 1992):

              tm = k1 D-r

    or, equivalently,

              log tm = -r log D + log k1 .

         In these expressions, k1 is a constant representing both the
    susceptibility of the mouse strain and the effectiveness of the
    radiation spectrum administered, D is the daily dose of radiation
    and r is a numerical exponent giving the slope of the dose response
    curve. Estimates of r vary between 0.5 and 0.6 in most experiments,
    with broad-spectrum UV and broadband UVB; the value of 0.5 is
    typical for large tumours, and 0.6 for small tumours (Blum  et al.,
    1959; de Gruiji, 1983). In one experiment with UVC a value r=0.2 was
    found (Sterenborg, 1988). The relationship found by de Gruijl  et
     al. (1983) for the induction of skin tumours less than 1 mm in
    diameter by UVB in albino hairless mice is shown in figure 7.1; the
    value of r in this relationship is 0.6.

         A corresponding expression for the dose-response relationship
    in albino hairless mice is given by:

              Y = k2 Dc td

    where Y is the average number of tumours per mouse (the yield), k2
    is a constant, D is the daily dose of UV radiation, t is the number
    of days of exposure to UV and c and d are numerical exponents
    describing the dose response relationship (van der Leun & der
    Gruijl, 1993). By describing the response in terms of a measure of
    incidence of tumours rather than time to tumour development, this
    expression may be more useful for risk assessment in humans.

    7.1.5  Effect of pattern of exposure

         Tumours can be induced by a single dose of UV (Hsu  et al.,
    1975; Strickland  et al., 1979). In mice, this has required a dose
    that caused ulceration of the skin which is known to be a tumour
    promoter.

         Two experiments have reported on the effects of pattern of UV
    (mainly UVB) exposure in albino hairless mice. At constant
    instantaneous intensity and constant weekly dose of UV, increasing
    fractionation of the dose from once weekly to three times weekly to
    five times weekly increased the incidence of tumours (Forbes  et
     al., 1981). Similarly, with a constant daily dose of UV, incidence
    of tumours increased with extension of the period of delivery (and
    corresponding reduction in instantaneous intensity) from 1.25 to 4
    or 12 hours (Kelfkens  et al., 1991); there was no appreciable
    difference between exposures of 4 and 12 hours. Similar results had
    been obtained by Forbes and Davies (cited in Kelfkens  et al.,
    1991).

    7.1.6  Action spectrum

         Ideally, to determine the action spectrum of UV carcinogenesis,
    experiments would be carried out with monochromatic radiation to
    determine the relationship between dose and median time to first
    tumour for each wavelength. However, narrow-band monochromatic
    sources suitable for experiments of these types are difficult to
    achieve. An alternative approach is to conduct experiments with a
    number of overlapping broad band sources and to derive an action
    spectrum by appropriate mathematical analysis of the results.

         Recently two large collections of data on carcinogenicity of UV
    in albino hairless mice following exposure to multiple overlapping

    sources in the UVC, UVB and UVA ranges, supplemented with results of
    experiments using highly filtered UVA sources, have been combined to
    produce what is probably the best estimate to date of the action
    spectrum for skin carcinogenesis in any strain of animals (de Gruijl
     et al., 1993). As far as possible, the end point in the
    experimental results used was the appearance of SCC.

         The action spectrum produced is shown in figure 7.2. The upper
    and lower dashed ( - - ) curves result from a sensitivity analysis
    showing the boundaries of the effects of displacements of 5 nm
    segments of the action spectrum that do not increase the chi2 for
    the fit of the spectrum to the data by more than 1.0. Effectiveness
    rises as wavelength increases to a peak at about 295 nm, falls
    steeply to an initial minimum at about 350 nm, rises again to 380 nm
    and then appears to fall away sharply. The definition of this action
    spectrum is very good in the UVB range, there is substantial
    uncertainty in the UVA range and essentially no information below
    280 nm except at 254 nm. Even with the uncertainty, if there is not
    actually a second peak in the UVA range, there is at least a plateau
    from about 340 to 380 nm. This action spectrum is very similar to
    the action spectra for erythema in humans (McKinlay and Diffey,
    1987) and the induction of oedema in the skin of mice (Cole  et al.,
    1986).

         Recently, results have been reported on the action spectrum for
    the production of melanomas in hybrid fish (Setlow  et al., 1993).
    Groups of between 20 and 124 fish were exposed to single doses, at
    two to six exposure levels, of UV or visible light at wavelengths of
    302 nm, 313 nm, 365 nm, 405 nm and 436 nm. Narrow band radiation was
    produced by use of a grating monochromator and, for the higher
    wavelengths, various filters to eliminate any radiation at
    substantially lower wavelengths. The experiment was terminated after
    four months and all fish examined for melanomas which occurred in 5%
    to 24% of control fish and 24% to 45% of irradiated fish. The
    estimated action spectrum obtained is shown in figure 7.3
    superimposed on published action spectra for mammalian cell
    mutagenicity and cytotoxicity. Relative to an effectiveness in
    producing melanomas of 1.0 for UV at 302 nm, the effectiveness at
    313 nm was 0.16 and that at 365 nm, 0.32; for visible light, the
    relative effectiveness were 0.017 at 405 nm and 0.023 at 436 nm. The
    effectiveness of UVA in producing melanoma in this model relative to
    the effectiveness of 300 nm radiation was some 3 orders of magnitude
    greater than the effectiveness of UVA, relative to the same
    baseline, in producing SCC in hairless albino mice (figure 7.2; de
    Gruijl  et al., 1993). An action spectrum has not yet been
    determined for production of melanoma in  M. domestica.

    FIGURE 7.2

    FIGURE 7.3

    7.1.7  Interaction between UV and chemicals

          Interaction with chemical carcinogens

         A number of studies have been carried out in which UV has been
    administered before or after administration of a known chemical
    carcinogen. A period of irradiation with UVB before application of
    3,4-Benzo[a]pyrene to the skin of mice increased the carcinogenic
    response to high-dose 3,4-Benzo[a]pyrene (Gensler & Bowden, 1987;
    Gensler, 1988). A number of studies in which mice were irradiated
    with UV after application of 7,12-Dimethylbenz[a]anthracene (DMBA)
    to the skin showed an increase in tumour production over that
    produced by DMBA alone (Epstein & Epstein, 1962; Epstein, 1965;
    Reeve  et al., 1990; Husain  et al., 1991). Of particular interest
    in this regard is the production of melanoma-like lesions in mice by
    the combination of DMBA and UV. In a number of experiments, DMBA
    alone has produced benign pigmented naevi. Subsequent administration
    of UVB, UVB + UVA, and UVA alone caused these naevi to grow into
    lesions that had the appearance of malignant melanoma; these lesions
    did not develop in mice treated with DMBA alone (Epstein, 1965;
    Epstein  et al., 1967; Husain  et al., 1991). Irradiation with UVB
    from Westinghouse FS40 sunlamps before application of DMBA and the
    promoter 12-O-tetradecanoylphorbol 13-acetate (TPA) reduced the
    incidence and number of tumours produced in CDF1 mice both when
    the site of subsequent application of the chemicals was irradiated
    and when a distant site was irradiated (Gensler and Welch, 1992).
    This effect, therefore, appeared to be mediated systemically.

         Both croton oil and TPA, applied to skin after irradiation with
    UV, have been shown to increase the carcinogenic response to UV
    (Epstein & Roth, 1968; Pound, 1970; Stenbäck, 1975b; Strickland  et
     al., 1985). Of two other, suspected cancer promoters, one, methyl
    ethyl ketone peroxide increased tumour production (Logani  et al.,
    1984) and the other, benzoyl peroxide, did not (Epstein, 1988;
    Iversen, 1988).

          Interaction with other chemicals

         In several experiments, mice exposed to either solar- simulated
    radiation or UVA and topical application of 5- methoxypsoralen
    showed higher incidences of skin papillomas or SCC than did mice
    exposed to UV or 5-methoxypsoralen alone (Zadjela & Bisagni, 1981;
    Cartwright & Walter, 1983; Young  et al., 1983). Methoxsalen
    (8-methoxypsoralen) shows the same effects when given with UV (IARC,
    1980).

    7.1.8  Mechanisms of UV carcinogenesis

          DNA damage

         There is abundant evidence that UV causes DNA damage, both by
    direct photochemical effects (for example, the production of
    cyclobutylpyrimidine dimers and 6-4 photoproduct) and by oxidative
    effects leading to DNA strand breakage and DNA-protein cross-linkage
    (see Chapter 6).

         There is evidence that UV-induced DNA damage, and particularly
    the formation of pyrimidine dimers, is one step in the mechanism
    whereby UV can cause cancer in experimental animals. First, Hart  et
     al. (1977) showed, in the clonal fish,  Poecilia formosa, that
    the transplantation of UVC irradiated thyroid cells from one fish to
    another induced thyroid tumours in the host fish. If irradiation of
    the thyroid of the donor fish was followed immediately by
    application of photoreactivating light (which monomerizes the
    pyrimidine dimers), the development of thyroid tumours was almost
    completely prevented. Second, and similarly, Ley  et al. (1991)
    showed that exposure to photoreactivating light, after exposure to
    UV, delayed and reduced the yield of skin tumours produced in  M.
     domestica by irradiation with Westinghouse FS40 sunlamps. The same
    result was obtained for corneal tumours but, unexpectedly, exposure
    to photoreactivating light was as effective in reducing
    carcinogenesis when given immediately before UV as when given after.
    There is evidence that photoreactivating light may also inhibit the
    production by UV of melanocytic lesions in the skin of  M. domestica
    (Ley  et al., 1989). Third, Yarosh  et al. (1992) applied
    liposomes containing T4 endonuclease V, an enzyme that specifically
    repairs cyclobutyl pyrimidine dimers, to the skin of albino hairless
    mice, three times a week, after they had been irradiated with
    Westinghouse FS20 sunlamps. The incidence of skin tumours observed
    in the mice fell in proportion to the quantity of active liposomes
    applied.

         There is also evidence that activated oncogenes and a mutated
    tumour-suppressor gene are present in some skin cancers induced
    experimentally by UV. Husain  et al. (1990) reported studies of
    expression of the cHa-ras oncogene in three papillomas and three
    carcinomas from among lesions that developed in 17 of 90 Sencar mice
    following a single exposure, at 7 x 104 J m-2, to UVB from
    Westinghouse FS20 sunlamps. RNA preparations from all six tumours
    showed elevated levels of Cha- ras-specific messenger RNA
    sequences, suggesting overexpression of this oncogene. DNA from the
    carcinomas, but not the papillomas, was able to induce
    transformation of NIH-3T3 cells which demonstrated overexpression
    and amplification of the Cha-ras oncogene. Subsequently, in their
    studies of melanocytic tumours produced by DMBA and UV in hairless
    mice (Skh-hr 2), Husain  et al. (1991) found mutations in codon 61
    of the N- ras oncogene in three of eight precursor naevi and one

    melanoma. The base transitions, however, were not of a type to
    suggest that they had been caused by UV. Kress  et al. (1992)
    screened exons 5 to 8 of the p53 tumour-suppressor gene for mutation
    in 35 epidermal tumours induced in four strains of mice by UV from
    Westinghouse FS40 sunlamps. Mutations were found in seven tumours.
    All mutations occurred at dipyrimidine sequences. C -> T and CC ->
    TT transitions were present in five of the seven tumours with
    mutations, strongly suggesting that these mutations were due to UV.

          Oxidative processes

         Oxidative processes in the skin may also mediate the
    carcinogenicity of UV, particularly UVA (Morlière  et al., 1992),
    in skin, either by direct oxidative damage to DNA or by way of
    potentially carcinogenic intermediates, such as cholesterol-5,6-
    epoxide (Black, 1987; Morin  et al., 1991). A number of studies
    have shown that relative increases in dietary polyunsaturated fatty
    acids, which may lead to lipid peroxidation, can increase the tumour
    response to exposure to UV in experimental animals and that this
    effect can be inhibited by the simultaneous feeding of antioxidants
    (Reeve  et al., 1988; Black  et al., 1992). In addition, the
    feeding of antioxidants, such as tocopherol, can reduce the tumour
    response to UV in the absence of any manipulation of dietary fat
    (Black & Chan, 1975; Gerrish & Gensler, 1993).

          Immune-suppression

         UV-induced immune-suppression appears to play a major role in
    UV carcinogenesis in mice. A review of the literature on the
    association between UV-induced immune suppression and increased
    susceptibility to tumours may be found in section 7.3.2.

         There is evidence also that administration of immunosuppressive
    drugs will increase the carcinogenic response to UV in experimental
    animals (IARC, 1992). Experiments have been carried out with
    anti-lymphocyte serum, azathioprine, cyclophosphamide, cyclosporine
    and 6-mercaptopurine. Consistent increases in the carcinogenic
    response to UV over two or more experiments were seen with
    azathioprine and cyclosporine (Reeve  et al., 1985; Nelson  et al.,
    1987; Servilla  et al., 1987; Kelly  et al., 1987, 1989).

    7.1.9  Conclusions

         From a review of the animal studies one can conclude the
    following about UV exposure of the skin:

         Carcinogenesis by sunlight is a widespread phenomenon among
    domestic and food animals. Most tumours found are squamous cell
    carcinomas (SCC). In experimental animals skin cancer is mainly
    caused by UV radiation. Again, most tumours induced are SCC. For the

    induction of SCC in albino hairless mice, the effectiveness is a
    function of wavelength:

    -    the effectiveness is found to peak in the UVB range;
    -    UVA is also carcinogenic, but at a much lower level of
         effectiveness - similar to what is found for erythema and
         tanning;
    -    the effectiveness in the UVC range is unknown, except for one
         wavelength, 254 nm; at that wavelength the effectiveness is
         lower than that in the UVB peak
    -    it is still unknown if there is any effectiveness of visible
         light.

         For the induction of SCC in hairless mice by daily exposures to
    UVB the dose effect relationship was found to be a power function.
    It showed no indication of a threshold dose, even at doses as low as
    3% of that required for an acute reaction, such as oedema.
    Radiations of different wavelengths can cooperate in the induction
    of SCC.

         Melanomas are much less common among animals. Only two animal
    models have been found where exposure to UV induced melanomas:

         -    the South American opossum  M. domestica, and

         -    a hybrid fish, derived from the swordtail and platyfish.

         An initial action spectrum was determined for the hybrid fish.
    It peaks in the UVB range but also shows a high effectiveness in the
    UVA. An action spectrum for the induction of melanoma in  M.
     domestica has not been determined.

         Basal cell carcinomas are uncommon in animals. Suitable animal
    models for the induction of BCC are not available.

    7.2  Immune Responses

    7.2.1  Immune function assays

         The immune system depicted in figure 7.4 is composed of primary
    lymphoid organs (bone marrow and thymus), secondary lymphoid organs
    including spleen and lymph nodes, and several cell types. In
    addition a number of mediators including cytokines, antibodies and
    complement regulate and/or are produced by the immune system. A
    number of studies have shown that exposure to UV suppresses contact

    hypersensitivity (CHS) and other delayed-type hypersensitivity (DTH)
    responses of the immune system. Both of these responses provide a
    measure the competence of T lymphocytes, suppression of which would
    be expected to compromise host resistance to infectious agents such
    as viruses and mycobacterium. Unless otherwise indicated animal
    studies have utilized FS sunlamps (see Chapter 3 for spectral
    characterization) and exposed shaved skin.

         Two types of experiments have been described in the literature.
    In experiments designed to investigate local immune suppression,
    mice were exposed to UV and a hapten chemical such as
    dinitrofluorobenzene (DNFB) was applied to the site of irradiation.
    After an incubation period of several days, animals were challenged
    by painting the chemical on the ear, and the immune response was
    assessed by measuring ear thickness, which is an expression of CHS.
    Mice exposed (usually for 4 successive days prior to chemical
    application) to doses of UV, which might cause a minimal erythemal
    response in an untanned human, failed to show ear swelling (Toews
     et al., 1980; Noonan & De Fabo, 1990; Jeevan  et al., 1992a). The
    failure to respond appeared to be due to the development of
    suppressor T cells which rendered the mouse tolerant to the
    particular antigen. This concept was supported by the observation
    that the same mice could not be sensitized with the same chemical
    through unirradiated skin 14 days later. Moreover, antigen-specific
    unresponsiveness could be adoptively transferred to naive mice by
    spleen and lymph node cells obtained from mice skin-painted with
    DNFB through UV-irradiated skin (Elmets  et al., 1983). It should
    be noted that the precise nature of the activity of these suppressor
    cells remains the subject of some debate.

         In a second type of experiment the effects of UV on systemic
    immune responses were demonstrated by irradiating mice at one site
    and subsequently (3-5 days later) exposing them to a contact
    allergen (CHS) or injecting a protein antigen (DTH) at an
    unirradiated site. These mice failed to respond when subsequently
    challenged on the ear (CHS) or in the footpad (DTH) with the same
    antigen (Noonan  et al., 1981a, Kim  et al., 1990; Ullrich  et
     al., 1986a; Ullrich, 1986; Jeevan & Kripke, 1990; Howie  et al.,
    1986; Denkins  et al., 1989; Giannini, 1986a,b). Again
    unresponsiveness was attributed to the development of antigen
    specific "suppressor T cells" (Noonan  et al., 1981b; Ullrich
    1985). These experiments suggest that immune suppression could occur
    even when the site of entry for the antigen was not the same as the
    site of UV exposure.

    FIGURE 7.4

         Noonan  et al. (1981a) reported a 50% suppression of CHS in
    BALB/c mice after a single dose of 2 KJ m-2 UV. She also reported
    that there was no difference in the dose response curves for
    UV-induced local and systemic immunosuppression of CHS responses;
    however, the dose of UV required to suppress CHS in C57BL/6 mice was
    6.4 times less than that required to produce similar immune
    suppression in BALB/c mice (Noonan & De Fabo, 1990). Also, while
    local suppression could be demonstrated in mice sensitized
    immediately after irradiation, systemic suppression occurred only if
    the sensitizer was applied 3 or more days after radiation (Noonan &
    De Fabo, 1990).

         Fifty percent suppression of CHS and DTH has been reported at
    doses ranging from 2.3 kJ m-2 to 20 kJ m-2, (Kim  et al., 1990;
    Jeevan & Kripke, 1990; Jeevan  et al., 1992b) levels which produce
    minimal or no edema. The dose needed to produce 50% suppression
    varies depending on the strain of mouse and type of antigen used,
    and on the laboratory producing the data. Suppression of
    virus-specific DTH was observed in mice exposed to a suberythemal
    dose of UVB 3-7 days prior to infection with Herpes simplex virus
    (HSV) but not 14 days and persisted for at least 3 months after
    irradiation (Howie  et al., 1986). Elimination of wavelengths below
    315 nm with a mylar filter either eliminated or greatly reduced the
    suppression of CHS following exposure to UV indicating that most if
    not all the suppression is due to UVB (Noonan  et al., 1981b;
    Noonan and De Fabo, 1990). Using narrow bands of UV at 10
    wavelengths from 250 to 320 nm, it was reported that maximum
    suppression of CHS occurred between 260 and 270nm, there was a
    shoulder in the action spectrum from 280-290 nm and then a steady
    decline to 3% of maximum at 320 nm (De Fabo and Noonan, 1983).

         There are two subpopulations of T helper cells, designated Th1
    and Th2 which appear to be differentially affected by UV exposure.
    These two populations are thought to regulate different sets of
    immune responses. Th1 cells produce interleukin (IL) 2 and gamma
    interferon (gamma IFN) as well as other cytokines, promote
    delayed-type hypersensitivity (cell mediated, type IV) responses
    such as CHS, provide help for certain antibody subtype responses
    including complement-fixing antibodies, activate macrophages, and
    may be particularly important for dealing with antigens expressed on
    cell surfaces, such as viral and tumour antigens (Coffman  et al.,
    1988). Th2 cells produce a different array of cytokines including
    IL-4 and IL-5 which promotes antibody responses. Only Th2 cells can
    stimulate a primary IgE response which is mediated by IL-4 and
    inhibited by gamma IFN (Coffman  et al., 1988). Thus, Th2 cells may
    be particularly important in responding to certain parasitic
    infections, and also play an important role in immediate-type
    hypersensitivity including reactions to common allergens such as
    pollen and dust mite.

         Recent studies have suggested that both local and systemic
    immune suppression induced by UV may be the result of an inability
    to present antigen to and hence activate Th1 cells (Simon  et al.,
    1990, 1991; Araneo  et al., 1989).

         Suppression of systemic CHS and DTH responses similar to that
    observed following UV exposure have been obtained by injecting mice
    with supernatant from UV-irradiated keratinocytes (Kim  et al.,
    1990; Rivas & Ullrich, 1992; Jeevan  et al., 1992c). Among factors
    derived from keratinocytes, the cytokines tumour necrosis factor
    alpha (TNF alpha) (Yoshikawa & Streilein, 1990; Vincek  et al.,
    1993) and IL-10 (Rivas & Ullrich, 1992) are thought to be important
    in UV-induced immunosuppression. Also, cis-urocanic acid, the
    product of UV isomerization of urocanic acid located in the stratum
    corneum, when injected subcutaneously or painted on the epidermis,
    produces immunosuppressive effects which mimic UVB-induced
    immunosuppression (Ross  et al., 1986), and the action spectrum of
    UV-induced immune suppression closely follows the absorption
    spectrum of urocanic acid (De Fabo and Noonan, 1983). In all three
    cases it is thought that these mediators act by modifying antigen
    presentation (Vermeer & Streilein, 1990; Rivas & Ullrich, 1992;
    Noonan  et al., 1988).

         In summary, it appears that UVB causes the release of mediators
    from the skin which alter the antigen presenting capability of
    Langerhans cells as well as antigen presenting cells at other sites,
    resulting in the development of "suppressor T-cells". It may be that
    these suppressor T cells are Th2 cells. The net result is the
    failure to activate Th1 cells and suppression of DTH responses
    thought to play an important role in host defences against certain
    types of tumours and microbial infections. The immune suppression is
    antigen specific, (i.e. only responses to antigens administered
    within 7 days after irradiation are affected) and is long lasting
    (at least 3 months).

         In addition to DTH and CHS responses, UV affects several other
    immune functions frequently included in the standard protocols for
    immunotoxicity testing. Spleen cells taken from UV-exposed mice
    (single exposure, 54 kJ m-2) failed to respond in a mixed
    lymphocyte response assay  in vitro (Ullrich, 1985), and spleen
    cells from tumour-implanted UV-treated (10 kJ m-2, 3 times/wk, for
    3 months) mice were not cytotoxic  in vitro against UV-induced
    tumours (Fisher & Kripke, 1977). In contrast, exposure to
    approximately 2-3 kJ m-2 day-1, for 23 days did not affect
    responses of spleen or inguinal lymph node cells to the mitogens,
    phytohaemagglutinin, concanavalin A or bacterial lipopolysaccharide.
    Similarly, lymphocyte responses to these mitogens and peritoneal
    macrophage phagocytic and tumouricidal activities were unaffected by
    exposure to 10 kJ m-2, 3 times per week, for up to 6 months
    (Funnell & Keast 1985; Norbury  et al., 1977).

         Kripke  et al., (1977) reported no effect of UV (in mice
    exposed to 10 kJ m-2, 3 times per week for up to 6 months) on the
    primary haemagglutonin antibody response to sheep red blood cells;
    however in another study, the IgM and IgG plaque forming cell
    responses of lymph node cells to sheep red blood cells given
    intradermally (through irradiated skin) were suppressed (Funnell &
    Keast, 1985). Also, suppressor cells generated by irradiating mice
    with a single exposure of 30-40 kJ m-2, 5 days prior to
    sensitization with trinitrochlorobenzene prevented the development
    of hapten specific antibody-forming cells when injected
    intravenously along with hapten conjugated sheep red blood cells
    into syngeneic recipients (Ullrich  et al., 1986b).

         UV exposure enhanced poly I.C augmented natural killer cell
    (NK) activity in mice (Lynch & Daynes, 1983); however, suppression
    of NK activity in rats has been reported following exposure to
    suberythemal doses of UV (Garssen  et al., 1993). Hence, in
    laboratory animals UV exposure appears to affect mixed lymphocyte
    responses, cytotoxic T cell activity, possibly NK activity, and in
    some circumstances antibody responses. With the exception of NK
    activity all of these responses are antigen specific responses. In
    contrast UV does not appear to alter several non-specific responses
    including macrophage phagocytic or tumouricidal activities, or T or
    B cell responses to mitogens. However, unlike CHS and DTH, none of
    these responses has been studied in detail in rodents, and dose
    response information is not available.

    7.2.2  Susceptibility to tumours

         Interest in the immunosuppressive properties of UVB was first
    sparked by the observations of Kripke and associates that UV
    exposure induced highly antigenic tumours which did not grow when
    transplanted into syngeneic, immunocompetent mice but did grow when
    transplanted into immunosuppressed or UV treated mice (Kripke 1974;
    Kripke & Fisher, 1976). The precise nature of the tumour antigens
    has not been defined. Mice treated with 1.8 kJ m-2, 3 times per
    week, for 3 months were unable to reject UV-induced syngeneic
    tumours when challenged subcutaneously at any time from 2 weeks into
    the irradiation exposure regimen to as late as 5 months after the
    end of UV treatment (Kripke & Fisher, 1976). Hence susceptibility to
    tumour challenge was detectable well before the appearance of
    primary skin cancers induced by the UV and persisted long after
    exposure ended. Similarly, more UV-induced fibrosarcoma tumour
    colonies were detected in the lungs following i.v injection of
    tumour cells in mice treated with 7 kJ m-2, 3 times per week, for
    5 weeks (Kripke & Fidler 1980) as compared to unirradiated controls.
    Susceptibility to tumour challenge was directly proportional to dose
    of UV and a dose fractionated over time was no more effective than
    the same total dose given as a single treatment (Kripke & Fidler
    1980; De Fabo and Kripke 1979).

         In BALB/cAnN mice 50% tumour incidence following subcutaneous
    tumour challenge at an unirradiated site occurred at a dose of
    approximately 40 kJ m-2. Enhanced tumour incidence was observed in
    mice exposed to 21.6 kJ m-2 as much as 32 weeks before tumour
    challenge. A number of studies (Fisher & Kripke, 1977; Daynes &
    Spellman, 1977; Daynes  et al., 1977; Spellman & Daynes, 1978;
    Spellman  et al., 1977; Ullrich & Kripke, 1984, de Gruijl and van
    der Leun 1982, 1983) have demonstrated that susceptibility to
    challenge with UV-induced tumours following UV exposure, like
    suppression of CHS and DTH, is mediated by antigen specific
    "suppressor T cells" which are present in spleen and lymph nodes of
    UV-irradiated mice. As with suppression of DTH and CHS responses,
    enhanced susceptibility to UV-induced tumours is antigen specific
    and long lasting.

         While it is very clear that UV increases the susceptibility of
    mice to UV-induced tumours, much less work has been done on the
    effects of UV on susceptibility to other types of tumours. UV
    treated mice that were unable to reject syngeneic UV- induced
    tumours were able to reject 2 types of non-UV-induced tumours, B16
    melanoma tumours and a spontaneous mouse leukaemia tumour (Kripke
     et al., 1977). Roberts and Daynes (1980) reported that mice
    irradiated with subcarcinogenic doses of UV for 3 weeks prior to
    treatment with either benz[a]pyrene or methycholanthrene at an
    unirradiated site had reduced latency periods for the development of
    these chemically-induced tumours. Also, chemically induced tumours
    from these UV-treated mice appeared to be more antigenic than those
    induced in untreated mice in that they were incapable of progressive
    growth when transplanted into normal (immunocompetent) syngeneic
    mice but were capable of progressive growth in UV-treated mice. The
    authors suggested that decreased selective pressures exerted by the
    host (due to UV-induced suppressor cells) at the time of tumour
    induction allowed more immunogenic tumours to emerge and progress.

    7.2.3  Susceptibility to infectious disease

         Several types of infectious disease models have been developed
    to study the effects of UV exposure: 1) the infectious agent was
    injected through irradiated skin, 2) exposure to the infectious
    agent occurred at a site distant from the site of irradiation, or 3)
    in the case of some herpes simplex virus (HSV) studies, exposure to
    UV occurred after infection. The first exposure scenario is
    representative of vector-borne infections and some vaccinations
    while the second exposure regimen is representative of infections
    which do not necessarily enter the host via the skin. The third
    scenario was designed to mimic reactivation of HSV in humans
    following sun exposure. In all cases the focus has been on microbial
    agents that are controlled, at least in part, by DTH responses.

         An example of the first type of model is that of  Leishmania
     major infection in mice (Giannini 1986a, 1986b, 1987, 1992). Mice

    were irradiated only on the tail with suberythemal doses (0.06-6 kJ
    m-2), 3 times/week, for 1 month and were infected intradermally
    through the irradiated surface of the tail 24 hr after the first UV
    exposure. In UV-treated mice the DTH response to  L. major antigens
    2 and 6 weeks post infection was suppressed, the number of organisms
    recovered from skin at the injection site was (Giannini, 1986a) and,
    a larger number of parasites was observed in the draining lymph node
    (Giannini, 1987, 1992). Finally, mice which were infected through
    irradiated skin failed to develop protective immunity such that
    lesions following reinfection at an unirradiated site were
    significantly larger when compared to lesions of previously infected
    but unirradiated mice (Giannini, 1986b).

         In a similar model mice treated with 13 or 33 kJ m-2 for 4
    consecutive days and infected intradermally with HSV at the site of
    irradiation had a higher incidence of zosteriform lesions than
    unirradiated mice, and at the higher exposure level 100% mortality
    was observed as opposed to no mortality in unirradiated mice
    (Yasumoto  et al., 1987). A decreased DTH response to viral antigen
    was observed and appeared to result from the induction of suppressor
    T cells (Yasumoto  et al., 1987, Aurelian  et al., 1988). Also,
    following  in vitro exposure to UV, a defect in the ability of
    Langerhans cells to present HSV antigen to lymphocytes was
    demonstrated (Hayashi & Aurelian, 1986).

         In contrast, no effect of UV on either parasite specific DTH
    responses or recovery of parasite from internal organs was observed
    following percutaneous injection of  Schistosoma mansoni through
    UV-irradiated (0.4 kJ m-2, 4 consecutive days) skin (Jeevan  et
     al., 1992a). In this same study, suppression of microbe specific
    DTH responses were not observed following intradermal injection of
     Mycobacterium bovis (BCG) or subcutaneous injection of  Candida
     albicans at the site of irradiation; however, the number of viable
    mycobacteria recovered from the lymphoid organs of BCG-infected mice
    was increased significantly in the UV treated mice for a period of
    more than 2 months post infection.

         BCG has also been used as an infectious disease model to
    illustrate systemic effects of UV immunosuppression. In this model
    mice were irradiated on the back and injected with BCG
    subcutaneously in the footpad. Mice exposed from 1 to 15 times (3
    times/week for up to 5 weeks) to one minimal erythemal dose (2.25 kJ
    m-2) showed significant suppression in their DTH response to
    tuberculin (PPD) and increased numbers of live bacteria in the
    spleen and lymph node compared to unirradiated controls (Jeevan &
    Kripke, 1990). However, when exposures were continued beyond 5
    weeks, the DTH response recovered and mice challenged with bacteria
    at that point did not exhibit increased numbers of organisms in the
    spleen and lymph node, suggesting that eventually some sort of
    adaptation to exposure occurs. Significant suppression of DTH was
    observed in mice which received a single dose as low as 1.4 kJ m-2

    3 days prior to infection and significant increases in bacteria in
    spleen and lymph node were observed in mice which received as little
    as 0.7 kJ m-2. Similar effects were observed when supernatants
    from keratinocyte cultures exposed to UV  in vitro were injected
    intravenously 3 days prior to infection (Jeevan  et al., 1992c).

         In a similar model, mice treated with a single high UV dose (45
    kJ m-2) 3 days before infection with  Mycobacterium lepraemurium
    exhibited significant suppression of DTH responses to mycobacterial
    antigen 3 and 6 months after infection and had significantly more
    bacteria in the infected footpad, lymph node, and spleen 3-6 months
    post infection (Jeevan  et al., 1992b). This high dose also reduced
    the median survival time in mice infected i.v. With a lower exposure
    dose (2.3 kJ m-2) 50% suppression of the DTH response to
    mycobacterial antigen was observed 3 months post infection, and
    increased numbers of bacteria were observed in the footpad, spleen
    and lymph node of mice exposed to UV doses greater than or equal to
    5.6 kJ m-2. When mice were treated with 2.25 kJ m-2, 3
    times/week from 3-15 times, the DTH response to  M. lepraemurium
    was suppressed, but as with BCG, the DTH response was normal in mice
    which received more than 15 exposures.

         Effects of systemic UV-induced immune suppression have been
    demonstrated in mice infected with HSV, 3 days after irradiation
    with 1 kJ m-2, at a site distant from the site of infection. In
    this model suppression of DTH was mediated by modulation of
    epidermal antigen presenting cells and the development of suppressor
    T cells (Howie  et al., 1986, 1987). Subcutaneous injection or
    epidermal application of UV-treated urocanic acid produced similar
    immune suppression (Ross  et al., 1986). Effects of UV on the
    actual progression of disease in this model have not been reported.

         While most of the work to date has been done in mice, recent
    work in the rat suggests that suberythemal doses of UV also suppress
    immune responses. In addition rats exposed to suberythemal doses of
    UV had higher levels of microorganisms in target organs following
    oral infection with  Trichinella spiralis or intraperitoneal
    infection with cytomegalovirus (Garssen  et al., 1993).

         In summary, exposure to suberythemal doses of UV has been shown
    to exacerbate a variety of infections in rodent models. Both
    infections which are initiated at the site of UV exposure and
    infections initiated at distant sites have been affected. While most
    of the infections test, have evolved in some manner more recent work
    indicates that systemic infections without skin involvement may also
    be affected. Enhanced susceptibility appears to result from
    suppression of T-helper-1 cell activity. The mechanisms associated
    with this suppression appear to be the same as those identified in
    association with suppression to CHS responses.

         Reactivation of latent HSV infections can be induced in mouse
    and guinea pig by exposure of the previously infected site to sub-
    or minimal erythemal doses of UV. However, the role that immune
    suppression plays in this reactivation has not been established
    (Blyth  et al., 1976; Norval  et al., 1987; Laycock  et al.,
    1991).

    7.2.4  Susceptibility to immunologically-mediated
    diseases

         Several studies have suggested that stimulation of Th2 cells
    may remain intact or be enhanced in UV-exposed mice (Simon  et al.,
    1990; 1991; Araneo  et al., 1989). Since Th2 cells are the only
    cells that can stimulate a primary IgE response (Coffman  et al.,
    1988), it is possible that UV exposure may increase the risk of
    immediate-type hypersensitivity. This possibility has not yet been
    studied, but deserves attention in light of the increasing incidence
    of morbidity and mortality due to asthma, which is often triggered
    by allergic responses (NIH, 1991; Danielle, 1988).

         Acceleration of autoimmunity following UV exposure has also
    been reported in mice. Both acute (2 hr/day, for 7 days) and chronic
    (3 hr/wk, for 4 wks) exposure to 20 J m-2 s-1 enhanced mortality
    in an autoimmune strain of mice (NZB X BZW F1). These exposures
    also caused increased serum antibodies to single stranded DNA,
    enhanced polyclonal B-cell activity in the spleen, and caused more
    severe renal glomerular inflammatory changes (Ansel  et al., 1985),
    all hallmarks of autoimmunity. It is unclear whether there is a
    relationship between these reactions and the effects of UV on
    immunological processes described above.

    7.2.5  Conclusions

         Exposure of mice to UV radiation impairs certain immune
    responses, enhances susceptibility to a variety of infectious agents
    and UV-induced tumours. While fewer studies have been done in rats,
    similar effects have been observed. Suppression of contact and
    delayed type hypersensitivity responses have been studied most
    extensively; however, several other cell mediated responses are also
    affected. Suppression of these immune responses appears to be
    mediated by release of soluble mediators from UB exposed skin which
    alter antigen presentation by Langerhans (and other) cells such that
    they fail to activate TH 1 cells. The resulting immune suppression
    is antigen specific, can occur regardless of whether antigen is
    applied at the site of exposure or not, and is relatively long
    lasting.

         Enhanced susceptibility to a variety of infections in mice and
    rats corresponds to suppression of DTH responses to microbe specific
    antigens. UV exposure also prevents the development of protective
    immunity to these infections.

    7.3  Ocular Studies

    7.3.1  Introduction

         The large number of animal studies enhances our understanding
    of both acute and delayed ocular effect from UV exposure (Zigman
    1993; Andley 1987; Dillon  et al., 1990). Studies of the action
    spectrum of delayed or permanent effects such as cataract or
    retinopathy and UV damage mechanisms in the cornea, lens and retina
    are only possible in animal models.

    7.3.2  General effects

         Photokeratitis is an acute reversible radiation-induced injury
    of the corneal epithelium. It is analogous to acute sunburn of the
    skin. Animal studies have clearly demonstrated that exposure to an
    artificial source of UVB can lead to acute photokeratitis, and
    action spectra have been determined (Cogan & Kinsey, 1946; Pitts  et
     al., 1977; Zuclich and Kurtin, 1977). Pitts (1974; 1978) estimated
    the mean threshold of UVB (290-315 nm) for photokeratitis was 35 J
    m-2 at 270 nm.

         Damage to corneal endothelium has been reported in rabbits
    (Doughty & Cullen, 1990). The effects on the corneal stroma have
    also been observed in rabbits, where reversible damage to the
    stromal keratocytes has followed exposure to UV (Ringvold &
    Davangar, 1985). This histological study showed that keratocytes
    disappear following UV exposure and later reappear. More recent
    animal studies have reported the use of a UVB filtering contact lens
    or the application of UVB absorbing chromophores to the cornea of
    rabbit eyes protects against photokeratitis (Bergmanson  et al.,
    1988; Oldenburg  et al., 1990).

    7.3.3  Cataractogenesis

         There is substantial evidence that exposure to UV induces
    discrete opacities in the anterior lens of experimental animals.
    Several studies have shown that UVB, but not UVA, induces anterior
    opacities in animals (Bachem, 1956; Pitts  et al., 1977; Jose &
    Pitts, 1985; Söderberg 1990). Supra threshold doses of UVB at 300 nm
    cause a Na-K shift between the lens and the surrounding (Söderberg
    (1991). This leads to swelling and disruption of lens cells,
    probably causing zones of deviating refractive index within the lens
    observed as opacities (Söderberg 1989). Anterior lens opacities have
    developed in albino mice after daily exposure for 1-2 months to
    mixed UVA and UVB source (290-400 nm), but not when the source was
    filtered to remove radiation <320 nm (Jose & Pitts, 1985). Anterior
    cataract have also been produced in young albino mice exposed to
    black light (predominantly UVA, with some UVB) (Zigman  et al.,
    1974). With more prolonged exposure of mice, cortical and posterior
    subcapsular opacities have been induced (Zigman  et al., 1975;

    Jose, 1986). By contrast, a dose of 2 kJ m-2 applied every day
    produced anterior polar cataract within 8 weeks. Extension of the
    irradiation to 5 months created an opacity in the deeper cortex
    (Weager  et al., 1989). In testing possible cataractogenic efficacy
    of UVA and UVB, UVB alone was cataractogenic, whereas, UVA showed
    the same effect only when other cataractogenic factors were added
    (e.g. X-ray) (Schmitt  et al., 1988).

         Ham  et al. (1989) irradiated rhesus monkeys to 1 mW cm-2
    daily for three years to UVA radiation (UVB radiation was carefully
    excluded) and was unable to detect lenticular opacities.

    7.3.4  Retinal effects

         It is well accepted that short wavelength visible non-coherent
    and coherent light, as well as UVA, can cause photochemical,
    mechanical and thermal damage to the retina and the pigment
    epithelium (review: Ham  et al., 1984; Organisciak & Winkler,
    1993). Aphakic monkeys revealed retinal lesions after exposure to
    UVA (Guerry  et al., 1985). Exposure of aphakic albino rats to UVA
    created acute retinal lesions with 50-80% more effectiveness than
    blue light (Rapp & Smith 1992). Mechanical and thermal lesions are
    almost entirely limited to exposure to high energy laser light.
    Photochemical damage can occur through two different light exposure
    regimes: short term, high irradiance (minutes to hours) and
    long-term (days to weeks) low irradiance (Kremers & van Norren,
    1988; Remé  et al., 1991; Rapp & Smith, 1992). Whereas damage
    thresholds differ depending on the animal species, basic mechanisms
    at the molecular and structural level are essentially the same (for
    review: see Organisciak and Winkler, 1993). Photochemical lesions
    are strongly dependent on wavelength, i.e. the absorbing chromophore
    and its quantum efficiency, obey, at least within a certain time
    frame, the reciprocity of intensity and exposure duration (e.g.
    Kremers & van Norren, 1988).

         Action spectra for high irradiance short exposure duration
    lesions have been obtained in several studies (e.g. Ham  et al.,
    1984). Action spectra for prolonged exposure to relatively low light
    levels are less clear and were first evaluated by Noell  et al.
    (1966). Prolonged exposure to low light levels may elicit secondary
    tissue responses that may not be directly related to the initial
    action spectrum. These tissues responses include acute cellular
    necrosis and apoptotic cell death, an inflammatory response with
    oedematous changes and macrophage invasion, and tissue proliferation
    with scar formation and neovascularization (e.g. Hoppeler  et al.,
    1988; Yoshida  et al., 1993). Action spectra, tissue responses,
    preventive measures as well as enhancing factors, which are all
    intensively investigated in many laboratories, are likely to be
    similar to the human.

         Studies of photochemical retinal injury in aphakic rhesus
    monkeys have extended the action spectrum for short-wavelength light
    damage down to 310 nm (Ham  et al., 1984). These studies showed
    that the retina, if exposed, is up to six times more vulnerable to
    photochemical damage in the UV than in the visible. These results
    are therefore of direct relevance to pseudophakic individuals
    without a UV-protective lens, and may also indicate a contributory
    role to light damage from the small fraction of UV that reaches the
    retina of the normal eye.

    8.  HUMAN STUDIES: THE SKIN

    8.1  Characteristics

    8.1.1  Structure and optical properties

         The skin is a large organ with an area of more than 1.5 m2 in
    adults. It provides the first stage of protection for chemicals,
    radiations, xenobiotics and also prevents the evaporation of water
    and the loss of ions and proteins. The skin has developed specific
    mechanisms for photoprotection and biological responses to UV as
    discussed below.

         Skin is composed of three very different parts (as shown in
    figure 4.1): the epidermis, dermis and subcutaneous tissue. The
    epidermis is the outermost layer of the skin and varies in thickness
    from 50 µm to 600 µm (palmoplantor skin). The fibrous proteins known
    as keratin are produced in the keratinocytes of the epidermis. It is
    keratin that serves as a major protective tough substance in the
    skin; hair and nails are composed almost entirely of keratin. The
    daughter cells of the keratinocytes in the basal layer (stratum
    basale) of the epidermis differentiate and become "prickle cells" of
    the malpighian layer of the epidermis - the  stratum malpighii. As
    these cells migrate outward, changes continue to occur; granules
    appear in the cytoplasm of each cell; the cells tend to flatten and
    form the  stratum granulosum. Still later, the cells lose their
    nuclei, die, dehydrate and flatten out to form the tough  stratum
     corneum (horny layer). It is generally agreed that the entire
    process of cell migration from the basal layer of the epidermis to
    final shedding from the surface of the stratum corneum takes 28 days
    in normal skin. Of this 28-day period, the cell spends about 14 days
    in the epidermis and 14 days in the stratum corneum.

         The skin contains millions of tiny glands, including: apocrine
    (sweat) glands which discharge sweat into hair follicles, eccrine
    (sweat) glands which carry saline from the dermis and subcutaneous
    layer directly to the skin surface, and sebaceous glands which
    secrete sebum - an oily substance which lubricates hairshafts and
    maintains a slightly acidic, oily film over the stratum corneum. The
    collection of eccrine sweat glands plays a major role in the body's
    thermoregulatory mechanism, since evaporative cooling is the most
    efficient means the body has for removing excess heat, as long as
    the humidity is not 100%. More than two litres of sweat can be
    discharged in one day. On average, the skin contains about as many
    sebaceous glands as eccrine sweat glands, except that there are few
    sebaceous glands on the palms of the hands and soles of the feet.

         The dermis, or corium, is much thicker than the epidermis, but
    consists of much larger cells. The dermis is largely connective
    tissue which gives the skin its elasticity and supportive strength.
    Nerve cells, blood vessels and lymphatic glands are found in the

    outermost dermal layer - the papillary dermis. Unlike the epidermis,
    the thickness of the dermis is not at all uniform throughout the
    body; it varies from 1 mm to 4 mm.

         The basal layer of the epidermis is delineated from the dermis
    by a complex basement membrane. The keratinocytes are anchored to
    the basement membrane by semi desmosomes. This is the only layer
    where cell division takes place. The division of keratinocytes,
    under normal conditions, occurs every 17 to 38 days. The division
    rate may vary in different parts of the body.

         Melanocytes are formed in the basal layer of the epidermis
    loosely attached to the neighbouring keratinocytes (ratio 1
    melanocyte for 46 keratinocytes). The melanocytes synthesize the
    pigment melanin which matures into melanosomes. Then the melanosomes
    are transferred to the keratinocytes where they are digested if
    their size is below 1 cubic micron. Large melanosomes, as in dark
    skin, are not digested but transferred intact to the stratum corneum
    with the normal shedding of the epidermis.

         Melanocytes are cells of nervous origin which have migrated
    during the 10th week of embryonic life in the epidermis and in the
    hair roots. These cells divide very slowly (one division every 3-5
    years). There is a boost of division when new hairs are growing or
    after exposure to UV light.

         Chemically, melanins are bipolymers: red melanins contain
    sulphurs and are soluble at PH 7.2, black melanins are insoluble.
    The ratio of red and black melanins within a melanocyte is
    genetically determined. Mixed types of melanins are deposited on a
    protein matrix contained in the melanosomes. Both types absorb UV
    and participate in the screening effect of the whole epidermis. Two
    major steps in the melanin synthesis take place: the enzymatic
    (tyrosinase) oxidation of the tyrosine to produce dopa
    (di-hydroxyphenylalanine) and the spontaneous oxidation of dopa in
    quinones.

         Black melanin absorbs UV. In the presence of oxygen, it
    produces a free radical which, in sufficient quantities, can be
    deleterious to the melanocytes and the cellular environment,
    including keratinocytes or in the dermis, fibroblasts and fibres.

         A third type of cell is present in the upper layers of the
    stratum Malpighi: the Langerhans cell. It is a migrating dendritic
    cell which is able to recognise foreign or abnormal structures. This
    cell plays a major role in immunological recognition and its
    activity is very sensitive to UV. Its function is impaired by a UV
    dose as low as one half of a minimal erythema dose.

         The innermost layer of the skin is generally known as the
    subcutaneous layer. It is composed largely of fatty tissue that

    serves a shock-absorbing and insulating role. The thickness of this
    layer varies considerably from one body region to another and from
    one person to another.

         Skin optics are governed by two basic processes, the absorption
    and scattering of light. Absorption is the loss of a photon when its
    energy is reduced within the atom or molecule, and more specifically
    in a target species called a chromophore. UV absorption occurs in
    wavelength bands where the molecules have characteristic absorption
    spectra. The absorbed photon energy is dissipated as heat or
    reemission of light when the excited molecules return to their
    ground state, or the energy is spent on photochemical reactions.
    Scattering is a process where the direction of propagation of UV is
    altered, especially at boundaries of refractive index. Scattering
    and absorption of the photons limit the depth of penetration of UV
    in the skin (see section 4.2 and figure 4.1).

         In the white skin, the change in the refractive index causes 5%
    of the normally incident light to be reflected. The remaining 95% is
    absorbed or reflected at the dermis. The back-scattered light is
    reflected again at different interfaces. This phenomenon explains
    why, in the epidermis and in the upper papillary dermis, the
    strength of the light becomes several times that of the incident
    light itself.

         The transmission of UV through isolated epidermis is strongly
    dependent on the chromophores contained in the structure. Aromatic
    amino acids (tryptophan, tyrosin, phenyl-alanine) absorb strongly
    near 275 nm, urocanic acid and melanins play the role of endogenous
    sunscreens. The DNA absorbs 260 nm wavelengths. Melanin is the
    unique chromophore with an absorption extending into the UVA and
    visible regions.

    8.1.2  Skin types

         The sensitivity of skin to UV has been defined by six
    phototypes: types I to IV are characteristics of caucasoid
    populations; type V represents mongoloid Middle Eastern populations
    (Fitzpatrick  et al., 1974); and type VI represents African and
    American negroid populations. The capacities to acquire natural tan
    or to present naturally a deep pigmentation are keys to the response
    to UV exposure. Among caucasians, there is a general correlation
    between skin type and resistance to sunburn and capacity to tan.

    8.2  Beneficial Effects

    8.2.1  Vitamin D3

         An established beneficial effect of UV exposure is the
    synthesis of vitamin D3 (Adams  et al., 1982). In adults, the
    epidermis contains nearly 50% of the total concentration of

    7-dihydrocholesterol in the skin. UVB exposure causes the
    provitamins D3 to be isomerized to pre-vitamin D3. During continual
    exposure to sunlight, the pre-vitamin D3 forms by
    photo-isomerization the biologically inert photo-products lumisterol
    and tachysterol. Once formed, the pre-vitamin D3 spontaneously
    isomerizes into vitamin D3 (reaction with a maximum efficacy at
    37°C), a more stable form. After crossing the basal membrane of the
    epidermis, vitamin D3 is linked to a circulating alpha1-globulin:
    vitamin D-binding protein. The protein linked vitamin D3 is
    transformed to 25-OH-T3 which can be measured in the blood.
    Transported to the kidney, it is metabolised to 1,25 dihydroxy
    vitamin D3 which is the biologically active form.

         Vitamin D3 is required for the intestinal absoprtion of calcium
    (Davies 1985). After a single whole body exposure to one MED the
    circulating vitamin D3 level increases by an order of magnitude (2
    ng/ml to 24 ng/ml within 24 hours) and returns to normal levels
    within a week (Holick 1985). Since active vitamin D3 is metabolised
    by the kidney, it was expected that in the circulating blood it does
    not change dramatically after repeated solar exposures. Chronically
    reduced vitamin D3 synthesis may lead to a deficit in active vitamin
    D3, as was found in elderly subjects (Omdahl  et al., 1982).

         The use of sunscreens was found to suppress cutaneous vitamin
    D3 synthesis (Matsuoka  et al., 1988). This was also found in
    children, pregnant or lactating women and debilitated patients with
    poor intestinal absorption. Widespread use of sunscreens could lead
    to vitamin D deficiency in some groups and inadequate fixation of
    calcium (Prystowsky 1988). More recently, receptors for the active
    form for vitamin D3 were found in the keratinocytes of the
    epidermis. Vitamin D3 inhibits the proliferation of cultured
    keratinocytes and induces them to terminally differentiate (Smith
     et al., 1986). The topical or oral administration of 1, 25-OH-D3
    has proved to be effective for the treatment of psoriasis, in
    replacement or in addition to the classical PUVA treatment. This is
    a new approach for the treatment of this condition and a possible
    explanation for the success of the heliotherapy (Morimoto and
    Kunihiko, 1989).

         For the entire system of vitamin D3 production the amount of UV
    radiation reaching the skin is critical. The doses needed are small,
    and daily exposures of the face and hands to sun and light for 15
    minutes is considered sufficient. The minimum dose requirement was
    estimated to be equivalent to 55 MED per year (Health Council of the
    Netherlands, 1986). When too little UVB reaches the skin,
    deficiencies of vitamin D may occur, resulting in a weakening of the
    bones. Groups at risk are particularly dark-skinned children in high
    latitude cities and elderly people living entirely indoors.
    Supplementation of vitamin D3 in the diet is then recommended.

    8.2.2  Skin adaptation

         Another beneficial effect of modest exposure to UVB radiation
    is the maintenance of the ability of the skin to sustain further UV
    exposures. Loss of this adaptation forms an important component in
    photodermatosis, skin diseases where the lesions are caused by
    light. These patients can be treated by regular exposures to UVB.
    The doses required are in the same range as that needed for the
    synthesis of vitamin D3.

    8.2.3  Other benefits

         It has been suggested that beneficial effects of UV exposures
    may occur such as: improvement of cardiopathy and functions, and
    better microorganism defense. These effects have not been confirmed
    in well designed studies. Bright light therapy for winter depression
    is most likely to be the consequence of visible light stimulating
    the ocular system. Because of the high illumination needed for this
    treatment special care is needed to avoid the emission of UV light
    by these sources (Terman  et al., 1990).

         Treatment of portwine stains and pigment dyschromia by lasers
    are applications of selective absorption of some wavelengths by
    specific chromophores contained in the lesion. Phototherapy by
    specific dyes absorbed by tumour cells or specific structures, is
    currently a growing field.

    8.3  Acute Effects

    8.3.1  Erythema and sunburn

         In its mildest form, sunburn consists of a reddening of the
    skin (erythema) that appears up to about 8 h after exposure to UV
    and gradually fades after a few days. In its most severe form, it
    results in inflammation, blistering, and peeling of the skin. The
    degree to which a person will experience sunburn depends critically
    on skin type. For fair-skinned people, the relative effectiveness of
    UV for tanning and for erythema are approximately the same over the
    entire UVB and UVA ranges of wavelengths (Parrish  et al., 1982).

         The most important factors that define if a dose of UV will
    induce erythema are the wavelength of the radiation, the skin type,
    and the pigmentation of the subject. UVA, UVB and UVC are all able
    to induce the erythema.

         The relative effectiveness of the different wavelengths to
    induce erythema is expressed as an erythema action spectrum
    (McKinlay and Diffey 1987). For minimal erythema, the most
    erythemogenic wavelengths are in the 250-290 nm range and a decrease
    in effectiveness is observed as the wavelengths increase. Erythema
    occurs 3-5 hours after UV exposure and reaches a maximum intensity

    between 8 and 24 hours, fading over 3 days. The vasodilatation of
    capillary vessels within the papillary dermis can be observed before
    the erythema becomes visible, and occurs in the same way for
    children, adults and elderly. However, the exposure time required to
    produce UVB erythema increases after about 60 years of age.

         Histologic alterations from erythema are observed in the
    photodyskeratotic keratinocytes as well as intercellular edema with
    exocytosis (lymphocytes in the epidermis). Superficial vascular
    plexus, endothelial cell enlargement, perivenular edema, red blood
    cells in the capillary are characteristic features observed between
    3 hours and 72 hours after UV exposure. Dermal neutrophils appear
    immediately after irradiation reaching a peak level at 24 hours.

         DNA may be the primary chromophore involved in the induction of
    erythema (Ley, 1985, Wolf  et al., 1993). Subsequently, a number of
    inflammatory mediators are induced. Cutaneous blisters have been
    used extensively to study these mediators. Prostaglandin (E2 and
    F2) levels were elevated in these blisters within 6 hours, peaked
    at 24 hours and returned to control levels by 48 hours after UV
    exposure, although erythema persisted beyond that time. Indomethacin
    suppressed prostaglandin formation, however, blood flow was only
    slightly altered by this treatment suggesting that other factors
    must play an important role in UVB induced inflammation (Greaves  et
     al., 1978). Similarly while elevated histamine levels have been
    observed, antihistamines have little effect in diminishing UV
    induced erythema (Gilchrest  et al., 1981).

         Finally, UV exposure causes keratinocytes, to release the
    cytokines interleukin-1 (IL-1) and tumour necrosis factor alpha (TNF
    alpha) both of which are potent mediators of inflammation (Oxholm
     et al., 1988; Rasanen  et al., 1989). UVB exposure also causes
    up-regulation of adhesion molecule such as ELAM-1 which facilitates
    inflammatory cell infiltration (Murphy  et al., 1991).

         Irradiation of human skin with 3 MED is associated with
    increased levels of transforming growth factor, suggesting a role of
    this molecule in keratinocyte proliferation, epidermal hyperplasia
    and angiogenesis.

    8.3.2  Skin pigmentation and tanning

         When skin is exposed to UV, two distinct tanning reactions
    ensue. Immediate pigment darkening (IPD) begins immediately on
    exposure to UV and is caused by the darkening of the pigment melanin
    that is already present in the skin; it is normally seen only in
    people who have at least a moderate constitutive tan. Such
    pigmentation begins to fade within a few hours after cessation of
    exposure. UVA is regarded as being most effective for IPD.

         Delayed tanning (melanogenesis) takes about three days to
    develop and is more effectively produced by UVB than by UVA (Parrish
     et al., 1982; Gange  et al., 1985). Delayed tanning is more
    persistent than IPD and results from an increase in the number, size
    and pigmentation of melanin granules. Exposure to UVB results also
    in an increase in the thickness and scattering properties of the
    epidermis (outer layer of the skin). Because UVA does not produce
    thickening of the epidermis, the tan obtained from it, while perhaps
    cosmetically acceptable, is not as effective in protecting against
    further exposure to UV as the equivalent pigmentation induced by
    exposure to UVB or solar radiation.

    8.3.3  Photosensitization

         The use of certain medicines may produce a photosensitizing
    effect on exposure to UVA as may the topical application of certain
    products, including some perfumes, body lotions, etc. Many
    medications and other agents contain ingredients that may cause
    photosensitivity, which is defined as a chemically induced change
    that makes an individual unusually sensitive to light. An individual
    who has been photosensitized may develop a rash, sunburn, or other
    adverse effect from exposure to light of an intensity or duration
    that would normally not affect that individual.

         Reactions to photosensitizing agents involve both photoallergy
    (allergic reaction of the skin) and phototoxicity (irritation of the
    skin) after exposure to ultraviolet radiation from natural sunlight
    or artificial lighting (particularly from tanning booths). This
    photosensitization of the skin may be caused by creams or ointments
    applied to the skin, by medications taken orally or by injection, or
    by the use of prescription inhalers.

         In addition to an exaggerated skin burn, itching, scaling,
    rash, or swelling, exposure to UV in combination with certain
    medications may result in (FDA 1992); Skin cancer, Premature skin
    aging, Skin and eye burns, Allergic reactions, Cataracts, Reduced
    immunity and Blood vessel damage.

         This can result in photoallergic or phototoxic reactions that
    are accelerated by UV exposure. Phototoxic contact dermatitis often
    occurs clinically as exaggerated sunburn but occasionally blisters
    may also occur on the erythematous areas. Most phototoxic
    sensitizers have an action spectrum in the UV from 280-430 nm.
    Window glass which absorbs UV below 320 nm will protect patients
    from phototoxic compounds with absorption below 320 nm, but fails to
    protect against photosensitizers such as tar and psoralens which are
    efficient at longer wavelengths. Examples of photocontact drugs and
    substances are given below in table 8.1.

    Table 8.1  Some photosensitizing substances
               (CIE, 1990, FDA, 1992)

         Sulphonamides
         Salicylanilides
         Coal-Tar derivatives
         -    acridine
         -    anthracene
         -    phenanthrene
         Dyes
         -    anthraquinone
         -    eosin
         -    methylene blue
         -    rose bengal
         Psoralens
         Some fragrances
         Cyclamate (artificial sweetener)
         Non-steroidal anti-inflammatory drugs (pain reliever,
         antiathritics)
         Deodorant and bacteriostatic agents in soaps
         Fluorescent brightening agent for cellulose, nylon or wool
         fibres
         Phenothiazines (major tranquilizers, anti-emetics)
         Sulfonylureas (oral anti-diabetics, hypoglycemics)
         Sunscreen ingredients
              -    6-Acetoxy-2,4,-dimethly-m-dioxane (preservative)
              -    Benzophenones
              -    Cinnamates
              -    Para-aminobenzoic acid (PABA)
              -    Paba esters
         Tetracyclines (antibiotics, anti-infectives)
         Tricyclic antidepressants

         The concentration of drugs needed to elicit a photoallergic
    reaction is much lower than that needed to cause a phototoxic
    reaction. On the other hand, photoallergic reactions occur only in a
    small proportion of exposed individuals while phototoxic reactions
    may occur in anyone, given sufficient exposure. Generally, no
    clinical reaction occurs on first exposure to an agent causing
    subsequent photoallergy. Even sunscreen agents used to protect
    against photocontact dermatitis may be photoallergenic. A small
    number of individuals who develop photocontact dermatitis may retain
    a persistent reactivity to light (including UV) long after exposure
    to the photosensitizing compound.

    8.4  Chronic Effects on the Skin Other than Cancer

         UV radiation causes a number of chronic degenerative changes in
    the skin, mainly in caucasian populations, as a result of its action

    on keratinocytes, melanocytes and components of the dermal stroma
    including fibrous tissue (collagen and elastin) and blood vessels.
    These changes include freckles (ephelides), melanocytic naevi,
    lentigines, telangiectasia, skin wrinkling and atrophy, yellow
    papules and plaques on the face, colloid milium (firm, small,
    yellow, translucent papules on the face, forearms and hands),
    diffuse erythema, diffuse brown pigmentation and ecchymoses
    (Goldberg & Altman, 1984). These or related changes are sometimes
    grouped into syndromes such as cutis rhomboidalis nuchae (thick,
    yellow, furrowed skin, particularly on the back of the neck),
    Favre-Racouchot syndrome (yellow, thick comedones and follicular
    cysts of the periorbital, malar and nasal areas) and reticulated
    poikiloderma (reddish brown reticulated pigmentation with
    telangiectasia and atrophy and prominent hair follicles on the
    exposed chest and neck). With the exception of freckles and
    melanocytic naevi, these changes are also referred to collectively
    as "photoageing" (Gilchrest, 1990) because of their association with
    increasing age but presumed correlation with total accumulated
    exposure to the sun rather than with age per se.

         In the US National Health and Nutrition Examination Survey
    (Engel  et al., 1988), age-adjusted prevalence proportions of
    senile elastosis, actinic (solar) keratosis, fine telangiectasia,
    localized hypermelanism, senile (solar) lentigines and freckles, in
    whites 1 to 74 years of age, were associated with lifetime exposure
    to the sun as estimated by dermatologists.

         Freckles and solar lentigines

         Freckles and solar (also called senile) lentigines are
    pigmented macules occurring on the sun exposed skin of caucasians.
    Their prevalence is increased in those with highly sun-sensitive
    skin (Azizi  et al., 1988). Freckles occur most commonly in
    children while the frequency of solar lentigines increases with age
    and is greatest in those over 60 years of age (estimated at 75% in
    the USA: Rhodes  et al., 1991). They show similar histological
    patterns: their are increased numbers of melanocytes and an
    increased concentration of melanin in the basal layer of the
    epidermis (Rhodes  et al., 1991). Melanocytic atypia has been
    observed in both. An increased risk of melanoma has been observed in
    relation to freckling in childhood and an increased risk of
    non-melanocytic skin cancer in relation to both freckling and
    prevalence of solar lentigines (see sections 8.4.1 and 8.4.2).

         Melanocytic naevi

         Melanocytic naevi are benign proliferations of melanocytes
    usually beginning in the basal layer of the epidermis and later
    extending into the dermis. They are common in white populations and
    rare in black and Asian populations (Armstrong & English, 1988;
    Gallagher  et al., 1991), are associated, in white populations,

    with phenotypic indicators of constitutional sensitivity to the sun,
    particularly fair skin colour (Green  et al., 1988b; Gallagher  et
     al., 1990b), occur mainly on body sites that are maximally or
    intermittently exposed to the sun (Kopf  et al., 1978, 1985;
    Augustsson  et al., 1990; Gallagher  et al., 1990a), occur more
    commonly in Australian than British children (Green  et al., 1988b)
    and in persons born in Australia than immigrants who arrived in
    Australia after about 15 years of age (Armstrong  et al., 1986),
    and somewhat inconsistently with measures of sun exposure, including
    sunburn in early life (Armstrong  et al., 1986; Gallagher  et al.,
    1990c; Coombs  et al., 1992). They are associated with an increased
    risk of cutaneous melanoma (see section 8.4.2).

         Solar keratoses

         Solar keratoses are benign proliferations of epidermal
    keratinocytes. They are very common on exposed body sites in older
    people in caucasian populations living in areas of high ambient
    solar irradiance (see for example, Marks  et al., 1983 and Holman
     et al., 1984a). Solar keratoses have been reported to be
    associated with phenotypic indicators of cutaneous sun sensitivity
    (Vitasa  et al., 1990), to be more common in people born in
    Australia than in migrants to Australia (Goodman  et al., 1984), to
    be associated with estimates of total and occupational sun exposure
    (Goodman  et al., 1985; Vitasa  et al., 1990), and to be
    associated with other benign indicators of cutaneous sun damage
    (Holman  et al., 1984b; Green 1991). Their number on the skin is
    strongly associated with risk of non-melanocytic skin cancer (see
    section 8.4.1).

    8.5  Cancer

         Epidemiological evidence relevant to the effects of UV on risk
    of cancer in humans derives mainly from study of the effects of sun
    exposure (presumably solar UV but not separable from other solar
    radiation) on cancer risk. There are four general lines of evidence
    available from which it may be inferred that sun exposure causes any
    particular cancer. They are that: the cancer in question occurs more
    frequently in people who are sensitive to the sun, occurs mainly at
    sun-exposed body sites, is increased in residents of areas of high
    ambient solar irradiance, and is increased in people with high
    personal sun exposure.These lines of evidence will not apply to all
    cancers that may be linked to sun exposure. They form nonetheless, a
    useful framework within which most of the relevant evidence can be
    described. To this framework will be added a consideration of
    evidence relating to artificial sources of UV when it is available.

         The most direct evidence of the carcinogenicity of UV in humans
    should come, in principle, from observation of the effects of
    personal exposure to the sun. In practice, it is very difficult to
    make measurements of personal sun exposure accurately. Most often

    they are made by questionnaire and require recall of rather
    non-salient details of life over 60 or more years. This is a very
    difficult task (Kricker  et al., 1993). A difficulty is presented
    by the fact that people who have sun sensitive skin and are at
    higher risk of skin cancer will tend to expose themselves less to
    the sun. To obtain an accurate measure of the effects of personal
    sun exposure, this confounding with sun sensitivity should be
    controlled - this has not always been done. It may not be
    surprising, therefore, that measures of personal exposure to the sun
    have not been consistently associated with risk of cancers thought
    to be related to the sun and that more indirect evidence has proved
    to be stronger.

    8.5.1  Nonmelanocytic skin cancer

         Introduction

         There are two major histopathological types of nonmelanocytic
    skin cancer: basal cell carcinoma (BCC) and squamous cell carcinoma
    (SCC). BCC is the commoner type in white populations.

         The epidemiology of nonmelanocytic skin cancer is difficult to
    describe accurately. Its routine recording is often not attempted by
    cancer registries because of the large numbers of cases involved
    and, if attempted, is invariably incomplete because of the rarity
    with which primary nonmelanocytic skin cancers require hospital
    treatment and the frequency with which probable nonmelanocytic skin
    cancers are not sent for histopathological verification of the
    diagnosis (Muir  et al., 1987).

         These difficulties have influenced the quality of the
    epidemiological evidence relating nonmelanocytic skin cancer to sun
    exposure. In addition, most of the earlier cross-sectional and
    case-control studies of nonmelanocytic skin cancer , and some of the
    more recent ones, (see, for example, table 11 in IARC, 1992) are
    deficient in that the control series were rarely population-based,
    appropriate effect measures or p values were often not estimated,
    and confounding by age and sex were not controlled in the analysis.
    In the narrative that follows, little reliance will be placed on
    these studies which include those reported by Lancaster & Nelson
    (1957), Gellin  et al. (1965), O'Beirn  et al. (1970), Urbach  et
     al. (1974), Aubry & MacGibbon (1985), O'Loughlin  et al. (1985),
    Herity  et al. (1989), Hogan  et al. (1989) and Gafa  et al.
    (1991).

         The results of most of the studies referred to below have been
    described and their results tabulated in detail in other
    publications (e.g., IARC, 1992) thus only the immediately salient
    features will be described here.

         Sun-sensitivity

          Race

         Nonmelanocytic skin cancer is much less frequent in populations
    with dark skins than those with light skins (Fitzpatrick & Sober,
    1985; Hoffman, 1987; Urbach, 1987). Data from cancer registries
    represented in Cancer Incidence in Five Continents Volumes II-VI
    (Doll  et al., 1970; Waterhouse  et al., 1976, 1982; Muir  et al.,
    1987; and Parkin  et al., 1992) in which direct comparisons of
    nonmelanocytic skin cancer incidence in different ethnic groups can
    be made within a single geographical area show that the incidence
    rates in the light skinned populations are consistently the highest.
    Similarly, in the 1977-78 US survey of nonmelanocytic skin cancer;
    rates in whites were 232.6 per 100 000 person years compared with
    3.4 in blacks (Scotto  et al., 1983).

         Available evidence suggests that BCC occurs less frequently
    than SCC among dark skinned populations. The opposite is the case in
    light skinned populations. BCC was the most common nonmelanocytic
    skin cancer reported in South African whites in 1949-1975 but was
    rare among black Africans, occurring mainly in albinos (Oettlé,
    1963; Oluwasanmi  et al., 1969; Rippey and Schmaman, 1972; Isaacson
     et al., 1978). SCC, principally on the lower limb and associated
    with previous trauma, was the commoner of the two in blacks (Oettlé,
    1963; Oluwasanmi  et al., 1969; Rippey and Schmaman, 1972; Isaacson
     et al., 1978; Rose, 1973). Melanesians (Foster & Webb, 1988) and
    Polynesians (Paksoy  et al., 1991) had less BCC than SCC, while no
    BCCs were reported in the Melanesians of North Samoa, a particularly
    heavily pigmented people (Foster & Webb, 1988).

         Ethnic background is an important determinant of risk of
    nonmelanocytic skin cancer in Caucasians. In people of southern
    European ethnic origin born in Australia, relative to other people
    born in Australia, risk of BCC was 0.56 (95% CI 0.14-1.65) for those
    with one southern European grandparent, 0.17 (0.00-1.05) for two and
    0.00 (0.00-0.86) for three or four (p for trend 0.002; Kricker  et
     al., 1991a). No subject with SCC in this study had any southern
    European grandparents (RR 0.00, 0.00-1.23).

          Individual sun sensitivity

         Among recent cross-sectional, case-control and cohort studies
    several have reported significantly elevated relative risk (RR)
    estimates for red or light hair colour with BCC (RRs between 1.5 and
    2.9; Hunter  et al., 1990; Green & Battistutta, 1990; Kricker  et
     al., 1991a) and SCC (RR=2.4; Kricker  et al., 1991a). However, a
    light complexion was significantly associated only with SCC (RR=3.3;
    Kricker  et al., 1991a), and high RRs (of 3.4) for light hair
    colour with BCC and SCC were reported in only one study (Vitasa  et
     al., 1990).

         Sensitivity of the skin to the sun, as measured by ability to
    tan and susceptibility to sunburn, was more consistently related to
    risk of BCC and SCC than hair, skin and eye colour. RRs of 2.0 or
    more for BCC were found with a skin that burns rather than tans
    (Marks  et al., 1989; Vitasa  et al., 1990; Hunter  et al., 1990;
    Kricker  et al., 1991a). For SCC, the evidence of an increased risk
    with sun-sensitive skin was somewhat weaker, with RRs around 2.0 and
    95% confidence intervals (CI) that were wide and included 1.0, in
    the two studies that reported RRs adjusted for age, sex, and other
    relevant confounders (Vitasa  et al., 1990; Kricker  et al.,
    1991a).

          Xeroderma pigmentosum and albinism

         Xeroderma pigmentosum (XP) is a recessively inherited syndrome
    characterised by clinical and cellular hypersensitivity to solar
    radiation and a defect in the capacity to repair UV-induced damage
    in DNA (Fitzpatrick  et al., 1963, Cleaver, 1973). Evidence of
    cutaneous sun damage may appear as early as 1-2 years of age in the
    absence of specific protection from the sun, and skin cancers are
    very frequent (Kraemer  et al., 1987). In a survey of 830 cases of
    XP located through published case reports, 45% were reported to have
    had skin cancers (Kraemer  et al., 1987). The median age of
    diagnosis of the first skin cancer was 8 years. Ninety seven per
    cent of BCC and SCC were on constantly exposed sites, that is, the
    face, head, and neck, compared with an estimated 80% in the general
    US population. In 220 XP patients in whom the number of skin cancers
    was stated, half had more than two cancers and 5% had more than 30;
    79 patients were described as having BCC and 112 SCC (Kraemer  et
     al., 1987).

         Albinism is an inherited disorder of melanin metabolism with a
    decrease in or complete absence of melanin; as a result, the skin of
    albinos is highly sensitive to the sun. The most common type of
    albinism occurs in 1 in 15,000 American blacks, 1 in 40,000 European
    or American Caucasians, and has estimated frequencies as high as
    1:3,900 in Soweto, South Africa (Kromberg and Jenkins, 1982), and 1
    in 1,000 in Nigeria (Cervenka  et al., 1979). African albinos have
    been reported to have a high incidence of SCC and a somewhat lower
    rate of BCC (Cervenka  et al., 1979; Luande  et al., 1985;
    Kromberg  et al., 1989).

         Body-site distribution

         It is commonly stated that the site distributions of
    non-melanocytic skin cancers, particularly SCC, correspond well to
    what would be expected from the exposure of different body sites to
    the sun. They do generally conform to this pattern with more than
    60% of lesions occurring on the head and neck (Kricker  et al.,
    1993). It is consistently observed, however, that the proportion of
    SCC on the upper limbs is higher than that of BCC while BCC has the

    higher proportion (some 10% or more) on the trunk (Scotto  et al.,
    1983; Osterlind  et al., 1988a; Giles  et al., 1988; Glass &
    Hoover, 1989; Levi  et al., 1988; Karjalainen  et al., 1989;
    Kricker  et al., 1990; Gallagher  et al., 1990a; Roberts, 1990;
    Magnus, 1991; Serrano  et al., 1991; Kricker  et al., 1993). At
    the sub-site level, BCCs are almost completely absent on the heavily
    exposed backs of hands, and infrequent on the forearms compared with
    the upper arms (Brodkin  et al., 1969; Scotto  et al., 1983;
    Goodman  et al., 1984; Kricker  et al., 1990); in addition, this
    subsite distribution on the face is not highly correlated with the
    distribution of UV erythema on the face (Diffey  et al., 1979).

         Residence in areas of high ambient solar irradiance

          Geographical variation

         Annual incidence rates of nonmelanocytic skin cancer in 29
    populations of mainly Western European origin in Cancer Incidence in
    Five Continents, volume 6 (Parkin  et al., 1992) show little
    evidence of any consistent relationship between incidence and
    latitude. The highest incidence rates were in the populations of
    Tasmania, Australia (42°S) (213.2 per 100,000 in men and 113.1 per
    100,000 in women) and British Columbia, Canada (49°N) (134.1 in men
    and 91.2 in women), and were also comparatively high in Switzerland
    (47°N) (78.7 in men and 50.2 in women) and Southern Ireland (53°N)
    (71.5 in men and 48.0 in women).

         In contrast to these international patterns, incidence rates of
    nonmelanocytic skin cancer within countries do appear to increase
    with proximity to the equator as indicated by broad place of
    residence, latitude, or measures of solar irradiance. Geographical
    variation in nonmelanocytic skin cancer incidence in the USA has
    been described in three National Cancer Surveys (Mountin & Dorn,
    1939; Dorn, 1944a, 1944b; Auerbach, 1961; Haenszel, 1963; Scotto  et
     al., 1974) and several related studies (Scotto  et al., 1982,
    1983; Serrano  et al., 1991). Incidence of all types increased with
    increasing proximity to the equator, with similar gradients for men,
    women and all ages. The same pattern is seen in Australia (Marks  et
     al., 1993)

          Migrants

         The nonmelanocytic skin cancer experience of light-skinned
    migrants from areas of low to areas of high ambient solar irradiance
    has generally been consistent with an effect of sun exposure on skin
    cancer incidence. In Australia, incidence and mortality were found
    to be lower in migrants, most of whom had come from the UK, an area
    of lower sun exposure, than in those born in Australia (Armstrong
     et al., 1983; Giles  et al., 1988). Kricker  et al. (1991a)
    examined in some detail the relationship between BCC and SCC and
    migration to Australia. Migrants (excluding those from southern

    Europe who may be at lower constitutional risk for skin cancer) had
    a lower risk of BCC and SCC (RRs around 0.3) than did those born in
    Australia. In addition, for BCC (which had sufficient cases to
    analyze) the RR in those who migrated in the first 10 years of life
    (1.05) was the same as that in those born in Australia (1.0) but
    then fell to low levels in those who migrated later in life (RR
    about 0.2). Hunter  et al., (1990) observed an association between
    risk of BCC and residence in southern parts relative to elsewhere in
    the USA (RRs of 1.6 for residence in California and 2.1 for
    Florida).

         Personal sun exposure

          Total sun exposure

         Reported measurements of total current or accumulated sun
    exposure in studies of nonmelanocytic skin cancer are likely to be
    subject to substantial error not only because of the difficulties in
    recalling sun exposure over periods of 60 years or more (Kricker  et
     al., 1993) but also because of the use of broad summary variables
    such as "estimated average daily outdoor exposure" (Gellin  et al.,
    1965). The two studies which attempted a more quantitative measure
    found no evidence of a positive association between estimates of
    total sun exposure and risk of BCC (RRs of 1.0 or less) (Hunter  et
     al., 1990; Vitasa  et al., 1990). Only one of these studies
    included SCC (with only about 50% confirmed histopathologically);
    for sun exposure above the 75th centile the estimated RR was 2.5
    (95% CI 1.2-5.4) (Vitasa  et al., 1990).

          Occupational sun exposure

         Early clinical reports associated skin cancer with outdoor
    occupation (Molesworth, 1927; Blum, 1948; Emmett, 1973). However,
    the relationship between occupation and skin cancer has been
    examined adequately in relatively few studies and the evidence for
    an association is weak. At the population level, the best conducted
    studies which attempted to classify occupational sun exposure on the
    basis of occupational title found only small differences in skin
    cancer incidence between outdoor and indoor workers. Several
    population studies reported associations with employment in
    agriculture (Atkin  et al., 1949; Whitaker  et al., 1979; Teppo
     et al., 1980) and outdoor employment generally (Beral & Robinson,
    1981; Vågerö  et al., 1986) but their interpretation is complicated
    by incomplete and potentially biased case ascertainment (Vågerö  et
     al., 1986; Teppo  et al., 1980), inclusion of other cancers
    (Atkin  et al., 1949), and confounding between occupation and
    social class (Beral & Robinson, 1981).

         There have been few well-conducted studies of occupational sun
    exposure at the individual level. They show a generally consistent
    but not strong association between BCC and SCC and various crude

    measures of outdoor employment, e.g., "indoors", or "outdoors"
    occupation. The RRs have generally been under 2.0 for BCC (Hogan  et
     al., 1989; Marks  et al., 1989; Green and Battistutta, 1990; Gafa
     et al., 1991). One study reported a high RR for SCC with an
    "outdoors" occupation but with an extremely wide confidence interval
    (Green and Battistutta, 1990). Other studies found RRs close to 1.0
    (Aubry and MacGibbon, 1985, Hogan  et al., 1989; Marks  et al.,
    1989; Gafa  et al., 1991) and only one appeared to be statistically
    significant (Hogan  et al., 1989).

          Recreational exposure

         The relationship of exposure to sunlight in non-working hours
    (largely recreational sun exposure) with nonmelanocytic skin cancer
    has been described in only two studies. One study reported an RR
    below 1.0 for BCC in those with "mainly outdoor leisure" and a
    positive but not statistically significant association for SCC: RR
    of 3.9 (95% CI 0.5-30.9) (Green & Battistutta, 1990). In the study
    of Aubry & MacGibbon (1985), the highest category of a
    "non-occupational exposure score" showed an elevated risk of SCC (RR
    1.6; p-value=0.07).

          Sunburn

         Risk of BCC was significantly increased in subjects with a
    history of sunburn in one of two recent studies. Hunter  et al.
    (1990) observed an increasing risk with increasing frequency of
    painful sunburns (RR 2.91, 95% CI 2.37-3.58, for 6+ occasions).
    Green and Battistutta (1990), on the other hand, found no
    discernible association between BCC and number of painful sunburns.
    The risk of SCC, however, was increased with any history of painful
    sunburn in this study: RRs 3.3 (95% CI 0.9-12.3) for 1-5 sunburns
    and 3.0 (0.7-12.2) for 6+.

         Other sun-related skin conditions

         Cutaneous microtopography (a measure of loss of the fine skin
    markings on the backs of the hands presumed due to loss of dermal
    collagen; Holman  et al., 1984a), prevalence of solar elastosis of
    the neck, solar telangiectasia and solar lentigines and a history of
    solar keratoses have been taken as indicators of a high level of
    total accumulated exposure to the sun in those who have BCC or SCC.
    The evidence that sun exposure causes these conditions, however, is
    no stronger, and may be weaker, than the evidence that it causes
    BCC, SCC or cutaneous melanoma.

         Increasingly severe sun-related skin damage as measured by
    cutaneous microtopography was associated with increasing risks of
    both BCC (RR 3.1, 95% CI 1.5-6.4 for the highest grade) and SCC (RR
    1.8, 95% CI 0.8-4.2 for the highest grade) (Kricker  et al.,
    1991a). Other indicators of sun damage to the skin (e.g. freckles,

    telangiectasia, and elastosis) were also strongly related to risk of
    both BCC and SCC, whether considered separately (Green and
    Battistutta, 1990; Kricker  et al., 1991a) or together (Holman  et
     al., 1984a; Green  et al., 1988a), and including when adjusted
    for cutaneous sun sensitivity (Kricker  et al., 1991a). The RRs
    associated with the presence of 40 or more solar keratoses were 10.4
    (95% CI 5.8-18.8) for BCC and 34.3 (95% CI, 14.0-84.0) for SCC
    (Kricker  et al., 1991a).

         Exposure to artificial sources of UV

         Any use of a sunlamp was associated with a statistically
    significantly increased risk of SCC in the study of Aubry & McGibbon
    (1985) but three other studies (O'Loughlin  et al., 1985; Herity
     et al., 1989; Hogan  et al., 1989) found no association with
    "artificial sunlight" or use of sunlamps or sunbeds.

    8.5.2  Cutaneous Melanoma

         Introduction

         Cutaneous melanoma began to be studied epidemiologically in
    1948 when it was separated from other primary malignancies of the
    skin in the 6th Revision of the International Statistical
    Classification of Diseases (WHO, 1948), and Eleanor MacDonald
    published the first population-based study, covering 272 incident
    cases of melanoma in Connecticut (MacDonald, 1948). Because of its
    likely fatality if not treated, melanoma is much more likely to come
    to medical attention and to be diagnosed histopathologically than
    nonmelanocytic skin cancer. It has, therefore, been more readily
    recorded on a population basis than nonmelanocytic skin cancer and
    better studied epidemiologically. Thus, while generally considered
    less clearly associated with sun exposure, a series of large and
    well-conducted studies in different countries over the last 10 years
    have served to clarify this issue.

         Sun-sensitivity

          Race

         Melanoma is a disease primarily of light-skinned populations
    and occurs much less frequently in people with darker skins. In the
    United States, the incidence in whites is ten fold or more higher
    than in blacks living in the same areas (Parkin  et al., 1992).
    Rates are lowest (less than 0.7 per 100 000 person years) in parts
    of Asia (eg Japan, Singapore, India, China, Philippines) while in
    Los Angeles, USA, incidence was less than 1 per 100 000 in Japanese
    and Chinese people compared with slightly higher rates, around 2.0,
    in blacks (Parkin  et al., 1992).

         Among Caucasians, ethnic background is an important determinant
    of melanoma incidence. The incidence is substantially lower among
    Hispanics than among other whites in the United States. For example,
    the incidence among Hispanics in New Mexico is less than 2 per
    100,000 person years, but in other whites it is about 11 per 100,000
    (Muir  et al., 1987). In several case-control studies, subjects
    with a southern or eastern European background had substantially
    lower risks than those of northern European or United Kingdom
    origins (Elwood  et al., 1984; Holman and Armstrong, 1984a; Graham
     et al., 1985).

          Individual sun sensitivity

         Blue eyes, fair or red hair, and pale complexion in people of
    European origin are well established risk factors for melanoma.
    These pigmentary characteristics were documented in most melanoma
    case-control studies (see table 16 in IARC, 1992 and, as examples,
    Elwood  et al., 1984, Holman & Armstrong, 1984a, and Green  et al.,
    1985). Relative risk for light skin colour ranged from a little more
    than 1.0 to about 3. Compared with those with dark brown to black
    hair, those with fair hair generally had a less than two-fold
    increase in risk, but those with red hair usually had a 2 to 4-fold
    increase in risk. Eye colour was generally a weak risk factor with
    relative risks less than 2, and what increased risk there was
    generally disappeared after adjustment for the other traits.

         Increased risks of melanoma in those with a reduced ability to
    tan and an increased tendency to sunburn were observed in all
    case-control studies in which these sun-sensitivity measures were
    examined (see, for example, Elwood  et al., 1984, Holman &
    Armstrong, 1984a, and Green  et al., 1985).

          Xeroderma pigmentosum and albinism

         In a review of reports of 378 patients with XP in which cancer
    was mentioned, 37 patients had a melanoma (Kraemer  et al., 1987).

         African albinos have been reported to have a low rate of
    melanoma (Cervenka  et al., 1979; Luande  et al., 1985; Kromberg
     et al., 1989). Levine  et al. (1992) state that there have been
    only 16 documented cases of melanoma in albino patients reported in
    the English language literature. Since albinos generally have a
    normal number of melanocytes (Dargent  et al., 1992), the rarity of
    melanoma in them may indicate that melanin plays an important role
    in the genesis of this cancer.

         Body site distribution

         The distribution of melanoma appears to favour sites which are
    less heavily sun exposed, that is the back and face in Caucasian men
    and the lower limbs in women (Crombie, 1981). However, when

    whole-population series of cancer cases have been examined and body
    surface area taken into account, sites that were usually covered by
    clothing had lower rates than those usually exposed, with the
    exception of the forearms and hands for which the rates per unit
    area were low (Elwood & Gallagher, 1983; Green  et al., 1993). The
    anatomic site distribution of melanoma in blacks is quite different
    from that in Caucasians with the majority of melanomas on the soles
    of the feet (Higginson and Oettlé, 1959; Lewis, 1967; Fleming  et
     al., 1975).

         Residence in areas of high ambient solar irradiance

          Geographical variation

         Internationally, the incidence of melanoma varies over
    100-fold. Among countries included in Cancer Incidence in Five
    Continents, Volume VI, the lowest rates reported around 1983-87 were
    0.1-0.2 per 100 000 person years in China (Qidong), Japan (Osaka),
    India (Bombay), and in Kuwaitis in Kuwait, while the highest were
    about 25 per 100 000 person years in parts of Australia (Parkin  et
     al., 1992). In the USA and Australia, which have reasonably
    homogeneous populations of mainly European origin distributed across
    a wide latitude range, melanoma incidence increases with increasing
    proximity to the equator or with increasing measured, annual ambient
    UV irradiance (Scotto & Fears, 1987; Jones  et al., 1992).

         A simple latitude gradient for melanoma is not evident in
    Europe. Armstrong (1984) showed that the incidence of melanoma in
    Europe decreased from about latitude 35°N to a minimum around 55o N
    and then rose with increasing latitude because of high rates in
    Scandinavian and Scottish populations. The gradient of risk for
    melanoma from north to south in northern Europe matches the
    increasing natural pigmentation of the skin and may also, in part,
    be due to differing patterns of sun exposure, particularly
    recreational and vacation sun exposure.

          Migrants

         If a person migrates in childhood to a country of high ambient
    solar radiation such as Australia, Israel, and New Zealand, their
    risk of melanoma is observed to be similar to that in those born in
    the country to which they migrate, whereas if they migrate after
    this age their risk is substantially less than that in the
    locally-born population (Katz  et al., 1982; Cooke and Fraser,
    1985; McCredie & Coates 1989; Steinitz  et al., 1989; Khlat  et
     al., 1992). Overall, British immigrants to Australia and New
    Zealand, where the populations are of predominantly British origin,
    had incidence and mortality rates of melanoma of about a half the
    levels in those born in these countries (Cooke and Fraser, 1985;
    Khlat  et al., 1992).

          Residence history

         Studies of individual lifetime ambient solar radiance, based on
    latitude, location (eg, tropical, near coast), or average sunshine
    hours of all places of residence are consistent in showing an
    increased risk of melanoma with increased average radiance (Green &
    Siskind, 1983; Holman  et al., 1984b; Graham  et al., 1985;
    Osterlind  et al., 1988b; MacKie  et al., 1989; Weinstock  et al.,
    1989; Beitner  et al., 1990). RRs for the highest category of
    ambient solar radiance were 2.8 for mean annual hours of bright
    sunlight at places of residence in Australia (Holman  et al.,
    1984b), between 1.4 and 2.6 for a southerly latitude in the USA
    (Graham  et al., 1985; Weinstock  et al., 1989), or a tropical or
    Mediterranean residence (MacKie  et al., 1989; Beitner  et al.,
    1990). RRs differed greatly for residence near the coast in Denmark
    (RR=1.7; Osterlind  et al., 1988b) and Australia (RR=5.0; Green &
    Siskind, 1983).

         Personal sun exposure

          Lifetime total exposure

         Inconsistent results have been obtained in studies in which
    lifetime total sun exposure were assessed by questionnaires. RRs for
    the highest exposure category were between 0.6 and 5.3 (see table 18
    in IARC, 1992). Two of five studies showed a statistically
    significant positive association with RRs of 3.4 and 5.4 (Grob  et
     al., 1990; Lê  et al., 1992) while one study showed a significant
    negative association (RR 0.6; Graham  et al., 1985). The other two
    RRs were 1.1 (Dubin  et al., 1986) and 5.3 (95% CI 0.8-30.8; Green,
    1984).

          Usual or recent total exposure

         In this context, total means exposure from all sources (i.e.
    occupational and non-occupational) rather than the total over some
    period of time. This concept has been measured in a variety of ways.
    For example, it has been common to enquire about present "usual"
    exposure to the sun or usual exposure in some time period in the
    fairly recent past. The evidence of any effect of such exposures on
    risk of melanoma is weak. Of five studies, four had relative risks
    for the highest categories of exposure ranging from 0.7 to 1.2
    (Elwood  et al., 1985b; Dubin  et al., 1986; Holman  et al.,
    1986; Cristofolini  et al., 1987). The remaining study found a
    relative risk of 2.5 (p < 0.001) for the highest category of
    exposure of average daily sun exposure 10-20 years ago (Rigel  et
     al., 1983).

          Occupational exposure

         Four studies of melanoma have shown statistically significant
    positive associations with estimated lifetime occupational exposure
    to the sun (see table 20 in IARC, 1992) with RRs ranging up to 6.0
    (95% CI, 2.1 to 17.4) for outdoor employment (Paffenbarger  et al.,
    1978; Dubin  et al., 1986; Garbe  et al., 1989; Grob  et al.,
    1990). On the other hand, four studies showed statistically
    significant results in which the RR for the highest category of
    exposure was < 1.0 (Elwood  et al., 1985b; Holman  et al., 1986;
    Osterlind  et al., 1988b; Beitner  et al., 1990). The remaining
    studies have shown RRs around 1.0.

          Recreational exposure

         Recreational exposure to the sun has generally been measured by
    either the type of recreational activity or the frequency or
    duration of outdoor recreation (see table 22: in IARC, 1992). Some
    studies have also recorded sun exposure during vacations separately
    from other recreational sun exposure. At least one statistically
    significant positive association was found in 11 of 16 studies in
    which recreational exposure was considered (Adam  et al., 1981; Lew
     et al., 1983; Rigel  et al., 1983; Elwood  et al., 1985b; Dubin
     et al., 1986; Holman  et al., 1986; Osterlind  et al., 1988b;
    Zanetti  et al., 1988; Beitner  et al., 1990; Grob  et al., 1990;
    Nelemans  et al. 1993). The RRs in the highest categories of
    exposure were generally between 1.5 and 2.5. Only one study has
    shown a negative association with recreational exposure: MacKie &
    Aitchison (1982), in Scotland, observed an RR of 0.4 (95% CI
    0.2-0.9) for the highest category of hours a week in outdoor
    recreation. Their statistical model, however, included socioeconomic
    status and history of sunburn, both of which may measure sun
    exposure; thus overadjustment of the estimate of RR is likely. All
    measures of number or frequency of vacations in sunny places were
    positively associated with risk of melanoma. The RRs for the highest
    category of this variable ranged from 1.2 to 5 and the finding was
    statistically significant in six of eight studies (Lew  et al.,
    1983; Elwood  et al., 1985b, Osterlind  et al., 1988b, Zanetti  et
     al., 1988; Beitner  et al., 1990; Nelemans  et al., 1993).

         Sunburn

         Fifteen of seventeen studies of melanoma in which sunburn was
    recorded (see table 23 in IARC, 1992) showed a statistically
    significant, moderately to strongly positive association with a
    history of sunburn (MacKie & Aitchison, 1982; Lew  et al., 1983;
    Elwood  et al., 1985a; Green  et al., 1985; Sorahan & Grimley,
    1985; Elwood  et al., 1986; Holman  et al., 1986; Holly  et al.,
    1987; Osterlind  et al., 1988b; Zanetti  et al., 1988; MacKie  et
     al., 1989; Weinstock  et al., 1989; Beitner  et al., 1990;
    Elwood  et al., 1990; Nelemans  et al., 1993) (see table 23: IARC,

    1992). Twelve of these studies showed RRs greater than 2.0 for the
    highest category of sunburn. Positive associations were obtained for
    both sunburn in childhood and sunburn at any age. The greater
    consistency of the relationship of melanoma with sunburn compared to
    that with other exposure variables may indicate a specific
    association with sunburn per se or simply that sunburn is a more
    accurately measured indicator of sun exposure.

         Presence of other sun-related skin conditions

          Cutaneous microtopography

         Severity of sun damage to the skin as measured by cutaneous
    microtopography was strongly associated with risk of melanoma in
    Western Australia. The RR for the most severe grade of damage was
    2.7 (95% CI 1.4-5.0; p-value for trend=0.003; Holman  et al.,
    1984b). There was no similar relationship, however, in a study in
    Denmark, in which similar methods were used (Osterlind  et al.,
    1988b).

         Freckling of the skin has been shown to be associated with an
    increased risk of melanoma in several studies (see for example,
    Elwood  et al., 1986, 1990, Osterlind  et al., 1988b; Dubin  et
     al., 1986).

          Solar keratoses and other skin cancers

         The risk of melanoma was significantly increased in association
    with a past history of nonmelanocytic skin cancer with RRs of 3.7
    (95% CI 2.1-6.6) in Australia (Holman  et al., 1984b) and 3.8
    (1.2-12.4) in the USA (Holly  et al., 1987). Relative risks for a
    history of solar keratoses or "actinic tumours on the face" were
    similarly high (Green & O'Rourke, 1985; Dubin  et al., 1986).

         Exposure to artificial sources of UV

         Several studies have reported increased risks of cutaneous
    melanoma in users of sunlamps or sunbeds (IARC, 1992). In the most
    recent of these studies (Walter  et al., 1990), the relative risks
    for any use of sunbeds or sunlamps were 1.9 (95% CI 1.2-3.0) in men
    and 1.5 (95% CI 1.0-2.1) in women. Relative risk increased with
    increasing duration of use: for more than 12 months use the relative
    risks were 2.1 (0.9-5.3) in men and 3.0 (1.1-9.6) in women. Positive
    associations were found in two of four other case-control studies in
    which these exposures were studied (Swerdlow  et al., 1988; MacKie
     et al., 1989; see also IARC, 1992). Sun exposure is a potential
    confounding variable in studies of sunlamps and sunbeds but has not
    been taken into account. Specifically it was considered by Walter
     et al. (1990). Exposure to a group of other artificial sources of
    UV, including plan printers, laboratory equipment emitting UV,
    insect tubes, black lights and photocopiers, was not associated with

    melanoma in one case control study in Australia (Holman  et al.,
    1986) but for a similar group plus welding the RR was 2.2 (95% CI
    1.0-4.9) in Canada (Elwood  et al., 1986). Siemiatycki (1991) found
    no association between arc welding or other occupational exposure to
    UV in a study of occupation and cancer in Canada.

         A number of case-control studies have examined the association
    between melanoma and exposure to fluorescent lighting. Only the
    first such study found a statistically significant positive result
    (Beral  et al., 1982). Six subsequent studies found little or no
    evidence of a positive association (IARC, 1992).

    8.5.3  Cancer of the lip

         Cancer of the lip is defined as cancer of the vermilion border
    and adjacent mucous membranes and thus excludes cancers of the skin
    of the lip (WHO, 1977). Most are SCC and occur on the lower lip
    (Keller, 1970; Lindqvist, 1979), a site more heavily exposed to the
    sun than the upper lip (Urbach  et al., 1966). The inclusion of
    adjacent oral mucosa in the definition of cancer of the lip raises
    the possibility of confounding of tobacco and alcohol use (well
    established causes of cancer of the mouth; Tomatis, 1990).

         The incidence of cancer of the lip is much more common in men
    than women. Among men in the USA, it is some 20 times more common in
    whites than blacks and it is rare in black and Asian populations
    worldwide (Parkin  et al., 1992). Incidence of and mortality from
    cancer of the lip are substantially lower in migrants to Australia
    and Israel, who come from places with potentially lower sun
    exposure, than those born in these countries (Armstrong  et al.,
    1983; McCredie & Coates, 1989; Steinitz  et al., 1989).

         The incidence of cancer of the lip is higher in rural than
    urban areas (Doll, 1991) and descriptive studies have consistently
    reported higher incidence or mortality rates of this cancer in men
    with outdoor occupations such as farmers, agricultural labourers and
    fishermen. (Atkin  et al., 1949; Clemmesen, 1965; Gallagher  et
     al., 1984; Olsen & Jensen, 1987; Lynge & Thygesen, 1990). Four
    case-control studies of varying quality have examined the
    association between outdoor work and cancer of the lip. Keller
    (1970) compared 301 men with cancer of the lip with 301 oral cancer
    controls. Crude RRs for employment as a farmer or in any outdoor
    work were 4.0 and 2.6 respectively. The use of patients with oral
    cancer as controls allowed a rough adjustment for confounding with
    tobacco and alcohol use in this study. Spitzer  et al. (1975)
    compared 339 men with cancer of the lip with 199 matched population
    controls. The relative risk for any outdoor work was 1.52 and, for
    employment as a fisherman, 1.50 (p < 0.05 in both cases). Pipe
    smoking but not alcohol drinking was controlled in the analysis. The
    other two studies also showed a positive association between outdoor
    work and cancer of the lip (Lindqvist, 1979; Dardanoni  et al.,

    1984) but did not control for smoking and alcohol use and showed
    other methodological difficulties (IARC, 1992).

    8.5.4  Ocular cancers

         Ocular melanoma and other ocular cancers are dealt with in
    Chapter 10.

    8.5.5  Other cancers

         Observations of increasing mortality with increasing latitude
    have formed the basis of suggestions that cancers of the breast,
    colon and prostate may be prevented by increasing exposure to the
    sun (Garland & Garland, 1980; Gorham  et al., 1989, 1990; Garland
     et al., 1989, 1990; Hanchette & Schwartz, 1992), perhaps through
    proposed anti-carcinogenic actions of vitamin D (Eisman  et al.,
    1980; Colston  et al., 1989; Schwartz & Hulka, 1990; Ainsleigh,
    1993). Variation with latitude in medical care, death certification
    practices, diet and other lifestyle factors are alternative
    explanations for the latitude gradients in these cancers. These
    observations, therefore, can be taken only to raise hypotheses about
    indirect anti-carcinogenic effects of UV which require testing by
    other means.

    8.5.6  Action spectrum

         There are few data from which the action spectrum of UV
    carcinogenesis in humans can be inferred. For practical purposes,
    therefore, reliance is placed on action spectra which have been
    determined in experimental animals (see chapter 7).

         If, as is argued below, it is accepted that the formation of
    cyclobutylthymidine dimers is a step in the production of mutations
    that are associated at least with BCC and SCC, then it would be
    reasonable to accept the action spectrum for production of thymidine
    dimers in humans as possibly indicative of the action spectrum for
    production of nonmelanocytic skin cancer (figure 8.1). This action
    spectrum was determined by Freeman  et al. (1989) who irradiated
    untanned gluteal skin of 30 caucasian volunteers with 1 or 2 MED of
    narrow band UV at 275 nm, 282 nm, 290 nm, 296 nm, 304 nm, 334 nm and
    365 nm. At least five subjects were irradiated at each wavelength.
    Following irradiation, shave biopsies were taken of the irradiated
    skin, DNA extracted and incubated with Micrococcus luteus
    endonuclease to determine the frequency of endonuclease sensitive
    sites which indicate the presence of thymidine dimers. Figure 8.1
    shows that the effectiveness of UV in producing dimers increased

    from 275 nm to a peak at 296 to 304 nm and then fell by four orders
    of magnitude to the lowest measured level at 365 nm. This fall is
    quite consistent with that observed in the relative effectiveness of
    UV for carcinogenesis in mouse skin over the same wavelength range.
    The authors reported, also, that observations had been made at 385
    nm and 405 nm and that these wavelengths were "largely ineffective"
    in net production of dimers.

         Some indication of the effectiveness of UVA in causing skin
    cancer could be obtained from the experience of people exposed to
    high doses of UVA from sunbeds or in tanning salons. There is some
    evidence to suggest that these exposures are associated with an
    increase in risk of nonmelanocytic skin cancer, cutaneous melanoma
    and ocular melanoma (see above). In none of the studies carried out,
    however, has exposure to mainly UVA sources been distinguished from
    sources that emit both UVA and UVB. In addition, confounding with
    sun exposure is a possibility in all these studies.

    8.5.7  Dose-response

         In theory, the best way to estimate the dose-response
    relationship between UV and cancer in humans would be to measure the
    lifetime exposure to the sun of each member of a population and
    relate this measurement, by way of a cohort or a case-control study,
    to their probability of developing skin cancer. In practice this is
    very difficult. First, measurement of lifetime exposure to the sun
    generally requires recall of amount and pattern of sun exposure and
    use of protective measures against the sun for as long as 40 or more
    years in the past.

         Second, exposure in early life may be particularly important in
    determining risk of skin cancer, at least, and this is to a large
    extent outside the range of accurate recall. Third, while problems
    of recall could be solved by conduct of a cohort study, such a study
    would require measurements of sun exposure beginning in the first
    few years of life and repeated periodically throughout life to the
    age at which UV-related cancers become common. Such a study would be
    logistically and economically impracticable. The whole process is
    complicated by the fact that a simple measure of lifetime exposure
    may be insufficient for adequate description of dose-response. The
    evidence that period and pattern of exposure may be important
    determinants of some UV-related cancers (see below) means that to
    model incidence correctly as a function of exposure it would be
    necessary to collect data on dose-rate of UV and its variation
    throughout life.

    FIGURE 8.1

         One attempt has been made to estimate the dose-response
    relationship between UV and SCC and BCC in a retrospective cohort
    study of Chesapeake Bay watermen (Strickland  et al., 1989). The
    watermen work mainly in traditional fishing tasks that are regulated
    by law and have undergone few changes in the past 80 years. They
    presented, therefore, better prospects than most populations for the
    accurate recall of past sun exposure. Briefly, 808 caucasian
    subjects among 1 250 licensed watermen aged 30 years or more
    participated in an interview and skin examination. Present skin
    cancers were identified at examination and a history of past skin
    cancer was taken and; as far as possible, histological confirmation
    of the diagnosis was obtained for both present and past skin cancer.
    Totals of 47 SCC in 35 subjects, 60 BCC in 33 subjects and 344 solar
    keratoses in 202 subjects were identified. About a half of the SCC
    and BCC had been diagnosed before the survey examination while over
    90% of the solar keratoses were diagnosed at the examination.
    Histological confirmation was available for at least one lesion in
    51% of subjects with SCC, 72% with BCC and 13% with a solar
    keratosis (Vitasa  et al., 1990). The average annual exposure of
    facial skin to solar UVB was estimated for each subject by the
    combination of data from a personal history of mainly occupational
    outdoor exposure from 16 years of age to the interview, estimates of
    ambient UV radiance in the area in which the watermen lived and
    worked, based on the US network of Robertson Berger meters (Berger &
    Urbach, 1982), and field measurements of individual exposure to UVB
    under working conditions as recorded by polysulphone film
    dosimeters. Annual average ambient UVB for the area was estimated at
    about 1.14 x 106 J m-2 or 3260 MED assuming 350 J m-2 to be
    equivalent to 1 MED. Personal annual average exposure of facial skin
    ranged from about 1% to 8% of ambient UVB, i.e., 33 to 260 MED a
    year.

         The dose-response relationships obtained are shown in figure
    8.2. They were derived from the following mathematical model:

         prevalence = (exposure)a (age)b

         or, equivalently,

         log (prevalence) =  a log (exposure) +  b log (age).

    FIGURE 8.2

         While prevalence is stated in the above expressions and the y
    axis in figure 8.2 is labelled as prevalence, it is believed that
    the figures are cumulative incidence proportions in which each
    subject with more than one of a particular type of lesion was
    counted only once. Only cumulative incidence of SCC appeared to
    increase with increasing estimated exposure to solar UVB. The value
    of  a for SCC was 1.7. It should be noted however, that 4 subjects
    with SCC among 64 men in the lowest sun exposure category were
    excluded from the regression because of the anomalously high
    incidence in this group and because of the possibility that these
    subjects may have been hypersusceptible to UV. Apart from all having
    solar keratoses, however, they did not consistently show phenotypic
    features of high susceptibility to UV and it was not possible on any
    basis, therefore, to remove hypersusceptible subjects from the
    higher dose categories. The values of  a for BCC and solar
    keratoses were -0.2 and 0.005 respectively. It was suggested that
    because of the high sun exposure of all men in the study the
    cumulative incidence of BCC may have reached a saturation point at
    the lowest level of sun exposure such that further increases in
    exposure did not lead to further increases in incidence.

         An alternative approach to estimating dose-response
    relationships has been to examine the relationship, geographically,
    between incidence or mortality of skin cancers in whole populations
    and estimated or measured ambient UV (Armstrong, 1993). This work
    has been done in the context of estimating the increase in skin
    cancer that might be expected from some increment in ground-level UV
    caused by depletion of stratospheric ozone. It has made the
    fundamental assumption that the observed geographical variations in
    incidence or morbidity rates of skin cancer is due largely to
    geographical variations in UV. The results have commonly been
    expressed in terms of the biological amplification factor (BAF)
    defined as follows (de Gruijl and van der Leun, 1980).

         There are substantial uncertainties regarding the dose-response
    relationships shown in figure 8.2. First, the estimates of dose were
    based on exposure of the face only, excluded dose in the first 16
    years of life and did not take account of dose-rate or pattern of
    exposure. Second, there is likely to have been appreciable
    underascertainment of cancers diagnosed before the survey and high
    proportions of all lesions were not confirmed histopathologically.
    Third, the exclusion of four men with SCC must have had an

    appreciable effect on the value of  a for SCC. Fourth, no mention
    is made of having controlled in the analysis for potential negative
    confounding between cutaneous sun sensitivity and sun exposure. Such
    confounding might be expected to reduce the slope of the
    dose-response relationships. Finally, no measure of precision was
    given for the values of a; given the small numbers of subjects with
    SCC and BCC, those for these two cancers must have had quite wide
    95% confidence intervals. This study gives only a hint, therefore,
    as to what the dose-response relationship between solar UVB and
    nonmelanocytic skin cancer might be at one end of the human exposure
    range. No similar observations are available for cutaneous melanoma.

         BAF = (dI/I)/(dD/D)

    where dI equals a small increment in the existing incidence of skin
    cancer, I, which results, in the steady state, from a small
    increment dD in the existing biologically effective ambient level,
    D, of solar radiation (i.e., spectral dose weighted by the action
    spectrum for production of skin cancer).

         The best known and most commonly used geographical
    relationships between skin cancer incidence and ambient UV used for
    dose-response estimation are those established from data on
    nonmelanocytic skin cancer collected in a special survey in the USA
    in 1977 and 1978 and UV measurements collected through the US
    network of Robertson-Berger meters (Scotto  et al., 1983).
    Corresponding relationships were later established between melanoma
    incidence measured by the US SEER cancer registries and the
    Robertson-Berger meter data (Scotto and Fears, 1987). The
    relationship for melanoma is shown in figure 8.3. In this figure,
    the logarithm of age-adjusted annual incidence of melanoma is
    plotted against the logarithm of the annual average Robertson-Berger
    meter counts at each location. The lowest annual average meter count
    was 95 x 104 units at Seattle and the highest, Albuquerque, New
    Mexico, at 197 x 104 units. These values translate, approximately,
    into 0.75 x 106 and 1.56 x 106 J m-2 respectively. The
    estimated slopes of the regression lines were 0.7 in males and 0.8
    in females which are also the estimated BAFs for melanoma from these
    data.

    FIGURE 8.3

         Table 8.2 summarises the most recent estimates of the BAFs for
    nonmelanocytic skin cancer (BCC and SCC) and cutaneous melanoma from
    data in the USA and Scandinavia. These estimates were based on
    various action spectra: the response spectrum of the Robertson-
    Berger meter (Scotto  et al., 1983; Scotto & Fears, 1987; Pitcher &
    Longtreth, 1991), the CIE reference spectrum of McKinlay and Diffey
    (1987), Moan  et al. (1989), Moan and Dahlbach, (1992), and the
    Sterenberg-Slaper action spectrum for UV radiation of skin cancer in
    albino hairless mice (de Gruijl and van der Leun, 1991). Generally,
    the BAF for BCC lay between about 1.5 and 2.0. For SCC and melanoma,
    there was much greater variation in estimates. Those for SCC were
    between about 2.0 and 4.0 when based on the US data but between 1.0
    and 2.0 when based on Norwegian data. Similarly, BAFs for melanoma
    were between 0.3 and 0.5 when based on US data but between about 1.0
    and 3.0 when based on Scandinavian data. While an attempt had been
    made to adjust the US estimates for confounding with population
    constitutional sensitivity to the sun and sun-related behaviour,
    even the unadjusted estimates (ranging from 0.6 to 1.0; Scotto &
    Fears, 1987; Pitcher & Longstreth, 1991) were considerably less than
    those for Scandinavia. The most likely reasons for the differences
    between the estimates is differences in their error.
        Table 8.2  Recent estimates of the biological amplification factors (BAF) for
    nonmelanocytic skin cancer and cutaneous melanoma based on geographical correlations
    between average annual ambient UV and skin cancer incidence or mortality.
                                                                                              
    Authors                           Region              Sex     BAFs for BCC    BAFs for SCC
                                                                  Incidence       Incidence
                                                                                              

    Nonmelanocytic skin cancer

    Scotto et al, 1983                USA (8 centres)     M       1.3-2.6a        2.1-4.1
                                                          F       1.1-2.1         2.2-4.3

    de Gruijl & van der Leun 1991     USA (8 centres)     MF      1.4             2.5

    Moan et al 1989                   Norway (6 areas)    M       1.5-2.0a        1.2-1.5
                                                                                              

    Melanoma

    Scotto & Fears, 1987              USA (7 areas)       M       0.4c
                                                          F       0.5c

    Pitcher & Longstreth, 1991        USA 215 SMAs        M                       0.4d
                                                          F                       0.3d

    Table 8.2 (contd).
                                                                                              
    Authors                           Region              Sex     BAFs for BCC    BAFs for SCC
                                                                  Incidence       Incidence
                                                                                              

    Moan & Dahlback, 1992             Norway              M       1.9
                                                          F       3.2

                                      Finland             M       1.3
                                                          F       2.2

                                      Sweden              M       1.9
                                                          F       2.3
                                                                                              

    a  Exponential model used in which the value of the BAF varies with ambient UV, thus range
       of values given.
    b  Same data as used by Scotto et al (1983) except that a power model was used instead of
       an exponential model and the most recent action spectrum for UV carcinogenesis in mouse
       skin was assumed.
    c  Adjusted for population estimates of ethnic origin, pigmentary characteristics, use of
       sunscreens, and hours per week of outdoor exposure.
    d  Adjusted for population estimates of ethnic origin, household income, outdoor occupation
       and education.
             This approach to estimation of dose-response has assumed, among
    other things, that: the correct action spectrum has been used to
    weight spectral UV irradiance when producing a single figure for
    ground level UV in each area; all members of the populations giving
    rise to the incidence rates have lived their whole lives in their
    present environment; the skin cancer incidence rates have been
    measured accurately and, in particular, that their error does not
    correlate with ambient UV radiance; and that possible confounding of
    ambient UV with constitutional sensitivity to the sun and
    sun-related behaviour is either unimportant or has been taken
    adequately into account. None of these assumptions is likely to be
    correct in any of the estimates of BAF made so far and the estimates
    are all likely, therefore, to be inaccurate (Armstrong, 1993).

         It should be noted that even if dose-response relationships
    between UV and skin cancer incidence determined at the population
    level are accurate, they may not reflect dose-response relationships
    at the individual level. This is because of the averaging of
    exposure and risk that occurs at the population level. The actual
    population exposure of the skin at any level of ambient UV is an
    average of many different exposures and the actual population
    incidence of skin cancer is an average of many individual risks of
    skin cancer. The association of individual risk to individual
    exposure may be quite complex (as has been postulated for melanoma;
    Armstrong, 1988) but at the population level this could still

    translate into a simple exponential or power relationship between
    ambient exposure and incidence.

         No attempts have been made to estimate the dose-response
    relationships for UV and cancers other than skin cancers that may be
    caused by UV.

    8.5.8  Effects of pattern of exposure

         There are a number of aspects of the epidemiology of cutaneous
    melanoma in European populations that appear inconsistent with a
    simple relationship between it and sun exposure (English &
    Armstrong, in press). First, in many populations, melanoma occurs as
    commonly in women as in men, although men are more likely to work
    outdoors; second, there is a relative peak in incidence in middle
    life, which is not the pattern to be expected from life-long
    exposure to an environmental agent; third, it most frequently occurs
    on the back in men and lower limbs in women, sites which are not
    maximally exposed to the sun; finally, melanoma is more common in
    indoor than outdoor workers and in those of higher socioeconomic
    status than those of lower status.

         These observations led to postulation of the "intermittent
    exposure hypothesis" for the relationship of sunlight to melanoma.
    This hypothesis states that incidence of melanoma is determined as
    much (or more) by the pattern of sun exposure as by the total
    accumulated dose of sun exposure and, specifically, that infrequent
    or intermittent exposure of untanned skin to intense sunlight is
    particularly effective in increasing incidence of melanoma
    (Armstrong, 1988). If, as seems plausible, this postulated effect of
    intermittency of sun exposure is due to protective thickening and
    pigmentation of the epidermis with more continuous sun exposure,
    then intermittent exposure would be maximally effective if it
    occurred at intervals longer than the time it takes for the skin to
    return to its prior level of sensitivity after a single episode of
    sun exposure (two to three weeks).

         The epidemiological evidence, summarised above, that risk of
    melanoma is apparently not increased with total or occupational
    exposure to the sun but is increased with increasing
    non-occupational exposure to the sun, in particular increasing time
    spent in sun-related vacations in populations with otherwise
    comparatively low sun exposure (Klepp & Magnus, 1979; Lew  et al.,
    1983; Elwood  et al., 1985b; Osterlind  et al., 1988b; Beitner
     et al., 1990; Zanetti  et al., 1992), and episodes of sunburn
    (IARC, 1992) is consistent with this hypothesis.

         There are less grounds for believing that pattern of UV is
    important in causing nonmelanocytic skin cancer than there are for
    melanoma. Specifically, nonmelanocytic skin cancer as a whole, and
    BCC and SCC separately, are more common in men than women (Kricker

     et al., 1990; Parkin  et al., 1992; Marks  et al., 1993) and, in
    the largest, well-collected, population-based incidence series
    available, they showed steadily increasing incidence rates with age,
    except for a downturn in older age-groups possibly due to
    underascertainment (Scotto  et al., 1983). On the other hand,
    recent series of BCC have shown unexpectedly high proportions on the
    trunk (Kricker  et al., 1990; Magnus, 1991) and, as reviewed above,
    the relationship of nonmelanocytic skin cancer with occupational
    exposure to the sun is not at all clear. Moreover, there is some
    evidence of an association between sunburn and other indicators of
    intermittent exposure and BCC (Hunter  et al., 1990; Kricker  et
     al., 1991b).

    8.5.9  Interactions between UV and other agents

         5-methoxypsoralen and 8-methoxypsoralen (methoxsalen)

         Methoxsalen, 5-methoxypsoralen and some other psoralens
    (hereafter referred to collectively as psoralens) are used in
    combination with UVA (referred to as PUVA) for the treatment of
    psoriasis and to produce repigmentation in vitiligo. A working group
    of International Agency for Research on Cancer concluded in 1987
    that there was sufficient evidence that PUVA (defined as
    8-methoxypsoralen plus long-wave UV radiation) causes cancer in
    humans, based mainly on evidence of its association with SCC.

         Recent studies have confirmed and extended this conclusion.
    Stern & Lange (1988) reported on 5 to 10 year follow-up of patients
    treated with PUVA in the USA. In 13 384 person years of follow-up of
    1 380 people, 100 people had developed 391 SCC and 94 had developed
    218 BCC. There were statistically significantly increasing risks of
    both BCC and SCC occurring more than 58 months after entry into the
    cohort with increasing number of treatments with PUVA. Risk of BCC
    rose to 6.9 (95% CI 3.2-13.1) with 260 or more treatments and risk
    of SCC rose to 50.1 (95% CI 24.9-89.5). Therapeutic exposure to
    ionizing radiation, high doses of tar and UVB was unrelated to
    number of treatments with PUVA and, therefore, did not confound the
    observed dose-response relationship. Lindelöf  et al. (1991)
    reported on follow-up for an average of 6.9 years of 4 799 Swedish
    patients treated with PUVA. Risk of SCC increased with increasing
    number of treatments with PUVA. In those who received 200 or more
    treatments, the relative risk, with reference to the general Swedish
    population, was 30.7 (14.7-56.5) in men and 18.5 (0.5-103.2) in
    women. BCC were not recorded in this study.

         Cases of melanoma have been reported in patients treated with
    PUVA but the rates of melanoma in US and Swedish PUVA cohorts were
    little higher than those expected on the basis of population
    incidence rates. In the US cohort the relative risk was 1.46 (95% CI
    0.3-7.3; Gupta  et al., 1988), while in the Swedish cohort it was

    1.1 (0.2-3.2) in men and 0.8 (0.1-3.0) in women (Lindelöf  et al.,
    1991).

         It is not possible in any of these studies to estimate what the
    risk of skin cancer would have been if the patients had been treated
    with UV alone. It would seem reasonable to assume, however, on the
    basis of the evidence from experimental animals (see above) that
    administration of psoralens potentiated the effect of UV. In
    addition analysis of clinical data even suggests that this
    carcinogenicity of PUVA exceeds that of UVB treatment by an order of
    magnitude (Slaper, 1987).

         Ionizing radiation

         Ionizing radiation causes nonmelanocytic skin cancer, mainly
    BCC, generally after a long latent period (Shore, 1990; Fragu  et
     al., 1991; Hogan  et al., 1991; Sadamori  et al., 1991; Moller
     et al., 1993). Incidence in the atomic bomb survivors was related
    strongly to exposure, although not for those closest to the blast
    (Sadamori  et al., 1991). There is evidence that ionizing radiation
    and UV may act synergistically in causing skin cancer. In a study of
    skin cancer following irradiation of the scalp for ringworm, Shore
     et al. (1984) found a substantial excess of BCC in white subjects.
    In blacks, no skin cancers were observed when 13 would have been
    expected from the rate on the UV-shielded scalp in whites. This lack
    of skin cancers in irradiated blacks suggest that UV radiation
    contributes more than additively to the production of these cancers
    in irradiated whites. This inference is supported by the observation
    that the excess of BCC around the UV-exposed fringe of the scalp in
    irradiated whites was some four to five times higher per unit dose
    and area than that on the UV-shielded hairy scalp (Shore  et al.,
    1984). Further, Modan  et al. (1993) observed, in a separate cohort
    of people irradiated for ringworm of the scalp, that sunbathing
    increased the risk of skin cancer 2.6 times (95% CI 1.1-6.1). All
    except one of 39 cancers were BCCs and 60% were on the scalp, thus
    suggesting that the ionizing radiation had also contributed to their
    causation. However, no formal analysis of interaction was reported
    from this study.

         Other agents

         One or more SCCs occur in 30% to 50% of patients with
    epidermodysplasia verruciformis, a rare chronic skin disease in
    which there are multiple benign skin lesions caused by a variety of
    different types of human papilloma viruses (HPV; Quan & Moy, 1991).
    HPV DNA sequences have been identified in the SCC occurring in these
    patients and most of the cancers occur on sun-exposed sites. This
    coincidence suggests that the UV interacts with the virus in causing
    these cancers. Benign warts are also increased in frequency in renal
    transplant patients and it is possible that they, with UV, are
    involved in causing the excess of nonmelanocytic skin cancers that

    occur in these patients (Bouwes-Bavinck, 1992). These cancers, too,
    occur mainly on sun-exposed sites.

    8.5.10  Mechanisms of UV carcinogenesis

         Precursor lesions

         Solar keratoses are very strongly related to risk of
    nonmelanocytic skin cancer (Kricker  et al., 1991a) and it is a
    matter of clinical observation that SCC often occurs in lesions
    which, previously, had the clinical appearances of solar keratoses.
    There is evidence that solar keratoses are, themselves, caused by
    exposure to the sun. Thus solar keratoses could be the visible
    expression of an early, UV-caused, mutational step in the pathway to
    skin carcinogenesis.

         Analogously with solar keratoses, benign melanocytic naevi are
    commonly observed in clinical or histopathological contiguity with
    melanomas and have been found, epidemiologically to be strongly
    related to risk of cutaneous melanoma (Holman & Armstrong, 1984a;
    Armstrong & English, 1988). Naevi may, therefore, represent an
    early, UV-caused, mutational step in the development of melanoma. In
    this regard, it is interesting to note that the most active
    development of naevi occurs during the first 15 years of life, thus
    providing a possible link to early life sun exposure as suggested by
    migrant data (see above). The appearance of naevi in children who
    have received chemotherapy for cancer may indicate that they can be
    caused by mutagens other then UV (Hughes  et al., 1989; Baird  et
     al., 1992; Green  et al., 1992).

         DNA damage

         The evidence of much increased incidence rates of BCC, SCC
    cutaneous melanoma and tumours of the anterior eye in patients with
    xeroderma pigmentosum (XP) (see above) strongly suggests that
    unrepaired, UV-induced photoproducts in DNA form part of the genesis
    of these cancers. Almost all forms of XP have an inherited
    deficiency in excision repair of DNA photoproducts (Kraemer  et al.,
    1987). The position is made somewhat less clear, however, by the
    existence of two other inherited syndromes involving deficiency in
    DNA excision repair, trichothiodystrophy and Cockaynes syndrome,
    that are not associated with increased incidence rates of skin
    cancer (Bridges, 1990; Barrett  et al., 1991).

         There is evidence also that low levels of DNA repair may be
    associated with skin cancer in the general population, although
    there is some disagreement in the results so far obtained .
    Munch-Petersen  et al. (1985) studied 29 patients with multiple
    skin neoplasms, 19 with multiple BCC only and 10 with both BCC and
    squamous lesions (SCC, solar keratosis, or Bowen's disease), and 25
    control subjects with roughly the same age and sex distributions.

    Patients with both BCC and squamous lesions had much higher
    UV-induced DNA synthesis in lymphocytes, as measured by
    incorporation of [3H]thymidine following irradiation with a Philips,
    TUV, 6 watt lamp (peak emission at 254 nm), than did controls.
    Patients with multiple BCC were not different from controls. The
    authors attributed this difference to possible defects in the DNA
    ligation process. Roth  et al. (1987) studied the rate of loss of
    dithymidine dimers in DNA of cultured fibroblasts that had been
    irradiated with a Philips 6-V germicidal lamp with peak emission at
    254 nm. Dimers were measured with a monoclonal antibody highly
    specific for the conformational change in DNA caused by the dimers.
    No significant difference was seen between 16 BCC cases and 30
    controls in the rate of loss of dimers. In ten melanoma patients,
    however, the mean percentage of bound antibody at 60 minutes was
    50.5 (SD 18.2) compared with 29.8 (SD 5.7) in controls (p=0.001)
    thus suggesting that average dimer repair capacity was less in the
    melanoma patients than the controls. Alcalay  et al. (1990)
    compared 22 patients with a history of one (13 patients) or more BCC
    in the past with 19 healthy volunteers. Each subject was irradiated
    at two locations on the lower back with 1 MED of UV from a 150 W
    xenon arc solar UV simulator from which wavelengths < 295 nm had
    been removed by a 1 mm WG 320 filter. Shave biopsies of skin were
    taken from one of the sites immediately and the other 6 hours after
    irradiation and the concentrations of pyrimidine dimers in DNA
    measured by the micrococcus luteus endonuclease assay. Dimer yields
    immediately after irradiation were similar in BCC patients and
    controls but, after 6 hours, the proportion of dimers that had been
    repaired was 22% in the BCC patients compared with 33% in the
    controls (p=0.06). Wei  et al. (1993) measured DNA repair capacity
    in 88 patients with a history of one or more histologically
    confirmed BCCs and 135 control subjects. Repair capacity was
    measured by assaying the capacity of lymphocytes to repair
    UV-induced damage in DNA of a nonreplicating recombinant plasmid.
    Damage was induced by irradiating the plasmid with either 3500 J
    m-2 or 7000 J m-2 of UV at 254 nm. The BCC cases had a
    significantly lower mean DNA repair capacity (p=0.047) than controls
    only when 29 controls with a family history of BCC or who themselves
    had an actinic keratosis were removed from the comparison.

         It now seems highly likely that UV can mutate the p53 tumour
    suppressor gene in human skin and that UV-induced mutations in this
    gene may be involved in the aetiology of some human cancers. First,
    a recent study has shown that mutations that are strongly suggestive
    of a UV effect, CC-TT changes at dipyrimidine sites, could be found
    in 17 of 23 (74%) samples of sun-exposed normal skin in Australian
    skin cancer patients compared with 1 of 20 (5%) samples of skin from
    sites not exposed to the sun (Nakazawa  et al., in press). Focal
    overexpression of p53 protein has also been observed in normal, sun
    exposed skin adjacent to BCC in 37 of 39 subjects but only in one
    keratinocyte in 14 samples of buttock skin from the same subjects
    (Shea  et al., 1992). Second, several studies have found mutations

    at dipyrimidine sites in the p53 gene in 36% to 56% of BCC (Rady  et
     al., 1992; Molès  et al., 1993; Ziegler  et al., 1993). In three
    similar studies of SCC one found these mutations in 14 (57%) of 24
    cases (Brash  et al., 1991) one in two of 10 cases (Pierceall  et
     al., 1991) and, the other, none of 13 (Molès  et al., 1993). The
    more detailed findings of these five studies are summarised in table
    8.3. Mutations were found in the p53 gene, following examinations of
    varying extent, in 46% of BCC and SCC, the majority of which (88%)
    were base substitutions or one or two base deletions at dipyrimidine
    sites (the proportion expected by chance is 75%). The highest
    proportions of these mutations resulted from C -> T, CC -> TT and
    C -> A base changes which is consistent with what would be expected
    if they had been caused by UV (Brash  et al., 1991). Several
    studies have found p53 protein by immunostaining (evidence of
    overexpression of possibly abnormal p53 protein) in from 42% to 83%
    of BCC and 56% to 65% of SCC (Barbareschi  et al., 1992; Shea  et
     al., 1992; Stephenson  et al., 1992, Ro  et al., 1993) and are
    thus reasonably consistent with the observations as to prevalence of
    mutations.

         The position with respect to p53 mutation is much less clear
    for other cancers possibly caused by UV. A number of studies have
    reported on the prevalence of detectable p53 protein in primary
    melanomas: the proportions have varied from 3.6% to 97% (see, for
    example, Stretch  et al., 1991; Akslen & Morkve, 1992; Lassam  et
     al., 1993; Ro  et al., 1993). There has been only one report of a
    p53 mutation in melanoma, a C -> T transition in codon 248 of the
    gene in the melanoma cell line SK-MEL-13 (Volkenandt  et al.,
    1991). The comparative lack of evidence of expression of p53 protein
    in melanocytic naevi (Yu  et al., 1992; Lassam  et al., 1993)
    would suggest that if p53 gene mutation is important in the genesis
    of melanoma it is not the earliest mutation. The protein product of
    p53 was detected in 12 of 18 (67%) choroidal melanomas (Tobal  et
     al., 1992) but none of 7 choroidal naevi. The DNA of exons 5, 7
    and 8 of the p53 was sequenced from two of the strongly positive
    choroidal melanomas: a mutation was found in each, one G -> T and
    one C -> G.

         Mutations in the ras oncogenes appear to be less frequent than
    in the p53 gene in nonmelanocytic skin cancers. Mutations were
    sought in one or more codons of Ha-ras, Ki-ras and N-ras in 77 BCC
    and found in five (6%) (van der Schroeff  et al., 1990; Lieu  et
     al., 1991; Campbell  et al., 1993). There were two mutations, one
    in a dipyrimidine site, in codon 61 of Ha-ras and three mutations at
    dipyrimidine sites in codon 12 of Ki-ras; two mutations were C ->
    A, one was C -> T and one was A -> T. These mutations were also
    sought in 79 SCC and found in 7 (9%) (van der Schroeff  et al.,
    1990; Corominas  et al., 1991; Lieu  et al., 1991; Campbell  et
     al., 1993). There were three mutations in dipyrimidine sites in
    codon 12 of Ha-ras, three, but none in dipyrimidine sites, in codon
    61 of Ha-ras and one at a dipyrimidine site in codon 12 of Ki-ras;

    three mutations were C -> A, one was C -> T and three were T ->
    A. Overall, 8 of 12 (67%) mutations were at dipyrimidine sites.
    There is nothing about the pattern of base changes that would favour
    UV as the cause of the mutations. The comparative rarity of ras gene
    mutations in nonmelanocytic skin cancers is supported by a study of
    26 such tumours in Japanese patients with xeroderma pigmentosum
    (Ishizaki  et al., 1992). Only one mutation was found, an A -> T
    change in codon 61 of Ki-ras.

         Mutations in ras oncogenes have been demonstrated in some human
    cutaneous melanomas. In three series totalling 59 primary melanomas
    (van't Veer  et al., 1989; Shukla  et al., 1989; Albino  et al.,
    1991), ras gene mutations were found in 11 (19%). The mutations
    occurred in codon 61 of N-ras (6), codon 13 of N-ras (4), codon 12
    of Ki-ras (3) and codon 12 of Ha-ras (1) (two mutations were found
    in each of 3 melanomas). Ten of the 13 mutations (77%) were at
    dipyrimidine sites and nine were C -> A, three C -> T, one T -> C
    and one T to A or G (IARC, 1992). The pattern of base changes is not
    such as to suggest that UV was responsible for the mutations.
    Mutations were sought but not found in codons 12, 13 and 61 of
    Ha-ras, Ki-ras and N-ras in up to 68 uveal melanomas (Mooy  et al.,
    1991; Soparker  et al., 1993).

    Table 8.3  Summary of results of detection of mutations in the p53d
    gene of BCC and SCC (Brash et al, 1991; Pierceall et al 1991; Rady
    et al., 1992; Molès et al., 1993; Ziegler et al., 1993).
                                                                      
    Mutations Detected                                BCC     SCC
                                                                      

                                                      (60)a   (46)

    Any mutation                                      52%     39%

    Base substitutions or 1 or 2 base deletions       45%     35%
    at a dipyrimidine site

    Types of base substitutions or 1 or 2 base        52%c    38%
    deletions at dipyrimidine sites b

              C -> T                                  13%     19%

              CC -> TT                                13%     31%

              C -> A                                  10%     6%

              C -> G                                  3%      0%

    Table 8.3 (contd.)
                                                                      
    Mutations Detected                                BCC     SCC
                                                                      

              T -> C                                  3%      0%

              T -> A                                  3%      0%

              G -> C                                  0%      6%

              C deleted                               3%      0%

              CC deleted
                                                                      

    a  Total numbers of cancers examined.
    b  In consistency with the format adopted by Brash et al (1991) and
       Ziegler et al (1993), four base substitutions reported as G -> A
       (Pierceall et al, 1991; Rady et al, 1992) were represented as
       C -> T and two reported as G -> T (Pierceall et al, 1991; Molès
       et al, 1992) were represented as C -> A.
    c  Proportions of all point mutations at dipyrimidine sites.

         Oxidative processes

         Oxidative effects are an important consequence of exposure to
    UV, especially UVA (see Chapter 6) and if they are involved in
    mediating its carcinogenic effects in humans it might be expected
    that anti-oxidant vitamins would reduce the risk of UV-related
    cancers. There is some evidence of such effects, mainly for
    cutaneous melanoma. Risk of melanoma was slightly but not
    significantly reduced by high intakes of vitamin E and carotene and
    high plasma concentrations of -carotene and -tocopherol in a
    case-control study in the USA (Stryker  et al., 1990). More
    persuasively, a nested case-control study in a cohort study in
    Finland showed significantly lower serum concentrations of -carotene
    and -tocopherol in 10 cases of melanoma and 18 controls (Knekt  et
     al., 1991). Melanoma was the only cancer in this study to be
    associated with low levels of anti-oxidant vitamins. A similar
    result was obtained for -tocopherol in a case-control study of
    melanoma carried out in Moscow (crude relative risk for highest
    tertile of -tocopherol was 0.08, 95% CI 0.01-0.38; Zaridze  et al.,
    1992); risk was not reduced by high serum concentrations of
    -carotene. As to nonmelanocytic skin cancer, in a hospital-based
    case-control study covering 53 cases of BCC and 35 of SCC, high
    intakes of vegetables were protective and cases had a significantly
    lower mean level of -carotene than did controls (Kune  et al.,
    1992). It should be noted, however, that cases were on average, five
    years older than controls. No evidence was found for a protective
    effect of dietary carotenoids with vitamin A activity, vitamin C or

    vitamin E in a study of diet and BCC in a cohort of US nurses
    (Hunter  et al., 1992).

         Immune-suppression

         UV has been shown to suppress immune functions in humans (see
    Chapter 9). In addition there are a number of lines of evidence that
    link immune suppression from other sources with skin cancer in
    humans.

         As early as 1980 it was clear that renal transplant patients,
    who receive long-term immunosuppressive therapy, had an increased
    incidence of nonmelanocytic skin cancer (Kinlen  et al., 1979;
    Hardie  et al., 1980). Among 290 transplant patients in Queensland,
    Australia, it was estimated that the incidence of nonmelanocytic
    skin cancer was 21 times that in the general population Hardie  et
     al., 1980); the ratio of BCC to SCC was reversed from 4:1 in the
    general population to 1:1.7 in the transplant patients suggesting
    that the incidence of SCC was particularly increased in them.
    Similar results were obtained in Canadian and Dutch series of
    transplant patients (Gupta  et al., 1986; Hartevelt  et al., 1990)
    In the Dutch study it was estimated that the incidence of SCC was
    250 times higher than that in the general population and that of BCC
    10 times higher.

         There is evidence to suggest that the nonmelanocytic skin
    cancers occurring in renal transplant patients are caused by sun
    exposure. First their site distribution is similar to that of all
    nonmelanocytic skin cancers and strongly favours sun exposed sites
    (Hartevelt  et al., 1990). Second, all 5 nonmelanocytic skin
    cancers found by Boyle  et al. (1984) in 94 transplant patients
    occurred in the 17 patients judged to have high sun exposure. In a
    more thorough study based on 137 Dutch transplant patients, 20 of
    whom had SCC, 7 BCC and 9 both types of cancer, Bouwes-Bavinck
    (1992) found a relative risk for 20 000 or more cumulated hours of
    sun exposure compared with 10 000 or less of 97.5 (95% CI 6.6-1444)
    for SCC and 49.3 (2.8-878) for BCC thus suggesting that sun exposure
    had a strong effect on development of the transplant related skin
    cancers.

         The incidence of cutaneous melanoma has also been observed to
    be increased in patients with renal transplants (relative risk 3.9,
    95% CI 1.4-8.5; Hoover, 1977; Greene  et al., 1981). All 14 renal
    transplant patients with melanoma reported by Greene  et al. (1981)
    had fair complexions, light-coloured hair, light-coloured eyes and a
    tendency to freckle thus suggesting that there may have been an
    interaction between sun exposure and immune suppression in causing
    their melanomas. The rapid appearance of new melanocytic naevi has
    been reported in two renal transplant patients (Barker & MacDonald,
    1988; McGregor  et al., 1991). Melanoma incidence is also increased
    in patients with lymphohaematopoetic neoplasms which are associated

    with immune suppression (Greene & Wilson, 1985; Tucker  et al.,
    1985b; Travis  et al., 1991, 1992) and has been reported to be
    increased in patients with human immunodeficiency virus infection
    (McGregor  et al., 1992; Reynolds  et al., 1993). Rapid appearance
    of melanocytic naevi has also been reported in HIV infection (Duvic
     et al., 1989).

    8.6  Conclusions

         UV exposure of the skin has both beneficial and harmful
    effects. The beneficial effects include photochemical production of
    vitamin D and widely believed but poorly documented effects on
    general well-being. The harmful effects include sunburn,
    phototoxicity, photoallergy, benign abnormalities of melanocytes
    (freckles, melanocytic naevi and solar or senile lentigines) a range
    of other chronic abnormalities resulting from UV injury to
    keratinocytes, blood vessels and fibrous tissue, often described
    together as "photoageing", skin cancer (melanoma and non-
    melanocytic cancer) and possibly cancer of the lip.

         The indirect epidemiological evidence that sun exposure causes
    cutaneous melanoma and non-melanocytic cancer is strong. Their
    incidence is less in darker-skinned ethnic groups than in those with
    lighter skins residing in the same geographic area and, within
    populations that are reasonably homogeneous ethnically, risk of skin
    cancer increases with decreasing pigmentation of the skin and
    reduced ability to produce a protective tan. Albinos, who lack
    cutaneous pigmentation, appear to have an increased risk of
    non-melanocytic skin cancer but not melanoma. The anatomic site
    distibution of SCC favours the head and neck and upper limbs, sites
    that are more or less continuously exposed to the sun when outdoors.
    This concentration is less pronounced for BCC and melanoma, but both
    are rare on sites that are rarely exposed to the sun.

         Within countries covering an appreciable span of latitude, an
    inverse relationship within latitude and incidence of both
    non-melanocytic skin cancer and cutaneous melanoma has generally
    been observed. The incidence of both cancers is substantially lower
    in migrants from the United Kingdom (an area of low solar
    irradiance) to Australia (an area of high solar irradiance) than it
    is in persons of similar ethnic origin born in Australia. This
    difference is not observed for BCC and melanoma in those who migrate
    to Australia within the first ten years of life; this observation
    may suggest that sun exposure in childhood is particularly important
    in determining subsequent risk of skin cancer. Similar observations
    have been made for melanoma in migrants to other countries with high
    solar irradiance. Risk of melanoma has been shown to increase with
    increasing lifetime average ambient solar irradiance at an
    individual's places of residence.

         The direct epidemiological evidence linking sun exposure and
    skin cancer is weaker. Estimated total sun exposure of individuals
    has not been consistently associated with risk of either
    non-melanocytic skin cancer or melanoma. Indicators of benign sun
    damage to the skin, however, are positively associated with risk of
    both types of skin cancer. Occupational sun exposure has been found
    to be weakly associated with non-melanocytic skin cancer in a few
    studies but not consistently associated, either positively or
    negatively, with melanoma. Non-occupational or recreational sun
    exposure is consistently and quite strongly associated with risk of
    melanoma whereas there are insufficient data to draw a conclusion
    regarding non-occuptional exposure and non-melanocytic skin cancer.
    The same is substantially true of history of sunburn although a few
    observations suggest that it is associated with non-melanocytic skin
    cancer.

         Together, the indirect and direct evidence is sufficient to
    conclude that sun exposure causes both melanoma and non- melanocytic
    skin cancer.

         Use of sunlamps or sunbeds has not been consistently associated
    with risk of non-melanocytic skin cancer in several rather poorly
    conducted studies. There is stronger evidence that it may be
    associated with risk of melanoma in a number of better studies but
    possible confounding of use of sunlamps and sunbeds with sun
    exposure has not been consistently controlled.

         Cancer of the lip is much more common in white than black
    populations and is less frequent in migrants from areas of low
    ambient solar irradiance to Australia and Israel than in those born
    in these countries. It is associated with outdoor work but possible
    confounding with tobacco and alcohol use has not been adequately
    controlled in any study.

         There are no data in humans from which the action spectrum for
    production of either non-melanocytic skin cancer or cutaneous
    melanoma can be inferred directly.

         Problems in the quantitative measurement of the sun exposure of
    individuals have prevented the determination, in humans, of the
    relationship between individual dose of UV and risk of any form of
    skin cancer. The observed geographical relationship between ambient
    UV irradiance and incidence of skin cancers has been used to
    estimate the quantitative relationship between ambient UV irradiance
    and population risk of these cancers. Exponential or power
    relationships have generally given an adequate fit to the data, but
    provide rather variable estimates of the proportional increase in
    incidence of skin cancer per unit proportional increase in
    biologically effective UV irradiance. The accuracy of these
    estimates is uncertain, particularly because of difficulties in the
    accurate measurement of skin cancer incidence and difficulties in

    control of likely confounding with cutaneous sensitivity to the sun
    and sun-related behaviour.

         A number of features of the epidemiology of melanoma suggest
    that infrequent or intermittent exposure of skin that is unadapted
    to sun exposure may be particularly important in its causation. This
    hypothesis is supported by the relatively strong evidence relating
    non-occupational (recreational) sun exposure and sunburn to
    melanoma. There is little evidence that pattern of sun exposure is
    important in the aetiology of non-melanocytic skin cancer but few
    directly relevant observations exist.

         The much increased rates of all skin cancers in patients with
    xeroderma pigmentosum, who are deficient in the capacity to repair
    UV-induced DNA damage, suggest that direct UV damage of DNA may be a
    step in the causation of these cancers. This suggestion has been
    supported by observation of UV specific mutations of the p53 tumour
    suppressor gene in a proportion of patients with non-melanocytic
    skin cancer. No such evidence is available for melanoma. Oxidative
    and immune suppressant effects may also contribute to the capacity
    of UV to cause skin cancers.

    9.  HUMAN STUDIES: IMMUNE FUNCTION

    9.1  Immune Function Assays

         Several investigators have examined the effect of UV on contact
    hypersensitivity (CHS) responses to dinitrochlorobenzene (DNCB) or
    other contact sensitizers in humans (Hersey  et al., 1983a, 1983b;
    O'Dell  et al., 1980; Kalimo  et al., 1983; Halprin  et al.,
    1981; Friedmann  et al., 1989; Sjovall  et al., 1985; Yoshikawa
     et al., 1990; Vermeer  et al., 1991; Cooper  et al., 1992). Some
    of these studies are inconclusive because they used patients with
    skin diseases or recent UV-exposure, had insufficient subjects for
    statistical analyses, or used subjective assessments of CHS. Hence
    only 3 of these studies are discussed here.

         Yoshikawa  et al. (1990) exposed human buttock skin to 4 daily
    UV doses of 1440 J m-2 using a high pressure mercury vapour lamp
    (290-320 nm).Immediately after the last exposure, the irradiated
    site was sensitized with 2000 µg DNCB. The inner surface of the
    forearm was challenged 30 days later with 50 µg DNCB and CHS was
    assessed. In this study 40-45% of healthy irradiated adults failed
    to develop CHS. They designated these subjects UVB-S and suggested
    that, as in mice, susceptibility to UV is genetically controlled.
    Among biopsy proven skin cancer patients a much higher percentage
    (92%) of UV-exposed individuals failed to developed CHS. The UVB-S
    phenotype appeared to be a risk factor for the development of skin
    cancer. Suppression of CHS was also demonstrated in 50% of
    black-skinned individuals indicating that melanin does not protect
    against the deleterious effects of UVB on the development of CHS
    (Vermeer  et al., 1991). Attempts to resensitize healthy UVB-S
    individuals through normal skin following primary sensitization to
    DNCB through irradiated skin were generally successful; however, 50%
    of skin cancer patients failed to respond to resensitization
    attempts suggesting that these patients developed immunological
    tolerance similar to that demonstrated in experimental animal models
    (Yoshikawa  et al., 1990; Vermeer  et al., 1991).

         Cooper  et al. (1992) exposed human buttock skin (using FS
    lamps) to 0.75 or 2 MEDs of UVB (1 MED = 291 J m-2 to 325 J m-2
    depending on the individual) for 4 days and sensitized with 30 µg
    DNCB through irradiated skin immediately after the last exposure.
    Subjects were challenged 3 weeks later with 4 serial 2-fold
    dilutions of DNCB, the highest of which was 12.5 µg. In addition,
    some subjects were exposed to 4 MED (moderate sunburn) and
    sensitized 3 days later with DNCB. Analysis of overall individual
    responses revealed decreased frequencies of fully successful
    immunizations in all UVB exposed groups. Increasing doses of UVB
    resulted in a linear decrease in immunological responsiveness to
    DNCB. Only 5% of individuals exposed to 2 MED had strong positive
    responses as opposed to 73% of unexposed individuals. The rate of
    immunologic tolerance to DNCB (lasting up to 4 months) in the groups

    that were initially sensitized on skin receiving erythemagenic doses
    of UVB was 31% compared to 7% in controls. Similar results were
    observed with UVA (320-340 nm) exposure (Cooper, 1993). The
    differences in the Cooper and Yoshikawa studies may have to do with
    different sensitization and challenge regimens, and differences in
    methods used to quantitate the CHS response. Also, Cooper controlled
    for diminished levels of CHS that occur in menstruating women except
    during midcycle (Oberhelman  et al., 1992) by sensitizing all
    female subjects 14 days after the onset of menses.

         Despite differences in the 2 studies it appears that
    unresponsiveness following the application of a contact sensitizer
    on UV-exposed skin can be induced in some parts of humans following
    exposure to moderate levels of UV and that some individuals become
    immunologically tolerant to the sensitizer in a manner reminiscent
    of animal studies. Cooper  et al. (1992) also demonstrated
    modulation (although not clear-cut suppression) of contact
    sensitivity to diphenylcyclopropenone (DPCP) in subjects exposed to
    4 MED and sensitized 3 days later through unirradiated skin. A
    similar effect was not observed with the 2 MED exposure regimen,
    either because the dose was lower or the sensitization was
    immediately following the last exposure. Hence the systemic effects
    observed in mice may also occur in humans, although currently the
    experimental data to support such effects are minimal.

         As in the mouse, UV treatment of human skin resulted in altered
    antigen presentation. UV caused depletion of Langerhans (CD1a+DR+)
    cells followed by an influx of CD1a- DR+ macrophages that
    preferentially activate CD4+ cells (suppressor-inducer) which, in
    turn, induce maturation of CD-8+ suppressor T lymphocytes and
    deregulate lymphocyte activation (Baadsgaard  et al., 1990; Cooper
     et al., 1986). Rasanen  et al. (1989) reported a 70-80%
    suppression in the ability of epidermal cells exposed to 2000 J
    m-2 UVB  in vivo to present PPD or HSV to lymphocytes  in vitro.
    In these studies recovery was observed 3 and 7 days post exposure.
    The UV wavelengths responsible for induction of CD1a-DR+ cells were
    found to lie predominantly within the UVB band and to a lesser
    extent in the UVC band; UVA was a poor inducer of these
    non-Langerhans cell antigen presenting cells (Baadsgaard  et al.,
    1987,1989). Hersey  et al. (1983b) reported an increase in CD8+ T
    cells and a decrease in CD4+ T cells in subjects exposed for 1
    hour/day for 12 days to natural sunlight and a suppressor T cell
    activity which lasted in many subjects for up to 2 weeks. Filtration
    of solarium radiation through mylar prevented these changes
    suggesting effects were due to UVB (Hersey  et al., 1988). Hence in
    humans as in mice, UVB appears to act by altering antigen
    presentation in ways that favours suppressor cells.

         Robinson & Rademaker (1992) studied 61 patients with two or
    more BCC for occurrence of a further BCC. Patients were counselled
    to avoid sun exposure, however there was a clear distinction between

    those who reduced their sun exposure and those who did not. The
    numbers of new BCC occurring after 36 months was determined by
    regular examination of all patients. Plasma lymphocyte
    subpopulations were measured at 0, 6, 12 and 18 months. At 36 months
    an index of sun exposure was developed from questions on sun-related
    behaviour before and after entry to the study. All 35 patients with
    high sun exposure had low T helper to T suppressor cell (CD4/CD8)
    ratios. The mean number of BCCs occurring during follow-up was 5.5
    in those with high sun exposure and low CD4/CD8 ratios, 2.2 in those
    with low sun exposure and low CD4/CD8 ratios and 1.1 in those with
    low sun exposure and high CD4/CD8 ratios. The difference between the
    first and the third groups was statistically significant. While
    there is total confounding between high sun exposure and low CD4/CD8
    ratios in this study it is consistent with the possibility that the
    effect of sun exposure on risk of recurrent BCC was mediated by way
    of its effect on cell-mediated immunity.

         UV from solaria suppressed natural killer(NK) cell activity in
    the blood of subjects exposed for 1 hr/day for 12 days and tested 1
    and 7 days after exposure. NK activity returned to normal 21 days
    post exposure. Filtration of UV through mylar did not affect this
    response; hence effects on NK activity were attributed to UVA
    (Hersey  et al., 1983a; Hersey  et al., 1988).

    9.2  Susceptibility to Tumours, Infectious and Autoimmune

    Diseases

         As previously indicated Yoshikawa  et al. (1990) found that
    suppression of contact sensitivity following UV exposure was more
    common among skin cancer patients than in healthy subjects. In their
    study only skin cancer patients developed an immunological
    tolerance. They suggested that individuals who were phenotypically
    sensitive to the immunosuppressive effects of UV were at greater
    risk of skin cancer.

         Adverse effects of UV on 4 types of human infections have been
    reported. Smallpox lesions were made larger by exposure to sunlight
    (Finsen 1901). Lesions from Herpes Simplex Virus(HSV) types I and II
    were reactivated by exposure to UV (Spruance, 1985; Klein, 1986).
    Using the criteria established by Yoshikawa  et al. (1990) for the
    UVB-S phenotype, Taylor  et al. (1993) reported that 66% of
    individuals who had a strong history of HSV lip lesions provoked by
    sun exposure were UVB-S as compared to 40-45% in the general
    population and 92% in skin cancer patients (Yoshikawa  et al.,
    1990). Also, exposure of immunosuppressed patients to sunlight led
    to an increased incidence of viral warts caused by papilloma virus,
    presumably due to UVB exposure (Boyle  et al., 1984; Dyall-Smith &
    Varigos, 1985). Hence there is some indication that UV may
    exacerbate certain infections in humans.

         Recent attention has been paid to the effects of UV on immune
    response in patients with human immune deficiency virus (HIV)
    because UV has been shown to activate HIV  in vitro (reviewed by
    Zmudzka & Beer, 1990). It has been suggested that UV exposure could
    progress these patients to full blown AIDS by interfering with
    protective immunity, since Th1 responses appear to be protective
    against AIDS whereas Th2 responses are not (Shearer & Clerici,
    1992). However, Warfel  et al. (1993) found no difference in CD4+
    cell counts in HIV patients before and after treatment with 50-60%
    of their MED for dermatological disease. More research in this area
    is needed to resolve these issues.

         Finally, it has been known for some time that UV exposure
    adversely affects the clinical course of systemic lupus
    erythematosus, an autoimmune disease (Epstein  et al., 1965);
    however, the relationship of UV in this immune response is unclear.

    9.3  Conclusions

         The above studies suggest that UV exposures at environmental
    levels suppress immune responses in both rodents and man. In rodents
    this immune suppression results in enhanced susceptibility to
    certain infections with skin involvement as well as some systemic
    infections. The mechanisms associated with UV-induced
    immunosuppression in rodents and man are similar. Also, host defense
    mechanisms which provide protection against infectious agents are
    similar in rodents and man as shown in figure 9.1. It is therefore
    reasonable to assume that exposure to UV may enhance the risk of
    infection and decrease the effectiveness of vaccines in humans.

         Additional research is needed to substantiate these
    assumptions. In particular, experimental studies are needed in
    rodents and man to access effects on immune function parameters in a
    fashion that allows quantitative comparisons between the two
    species. Results of experiments on rodents showing the effects of UV
    exposure on the susceptibility to various infections could be used
    to extrapolate the risk of infectious disease in humans. Ultimately
    epidemiological studies on the effects of UV exposure on
    susceptibility to infection and vaccine effectiveness are needed to
    validate this hypothesis.

    FIGURE 9.1

    10.  HUMAN STUDIES: THE EYE

    10.1  Introduction

         Since the early part of this century ophthalmologists have
    suggested an association between sunlight exposure and UV and the
    development of cataracts and other ocular effects (Widmark 1889,
    1901; Hess 1907; Martin 1912; Birch-Hirschfeld, 1914; Verhoeff and
    Bell, 1916; Duke-Elder, 1926b); however, only in the past twenty
    years have epidemiological studies provided a scientific link. UV is
    probably one of a number of factors associated with the development
    of cataract. Despite the number of animal and human studies, many
    questions remain as to the validity of interpretations of past data
    and the biological and physical factors that influence the outcome
    of UV exposure of the eye.

    10.2  The Eye

         The eyeball is deeply set in a bony orbital cavity, and the
    upper bony ridge provides not only protection from mechanical injury
    but also serves as a shield from overhead sky light. The upper and
    lower eyelids (figure 10.1) also provide considerable protection by
    serving as " shutters" against bright light. The eyeball consists
    of three layers of tissue: (a) an outer protective layer, the sclera
    and cornea; (b) a middle layer of blood vessels, pigment cells and
    muscle fibres called the uvea; and (c) an inner, light sensitive
    layer called the retina.

         The sclera, the outer posterior layer, is a tough, thick,
    opaque tissue formed of collagen fibres. The cornea, the anterior
    transparent part of the eyeball, consists of multiple layers. The
    surface epithelium continues with the surface epithelium of the
    conjunctiva. The epithelial cells are known to be constantly
    changing and the basal layer overlies the Bowman layer. The
    outermost epithelial cells undergo rapid turnover, having a lifetime
    at the surface of approximately 48 hours.

         The main bulk of the cornea, the stroma, consists of highly
    organized collagen fibres in a pattern that makes the cornea
    transparent. The innermost layer, the endothelium, is a single layer
    of an active ion pump (Na-K) that maintains the hydration state of
    the stroma, an important factor for corneal transparency.

         The middle layer of the eyeball consists of the iris
    anteriorly, the ciliary body, and the choroid. The iris is a
    diaphragm that adapts by changing the pupillary size according to
    ambient light level. The iris is formed of mainly pigmented cells of
    various densities, blood vessels and smooth muscle fibres that are
    attached to the anterior part of the ciliary body. The smooth muscle
    fibres form the sphincter and dilator of the pupil.

    FIGURE 10.1

         As part of the middle layer lining the sclera, the choroid
    consists of a meshwork of blood vessels, nerves and pigment cells
    that contribute to the nutrition of the retina and support the
    function of the innermost layer, the retina, that contains the
    photoreceptors and the neuronal network.

         The lens, embryologically of ectodermal (skin) in origin,
    consists of closely and orderly packed transparent elongated lens
    cells that are enclosed in a capsule. New lens cells are constantly
    formed at the so-called lens equator and old ones are displaced
    towards the centre of the lens. The lens is suspended to the ciliary
    body by fine ligaments. The lens shape can be changed by contraction
    or relaxation of the ciliary muscles. This change provides a
    focusing power to the eye that is called accommodation and is made
    possible by the elasticity of the young lens. The lens is avascular
    and obtains its nutrient from the aqueous fluid in front of it and
    the vitreous body that fills the posterior cavity behind it.
    Notably, the lens proteins cannot be renewed and therefore
    accumulate lesions inflicted on them throughout life.

         The photoreceptors, the rods and cones of the retina,
    constitute the primary light receptor with rods functioning at low
    light levels (scotopic vision) and cones operating at high light
    levels (photopic vision). Thus the rod system subserves the function
    of retinal sensitivity, whereas cones provide colour vision, high
    resolution visual acuity and motion perception. The retinal pigment
    epithelium is essential for the maintenance of photoreceptor
    metabolism including, the transport, storage and regeneration of the
    visual pigments.

         Visible light (400-760 nm) incident upon the eye is strongly
    refracted at the cornea, then transmitted through the aqueous humour
    in the anterior chamber to the lens where it is refracted further.
    After transmission through the vitreous gel-like structure, it
    finally reaches the light-sensitive receptors in the retina. It is
    these structures that are primarily damaged by UV and visible
    radiations. The neuronal layers perform the complex task of
    information processing. The primary visual signal is transformed and
    ultimately transmitted to the visual cortex in the brain, thus
    providing the image seen by the eye.

         Most of the UV incident on the eye is absorbed in the tear
    film, the cornea and the lens. The lens and the tissues in the
    anterior part of the eye may however, be exposed to UV at
    wavelengths above 295 nm and the retina is exposed to a fraction of
    the incident UVA. Absorption of UV in the ocular media is given in
    Fig 10.2 (Sliney & Wolbarsht, 1980). Boettner & Wolter (1962)
    measured the transmission of direct forward scattering UV in the
    cornea, aqueous humour, lens and vitreous humour from freshly
    enucleated normal human eyes. The cornea absorbed all UV with
    wavelengths <300 nm, while above 300 nm some UV was transmitted

    through the cornea. About 60% of UV at 320 nm and 80% at 380 nm was
    transmitted through the cornea. The aqueous humour transmitted most
    incident UV (90% transmission at 400 nm) with no evidence of
    scattering.

         Recently Barker and Brainard (1993) quantified the change in UV
    transmittance of the human lens with age. All of these studies
    clearly show the steady decrease in UV transmittance of the lens
    age. At birth there is a small window of transmission to the retina
    at 320 nm. This window almost disappears by the second decade due to
    an age-related yellowing of the human lens. As shown in Fig 10.3 the
    lens absorption is strongest in the 340-380 nm band with somewhat
    less absorption in the 310-320 nm range (Rosen, 1986; Barker and
    Brainard 1993). The human lens is unique in that it contains a UVA
    absorbing filter (O-beta of 3-hydroxykynurenine) which protects the
    retina.

    10.3  Study Design

         This chapter reviews the epidemiological evidence for a causal
    association between exposure to UV and development of specific eye
    diseases which have at some point been linked with exposure to UV.
    Two distinctive types of UV exposure assessments have been used in
    the epidemiological studies. Some studies have related the
    occurrence of eye disease to non-personal factors associated with
    place of residence, such as meteorological data on average annual UV
    dose or average annual hours of sunlight. Other studies have
    obtained estimates of exposure at the individual level (e.g. hours
    of sunshine exposure, lifetime exposure to UV) and related these to
    disease occurrence.

         Three types of epidemiological studies have been used to
    investigate an association with UV exposure: geographical
    correlation studies, cross-sectional studies, and case-control
    studies. In the geographical correlation studies the prevalence of
    eye disease in different areas has been related to non-personal
    factors associated with place of residence. These studies are useful
    for generating hypotheses but of limited value in testing a
    particular hypothesis because observed correlations may result from
    confounding by other factors which also vary geographically, and
    because the level of exposure for persons with the disease is not
    known.

    FIGURE 10.2

    FIGURE 10.3

         The second, and most common type of study design, has been the
    cross-sectional study in which a population or occupational group is
    surveyed and disease prevalence measured. Cross-sectional surveys
    identify all persons with the disease, some of which are new cases
    while other persons may have had the disease for a period of time.
    Some cross-sectional studies have related non-personal factors
    associated with place of residence to disease status. Other studies
    have collected detailed information from each study participant on
    personal exposure to UV or indices of exposure (e.g. hours of
    exposure to sunshine, occupation) and related these exposures to
    disease status.

         The third type of study design has been the case-control study,
    where differences in UV exposure between persons with the disease
    and those without have been compared. Cases have usually been drawn
    from a hospital or clinic. Controls have been drawn from other
    hospital or clinic patients or from the general population and have
    usually only included persons with good visual acuity, without the
    disease of interest, and without a disease that is associated with
    UV exposure. Some case-control studies have related disease status
    to non-personal factors associated with place of residence while
    others have used personal exposure information.

         Data from both cross-sectional and case-control studies can be
    useful in confirming a hypothesis, but have a number of limitations.
    If there is an excess risk of death associated with the disease, as
    has been suggested for cataract (Minassian  et al., 1992; Vitale
     et al., 1992), both types of study will be biased towards the
    survivors. In addition, disease status and prior exposure indices
    are measured at the same time and it may not be possible to
    differentiate between cause and effect, especially if the disease
    has a long latency period.

    10.4  Diseases of the External Eye

    10.4.1  Photokeratitis and photoconjunctivitis

         Cases of photokeratitis and photoconjunctivitis have occurred
    between 0.5 and 24 hours after prolonged exposure to intense solar
    radiation, often in highly reflective environments (Wittenberg,
    1986). The most severe cases are usually manifested as snow
    blindness, suggesting that UV is the cause of this condition.

         The action spectrum for UV photokeratitis produced in the
    rabbit was first measured by Cogan and Kinsey (1946). Pitts (1974,
    1978) in a series of laboratory studies on humans estimated the mean
    threshold of UVB (290-315 nm) for photokeratitis at 3500 J m-2.
    These laboratory data are supported by Blumthaler  et al. (1987),
    who estimated that the radiant exposures in clinically observed
    cases of photokeratitis ranged from 1200 to 5600 J m-2. It is
    estimated that 100 to 200 seconds of direct, unattenuated exposure

    to 295-315 nm solar radiation will result in photokeratitis (Sliney,
    1987; Wittenberg, 1986). Blotting out the solar disc would remove
    around 40% of the UV, still leaving a threshold of around 5.5
    minutes. Sliney (1986) has estimated that the reflected levels of UV
    from light sand should be sufficient to cause a threshold
    photokeratitis within exposure periods of 6-8 hours centred around
    midday, and within 1 hour for UV reflected from snow.

         Experimental data shows photokeratitis can be induced in
    animals by UVB exposure and that use of UVB absorbing contact lens
    or chromophores can prevent UVB induced photokeratitis in laboratory
    animals. Collectively, there is sufficient experimental and
    epidemiological evidence that exposure to intense UVB radiation
    causes photokeratitis and photoconjunctivitis.

    10.4.2  Climatic droplet keratopathy

         Climatic droplet keratopathy, among a variety of other names
    (Gray  et al., 1992), is also known as spheroidal degeneration from
    its histological appearance. It is a degenerative condition usually
    affecting both eyes symmetrically, and restricted to the exposed
    interpalpebral band of the cornea. This condition is of major
    significance for vision in some parts of the world, reducing vision
    to blindness levels in older people. For example, in Mongolia it has
    been found in an initial survey to be the third cause of blindness
    (Baasanhu  et al., in press).

         Climatic droplet keratopathy occurs throughout the world, but
    is more common in areas with snowfall persisting late into the
    summer in the northern hemisphere, such as parts of northern Canada,
    Siberia and Mongolia, and in areas of sand and desert in other
    latitudes, including Somalia, the Arabian peninsula, Iran, and
    Australia. It is also particularly common on sea coasts where there
    is coral sand or the sand is impregnated with salt, such as the
    islands of the Red Sea (Gray el al., 1992).

         In a cross-sectional study of Australian aborigines Taylor
    (1980a) found no correlation between the prevalence of climatic
    droplet keratopathy and ambient UVB levels, although the condition
    was more common among those working as stockmen. Johnson (1981)
    reported a geographical correlation with the calculated flux of
    reflected UV from snow and ice throughout the year in the eastern
    coast of Newfoundland and Labrador and the eastern Arctic of Canada.

         In a cross-sectional study of Chesapeake Bay waterman study
    Taylor  et al. (1989) examined the risk of climatic droplet
    keratopathy with chronic UVB exposure. Although a positive
    association was found (RR= 6.4, 95%CI=2.5-11.7) for those in the
    highest quarter of exposure compared to those with the bottom
    quarter, further analyses of this data (Taylor  et al., 1992)

    showed the risk of climatic droplet keratopathy was also related to
    UVA.

         There is strong evidence that the corneal degeneration is due
    to environmental factors. Circumstantial evidence exists that it is
    caused by solar UV, mainly reflected from ground surfaces such as
    snow and sand which are particularly reflective of UV.
    Histologically, the material deposited in the superficial corneal
    stroma as spheroidal droplets is most likely to be derived from a
    mixture of altered plasma proteins, including fibrinogen, albumin,
    and immunoglobulins (Johnson & Overall, 1978).

         Other proposed aetiological agents such as low atmospheric
    humidity, low temperature or high temperature have been excluded. It
    is possible that particulate injury by wind-blown ice or snow or
    sand particles may contribute to the development of the condition by
    causing inflammation and therefore outpouring of additional plasma
    proteins from the blood vessels of the limbs.

    10.4.3  Pinguecula

         Pinguecula is a fibro-fatty degeneration of the interpalpebral
    conjunctiva. The pathological changes that occur in pinguecula are
    similar to actinic elastosis of the skin, a condition thought to be
    linked to sunlight exposure. This indirect evidence suggests that
    exposure to sunlight may be a risk factor for pinguecula.

         Geographical variation in the occurrence of pinguecula has been
    reported, with higher prevalence in Arabs living near the Red Sea
    than in Eskimos from Greenland or Caucasians in Copenhagen (Norn,
    1982). Johnson  et al. (1981) in a study of pinguecula in Labrador
    found the size of pinguecula was correlated with the severity of
    climatic droplet keratopathy. Taylor  et al. (1989) in the study of
    Chesapeake Bay watermen found a weak association for the presence or
    absence of pinguecula with exposure to UVA and UVB. The relative
    risk for the top quartile of exposure was 1.4 (95%CI=0.9-2.2), less
    than for climatic droplet keratopathy or pterygium. Karai &
    Horiguchi (1984) in a study of 191 Japanese welders found no
    difference in the occurrence of pinguecula between welders and
    controls.

         It is concluded that there is currently insufficient
    epidemiological or experimental data for an assessment of the risk
    of pinguecula with exposure to UV.

    10.4.4  Pterygium

         Pterygium is a triangular shaped degeneration and hyperplastic
    process in which the bulbar conjunctiva encroaches on the cornea.

         A geographical association between variation in the occurrence
    of pterygium and variation in sunlight exposure was first suggested
    by Talbot (1948). Based on observation of pterygium in New Zealand
    and South Pacific Islands Elliott (1961) suggested pterygium in
    these locations resulted from UV exposure.

         Studies of non-personal factors associated with place of
         residence

         In a study of pterygium patients in US Veterans Administration
    hospitals during 1957-59 Darrell & Bachrach (1963) related mean
    daily UV (319nm) levels to the ratio of pterygium to all hospital
    discharges. A trend was found between UV level and pterygium ratio
    for persons born in rural counties. A similar association with UV
    level was seen for persons residing in a rural county at the time of
    the hospital discharge. Cameron (1965) examined the global pattern
    of pterygium and reported an inverse gradient with latitude. In a
    study of Australian aborigines Taylor (1980a) found pterygium was
    correlated with ambient UVB level and hours of sunshine at place of
    residence. Data from Canada indicates that pterygium is also common
    in arctic and sub-arctic environments (Johnson  et al., 1981).
    Moran & Hollows (1984) found a nonsignificant increase in the
    prevalence of pterygium among Australian aborigines residing in
    areas with higher ambient UV levels.

         Studies with personal exposure measurements

         Four studies have related occurrence of pterygium to personal
    measurements. Karai & Horiguchi (1984) examined 191 Japanese welders
    for the presence of pterygium. A trend of increasing risk of
    pterygium was found with years of employment as a welder, an
    indirect measure of cumulative occupational exposure to UV.

         Booth (1985) undertook a hospital-base case-control study of
    pterygium in Sydney. No difference between cases and controls was
    found in subjective assessment of exposure to sunlight in work or
    sport. However, a family history of pterygium was found to be a
    strong risk factor.

         Among Chesapeake Bay watermen, Taylor  et al. (1989) found a
    dose-response relationship between risk of pterygium and exposure to
    UVA and UVB. The relative risk for the top quarter of UVB exposure
    was 3.1 (95%CI=1.8-5.3) compared to the lowest quarter. However, it
    is noted that pterygium was equally associated with ocular exposure
    to UVA and visible light.

         Mackenzie  et al. (1992) undertook a hospital-based
    case-control study of pterygium in Queensland. A strong
    dose-response relationship was found with closeness of place of
    residence to the equator, type of outdoor work environment (e.g.
    sandy) and amount of time spend outdoors. The most striking finding

    was the magnitude of risk associated with spending most of the time
    outdoors was stronger when related to childhood exposure (RR=17.2,
    95%CI=6.2-47.6) than to adult exposure (RR=5.7, 95%CI=3.1-10.6). The
    risk associated with working at ages 20-29 in an outdoors
    environment of mainly sand or concrete was associated with a
    relative risk of 11.3 compared with indoor workers. Corneo (1993)
    has suggested that the cornea is acting as a side on lens focusing
    light and also UV across the anterior chamber to the nasal limbus.
    This hypothesis may explain why pterygia usually commence on the
    nasal side of the eye.

         Evaluation of epidemiological evidence

         While several geographical studies have reported an inverse
    trend with latitude, the common occurrence of pterygium in arctic
    and subarctic locations suggest that closeness to the equator does
    not fully explain the distribution of this disease.

         The strength of the association with time spent outdoors
    reported by Mackenzie  et al. (1992) suggests that the association
    may be causal. However, there is insufficient evidence to show that
    the observed association with UV exposure is not, in part, due to
    confounding. The findings from three of the studies lend support to
    a hypothesis that irritation by particulate matter is associated
    with pterygium. The Australian aborigines live in a dry, dusty
    environment and welders are occupationally exposed to a range of
    particles. Similarly, the Queensland study found highest risk among
    those who worked in sandy locations. The particulate matter
    hypothesis is also supported by a report (Dhir  et al., 1967) of
    higher prevalence of pterygium among Punjabi Indians working in
    sawmills (an indoor occupation) in New Delhi and British Columbia
    than Punjabi farmers (an outdoor occupation). Similar findings of
    increased risk of pterygium among sawmill workers in Thailand and
    Taiwan have been reported (Detels & Dhir, 1967). The evidence of
    possible confounding by particulate matter is inconsistent, with the
    Chesapeake Bay watermen study finding an association with sunlight
    exposure in a location that was neither hot, dry or dusty.

         It is not possible, based on available epidemiological data, to
    assess the risk of pterygium with exposure to UV because of possible
    confounding of observed associations by exposure to particulate
    matter or other factors.

    10.4.5  Hyperkeratosis, carcinoma-in-situ, and squamous cell
            carcinoma of the conjunctiva

         These conditions probably form a gradation of development and
    cannot necessarily be distinguished clinically. Invasive squamous
    cell carcinoma is often said to arise from a pre-cancerous lesion.
    Epithelial dysplasia and carcinoma-in-situ look the same, and are
    sometimes keratinized and present as leucoplakia in which case the

    term actinic keratosis may be applied (Naumann & Apple, 1986;
    Garner, 1989). The main argument for an actinic causation is that
    these tumours usually present in the exposed area of the eye between
    the lids (the interpalpebral fissure) and under conditions where
    they may be expected to be exposed to solar radiation.

         Xeroderma pigmentosum is a recessively inherited syndrome
    characterized by clinical and cellular hypersensitivity to solar
    radiation and a defect in the capacity to repair UV-induced damage
    in DNA (Fitzpatrick,  et al., 1963). Among reports of 337 patients
    with xeroderma pigmentosum for whom ocular findings had been
    described, Kraemer  et al. (1987) identified 88 ocular tumours of
    which 73 were specific to the corneal-scleral limbus (34), the
    cornea (24) or the conjunctiva (15). Of the non- melanomas for which
    histopathological type was specified, 28 were squamous cell
    carcinomas and 12 were basal cell carcinomas. Among 64 patients with
    ocular neoplasms whose age was stated, half the neoplasms had
    occurred before 11 years of age. While the eyelids are a site of
    preference for basal cell carcinoma, this tumour rarely, if ever,
    arises in the conjunctiva in otherwise normal individuals.

         Squamous cell carcinoma of the conjunctiva is an uncommon
    tumour. Garner (1989) reviewed all cases of tumours at the limbus
    sent over a 40 year period for examination to the Institute of
    Ophthalmology, London. The total was only 636 tumours, of which 73
    were squamous carcinomas. This amounts to less than 2 cases per year
    coming to the Pathology Laboratory, even though Moorfields Eye
    Hospital which the Laboratory serves, attracts patients from all
    over the country and from overseas.

         Lee & Hirst (1992) attempted to provide population-based
    figures and estimate the incidence of these tumours in metropolitan
    Brisbane (latitude 30° south). They surveyed the histological
    records of all ocular surface tumours examined in the pathological
    laboratories over the previous 10 years, serving a population of
    more than 745,000 in 1989. There were 139 cases of which 79 were
    corneal epithelial dysplasia, 28 carcinomas-in-situ and 32 were
    squamous cell carcinomas. There was a strong male preponderance. The
    incidence ranged from 1 per 100,000 in 1980 to 2.8 per 100,000 in
    1982. This is well below the rate for squamous cell carcinoma of the
    skin and melanoma of the skin in Queensland as a whole. On the other
    hand, it is a substantially higher rate than that recorded in London
    where the pathological laboratory referred to also covers a much
    larger population.

         Squamous cell carcinoma of the conjunctiva has been reported to
    form a greater proportion of eye tumours in Africans living in areas
    close to the equator (Templeton, 1967) than in the south of Africa
    (Higginson & Oettlé, 1959), and much higher than in Baltimore
    (39°N). The incidence (0.3 per 100,000) in Uganda (0°) has been

    reported to be twice that in Denmark (55°N) despite the potential
    underascertainment in Africa (Templeton, 1967).

         It is extremely rare for a neoplasm to arise  de novo in the
    corneal epithelium, where it may be called a corneal intra-
    epithelial neoplasm. Most such intra-epithelial sheets are connected
    at the corneo-scleral limbus to a conjunctival lesion, such as a
    papilloma or a leucoplakia over a pterygium or pinguecula (Waring
     et al., 1984). The only available evidence for an UV aetiology is
    the location of the lesion within the interpalpebral fissure, and
    the fact that it may arise from a lesion which is itself associated
    with UV. Three cases of corneal intra-epithelial neoplasia have been
    recently reported in people aged 31 to 38 who wore contact lenses
    and were considered to have had substantial exposure to artificial
    and solar UV (Guex-Crosier & Herbort, 1993).

    10.5  Diseases of the Lens

    10.5.1  Cataract

         For the purpose of this review, a cataract is defined as an
    opacity of the lens of the eye. The three major types of cataract
    are cortical, nuclear and posterior subcapsular (PSC). When a lens
    opacity interferes with vision, a clinically significant cataract is
    present. If left untreated, cataract will often progress to
    blindness. Cataract causes half of the world's blindness.

         Definitions of cataract and methods used to assess the presence
    and severity of cataract have not been uniform in epidemiological
    studies of cataract. Many studies include lens opacities that are
    not necessarily accompanied by a decrease in visual acuity. Some
    have combined all three major types of lens opacities into a single
    "cataract" category, while others have investigated associations for
    specific types of opacity. Methods of assessing the presence and
    sometimes the severity of opacities range from reviews of existing
    charts to clinical examinations using written definitions of
    cataract, and the use of standardized grading systems that have been
    found to be highly reliable.

         Occupational case series

         When cataracts result from occupational exposure to UV, it may
    be difficult to differentiate between the contribution of
    occupational and non-occupational factors to the development of the
    disease. Lerman (1980) described the onset of lens opacities in
    three persons who worked in a dental clinic and were exposed to UV
    (300-400 nm) from a dental curing unit. The lens damage varied from
    posterior subcapsular cataract in the dentist, who was reported to
    have received the highest dose, to zonular type opacities in one of
    the dental assistants. However, any retrospective reconstruction of

    the actual ocular exposure has a large degree of uncertainty, and
    the results from such an exercise must be interpreted with caution.

         Studies in which UV exposure was inferred from place of residence

         Selected studies of humans exposed to solar UV are presented in
    tables 10.1 and 10.2. Studies were selected for inclusion in the
    tables on the basis of the scientific quality of the published
    report and the overall contribution of the paper to the evaluation
    of the UV-cataract hypothesis. The tables do not include the
    relative risks for other factors, which in some instances are higher
    than for sunlight or UV exposure.

         Hiller  et al. (1977) investigated sunlight and cataracts
    using data from the large sources (blindness registries in 14 states
    and the cross-sectional Health and Nutrition Examination Survey
    (HANES) of 35 geographic areas of the US) and US Weather Bureau
    geographical data on annual hours of sunlight in each geographical
    area. Above the age of 65 the prevalence of cataract increased with
    annual hours of sunlight, with the highest prevalence found in
    locations with 3000+ annual hours of sunshine. At ages 45-64 there
    was some evidence of an association with hours of sunshine but the
    gradient was weaker. Below age 45 there was little association
    between annual hours of sunshine and prevalence of cataract. In a
    further analysis of HANES data Hiller  et al. (1983) reported a
    correlation between average daily UVB levels and prevalence of
    cataract. Analysis of HANES data by type of cataract (Hiller  et
     al., 1986) revealed UVB levels at location of residence were
    associated with pure cortical cataract but not with pure nuclear or
    posterior subcapsular cataract.

         The prevalence of cataract among Australian aborigines was
    found to be correlated with annual ambient UVB level at place of
    residence (Taylor  et al., 1980b). Hollows & Moran (1981) found the
    prevalence of cataract was highest among aborigines living in the
    north of Australia, an area with high average daily UVB radiation.
    Mao & Hu (1982) studied age related cataract in seven rural areas of
    China and found the prevalence of cataract was correlated with
    annual direct solar radiation.

         Residents of rural villages in Nepal had a prevalence of
    cataract related in different zones of the country to average hours
    of sunshine (Brilliant  et al., 1983). The prevalence was higher in
    the plains where there were 12 hours of direct sunshine compared to
    the mountains with 7-9 hours per day. Factors such as use of glasses
    and hats modify personal ocular exposure to UV and should be
    assessed.


    
    TABLE 10.1  SELECTED HUMAN CATARACT STUDIES WITH SUNLIGHT OR UV EXPOSURE BASED ON RESIDENCE
                                                                                                                                               
    AUTHOR           POPULATION             MEASURE OF            MEASURE OF               ASSOCIATIONS OBSERVED      COMMENTS
                                            OUTCOME               SUNLIGHT EXPOSURE
                                                                                                                                               

    Hiller et al.    MRA: 9110 persons      Blind from cataract,  Average hours of         RR= 3.3 (age 65-74)        Adjusted for age and
    (1977, 1983,     registered as blind    visual acuity (VA)    sunlight < 2400 vs                                  sex; blind registry data
    1986)            in 14 US States;       6/60                  3000+
                     whites;

                     NHANES: 3580 persons;  Lens opacity and VA   Average hours of         RR= 2.7 (age 65.74)        Adjusted for age and sex
                     probability sample of  < 6/7.5               sunlight < 2400 vs
                     US population;                               3000+

                     NHANES data            Lens opacity and VA   Average daily UVB        RR= 1.58 (age 45-74,       Adjusted analysis
                                            < 6/9                 count in area of         p<.05)
                                                                  residency 6000 vs 2600

                     NHANES data            Nuclear and cortical  Average daily UVB        RR= 3.6 for cortical       Adjusted analyses; pure
                                            opacity               count in area of         opacity; no association    opacity types only
                                                                  residency 6000 vs 2600   with nuclear opacity

    Taylor (1980b)   Survey of 350          Lens opacity with     Average daily sunlight   RR= 4.2 (95% CI=0.9-18.91  Unadjusted for age or
                     Australian Aborigines  good vision, poor     hours in area of                                    potential confounding
                                            vision or blindness   residence: 9.5+ vs < 8                              factors

                                                                  Annual mean UVB          RR= 1.8 (95%ci=0.9-3.4)1
                                                                  radiation level for
                                                                  area of residence:       1 95%CI estimated from
                                                                  3000 vs 2000             published data

    Hollows & Moran  Survey of 64,307       Lens opacity and VA   Average daily UVB        Significant positive       Wide age bands;
    (1981)           Aborigines and 41,254  < 6/6                 count in 5 zones of      correlation between        unadjusted analyses
                     non-Aborigines,                              Australia: 3000 vs       prevalence of lens
                     Australia                                    1000                     opacity and UVB counts
                                                                                           in Aborigines; no
                                                                                           association in
                                                                                           non-Aborigines

    TABLE 10.1 (contd).
                                                                                                                                               
    AUTHOR           POPULATION             MEASURE OF            MEASURE OF               ASSOCIATIONS OBSERVED      COMMENTS
                                            OUTCOME               SUNLIGHT EXPOSURE
                                                                                                                                               

    Brilliant et     Survey of 27,785       Lens opacities or     Average daily sunlight   RR= 3.8 (p<.001)           Adjusted for age and
    al. (1983)       Nepalese; national     aphakia               hours: 12 vs 712 vs                                 sex; RR decreased with
                     probability sample;                          7-9                                                 increasing altitude; sun
                     lifelong residents;                                                                              blocked by mountains at
                                                                                                                      high elevations

                                                                  12 vs 712 vs 7-9         RR= 2.6 (p<.005)

    Cruickshanks et  Cross-sectional        Nuclear, cortical     Average annual ambient   UVB exposure associated    Adjusted for other risk
    al. (1993)       survey of 4926         and PSC opacities     UVB exposure             with cortical opacities    factors; measure of
                     persons, Wisconsin,                                                   in men (RR=1.36,           exposure represents
                     USA                                                                   95%CI=1.02-1.79) but not   average potential
                                                                                           women, not associated      exposure at residency;
                                                                                           with nuclear or PSC        also Table 9.2
                                                                                           opacities
                                                                                                                                               

    TABLE 10.2  SELECTED HUMAN CATARACT STUDIES WITH INDIVIDUAL UV EXPOSURE ASSESSMENT
                                                                                                                                               
    AUTHOR        POPULATION             MEASURE OF         MEASURE OF SUNLIGHT          ASSOCIATIONS OBSERVED               COMMENTS
                                         OUTCOME            EXPOSURE
                                                                                                                                               

    Collmann et   Clinic-based case      Nuclear, cortical  Average annual sunlight      No significant association with     Low power to
    al. (1988)    control study of 113   and PSC opacities  exposure, based on           any type of opacity                 detect
                  cases and 168                             residential history and                                          association;
                  controls, North                           amount of time spent in sun                                      matched on age
                  Carolina, USA; whites                                                                                      and sex

    Taylor et     Cross-sectional        Nuclear and        Cumulative ocular exposure   Dose-response relationship in       High exposure
    al. (1988)    survey of 838          cortical           to UV since age 16, based    which a doubling of cumulative UVR  study population;
                  watermen, Maryland,    opacities          on life history and ocular   exposure increased risk of          detailed ocular
                  USA                                       exposure model               cortical opacity by 1.60            exposure model
                                                                                         (95%CI=1.01-2.64). RR= 3.30
                                                                                         (95%CI=0.90-9.97) for highest vs
                                                                                         lowest quartile. No association
                                                                                         between ocular exposure to UVR and
                                                                                         nuclear opacity

    Bochow et     Clinic-based           PSC cataract       Cumulative ocular exposure   Increased levels of UVB exposure    Adjusted analyses;
    al. (1989)    case-control study of  (surgical          since age 16, based on life  associated with increased risk of   association
                  168 cases and 168      patients)          history and exposure model   PSC cataract                        present when
                  controls, Maryland,                                                                                        adjusted for
                  USA                                                                                                        cortical cataract

    Dolezal et    Clinic-based           Cataract           Lifetime sunlight exposure,  No association between lifetime     Only partially
    al. (1989)    case-control study of  (scheduled for     based on life history,       sunlight exposure and risk of       adjusted for
                  160 cases and 160      surgery)           amount of time in sun and    cataract; use of head covering      potential
                  controls, Iowa, USA                       use of glasses and hat       reduced risk of cataract in males   confounding
                                                                                         (RR=0.48, 95%CI=0.25-0.94)          factors; crude
                                                                                                                             index of exposure,
                                                                                                                             low power

    TABLE 10.2 (contd).
                                                                                                                                               
    AUTHOR        POPULATION             MEASURE OF         MEASURE OF SUNLIGHT          ASSOCIATIONS OBSERVED               COMMENTS
                                         OUTCOME            EXPOSURE
                                                                                                                                               

    Italian-Am.   Clinical-based         Nuclear,           Work location in the         Cortical and mixed opacities        Adjusted for other
    study (1991)  case-control study of  cortical, PSC and  sunlight; leisure time in    associated with work location in    risk factors;
                  1008 cases and 469     mixed opacities    the sunlight; use of         sunlight (RR=1.75,                  crude indices of
                  controls, Italy                           glasses and hat              95%CI=1.15-2.65), leisure time in   exposure; also
                                                                                         sunlight (RR=1.45,                  Table 9.1
                                                                                         95%CI=1.09-1.93). Cortical, PSC
                                                                                         and mixed opacities associated
                                                                                         with use of a hat in summer
                                                                                         (RR=1.80, 95%CI=1.17-2.47). No
                                                                                         association between sun exposure
                                                                                         indices and nuclear opacities

    Leske et al.  Clinical-based         Nuclear, cortical  Work in sunlight; leisure    Work in sunlight significantly      Analyses adjusted
    (1991)        case-control study of  and PSC mixed      time in sun; residence and   reduced risk of nuclear opacity     for other risk
                  945 cases and 435      opacities          travel to areas of high sun  (RR=0.61, 95%CI=0.37-0.99); no      factor
                  controls,                                 exposure, use of hat and     significant associations between
                  Massachusetts, USA                        sunglasses                   exposure and cortical or PSC
                                                                                         opacities

    Cruikshanks   Cross-sectional        Nuclear, cortical  Leisure and work time        No associations with cortical       Adjusted for other
    et al.        survey of 4926         and PSC opacities  outside; use of glasses and  opacities; reduced risk of nuclear  factors; crude
    (1993)        persons, Wisconsin,                       hat                          and PSC opacities amount men for    indices of
                  USA                                                                    outdoor leisure time in winter;     exposure; also
                                                                                         use of hats and sunglasses          Table 9.1
                                                                                         significantly increased risk of
                                                                                         PSC opacity in women
                                                                                                                                               


    
         Age and sex adjusted prevalence for all types of cataract in
    persons aged 40 years and older was found to be 60% greater in Tibet
    than in Beijing (14.6% versus 9.1%, p>0.001) (Hu  et al., 1989).
    The authors suggested a relationship with higher UV at the higher
    altitudes of Tibet, but confounding factors could not be excluded
    and prevalences were higher in women than men.

         Studies with personal exposure measurements

         A number of studies have collected information from each study
    participant and estimated personal exposure to either sunlight or
    UV. Factors such as use of glasses and hats should be assessed. The
    characteristics of selected studies are outlined in Table 10.2.

         In a cross-sectional study of cataract in the Punjab related
    prevalence of cataract to work environment Chatterjee  et al.
    (1982) found a suggestion of lower cataract incidence among men
    whose main work location was outdoors (RR=0.7, 95%CI=0.5-1.1).

         Collman  et al. (1988) examined lifetime exposure to sunlight
    in a clinic-based case-control study of cortical, nuclear or PSC.
    Lifetime exposure to sunlight was estimated from intensity of solar
    radiation in area of residence, years of residence and average
    amount of time spent outdoors during daylight hours. A
    non-significant risk (RR=1.1) of cataract was found for the highest
    category of lifetime exposure to sunlight.

         Personal exposure history and ambient UVB data were combined to
    estimate an individual's lifetime annual ocular exposure to UVB
    after age 15 in a cross-sectional study of Chesapeake Bay watermen
    (Taylor  et al., 1988). This included information on occupational
    and leisure exposures, type of work surfaces, seasons, and use of
    head wear and eyewear. A moderate association with a trend of
    increasing risk with exposure to UVB was seen for cortical cataract,
    with a RR of 3.3 (95%CI=0.9-10.0) for the top quarter of exposure
    relative to the bottom quarter. A nonsignificant association was
    also found between exposure to UVA (320 -340 nm) and prevalence of
    cortical cataract. Little evidence was found for an association
    between UVA or UVB exposure and nuclear cataract. It is noted that
    UVA and UVB exposures were highly correlated and that the study
    would not have been able to differentiate between the effects of UVA
    and UVB. Further analyses suggested a significant difference between
    the cumulative lifetime ocular exposure among cases of cortical
    cataract compared to non- cataract controls. No threshold or latency
    period was observed.

         In a clinic-based study of PSC cataracts in Maryland, cases
    were persons who underwent PSC extraction in an ophthalmic practice
    (Bochow  et al., 1989). Controls, matched on age, sex, and type of
    referral were chosen from other patients on the appointment book of

    the same ophthalmic practice who did not have a PSC cataract or a
    previous cataract extraction. Annual and cumulative ocular UVB
    exposures were estimated for each individual using the same method
    as the studies of Chesapeake Bay watermen. Thirty-nine percent of
    cases had a pure PSC cataract, the remaining 61% had mixed PSC and
    other cataracts. Almost half the controls had a non PSC cataract
    (nuclear, cortical or other lens opacity). UVB exposure was
    significantly associated with PSC cataracts. Both the average
    cumulative exposure and average annual exposure were higher in cases
    than controls, after adjusting for steroid use, eye colour,
    education, diabetes and presence of cortical cataracts.

         In a hospital-based study of cataract patients in Iowa Dolezal
     et al. (1989) found little evidence of an association between
    individual lifetime sunlight exposure and cataract. Mohan  et al.
    (1989) in a similar study of cataract in New Delhi examined a range
    of environmental factors, including occupation. An increase in cloud
    cover was significantly associated (RR=0.8, 95%CI=0.7-0.9) with
    cataract when adjusted for each of the other environmental
    variables. The study did not quantify individual lifetime exposure
    to sunlight or UV.

         In a study of cataract patients in a Massachussetts hospital,
    Leske  et al. (1991) investigated occupational exposure to
    sunshine. No association was found for PSC cataract (RR=1.3,
    95%CI=0.7-2.3), cortical cataract (RR=0.9, 95%CI=0.6-1.3), or mixed
    cataract (RR=0.8, 95%CI=0.6-1.1) among those with at least 2 hours
    of exposure to bright sunshine per day for at least 2 months. The
    risk of nuclear cataract was reduced (RR=0.5, 95%CI=0.3-0.9).

         A hospital-based study from Italy (Italian-American Cataract
    Study Group, 1991) found an excess of pure cortical and mixed
    cataract (RR=1.8, 95%CI=1.2-2.6) and a nonsignificant deficit of
    nuclear (RR=0.6) and PSC cataract (RR=0.8) among those with a work
    location in the sunlight. Leisure time spent in the sunlight was
    associated with an excess of cortical and mixed cataract (RR=1.4,
    95%CI=1.1-1.9) and a nonsignificant deficit of posterior subcapsular
    cataract (RR=0.6).

         In an Indian clinic-based study of cataract and history of
    severe diarrhoeal diseases Bhatnagar  et al. (1991) found an
    elevated risk of cataract (RR=2.1, 95%CI=1.2-3.6) for outdoor
    occupations compare with indoor occupations. However Zaunuddin &
    Saski (1991) found no relationship between hours of exposure to
    sunshine and prevalence of nuclear or cortical cataract in Sumatra
    (0° latitude). In Beaver Dam, Wisconsin, Cruickshanks  et al.
    (1993) found no association between average annual exposure to UVB
    and cortical, PSC or nuclear cataract. Wong  et al. (1993) surveyed
    fishermen in Hong Kong. A sun exposure score was calculated based on
    daily sunlight exposure, and protection from use of a canopy, hat,
    and glasses. The highest grades of cataract of all types considered

    together were more common in subjects with the highest sun exposure
    scores, but none of these associations was significant at the 5%
    level. A population-based case-control study (Shibati  et al.,
    1993) reported an increased risk of cortical cataract among men aged
    40-50 years who spent 5 or more hours per day outdoors compared with
    those who spent less time outdoors (RR = 6.89; 95%CI = 1.22-39).

    Evaluation of epidemiological evidence

         An association has been demonstrated between prevalence of
    cataract and residence in areas at low latitudes, with long hours of
    sunlight or high ambient UV radiance in several studies undertaken
    in different parts of the world. However, in each study, the
    observed association may be confounded by other possibly causal
    factors. Certain of the earlier studies did not classify the lens
    opacities into types of cataract.

         Cortical cataract was examined separately in four studies. Only
    one study assessed individual exposure. Taylor  et al. (1988) found
    a dose-response relationship with exposure to UVB radiation. The
    relative risk for the highest exposure category was three times that
    for the lowest exposure category. It is unlikely that the exposure
    assessment was able to distinguish between UVA and UVB exposure. The
    other two used simple measures of sun-related behaviour. Leske  et
     al. (1991) found no association between exposure to bright
    sunshine and cortical cataract, while in the Italian-American
    Cataract Study (1991), a work location in the sunlight was related
    to cortical and mixed cataract. The Italian study also found an
    association between leisure time outdoors and cortical and mixed
    cataract. The other two studies showed non-significant trends in
    opposite directions. More recently, Cruickshanks  et al. (1993)
    found annual UVB exposure was associated with cortical opacities
    among men, but no association was found for women.

         Four studies report risk estimates for posterior subcapsular
    (PSC) cataracts. Bochow  et al. (1989) measured individual exposure
    and found PSC cataract patients had higher annual and cumulative
    exposures to UVB than controls, even after allowing for the effects
    of several other factors. The other two studies used simple measures
    of sun-related behaviour and showed non-significant trends in
    opposite directions. Leske  et al. (1991) found elevated risk for
    pure PSC cataract patients compared to controls. However, the
    Italian-American Cataract Study (1991) reported reduced risk for
    pure PSC cataract patients with a work location in the sunlight or
    who spent leisure time in the sunlight. Cruickshanks  et al. (1993)
    found no association between annual UVB exposure and risk of PSC
    cataract.

         Five studies provide risk estimates separately for nuclear
    cataracts. These studies are consistent in showing no association
    between UV exposure and nuclear cataract (Taylor  et al., 1988;

    Dolezal  et al., 1989; Leske  et al., 1991; The Italian-American
    Cataract Study, 1991; and Cruickshanks  et al., 1993). Collectively
    these studies are consistent in showing no association between UV
    exposure and nuclear cataract.

         All of the published epidemiological studies of UV and cataract
    have been challenged by the enormous difficulty of determining
    ocular exposure in different climates. As noted previously, the
    cornea and lens are seldom directly exposed to light rays from much
    of the sky; hence the sunlight scattered from the ground and the
    horizon determine the actual accumulated UV dose.

         These studies clearly demonstrate that UV is at least one
    aetiologic factor in cataractogenesis. However, extrapolation of
    strong associations found in a mid-latitude population where no
    serious nutritional problems are present (e.g. Taylor,  et al.,
    1988) to a tropical population in less developed regions, may not be
    valid, since the contribution of UV relative to other factors such
    as malnutrition and dehydration may be far more important.

    10.5.2  Exfoliation syndrome

         The exfoliation syndrome (pseudoexfoliation of lens capsule)
    consists of abnormal material deposited on or arising from various
    parts of the anterior eye. This condition was originally described
    from Finland by Lindberg (1917). In the Nordic countries it
    contributes to a high proportion of glaucoma in the older
    population. This appears as a round area in the centre of the
    anterior lens capsule, corresponding to normal pupil size, on which
    bluish-grey flakes are deposited. This is surrounded by a clear
    zone, which in turn is surrounded by a peripheral band of
    involvement as well. On the border of the pupil it looks like
    "dandruff". Similar material is trapped in the pores of the
    trabecula meshwork and may be seen on the ciliary processes, on the
    zonulas, surrounding the conjunctival vessels and in retro-orbital
    tissues. It is a basement membrane material, akin to amyloid in some
    respects, although many histochemical studies do not support this
    identification.

         The prevalence varies enormously from country to country, and
    even within countries. The highest prevalence was found in the
    Navajo Indians of New Mexico, in which 38% were over 60 years of
    age. At the other extreme, only 2 cases have ever been recorded in
    Eskimos, and these were two Greenlanders, possibly of mixed
    ancestry, aged over 70 years (Ostenfeld-Åkerblom, 1988). The average
    prevalence in central Europe is around 2% on the basis of figures
    from several authors (Forsius, 1988).

         The possibility of environmental factors was proposed by Taylor
    (1979) based on observations of exfoliation in Australian

    aborigines. The distribution of exfoliation was linked to annual
    global radiation and to climatic droplet keratopathy.

         Exfoliation syndrome sometimes occurs in other areas of high
    UV, and high prevalence of climatic keratopathy. Examples include
    Somalia, Djibouti and Saudi Arabia. There is, however, considerable
    evidence to suggest that UV is not the main factor associated with
    the development of exfoliation. The geographic distribution does not
    consistently correspond with that of climatic keratopathy. There may
    be wide differences in prevalence of exfoliation syndrome at similar
    latitudes. For example, it is frequent in parts of East Africa, but
    rare in West Africa. Similarly it may be seen at high prevalence in
    the Lapps of Finland and Sweden, but not in Eskimos at the same
    latitude. The prevalence may vary within the same country. A total
    of 4,042 patients aged over 50 were examined in clinics in 6 areas
    in different parts of France over a 2 week period. The prevalence
    was high in Brittany (20.6% in those over 60 years) and extremely
    rare in Picardy at a similar latitude (Colin  et al., 1985).
    Exfoliation is usually more frequent in females than males. It is
    found in parts of the eye, such as the ciliary body and in the
    orbit, remote from the influence of light. Forsius (1988) has
    reviewed the evidence for genetic aetiology for the condition.

         The present conclusion is that environment, at least in the
    form of UV, is not the primary cause. There is not a consistent
    direct relationship with solar radiation. There is at least a major
    racial or genetic predisposition, but it is possible that light or
    some other environmental factor activates or induces the development
    of the exfoliation syndrome in those who are genetically
    predisposed.

    10.5.3  Anterior lens capsule

         In 1989 a previously unrecorded condition was reported from
    Somalia (Johnson  et al., 1989). This consisted of alterations of
    the pupillary area of the anterior capsule of the lens. The first
    stage appeared to be an opalescence of the capsule, which then
    became a plateau-shaped elevation above the surrounding contour of
    the anterior lens. In its most developed form it was a bagging of
    the anterior lens capsule and contents through the pupil, appearing
    like a hernia. This condition was invariably associated with
    climatic droplet keratopathy, but not necessarily with cataract. In
    fact, there appeared to be an inverse relation with cataract.

         The absolute association with climatic keratopathy suggests
    that it also may be due to excessive UV exposure. Attempts to secure
    histology on extracted lenses with this condition were difficult
    because the lens capsule so frequently tore from the rest of the
    lens as it was extracted by the cryoprobe. The capsules examined
    showed thinning and splitting of the layers, and death of many of
    the nuclei of the epithelial cells. However, there were no controls

    from the same geographical and ethnic area of the same ages for
    comparison.

    10.6  Diseases of the Choroid and Retina

         Among adults, only extremely small amounts of UVA and UVB at
    wavelengths below 380 nm reach the retina, because of the very
    strong absorption by the cornea and lens. Less than 1% of radiation
    below 340 nm and 2% of radiation between 340 and 360 nm reaches the
    retina (Barker and Brainard, 1993). Even in early childhood the
    highest spectral transmittance reaches about 4% in the UVB and is
    generally of the order of 1%. However, because of the biological
    activity of the shorter wavelengths of UVB, the biological
    importance of the small amount of this radiation that does reach the
    retina cannot be completely neglected. As children age, UV is
    increasingly absorbed by the cornea and lens, and the proportion
    reaching the retina decreases. This suggests firstly, that exposure
    to UV during childhood may be of more importance than exposure to UV
    during adult life, and secondly, that exposure to longer wavelength
    radiation (e.g. visible light) may be of more importance in
    adulthood.

    10.6.1  Uveal melanoma

         Exposure to solar radiation is considered to be causally
    associated with the development of cutaneous malignant melanoma
    (IARC, 1993). There is a possibility that exposure to UV may also
    cause melanoma of the uveal tract. There is no separate ICD code for
    intra-ocular melanoma, so descriptive studies have generally been
    based on cancer of the eye (ICD-9 190), of which it has been
    estimated that 80% are intra-ocular melanomas (Osterlind, 1987). In
    the case-control studies cases of uveal-tract melanomas were
    confirmed histologically, but also included tumours of iris and
    ciliary body with those of the choroid.

         The incidence of cancer of the eye is higher among white than
    black or Asian populations residing at the same latitude. For
    example, in US whites the incidence rates are 0.7 per 100 000 person
    years in males and 0.6 in females compared with 0.2 in both sexes in
    blacks (Parkin  et al., 1992). Among people of European ancestry,
    risk of ocular melanoma was observed to be least in those of
    southern European ethnic origin; for example, in comparison with an
    RR of 1.0 in those of southern European origin, the RR in people of
    northern European origin was 6.5 (95% CI 1.9-22.4; Seddon  et al.,
    1990). Risk of ocular melanoma was observed to be increased in those
    with light eye colour, with RRs of 1.7 to 2.1 (Gallagher  et al.,
    1985; Tucker  et al., 1985c; Holly  et al., 1990), but not when
    ethnicity was taken into account (Seddon  et al., 1990). Kraemer
    (1987) found five cases of ocular melanoma among reports of 337
    patients with xeroderma pigmentosum for whom ocular findings had

    been described. The defect in this condition is failure to repair
    DNA after damage by UV.

         There is no evident latitude gradient in incidence of ocular
    melanoma in white populations of the northern hemisphere or
    Australia (IARC, 1992) and, within the USA, its risks in those born
    in southern parts of the country, where ambient solar radiation is
    highest, has variously been reported to be more (Tucker  et al.,
    1985c), less (Seddon  et al., 1990) or the same (Schwartz & Weiss,
    1988; Mack & Floderus, 1991) as that in those born elsewhere in the
    country. Similarly Gallagher  et al. (1985) in Canada found no
    association with latitude of residence.

         Two studies have examined place of birth and risk of uveal
    melanoma, but the findings are inconsistent. Tucker  et al. (1985)
    found an excess of cases were born south of latitude 40°N, whereas
    Seddon  et al. (1990) found a deficit.

         Indicators of personal sun exposure have been inconsistently
    associated with risk of cancer of the eye or ocular melanoma. A
    small rural excess in incidence of cancer of the eye has been
    reported (Doll, 1991). Two descriptive studies reported an
    association with farming (Saftlas  et al., 1987; Gallagher, 1988)
    but this was not found in several other such studies (Milham, 1983;
    Office of Population, Censuses and Surveys, 1986; Vågerö  et al.,
    1990) or two case-control studies of ocular melanoma (Gallagher  et
     al., 1985; Seddon  et al., 1990). Some high exposure activities
    such as gardening (RR 1.6, 95% CI 0.7-1.6) and taking sunny
    vacations (RR 1.5, 95% CI 1.0-2.3, for highest category) were
    significantly associated with increased risks of ocular melanoma in
    one case-control study (Tucker  et al., 1985c) but no similar
    associations with personal sun exposure at work in leisure time, or
    in vacation were found in three other studies (Gallagher  et al.,
    1985; Holly  et al., 1990; Seddon  et al., 1990). Indeed,
    Gallagher  et al., (1985) found an elevated risk for government
    workers, a predominantly indoor managerial group. The lack of use of
    protective eyewear (sunglasses, visors, headgear) was associated
    with an increased risk of ocular melanoma in one study (Tucker  et
     al., 1985c) with an RR for infrequent or rare use of 1.6 (95% CI
    1.2-2.2). Weak evidence of a similar effect was found by Seddon  et
     al. (1990).

         No statistically significant association has been observed
    between ocular melanoma and a personal history of skin cancer in
    several studies of cancer registry or other data (Osterlind  et al.,
    1985; Tucker  et al., 1985a; Holly  et al., 1991; Lischko  et al.,
    1989; Turner  et al., 1989).

         There is evidence of associations between exposure to sunlamps
    and some other artificial sources of UV and risk of ocular melanoma
    in the three case-control studies in which they have been examined.

    Tucker  et al. (1985c) found a relative risk of 2.1 (95% CI
    0.3-17.9) for frequent use of sunlamps compared with no use (p=0.10
    for trend over four categories of use); Holly  et al. (1990) found
    a relative risk of 3.7 (95% CI 1.6-8.7) for ever having an exposure
    to "artificial UV or black light" and with welding burn, sunburn to
    eyelids, or snow-blindness, RR 7.2 (95% CI 2.5-20.6); and Seddon  et
     al. (1990) found a relative risk of 3.4 (95% CI 1.1-10.3) for
    frequent or occasional use of sunlamps compared with never used. In
    one of these studies, there was also a strong association with
    employment as a welder (RR 10.9, 95% CI 2.1-56.5; Tucker  et al.,
    1985c). No similar association was found by Seddon  et al. (1990)
    but an increased risk in welders (RR 8.3, 95% CI 2.5-27.1) was found
    in an occupational study of French Canadians (Siemiatycki, 1991).

         Overall, the epidemiological studies do not provide convincing
    evidence of an association between exposure to solar UV and uveal
    melanoma. None of the studies has developed a practical assessment
    of individual cumulative ocular exposure to UVB. They have all used
    various simple estimates of sun-related behaviour.

         On the other hand, the use of a sunlamp, an artificial source
    of UV, was significantly associated with uveal melanoma in the two
    case-control studies that examined their use. Another study found
    elevated risk of uveal melanoma with exposure to UV or black lights.
    Collectively, these studies suggest frequent use of a sunlamp may be
    associated with a 2-4 fold increase in risk of developing uveal
    melanoma. Sunlamp use can produce over five-fold more DNA damage per
    unit of erythema than the sun (Nachtwey & Rundel, 1981).

         The large number of ocular melanomas in xeroderma pigmentosum
    patients also means that exposure to UV cannot be ruled out as a
    causative factor.

    10.6.2  Age-related macular degeneration

         Age-related macular degeneration (AMD) is one of the leading
    causes of blindness in the industrialized world. Visual loss can
    occur because of the development of geographic atrophy (loss of the
    outer retinal segments and retinal pigment epithelium), retinal
    pigment epithelial detachment or sub-retinal neovascularization
    (exudative AMD). Prior to visual loss AMD is characterized by the
    presence of drusen (lipofuscin and other material deposited between
    the retinal pigment epithelial cells and Bruch's membrane and
    appearing as yellow-white nodules with distinct and indistinct edges
    on retinal examination).

         There is evidence for association of AMD with UV exposure.
    Photochemical retinal damage can occur from prolonged exposure to
    high intensity light. Whether such damage is directly related to AMD
    is unknown. Although aged Rhesus monkeys have drusenoid deposits, no
    good experimental animal models for AMD currently exist.

         In a case control study, Hyman  et al. (1983) found no
    association of AMD and light exposure based on residential history.
    They also found no association of AMD to occupational light
    exposure.

         In studies based on individual exposure data; the results are
    equivocal. The initial evaluation of the association of AMD and UV
    exposure in the cross-sectional study of Chesapeake Bay watermen
    revealed no statistically significant association (West  et al.,
    1989). However, a reanalysis based on the small number of cases of
    AMD with exudative disease or geographic atrophy suggested an
    association with 20 year exposure to blue light but not UVA or UVB
    (Taylor, 1992).

         The Beaver Dam Eye Study found an association of late stage AMD
    (exudative AMD or geographic atrophy) and summer leisure time
    outdoors (RR=2.2 CI=1.1-4.2). It was also suggested that an
    association existed in men only between early stage AMD and summer
    leisure time outdoors. The magnitude of the risk estimates were
    unchanged after adjusting for numerous possible confounding factors
    (Cruickshanks et.al. 1993).

         It can be concluded that there are very limited data
    demonstrating an association of AMD with UV exposure. The finding of
    an association with blue light exposure is consistent with the
    wavelengths of visible light reaching the retina and needs further
    investigation.

    10.7  Conclusion

         The causal links between UVB exposure and various ocular
    conditions were evaluated on the basis of the following definitions:

     Sufficient evidence for a causal association indicates that
    positive associations have been observed between human exposure to
    UV and the effect in which chance, bias and confounding could be
    ruled out with reasonable confidence.

     Limited evidence for a causal association indicates that positive
    associations have been observed between exposure to UV and the
    effect for which a causal interpretation is considered to be
    credible, but chance, bias or confounding could not be ruled out
    with reasonable confidence.

     Inadequate evidence for a causal association indicates that the
    available studies are of insufficient quality, consistency, or
    statistical power to permit a conclusion regarding the presence or
    absence of a causal association between UV and the effect, or no
    data were available.

     Evidence for lack of causal association indicates that there are
    several adequate studies covering the range of exposure that humans
    are known to encounter which are consistent in not showing a
    positive association between UV and the effect.

         There is sufficient evidence to link photokeratitis to acute
    ocular exposure to UVB.

         Sufficient evidence exists to link the production of cortical
    and PSC cataracts to UVB exposure in animals. There is limited
    evidence to link cortical and PSC cataract in humans to chronic
    ocular exposure to UVB. Inadequate evidence is available to link PSC
    cataract in humans to chronic UVB exposure. Insufficient data have
    been collected upon which to evaluate the risk of cataract
    associated with childhood exposure to UVB. Half the world's
    35-million blind people are blind because of cataract. The
    proportion of cataract that results from UVB exposure is unknown,
    but may be as high as 20%.

         There is limited evidence to link sunlight exposure of the eye
    to the development of pterygium. It is unclear whether the observed
    association is specific for UV. The contribution from other
    environmental factors remains unclear.

         There is limited evidence to associate climatic droplet
    keratopathy with UV exposure and insufficient to link pinguecular
    and cancers of the anterior ocular structures. Insufficient evidence
    exists to link uveal melanoma to ocular exposure to solar UV
    radiation. However, several epidemiological studies have suggested
    that the use of sunbeds (an artificial source of UV) is associated
    with uveal melanoma.

         There is inadequate evidence of an association between ocular
    UV exposure and acute solar retinitis, age-related macular
    degeneration, acceleration of pigmentary retinopathies and
    exfoliation syndrome.

    11.  EFFECTS ON PLANT AND AQUATIC ECOSYSTEMS

    11.1  Introduction

         Human populations may be affected by direct and indirect
    consequences of increased solar UVB on aquatic food webs. Because
    more than 30% of the world's animal protein for human consumption
    comes from the sea (in many developing countries this percentage is
    even larger), a substantial decrease in biomass production would
    diminish fishery resources in the face of growing world populations.
    Reductions of leaf area, fresh and dry weight, lipid content and
    photosynthetic activity were typically found in UVB sensitive plant
    species. Additionally, alterations of leaf surface, epicuticular
    waxes, diffusion of water vapour through the stomata have been
    reported. For previous comprehensive publications see Caldwell  et
     al. (1989), Wellmann (1991) and SCOPE/UNEP (1993).

    11.2  Effects on Terrestrial Plants

    11.2.1  UV penetration into the leaf

         UVB has a direct effect on photosynthesis. Reductions in
    photosynthesis often accompany changes in leaf pigmentation,
    anatomy, and leaf thickness. After exposure to enhanced UVB, the
    internal light regime of leaves was altered (Bornman & Vogelmann,
    1991). In a recent study,  Brassica campestris (origin: northern
    latitudes) was subjected to 6.3 kJ m-2 day-1 of UVB and
    responded by increasing leaf thickness by 45% and UVB screening
    pigments by 21% relative to controls (Bornman & Vogelmann, 1991).
    Chlorophyll content (per leaf area) and photosynthetic activity
    decreased while scattered light within the leaves of UV-treated
    plants increased. Since the distribution of photosynthetically
    active radiation was altered at different depths within leaves after
    UV, these changes can also be expected to have an indirect effect on
    photosynthetic capacity.

         In a study on a group of 22 diverse plant species (including
    herbaceous and woody dicotyledons, grasses and conifers), widely
    varying UVB penetration was found. For instance, epidermal
    transmittance of the herbaceous dicotyledons ranged from 18% to 41%
    with penetration up to 140 µm, while conifer needles excluded a
    large percentage of the incident UVB. Penetration of UVB into leaves

    of the woody dicotyledons and grasses was in between that of the
    herbaceous dicotyledons and conifers (UNEP 1989).

    11.2.2  Changes in growth

         The growth of many plant species is reduced by enhanced levels
    of UVB. The main components of plants affected by UVB are shown in
    figure 11.1 (UNEP 1989). The ozone filter technique was used to
    simulate a relative solar UVB enhancement of 20% by providing 54.4
    kJ m-2 day-1 (unweighted) or 5.1 kJ m-2 day-1 of
    biologically effective radiation (UVBBE) through one cuvette and
    45.3 kJ m-2 day-1 (unweighted) or 3.6 kJ m-2 day-1 UVBBE
    through the other cuvette (Tevini et al., 1991b). These were average
    values measured from May 1990 to August 1990 and are equivalent to
    an ozone depletion of approximately 10%. Plant height, leaf area,
    and the dry weight of sunflower, corn, and rye seedlings were
    significantly reduced, while oat seedling remained almost unaffected
    (Tevini  et al., 1991b). The reduction of hypocotyl growth of
    sunflower seedlings under artificial UVB irradiation is associated
    with a UV dependent destruction of the growth regulator
    indole-3-acetic acid (IAA) and the formation of growth inhibiting
    IAA photoproducts. The inhibition of elongation in UV-irradiated
    sunflower seedlings might also be due to the action of peroxidases
    working as IAA-oxidase, causing a decrease in cell wall
    extensibility of the hypocotyl epidermis (Ros, 1990). Shading of
    shoot apex was shown to reduce UVB induced reduction in growth of
    Vigna seedlings (Kulandaivelu  et al., 1993).

    11.2.3  Effects on plant function

         When high UVB irradiances were used in combination with low
    levels of white light, such as commonly found in growth chambers,
    effects on photosynthesis were generally deleterious. However, even
    in the presence of higher levels of white light in green houses and
    in the field, reductions in photosynthesis of up to 17% were
    reported in the UVB sensitive soybean cultivar Essex when supplied
    with UVB equivalent to an 18% ozone depletion (Murali & Teramura,
    1987). Solar UVB also reduced net photosynthesis in sunflower
    seedlings by about 15% when a 12% ozone depletion was simulated by
    using the ozone filter technique (Tevini  et al., 1991c). One
    reason for the reduction in overall photosynthesis might be due to
    stomatal closure by enhanced UVB.

    FIGURE 11.1

         Recent studies reveal the effects of UVB radiation on tropical
    plants. Rice is among the most important tropical crops in the
    world. Sixteen rice ( Oryza sativa L.) cultivars from several
    different geographical regions when grown for 12 weeks in
    greenhouses with supplemental levels of UVB exposure equivalent to
    20% ozone depletion over the equator (15.7 kJ m-2 day-1 UVBBE)
    showed alterations in biomass, morphology, and photosynthesis.
    Approximately one-third of all cultivars tested showed a
    statistically significant decrease in total biomass with increased
    UVB exposure. Photosynthetic capacity declined for some cultivars,
    but only a weak relationship existed between changes in
    photosynthesis and biomass with increasing UVB exposure. In one of
    the rice cultivars tested, total biomass significantly increased by
    20% when grown under enhanced levels of UVB exposure. Therefore,
    despite the fact that the effects of UVB are generally damaging, in
    some cases, it has been reported to have a stimulating effect. Such
    positive growth effects are presently not easily explainable.
    Results from this experiment indicate that 1) a number of rice
    cultivars are sensitive to increases in UVB exposure; 2) the
    diversity exhibited by rice in response to increased levels of UVB
    suggests that selective breeding might be successfully used to
    develop UVB tolerant rice cultivars. Other preliminary screening
    studies on rice seedlings also corroborate these observations
    (Coronel  et al., 1990).

         In a three year field study (Sullivan & Teramura, 1991),
    photosynthetic capacity was generally reduced in loblolly pine trees
    exposed to supplemental levels of UVB simulating a 16% and 25% ozone
    depletion (11.5 and 13.6 kJ m-2 UVBBE). Absolute reductions
    varied from 0 to 40% between the seed sources and with needle age.
    For example, photosynthesis was significantly reduced by up to 40%
    in needles which had been exposed to UVB for an entire season, but
    only 19% on recently expanded needles. These reductions, however,
    were only transient in some plants because they could not be
    detected following the dormant winter period. This suggests that UVB
    repair mechanisms may exist. Measurements of chlorophyll
    fluorescence and the photosynthetic response to light indicated that
    the quantum yield was significantly reduced in some cases by direct
    effects on photosystem II. No significant effects were observed on
    stomatal conductance of transpiration, and chlorophyll
    concentrations were not generally altered by UVB exposure.

          In vitro studies, using isolated chloroplasts, indicate that
    UVB-induced damage to photochemical reactions is greater in C3
    plants ( Dolichos lab lab, Phaseolus mungo, and  Triticum vulgare)
    than in C4 ( Amaranthus gangeticus, Zea mays, and  Pennisetum
     typhoides). Such differences are associated with the polypeptide
    composition of the thylakoids (Kulandaivelu  et al., 1993).

         Studies in growth chambers with the ozone filter for
    attenuating solar UVB report significant reductions in net

    photosynthesis (measured under saturating light conditions) on a
    leaf area and whole plant basis in sunflower seedlings, when grown
    for three weeks at a daily maximum temperature of 28°C or 32°C under
    a 20% higher UVB level compared to controls (5.1 kJ m-2 day-1
    UVBBE vs. 3.6 kJ m-2 day-1). These represent average values
    from May 1990 to August 1990 and are equivalent to approximately a
    10% ozone depletion. In contrast, net photosynthesis was lower in
    maize seedlings only during the earliest stages of development at
    both temperatures (Tevini  et al., 1991c).

    11.2.4  Species competition

         Enhanced UVB exposure can cause changes in the growth of plants
    without necessarily decreasing plant production. Such changes
    include reduced leaf length, increased branching, and increased
    number of leaves (Barnes  et al., 1990). These changes seem to be
    general among different crop and weed species. Both graminoid and
    broad-leaf species respond in this fashion, with graminoids
    generally more responsive. Although these growth form changes do not
    lead to changes in the production of monocultures, in mixed species
    these alterations can lead to a change in the balance of competition
    for light.

         Multi-year field studies had shown that the balance of
    competition between wheat and wild oat (a common weed) began to
    favour wheat when mixtures of these species were subjected to
    increased UVB irradiation (Barnes  et al., 1988). In a recent study
    involving canopy light microclimate assessments and a detailed
    canopy radiation interception model, it was shown that the shift in
    growth-form of the two species was sufficient to quantitatively
    explain the change in the competitive balance (Ryel  et al., 1990).
    Thus, in many cases where plants are not necessarily depressed in
    overall growth by increased UVB exposure, changes in growth-form can
    have ecologically meaningful consequences. The direction of
    competitive balance changes are not easily predicted at present.
    However, altered competitive balance also has important implications
    for mixed-crop agriculture and species composition of
    nonagricultural ecosystems.

    11.2.5  Plant diseases

         Certain diseases may become more severe in plants exposed to
    enhanced UVB levels. Sugar beet ( Beta vulgaris) plants infected
    with  Cercospora beticola, and receiving 6.9 kJ m-2 day-1
    UVBBE, showed large reductions in leaf chlorophyll content, and
    fresh and dry weight of total biomass. In another study, three
    cucumber ( cucumis sativus) cultivars were exposed to a daily UVB
    dose of 11.6 kJ m-2 UVBBE in a greenhouse before and/or after
    infection with  Colletotrichum lagenarium or  Cladosporium
     cucumerinum, and analyzed for disease development (Orth  et al.,
    1990). Two of the three cultivars were disease resistant and the

    other was disease susceptible. Pre-infection treatment with UVB led
    to greater disease development in the susceptible cultivar and in
    one of the disease resistant cultivars. Post-infection treatment did
    not alter disease development. The increased disease development in
    UVB irradiated plants was found only on the cotyledons and not on
    true leaves, suggesting that the effects of UVB on disease
    development in cucumber vary according to the cultivar, timing of
    UVB exposure, and tissue age.

    11.2.6  UV-protection systems

         Epidermal pigments

         UVB induces flavonoid production (Wellmann, 1971), and may
    regulate the synthesis of UV protective pigments (Braun & Tevini,
    1991). In a study using two important crops (rye and oat),
    UV-fluence and wavelength dependent accumulation of isovitexin
    derivatives in the epidermal layer of rye seedlings prevented damage
    to chloroplast functions. In contrast, photosynthetic function was
    low without the accumulation of screening pigments (Tevini  et al.,
    1991a). Because the epidermal layer of oat seedlings already
    accumulates large amounts of UV-absorbing pigments during early
    development, the photosynthetic apparatus is better protected than
    rye seedlings against UVB (Braun, 1991). This inherently higher
    flavonoid production occurs even in absence of UVB irradiation, and
    appears to be constitutive in nature. UVB induction of flavonoids
    was demonstrated in two species of columbines,  Aquilegia caerulea,
    growing in alpine environments, and  Aquilegia canadensis, which
    grows at lower elevations (Larson  et al., 1990). In both species,
    flavonoid content increased upon UVB irradiation, even though the
    alpine species accumulated higher amounts in the UVB-free controls
    when compared to  A canadensis after UVB irradiation. This
    demonstrates that plants which are already genetically adapted to
    higher UVB environments can further increase their adaptation
    capacity.

         Photorepair

         A second protective mechanism in plants is photoreactivation.
    The UV-induced production of DNA pyrimidine dimers can be repaired
    by DNA photolysase. This enzyme was shown to increase with UVB
    irradiation in  Arabidopsis (Pang & Hays, 1991) but also by visible
    light via phytochrome in bean seedlings (Langer & Wellmann, 1990).
    This inducibility means that  de novo synthesis of DNA photolysase
    itself is a target for UV damage. Thus, the repair capacity of the
    cell may be reduced in the presence of increasing UVB (Wellmann,
    1991).

    11.3  Effects on Aquatic Ecosystems

         Aquatic ecosystems contribute more biomass (104 Gt a-1) than
    all terrestrial ecosystems (100 Gt a-1) combined. Recent work on
    UVB effects has concentrated on inhibition mechanisms and field
    studies in the subpolar waters of Antarctica, because of its high
    biomass production and the occurrence of the ozone hole over this
    region.

         Phytoplankton organisms orient within the water column using
    external factors. Mobility and orientation mechanisms are impaired
    by UVB exposure. Most organisms do not possess UVB receptors, and so
    cannot avoid deleterious effects of enhanced UVB that produces
    higher intensities deeper into the water column. New action spectra
    indicate that, in addition to DNA, other targets absorb UVB
    including proteins of the photoreceptor and photosynthetic
    apparatus. The inability of phytoplankton to adjust their position
    within the water column causes massive inhibition of photosynthesis.
    In only a few cases have UVB inducible screening pigments been
    identified.

    11.3.1  Effects on phytoplankton

         Recent UVB aquatic research has concentrated on phytoplankton
    and the Antarctic ecosystem. As shown in figure 11.2, phytoplankton
    is at the base of the aquatic food chain/trophic structure and
    serves as food for primary consumers (e.g., larvae of fish and
    shrimp), which in turn are consumed by secondary and tertiary
    consumers (e.g. fish). The production of phytoplankton has been
    estimated at about 6 x 1014 kg (UNEP, 1989). A loss of 10% would
    far exceed the gross national product of all countries in the world,
    assuming any reasonable price for biomass on the market. table 11.1
    gives the estimated annual biomass production for plankton and fish.

    Table 11.1  Estimated annual biomass production at different levels
    in marine food web and possible loss after 10% decrease at the
    phytoplankton level (adapted from UNEP 1989 report)
                                                                      
    Type                     Annual Production                 10% loss
                            (in million tonnes)
                                                                      
    Phytoplankton                 600,000                        60,000
    Zooplankton                    60,000                         6,000
    Small fish                      6,000                           600
    Large fish                        600                            60
                                                                      

    FIGURE 11.2

         Concentrations of phytoplankton in subpolar waters may be 103
    to 104 times greater than concentrations of phytoplankton found in
    tropical and subtropical seas (Jeffery & Humphrey, 1975). Any
    significant increase in UVB could well diminish growth and
    productivity of phytoplankton, subsequently affecting all higher
    trophic levels in the aquatic food web. Therefore, it is not
    surprising that a majority of recent research has looked at the
    effects of increased UVB exposure in Antarctic waters. Ongoing
    research activities include investigations of both direct
    (physiological and behavioural) and indirect (trophic implications)
    effects.

         Phytoplankton dwell in the top layers of the water column (the
    photic zone) because of their requirement for solar energy
    (Ignatiades, 1990). Their position within the column is maintained
    by precise orientation strategies using light, gravity and other
    external factors as guides. Phytoplankton in the photic zone would
    be exposed to any increase in solar UV. Most phytoplankton organisms
    do not possess UVB photoreceptors to guide them away from harmful
    UV, a situation similar to humans. Previous work demonstrated that
    mobility/orientation mechanisms in response to light are impaired by
    solar UV (Häder & Worrest, 1991; Baker & Smith, 1982). The ability
    of phytoplankton to adjust their position within the water column,
    in response to constantly changing conditions, may even be affected
    at ambient UVB levels. Although ambient UVB fluxes may cause damage
    to some species of phytoplankton, it should be emphasized that there
    are uncertainties regarding the magnitude of these effects. These
    included problems of extrapolating laboratory findings to the open
    sea and the nearly complete absence of data on long-term effects and
    ecosystem responses. Likewise, there is a need to investigate
    adaptation mechanisms. Before effects of exposure to solar UVB can
    be predicted, information is required on seasonal abundances and
    vertical distributions of marine organisms, vertical mixing, and the
    penetration of UVB into appropriate water columns.

         In their natural habitats, organisms are exposed to a wide
    range of UVB intensities. This radiation has been shown to affect
    growth, photosynthesis, nitrogen incorporation, and enzyme activity
    (Döhler & Alt, 1989; Döhler, 1990).

    11.3.2  UV increase and primary biomass production

         Recent results indicate that orientation mechanisms responsive
    to both light and gravity are affected by solar UV in a number of
    ecologically significant phytoplankton groups (Häder & Lui, 1991).
    Action spectra inhibitory effects are different from the DNA
    absorption spectrum and the action spectra calculated for higher
    plants, suggesting that UV exposure affects these organisms by a
    different mechanism. Proteins essential for specific functions such
    as orientation and photosynthesis are the primary targets of UVB.

         Biochemical analyses conducted to reveal the molecular targets
    of UVB inhibition show that specific photoreceptor proteins are
    degraded. Simultaneously photosynthetic pigments (responsible for
    converting solar energy) are bleached and destroyed by radiation
    (UNEP, 1989). The results of these biochemical studies are further
    supported by spectroscopic investigations showing losses in
    pigmentation.

         In order to evaluate the effects of enhanced UV, the vertical
    movement of natural phytoplankton was analyzed in 3m Plexiglas
    columns (Eggersdorfer & Häder, 1991). Most organisms moved to the
    surface during daytime hours, although some species avoided periods
    of intense UV during the midday hours by moving slightly down in the
    water column. However, this avoidance response is not sufficient to
    protect organisms under conditions of increased UVB irradiation
    (UNEP, 1989).

         The UVB irradiance in Antarctic waters significantly increased
    during the occurrence of the ozone hole (Bidigare, 1989; Lubin  et
     al., 1989; Karentz & Lutze, 1990). However changes in marine
    productivity accompanying UV flux changes have not been determined.
    Recent measurements show that UVB penetrates 65 meters deep into
    clear Antarctic waters (UNEP 1991). Consequently, measurements of
    photosynthetic biomass production in Antarctic waters under the
    ozone hole show a pronounced decrease in productivity by up to 25%
    (Holm-Hansen, 1990).

         Field studies indicate that photosynthesis is impaired first,
    followed by decreases in protein concentration and changes in
    pigment composition. As a result, a dramatic decrease in
    photosynthetic oxygen production can be measured after exposure to
    solar radiation (Smith  et al., 1980). Other spectral bands, such
    as UVA and visible radiation, may contribute to photosynthetic
    inhibition (Smith  et al., 1980). Likewise, photosynthetic
    inhibition has been detected in macroalgae at their natural depth
    (Bittersmann  et al., 1988; Nultsch  et al., 1990).

         In contrast to higher plants only a few photoplanktons produce
    UV absorbing substances (Carretto  et al., 1990; Karentz  et al.,
    1991b). However, all of these mycosporine-like amino acids have
    maximal absorption in the UVA range and only secondary peaks in the
    UVB. It is not clear whether the production of these potentially
    screening substances can be induced by exposing organisms to UV
    (Raven, 1991). One exception is cyanobacteria where a UVB inducible
    pigment has been found within the slime sheath surrounding the
    organisms, which absorbs up to 88% of the incident UVB
    (Garcia-Pichel and Castenholz, 1991).

    11.4  Conclusion

         Field and laboratory experiments on plant responses to
    increased UVB radiation underscore the concern for agriculture,
    forestry and natural ecosystems as the stratospheric ozone level is
    depleted.

         Growth and photosynthesis of certain crop plants can be
    inhibited even under ambient levels of UVB radiation. Certain
    environmental factors, both biotic (e.g. plant diseases and
    competition with other plants) and abiotic (e.g. carbon dioxide,
    temperature, heavy metals, and water availability) can alter UVB
    effects in plants. This increases the difficulty in making any
    quantitative predictions. Plants in temperate regions and certain
    tropical species were found to be adversely affected by enhanced UVB
    radiation.

         Marine ecosystem which provides the primary food for human
    consumption (in some countries) has been shown to be more sensitive
    to UVB than terrestrial plants. One consequence of loss in
    phytoplankton is reduced biomass production which would be
    propagated throughout the whole food web. The marine phytoplankton
    is a major absorber of atmospheric carbon dioxide. Any reduction in
    this population would decrease the uptake of carbon dioxide and so
    augment the greenhouse effect.

    12.  HEALTH HAZARD ASSESSMENT

    12.1  Introduction

         For the vast majority of people, the sun is the single largest
    source of exposure to UV. In some cases solar exposure will be
    elective such as from sunbathing, in others it will be adventitious
    as a result of outdoor recreational and/or occupational activity.
    Exposure may also occur from artificial sources of UV, either
    deliberately for example during medical treatment or the use of a
    sunbed for cosmetic purposes.

         The health risks associated with exposure to UV include those
    of both acute and chronic effects and will vary according to the
    nature of the exposure. Factors important in assessing such risks
    include: the biologically-effective irradiance of the UV incident on
    the person exposed; the duration and frequency of occurrence of
    exposures; and the individual sensitivity of the person to UV as
    determined by genetic and other factors.

         International guidelines on protection against UV given in
    chapter 13 are based on available scientific data (IRPA/INIRC,
    1991). The guidelines define occupational exposure limits (ELs)
    below which it is expected that nearly all people may be repeatedly
    exposed without adverse effects. The Els are intended to be used to
    evaluate potentially hazardous exposures from, for example, solar
    radiation, arcs, gas and vapour discharges, fluorescent lamps and
    incandescent sources. The Els are generally below levels which are
    often used for the UV exposure of patients required as part of
    medical treatment and below levels associated with sunbed exposure.
    IRPA/INIRC recommend that, where they are to be incorporated in
    regulations, the ELs should be considered as absolute limits for the
    eye, but only as advisory for the skin. This is because of the wide
    range of susceptibility to skin injury depending on skin type. The
    values were developed by considering lightly pigmented populations,
    with greatest sensitivity to sunburn and non- melanocytic skin
    cancer. In recommending similar limits, Threshold Limit Values
    (TLVs), the American Conference of Governmental Industrial
    Hygienists (ACGIH 1993) indicate that conditioned, tanned,
    individuals can tolerate skin exposure in excess of the exposure
    limits without sunburn effects, but that conditioning may not
    protect individuals against skin cancer. ELs are not intended to
    apply to exposure of pathologically photosensitive individuals, to
    people concomitantly exposed to photosensitising agents or to
    neonates.

         The threshold for adverse acute effects may be exceeded in
    certain people without exceeding the exposure limit. The risk of
    skin cancer increases with cumulative exposure over time. Therefore,
    it is recommended that all exposures should be reduced as far as is
    reasonably practicable.

         An estimate of the risk of cumulative exposure to UV can be
    expressed in terms of the cumulative incidence (I) of non-melanoma
    skin cancer in the form of the equation (Schothorst  et al., 1985;
    Diffey 1988).

                   I = gamma A Hß aalpha                            12.1

    where I is the total number of cases per 100,000 of the population
    in the age group up to age 'a' years:

    H is the annual carcinogenic-effective radiant exposure at the skin
    surface:

    A is the fraction of the body surface area exposed:

    gamma, ß and alpha are numerical constants reflecting the genetic
    susceptibility of the exposed population, the biological
    amplification factor and the age dependence of the cumulative
    incidence respectively.

         The above equation applies only to situations where the annual
    effective exposure remains constant year by year. This will not be
    the case, for example, for workers whose job may result in
    additional occupational exposure during their working lives (Diffey
    1988). In these circumstances the cumulative incidence I can be
    represented by the equation (Slaper  et al., 1986).

              I = gamma A Heß aalpha                                12.2

    where He is given by

              He = H + Ho(a-as)/a

    H is the annual effective radiant exposure of solar UV:

    as is the age at which the occupational exposure began:

    Ho is the annual effective radiant exposure from occupational
    exposure.

         In the following sections estimates of personal exposures
    likely to result from different situations are considered and, where
    the availability of data permits, examples of estimates of
    quantitative risk are provided. The exposure situations include
    elective exposure to solar UV, from medical treatment and from the
    use of sunbeds, as well as adventitious exposure resulting from
    solar radiation and from artificial sources of UV.

    12.2  Elective Exposures

    12.2.1  Medical exposure

         UV is used in medicine for both the diagnosis and treatment of
    disease. The highest and most extensive medical exposures to UV
    result from phototherapy and photochemotherapy.

         Medical treatment using UV carries the risk of acute side
    effects and long term risks of chronic effects, in particular skin
    cancer. The acute side effects of UVB phototherapy exposure include;
    reddening, swelling, blistering and desquamation of the skin. The
    acute effects following PUVA treatment include itching, skin pain
    and nausea (Green  et al., 1992).

         Chronic effects include those on both the eyes and the skin.
    However, the eyes of patients are generally well protected during
    PUVA treatment and only one case of PUVA induced cataract in humans
    appears to have been reported (Cyrlin 1980). Structural changes in
    the skin, akin to those characteristic of long term damage resulting
    from solar radiation exposure, have been observed as a result of
    both UVB and PUVA phototherapy.

         The risk of skin cancer from ultraviolet phototherapy has been
    reviewed by Green  et al. (1992). They conclude that data on the
    association between UVB phototherapy and skin cancer are very poor.
    Only one case control study has been reported and no correlation was
    found between UVB phototherapy and skin cancer in 85 patients who
    had received UVB phototherapy for up to 25 years. The association
    between PUVA treatment and skin cancer in psoriasis patients has
    been reported in ten studies. However, only one study (Stern  et
     al., 1979) indicated that PUVA acted as an independent carcinogen,
    the risk of developing skin cancer being 6-12 times that of a
    population survey. The results of other studies point to PUVA acting
    in the role of a co-carcinogen, other factors involved being a
    family history of skin cancer and treatment with antipsoriatic
    medication. Comparisons and interpretations of data from US and
    European studies have been complicated by differences in the
    exposure regimens used.

    12.2.2  Phototherapy of seasonal affective disorder (SAD)

         Typical treatment consists of sitting in front of a panel of
    fluorescent lamps and exposing the face, with the eyes open, to an
    illuminance of about 2500 lux for 2 to 6 hours per day during the
    winter months. This is equivalent to a daily erythemally-effective
    radiant exposure of up to 40 J m-2 (Diffey 1993). An annual
    cumulative exposure of up to 24 MEDs results, which, if continued
    for several decades is estimated to result in a risk of non-melanoma
    skin cancer of about 1.5, compared with someone (typical indoor
    worker) minimally exposed.

    12.2.3  Sunbeds

         Among lightly pigmented people, a suntanned skin is
    unfortunately still socially desirable. The establishment of the
    'suntanning industry' has enabled the acquisition of a 'suntan'
    irrespective of the availability of solar radiation. Suntanning is
    caused by UV and as with exposure to the sun, there are attendant
    risks in using sunbeds.

         Adverse health effects of sunbed use include acute effects on
    the skin, longer term structural damage of the skin and increased
    risk of skin cancer (Roza  et al., 1989). Provided that appropriate
    protective eyewear is worn, adverse health effects on the eyes will
    be avoided. The extent to which adverse, but relatively transient
    skin effects occur is demonstrated by the results of a survey of the
    use of sunbeds in the United Kingdom (Diffey 1986). Questionnaire
    replies from over 1000 sunbed users indicated that the incidence of
    acute adverse effects was substantial; 28% of users complained of
    itching and about 8% developed a skin rash or had felt nauseous at
    some time during or immediately after exposure. The incidence of
    such effects was higher among women taking oral contraceptives than
    in women who were not. No conclusion could be drawn about other
    medications as few individuals were taking them. The survey also
    revealed that 50% of sunbed users in the United Kingdom are female
    and aged between 16 and 30 years. 43% of users had skin types I and
    II, the most sensitive to the adverse effects of UV and the ones
    least likely to tan well.

         Structural damage to human skin from exposure to UVA, such as
    demonstrated experimentally in mice (Kligman  et al., 1987, Bissett
     et al., 1989) might be expected in people as the result of
    excessive use of sunbeds. Increased skin fragility and blistering
    (Farr  et al., 1988; Murphy 1989) and atypical melanocytic lesions
    (Jones  et al., 1987) have been observed in people who have used
    UVA sunbeds excessively. In some people, photodermatosis and
    polymorphic light eruption, is readily caused by exposure to UVA
    radiation from a sunbed (Rivers  et al., 1989). Certain
    photo-aggravated dermatoses, such as lupus erythematosus, are
    exacerbated by the use of UVA sunbeds (Stern and Docken 1986). The
    use of certain medications such as antihypertensives and antibiotics
    (Hawk 1984) and topical application of certain products, including
    perfumes, body lotions, etc., may produce a photosensitising effect
    on exposure to a sunbed.

         Some localised skin and systemic changes in immunological
    reactions result from exposure to UV and, of particular relevance to
    sunbeds, from exposure to UVA. There is also evidence that exposure
    to UV can accelerate the growth of human viruses (e.g. Otani and
    Mori 1987, Perna  et al., 1987), including human immunodeficiency
    virus (HIV) (Zmudzka and Beer 1990). At present, the significance of

    these observations with respect to the health of people exposed to
    UV from sunbeds is unclear.

         The risk of non-melanocytic skin cancer in northern Europeans
    who sunbathe and who use sunbeds has been estimated by Diffey (1987)
    using a mathematical model of skin cancer incidence that makes
    allowance for childhood, occupational and recreational solar
    radiation exposure. The model takes into account the fractions of
    total skin surface area that might normally be exposed during
    everyday activities and the larger fraction, normally unexposed, but
    exposed during sunbathing and sunbed use. The annual effective
    radiant exposures resulting from different activities and exposure
    scenarios used as the basis for risk calculations are expressed as
    representative minimal erythemal doses (MEDs) in table 12.1.

    Table 12.1  Representative annual minimal erythemal doses (MEDs) for
    various exposure scenarios in northern Europeans (Diffey, 1987,
    1993a).
                                                                      
         Scenario                                        Annual
                                                         MEDs
                                                                      

         Outdoor worker                                  270

         Indoor worker                                   90
         (including weekend exposure)

         Sunbathing holiday in Mediterranean             50-100
         area for 2 week period in summer

         UVA sunbed (low pressure fluorescent            20
         lamps, 30 x 30 minute sessions *
                                                                      

    *  Recommended maximum number of sessions per year (IRPA/INIRC,
       1991a) with estimated exposure of 0.7 MED per session
       (Diffey, 1987).

         Diffey (1987) concluded that the increased risk of
    non-melanocytic skin cancer associated with the use of a UVA sunbed
    for 10 sessions per year is negligible. However, frequent use of a
    sunbed, for example once per week from the age of 20 years will
    result in an estimated doubling of the risk of non-melanocytic skin
    cancer by age 45 years.

         Several studies have reported increased risks of cutaneous
    melanoma in users of sunlamps and sunbeds (IARC 1992).

    12.2.4  Sunbathing

         A significant contribution to the risk of non-melanocytic skin
    cancer that results from sunbathing (Diffey, 1987). In these
    calculations it was assumed that the subjects were indoor workers
    who began sunbathing at age 20 years, and that their annual exposure
    up to age 16 years is one half that of an outdoor worker. An indoor
    worker who does not sunbathe is estimated to have a 2-3% risk of
    non-melanoma skin cancer by age 70. Annual two week vacations spent
    sunbathing at Mediterranean latitudes (approx. 40°N) will increase
    this risk by a factor of about 5. Sunbathing for four weeks annually
    is estimated to result in a 10-20 fold increase in cumulative risk
    compared with non-sunbathers. It should be stressed that sunbathing
    during vacation periods was assumed to take place for nearly the
    whole day on every day of vacation. This is unlikely to be the case
    for most holiday makers and so the estimated risk factors for most
    people taking holidays in a sunny climate are likely to be one-third
    to one-half of those given.

    12.3  Adventitious Exposures

    12.3.1  Outdoor exposures

         Studies using personal dosimeters to monitor exposure of
    individuals to UV have shown that, on average the
    carcinogenic-effective UV exposure (excluding vacation exposure) of
    outdoor workers is about 3 times that of indoor workers (Holman  et
     al., 1983; Larko and Diffey 1983; Schothorst  et al., 1985).

    Assuming that up to the age of 16, indoor and outdoor workers
    receive the same carcinogenic-effective radiant exposure and that
    both receive equivalent vacation exposures, the relative risk for
    non-melanoma skin cancer for outdoor workers is estimated to be 3.7
    times that for indoor workers (Diffey 1987). However, recent
    population studies have found only small differences in skin cancer
    incidence between outdoor and indoor workers (see chapter 8).

         Armstrong & Kricker (1994) have estimated the proportion of
    cutaneous malignant melanomas that is caused by sun exposure. The
    estimated proportions varied from 0.97 in males and 0.96 in females
    in Queensland, Australia, when the incidence on the whole body was
    compared with the incidence on unexposed sites, to 0.68 when the
    incidence in people born in Australia was compared with that in
    migrants to Australia from areas of lower sun exposure. A comparison
    of whites and blacks in the US in which the incidence in blacks was
    taken as the incidence in unexposed whites, gave estimates of 0.96
    in males and 0.92 in females. It was estimated that some 59,000
    (65%) of about 92,000 malignant melanomas worldwide in 1985 were
    caused by sun exposure.

    12.3.2  Artificial sources

         General public exposure

         Exposure of the general public to potentially hazardous levels
    of UV from artificial sources is unlikely. However, an area that
    merits note is the increased use of tungsten halogen lamps for
    lighting. Very high filament temperatures combined with quartz
    envelopes results in the emission of significantly higher levels of
    UV compared with conventional incandescent lamps (Césarini & Muel,
    1989; McKinlay  et al., 1989). For many applications of such lamps
    the presence of a glass filter effectively attenuates the
    potentially harmful UV. However, for some others, and particularly
    when such lamps are incorporated in desk-top luminaires, no
    protective filter is present, and significantly high levels of
    erythemally effective UV have been measured. However, for lamps of
    similar design and having nominally the same power, there is a wide
    range of emitted levels of UV, (for example, between 2 and 56 mW
    m-2 effective, in the beam of the lamp at 30 cm distance)

         Occupational exposure

         Sources that emit UV are used for a wide range of applications
    in the workplace. High intensity discharge sources should be, and
    often are, contained in interlocked enclosures thus obviating
    hazardous exposure of people. Hazard evaluation surveys generally
    consist of measurements of the effective exposure levels of UV and
    comparison with recommended guidelines on limiting exposure. This
    approach addresses the adverse acute effects of exposure, but very
    few risk assessments have been carried out in relation to chronic
    effects. The risk of non-melanocytic skin cancer for any
    occupational situation can be calculated using multivariant
    analysis, provided the effective exposures related to the occupation
    are known. However, although data on levels of exposure exist for
    some occupational situations there is generally a paucity of such
    data. Where estimates of skin cancer risk have been made, they have
    related to specific occupational exposure situations (Diffey 1988,
    1989) or broadly with respect to general (fluorescent) lighting
    (Lytle  et al., 1993).

         Examples where occupational exposure to UV from artificial
    sources may occur include; electric arc plasma welding, the drying
    and curing of inks, resins, plastics and paints; printing, graphics
    arts, copying (using photographic processes) and photography,
    photoetching, projector lamps operation, medicine, scientific
    laboratory work, tanning and salon work and UV associated with
    general lighting fluorescent lamps and desk-top and other luminaires
    incorporating tungsten halogen lamps.

         Welders are likely to be the largest occupational group with
    potential exposure to UV (IARC 1992). The UV effective irradiance

    levels around an operating welding arc can exceed recommended
    exposure guideline levels by several orders of magnitude, and the
    occurrence of acute effects of exposure, photokeratitis and
    erythema, is common (Eriksen 1987).

         Many industrial photoprocesses use high intensity discharge
    lamps that emit copious quantities of UV. However, most of these
    sources are effectively shielded and interlocked to prevent human
    exposure. Some sources emit leakage radiation and, under these
    circumstances, a detailed measurement survey is required to assess
    the degree of hazard.

         Medical physiotherapists involved in the phototherapy of
    patients are occupationally exposed to UV and an analysis of
    probable risk has been published (Diffey 1988). During a working
    life of 40 years, the additional risk of non-melanoma skin cancer
    for a member of this group is estimated to be around 25% compared
    with that for non-exposed workers.

         General lighting

         Low pressure mercury vapour fluorescent lamps are ubiquitous in
    the workplace and the home and concern has been expressed regarding
    the potential role of the UV they emit in the aetiology of malignant
    melanoma. Epidemiological data on exposure to general lighting
    fluorescent lamps and malignant melanoma are few and inconsistent.
    IRPA/INIRC (1991) concluded that UV exposure from indoor fluorescent
    lighting should not be considered as a malignant melanoma risk.

         The exposure levels associated with the use of general lighting
    fluorescent lamps have been measured in a number of studies (Cole
     et al., 1986; McKinlay & Whillock, 1987; Muel  et al., 1988 and
    Lytle  et al., 1993). Exposure data vary considerably depending on
    the types of lamps examined and the exposure conditions considered.
    A recent study (Lytle  et al., 1993) serves to illustrate estimates
    of the risk associated with long term exposure to unfiltered general
    lighting fluorescent lamps used in the United States. It was
    estimated that 50 years exposure to typical unfiltered levels of UV
    from fluorescent lamps added approximately 4% (1.6 - 12%) to any
    risks associated with exposure to solar UV.

    13.  INTERNATIONAL GUIDELINES ON EXPOSURE TO ULTRAVIOLET RADIATION

         A number of national and international organizations have
    promulgated guidelines or standards on exposure to UV. Most are
    based upon the same basic criteria of ACGIH (1993) and IRPA/INIRC
    (1991).

         The basic exposure limit (EL) for both general public and
    occupational exposure to UV incident on the skin or eye is 30 J
    m-2 effective), when the spectral irradiance Elambda at the eye
    or skin surface is mathematically weighted with the hazard relative
    spectral effectiveness factor Slambda from 180 nm to 400 nm. This
    is given as follows:

         Eeff = Sigma Elambda Slambda Deltalambda

    where:

         Eeff = effective irradiance W m-2

         Elambda = spectral irradiance from measurements in W m-2
    nm-1

         Slambda = relative spectral effectiveness factor (unit-less)

         Deltalambda = bandwidth of the calculation or measurement in
    nm

         At 270 nm in the UVC range, Slambda is 1.0, but at 360 nm in
    the centre of the UVA range, its value falls to 0.00013, and
    continues to fall for longer wavelengths.

         For the UVA, the total radiant exposure incident on the
    unprotected eye should not exceed 104 J m-2 (1 J cm-2) within
    an 8 h period. The total 8 h radiant exposure incident on the
    unprotected skin should not exceed the values in table 13.1.

         The radiant UV exposure incident upon the unprotected skin or
    eye within an 8-hour period should not exceed the values given in
    table 13.1. The limits apply to sources whose emissions are measured
    with an instrument having a cosine response detector oriented
    perpendicular to the most directly exposed surfaces of the body when
    assessing skin exposure and along (or parallel to) the line(s) of
    sight when assessing ocular exposure. Although no measurement
    averaging aperture is recommended, 1 mm is commonly used.

    Table 13.1  International UV exposure limits and spectral weighting
                factor (IRPA/INIRC, 1991)
                                                                      
    Wavelengtha (nm)    EL          EL            Relative Spectral
                        (J m-2)     (mJ cm-2)     Effectiveness Slamda
                                                                      

    180                 2,500       250           0.012
    190                 1,600       160           0.019
    200                 1,000       100           0.030
    205                 590         59            0.051
    210                 400         40            0.075
    215                 320         32            0.095
    220                 250         25            0.120
    225                 200         20            0.150
    230                 160         16            0.190
    235                 130         13            0.240
    240                 100         20            0.300
    245                 83          8.3           0.360
    250                 70          7.0           0.430
    254b                60          6.0           0.500
    255                 58          5.8           0.520
    260                 46          4.6           0.650
    265                 37          3.7           0.810
    270                 30          3.0           1.000
    275                 31          3.1           0.960
    280b                34          3.4           0.880
    285                 39          3.9           0.770
    290                 47          4.7           0.640
    295                 56          5.6           0.540
    297b                65          6.5           0.460
    300                 100         10            0.300
    303b                250         25            0.190
    305                 500         50            0.060
    308                 1,200       120           0.026
    310                 2,000       200           0.015
    313b                5,000       500           0.006
    315                 1.0 x 104   1.0 x 103     0.003
    316                 1.3 x 104   1.3 x 103     0.0024
    317                 1.5 x 104   1.5 x 103     0.0020
    318                 1.9 x 104   1.9 x 103     0.0016
    319                 2.5 x 104   2.5 x 103     0.0012
    320                 2.9 x 104   2.9 x 103     0.0010
    322                 4.5 x 104   4.5 x 103     0.00067
    323                 5.6 x 104   5.6 x 103     0.00054
    325                 6.0 x 104   6.0 x 103     0.00050
    328                 6.8 x 104   6.8 x 103     0.00044
    330                 7.3 x 104   7.3 x 103     0.00041
    333                 8.1 x 104   8.1 x 103     0.00037
    335                 8.8 x 104   8.8 x 103     0.00034
    340                 1.1 x 105   1.1 x 104     0.00028

    Table 13.1 (contd).
                                                                      
    Wavelengtha (nm)    EL          EL            Relative Spectral
                        (J m-2)     (mJ cm-2)     Effectiveness Slamda
                                                                      

    345                 1.3 x 105   1.3 x 104     0.00024
    350                 1.5 x 105   1.5 x 104     0.00020
    355                 1.9 x 105   1.9 x 104     0.00016
    360                 2.3 x 105   2.3 x 104     0.00013
    365b                2.7 x 105   2.7 x 104     0.00011
    370                 3.2 x 105   3.2 x 104     0.000093
    375                 3.9 x 105   3.9 x 104     0.000077
    380                 4.7 x 105   4.7 x 104     0.000064
    385                 5.7 x 105   5.7 x 104     0.000053
    390                 6.8 x 105   6.8 x 104     0.000044
    395                 8.3 x 105   8.3 x 104     0.000036
    400                 1.0 x 106   1.0 x 105     0.000030
                                                                      

    a  Wavelengths chosen are representative; other values should be
       interpolated at intermediate wavelengths.
    b  Emission lines of a mercury discharge spectrum.

         The permissible exposure duration, tmax, for exposure (in
    seconds) to UV is calculated by:

              tmax = 30/ Eeff (W m-2)

         Examples are provided in table 13.2.

    Table 13.2  Limiting UV exposure durations based on exposure limits
    (IRPA/INIRC, 1991)
                                                                      
    Duration of exposure                 Effective irradiance       
    per day                       Eeff (W m-2)        Eeff (µW cm-2)

    8 hours                       0.001               0.1
    4 hours                       0.002               0.2
    2 hours                       0.004               0.4
    1 hour                        0.008               0.8
    30 minutes                    0.017               1.7
    15 minutes                    0.033               3.3
    10 minutes                    0.05                5
                                                                      

         The EL's were developed considering lightly pigmented
    populations with greatest sensitivity and predisposition to adverse
    health effects from exposure to UV. The limits apply to UV exposure
    of the working population, but with some precaution also apply to
    the general public. However, some rare, highly photosensitive

    individuals exist who may react adversely to exposure at these
    levels. These individuals are normally aware of their heightened
    sensitivity. Likewise, if individuals are concomitantly exposed to
    photosensitizing agents, an enhanced reaction can take place. Many
    individuals who are exposed to photosensitizing agents (ingested or
    externally applied chemicals, e.g., in cosmetics, foods, drugs,
    industrial chemicals, etc.) may not be aware of their heightened
    sensitivity. Lightly pigmented individuals conditioned by previous
    UV exposure (leading to tanning and hyperplasia) and heavily
    pigmented individuals can tolerate skin exposure in excess of the
    EL's without erythemal effects. However, repeated tanning may
    increase the risk of accelerated skin aging and even skin cancer.
    Such risks should be understood prior to the use of UV for medical
    phototherapy or cosmetic exposures.

    14.  PROTECTIVE MEASURES

    14.1  Introduction

         It is widely accepted by scientific and medical authorities
    throughout the world that UV is potentially carcinogenic and capable
    of producing other undesirable health effects. It is sensible
    therefore to take steps to minimise UV exposure. Many risks in life
    are completely beyond our control such as contracting a rare
    disease. However, risks to health associated with exposure to UV
    from both natural and artificial sources can be substantially
    reduced by taking appropriate control measures.

         Since UV exposure occurs externally, simple measures can be
    taken to reduce the exposures received. A high degree of protection
    can be afforded by protective clothing (including hats), UV
    protective eyewear (welding helmets, face shields, goggles,
    sunglasses, spectacles etc.) and by sunscreens for exposed skin.
    However, the degree of protection afforded can be reduced by
    ingestion of photosensitizing drugs or photoallergic/phototoxic
    reactions produced by chemicals or cosmetics in contact with the
    skin. Thus, education is also an important control measure.

    14.2  Education

         Concern about high incidences of skin cancer and eye damage
    have led to national educational campaigns in some countries to
    encourage people to protect themselves against excessive UV exposure
    from the sun and in the workplace. Educational programmes directed
    at both the workforce and the public are intended to create an
    awareness of the adverse health effects that can result from
    overexposure to UV.

         Presently, in several different countries around the world,
    daily environmental UV levels are supplied to the general public in
    the form of UV indices. Their provision is intended to educate the
    public on the basic climatology of UV, increase awareness of the
    hazards of UV and provide information necessary to plan protection.
    In a report commissioned to investigate Canadian attitudes to
    Environment Canada's UV index (Environment Canada, 1993), it was
    found that 73% of Canadians were aware of the UV index, 91% believed
    UV affected human health, and most important, 59% of respondents who
    were aware of the UV index had changed their sun exposure habits.

         Different countries have adopted their own national UV indices.
    Unfortunately, they are not all compatible. Clearly, international
    uniformity would help to prevent confusion.

         Educational programmes aim to produce a change in knowledge and
    attitudes, then a change in behaviour and eventually a reduction in
    the incidence and mortality rates of skin cancer. A survey of

    sunscreen use on beaches in Brisbane, Australia found about 70% of
    females and males applied sunscreen. Half of the sunscreens provided
    the maximum protection (SPF 15+) and almost 90% used a waterproof
    formulation. However, the sunscreen was not applied over the entire
    body with over half neglecting ears and lower limbs (Pincus  et al.,
    1991). Survey results show that sunburn is still occurring; in a
    randomly selected group of adults in Melbourne, Australia 16%
    reported sunburn over a summer weekend (Hill  et al., 1992). With
    regard to skin cancers it is still too early to evaluate the
    effectiveness of current educational/publicity campaigns.

         With respect to changes in behaviour, there is good reason to
    avoid exposure to the midday sun since it has been estimated that up
    to one third of the day's erythemally effective UV is received
    within the period one hour before noon to one hour after noon. Shade
    is a useful method of protection but its value should not be
    over-estimated, as one may be exposed to a quarter or more of the
    total solar UV while shaded from direct sunlight, depending on the
    prevailing exposure conditions.

    14.3  Protection Factors

         The concept of a protection factor is useful when attempting to
    quantify the UV protection that items such as sunscreens, clothing
    and eyewear can provide (Gies  et al., 1992). To determine the
    protection factor, the following procedure is conducted. An
    effective dose (ED) of UV to the unprotected skin or eye is
    calculated by summing the incident solar spectral power over the
    wavelength range 280 to 400 nm. In order to determine the effective
    dose (EDm) for the skin or eye when it is protected, the
    calculation is repeated with the spectral transmission of the
    protection item as an additional weighting. The protection factor
    (PF) is then defined as the ratio of ED to EDm and is given by the
    following equation:

                   ED        SigmaElambda.Slambda.Delta lambda
         PF   =                                                   
                   EDm    SigmaElambda.Slambda.Tlambda.Delta lambda

    where:

    Elambda = spectral irradiance (W m-2 nm-1) at wavelength lambda

    Slambda = relative spectral effectiveness

    Tlambda = spectral transmission of protective item at
              wavelength lambda

    Delta lambda = wavelength interval or bandwidth (nm)

    lambda = wavelength (nm)

         The inclusion of the spectral effectiveness function in the
    calculation ensures that sufficient weighting is given to the
    biologically effective wavelengths below 315 nm. A description of
    how the protection factors for fabrics, eyewear and sunscreens are
    determined is given in Roy and Gies (1993) and CIE (1991).

    14.4  Clothing

         Use of protective clothing provides one of the simplest means
    of reducing UV exposure. Hats have been shown to afford protection,
    to various degrees, to the forehead, scalp, ears and most of the
    neck. Their protective properties have been studied by Diffey and
    Cheeseman (1992). In this study hats were classified into four
    categories, small brim, medium brim, large brim and peaked cap. The
    protection afforded to different anatomical sites on the head and
    neck is shown in table 14.1.

    Table 14.1  Sun protection for various anatomical sites on the head
    and neck provided by different types of hat (Diffey & Cheeseman,
    1992).

                                                                      
                             Typical sun protection factor *
                                                                      
    Style of hat   Forehead   Nose       Cheek      Chin       Back of
                                                               neck
                                                                      

    Small brim     15         1.5        1          1          1
    < 2.5 cm

    Medium brim    >20        3          2          1          1
    2.5-7.5 cm

    Large brim     >20        7          3          1.2        5
    >7.5 cm

    Peaked cap     >20        5          1.5        1          1
                                                                      

    * In this table 'sun protection factor' is defined as the reciprocal
      of the fraction of UV exposure recorded relative to that of the
      unprotected head.

         The degree of protection provided by clothing depends on the
    penetration of UV through materials and this can vary considerably.
    Fabrics which are visibly opaque tend to be more highly absorbent of
    UV, but the structure or weave of a material is the most important
    factor in determining its protective value. Colour and thickness are
    a poor guide to UV protection. The transmission properties of some
    fabrics commonly used in the manufacture of clothing for everyday

    wear are given in table 14.2 (Welsh and Diffey 1981). Here the
    protection factor is an estimate of the protection afforded against
    biologically effective solar radiation. A high protection factor is
    associated with a tightly woven material. UV is transmitted and
    scattered through the interstices of the material itself rather than
    penetrating the fabric.

         Gies  et al. (1992) have recently extended the concept of
    protection factor (PF) to fabrics and proposed the Ultraviolet
    Protection Factor (UPF). The UPF scheme, as shown in table 14.3, is
    designed to give the general public information on the amount of UV
    protection available from fabrics and clothing and is now in use in
    Australia. UPF is analogous to SPF and a fabric of UPF 10 would, in
    principle, provide a similar level of protection as a sunscreen of
    SPF 10.
    
    Table 14.2  UVB transmission properties of common fabrics
                (from Welsh & Diffey 1981)
                                                                                            
    Fabric                        Structure   Colour      Thickness   % of        Protection
                                                          (mm)        incident    factor
                                                                      UV
                                                                      transmitted
                                                                                            

    Nylon-tricel                  Woven       Black       0.1         0.15         750
    Nylon-viscose jacquard        Woven       Black       0.2         0.20         500
    Nylon                         Woven       White       0.1         1.7          55
    Nylon-terylene mixture        Knitted     Blue        0.2         11           9
    Nylon-acetate jersey          Knitted     Pink        0.2         24           4
    Polyester-slub viscose        Woven       Pink        0.5         9           14
    Polyester-printed lawn        Woven       Red         0.2         7           11
    Polyester-jersey              Knitted     Fawn        0.3         7           14
    Polyester-jersey              Knitted     Cream       0.4         5           19
    Polyester-jersey              Knitted     Black       0.3         8           12
    Polyester-jersey              Knitted     Orange      0.3         5           23
    Polyester-jersey              Knitted     Turquoise   0.5         16          6
    Polyester-jersey              Knitted     Brown       0.7         1.6         68
    Polyester-jersey              Knitted     Black       0.5         4.4         23
    Polyester-brushed jersey      Knitted     Blue        0.4         5.2         19
    Polyester-brushed jersey      Knitted     Green       0.4         6.0         16
    Polyester-bouclette           Knitted     Orange      0.4         3.0         33
    Polyester-bouclette           Knitted     Purple      0.4         2.0         51
    Cotton-needlecord             Woven       Brown       0.5         < 0.1       > 1000
    Cotton-denim                  Woven       Blue        0.5         < 0.1       > 1000
    Cotton-printed                Woven       Brown       0.3         < 0.1       >1000
    Cotton-printed                Woven       Cream       0.3         2.7         36
    Wool-jersey                   Knitted     Fawn        0.7         0.7         150
                                                                                            
    
    Table 14.3  Summary of the ultraviolet protection
                factor scheme for fabrics
                                                                      
         Ultraviolet      Mean % UV         Protection
         protection       transmission      category
         factor UPF
                                                                      
         UPF 40+          less than 2.5     Maximum protection 40+
         UPF 30 to 39     3.3 to 2.5        Very high protection
         UPF 20 to 29     5.0 to 3.3        High protection
                                                                      

         Whether a material is wet or dry is important in relation to
    its UV transmission properties. The spectral transmittances of some
    cotton and polyester-cotton samples are shown in figure 14.1.
    Measurements show that the variation of UPF for wet and dry fabrics
    is consistent for cotton, all examples showing a decrease in UPF
    when wet (Gies  et al., 1992). The variation in UPFs between wet
    and dry was less consistent for polyester- cotton than for the
    cotton, some UPFs increasing for wet fabric, while others decreased.

    14.5  Sunscreens

         Sunscreens are physical and chemical topical preparations which
    attenuate the transmission of solar UV into the skin by absorption,
    reflection or scattering. Physical sunscreens (sunblocks), for
    example zinc oxide, titanium dioxide or red ferric oxide, function
    by reflecting and scattering and provide protection against a broad
    spectrum of UV and visible wavelengths. They are normally nontoxic
    and have few known adverse effects. Sunscreens based on chemical
    absorbers contain one or more colourless UV-absorbing ingredients
    which generally absorb UVB radiation more strongly that UVA.
    Para-aminobenzoic acid (PABA) and its derivatives, salicylates,
    cinnamates and camphor derivatives primarily absorb UVB and transmit
    UVA; benzophenones essentially absorb UV of wavelengths of less than
    360 nm. The use of solely UVA absorbers (di-benzoyl-methane) is
    allowed in only certain countries. These chemicals are all based on
    benzene (Moseley 1988).

         The application of any sunscreen normally changes the spectrum
    of UV that reaches the target cells. Although most sunscreens are
    designed to attenuate UV, some contain additives such as bergamot
    oil (containing 5-methoxypsoralen) to enhance pigmentation and
    photoprotection (Young  et al., 1991). The role of such
    preparations remains controversial.

    FIGURE 14.1

         The generally accepted parameter for evaluating the efficacy of
    sunscreen preparations is the sun protection factor (SPF), which is
    defined as the ratio of the least amount of UV required to produce
    minimal erythema after application of a standard quantity of the
    sunscreen to the skin to that required to produce the same erythema
    without sunscreen application. Several countries have published
    recommendations for the efficacy testing of propriety sunscreens,
    e.g. the US Food and Drug Administration (1978) and CIE (1991).

         Many factors influence SPF values; particularly important are
    the spectral power distribution of the source used for SPF testing
    and a clear definition of the end-point used for assessment (Urbach,
    1989). Variations in these factors can lead to considerable
    differences in measured SPF values for the same product.

         SPF values generally reflect the degree of protection against
    solar UVB radiation, but their protective capacity against UVA must
    also be defined. Several in-vivo and in-vitro methods have been
    proposed for defining protection against UVA but there is currently
    no international consensus on which is the most appropriate.
    However, this issue is currently being considered by a technical
    committee of the CIE.

         Correctly used, sunscreens are effective in preventing sunburn.
    Actual SPF values depend critically on the thickness of the
    application and on other factors such as absorption into the skin,
    sweating and contact with water (for example while swimming).

         If sun exposure causes skin cancer, it would be expected that
    the use of efficient sunscreens would prevent skin cancer. In animal
    experiments, sunscreens with a SPF of as low as 4 have been shown to
    be effective in preventing or delaying the onset of UV-induced skin
    tumours and cancers (Forbes  et al., 1989). There is, at present,
    neither direct epidemiological nor laboratory evidence to suggest
    that they prevent melanoma or basal cell carcinoma. The few studies
    conducted to date suggest either no effect or a causal rather than a
    protective effect. For example, Holman  et al. (1986) found a
    relative risk of melanoma of 1.06 (95% confidence interval
    0.71-1.57) with 1-9 years use of sunscreens and 1.15 (95% confidence
    interval 0.78-1.68) with 10+ years of use relative to "never use". A
    statistically significant positive association of "often or very
    often" use of sun protection agents with melanoma (relative risk
    1.8, 95% confidence interval 1.5-3.8) was found by Beitner  et al.
    (1990). In their cohort study of basal cell carcinoma in US nurses,
    Hunter  et al. (1990) found a higher risk in those who spent 8+
    hours per week outside and used sunscreen than in those who spent
    the same time outside but did not use sunscreen.

         These results cannot be taken at face value. First, there is
    likely to be negative confounding of the use of sunscreens with
    cutaneous sun sensitivity, people with highly sun sensitive skins

    are likely to use sunscreens more than people with less sensitive
    skin. Second, there is good evidence that the epidemiological
    studies have identified an enthusiastic suntanning population using
    sunscreens with minimal SPF to promote their suntan. Third, it is
    likely that people adopt or increase their use of sunscreens after
    their first skin cancer has been diagnosed, or their attention is
    drawn to their personal risk in some other way (for example,
    diagnosis of a solar keratosis). Thus present users of sunscreens
    are likely to be enriched with people at higher than average risk of
    skin cancer. While adjustment for confounding with skin sensitivity
    has been carried out in some studies (Holman  et al. 1986, Hunter
     et al. 1990), this third issue has not been addressed in any
    published analysis. Fourth, relatively little information is
    available on the mutagenic and carcinogenic potential of various
    sunscreens. The US National Cancer Institute (1989) recommended the
    following six compounds to be evaluated for chronic testing in
    rodents by the US National Toxicology Program: cinoxate;
    2-ethylhexyl 2-cyno-3,3-diphenyl-acrylate; 2-ethylhexyl
    para-methoxycinnamate; homosalate; methyl anthranilate; and
    oxybenzone. Neither epidemiological nor long term mammalian
    carcinogenicity data are available on these compounds. The results
    of in-vitro studies were assessed as either negative or inconsistent
    among systems or among batches of a compound (because of
    impurities). 2-ethylhexyl para-methoxycinnamate was implicated as a
    potential tumour initiator in one study in which hairless mice were
    painted with the compound over a nine-week period and subsequently
    treated with the tumour promoter, croton oil (Gallagher  et al.,
    1984a). Subsequent work by Reeve  et al. (1985), however, failed to
    confirm these results, and Forbes  et al. (1989) found no evidence
    of tumour initiation by the compound in an initiation-promotion
    experiment in mice.

         Sunscreens have been strongly promoted by the medical community
    such as in the United States, Canada, Australia, and Scandinavian
    countries. However, incidence rates of melanoma have risen steeply
    in recent decades, even after the introduction of sunscreens (Lee
    1989, Jensen and Bolander 1980, Magnus 1977, 1986; Gallagher  et
     al., 1986). The use of high SPF sunscreens is recommended by WHO
    and by the International Union Against Cancer (UICC) as a supportive
    part of a sun avoidance programme rather than its main thrust (Marks
    & Hill, 1992).

         Both UVA and UVB have been shown to mutate DNA and cause skin
    cancers in animals (Staberg  et al., 1983). UVA penetrates deeper
    into the skin than UVB and because of the energy distribution of
    sunlight and filtering by the outermost layers of the skin,
    melanocytes receive up to 70 photons of UVA for every photon of UVB.
    Sunscreens effectively block solar UVB. UVB is the normal stimulus
    for accommodation of the skin such as thickening and increased
    pigmentation.

         Sunscreens suppress normal warnings of overexposure such as
    erythema and sunburn and allow excessive exposure to wavelengths of
    sunlight they do not block. Due to lack of these natural signs
    sunscreens create a false sense of security and individuals tend to
    stay in sun longer. In view of these behavioural changes which
    increases individuals UVA exposure it has been suggested that,
    because of the rising incidence of melanoma, UVA may be associated
    with its occurrence (Garland  et al., 1992). While a recent study
    (Setlow  et al., 1993) in fish reported melanoma induction by UVA,
    the role of UVA in the causation of human malignant melanoma has yet
    to be established.

         Trans-urocanic acid, a natural compound of the stratum corneum
    which absorbs UVB and is used as an additive in some commercial
    sunscreen products, increased the yield of solar-simulated UV
    induced tumours in hairless mice (Reeve  et al., 1989). The
    significance of this finding for human exposure has not been
    evaluated.

         Reports on sunscreen protection against UV induced
    immunosuppression have been equivocal. Fisher  et al. (1989)
    reported no protection of hairless mice against UV-induced systemic
    immunosuppression following application of Padimate 0 (a PABA ester
    and UVB absorber) or oxybenzone (a UVB/UVA absorber) with sun
    protection factors (SPF) of 6 or 15. Similarly SPF 15
    octy-N-dimethyl-p-aminobenzoate (o-PABA) had no effect on UV-induced
    systemic immunosuppression; nor was susceptibility to UV-induced
    tumours altered in hairless mice (Reeve  et al., 1991). In this
    study however, 2-ethylhexyl-p-methoxycinnamate (2-EHMC, SPF-15) was
    fully protective. Morison (1984) also reported protection of mice
    with PABA against both immunosuppression and susceptibility to UV-
    induced tumours. In recent studies (Wolf  et al., 1993a, b),
    sunscreens with SPFs of 4-6, including o-PABA, 2-EHMC, and
    benzophenone all provided protection of mice against UV- induced
    local and systemic immunosuppression, although they were less
    effective in preventing systemic effects. Finally, in a human study,
    a sunscreen containing 8% octyl dimethyl PABA, 2% 2-hydroxy-
    4methoxybenzophenone, and 2% methoxydibenzoyl methane (SPF 15) did
    not protect against effects of solarium exposure on NK activity,
    recall antigen skin tests and immunoglobulin production  in vitro
    in mitogen stimulated cultures (Hersey  et al., 1987).

         While further studies are still needed to clarify concerns
    raised about the ingredients and protectiveness of sunscreens, broad
    spectrum sunscreens which absorb both UVA and UVB with an SPF of at
    least 15 are still recommended as an effective means of personal
    protection against UV exposure.

    14.6  Tanning Devices

         Our modern lifestyle has suggested that having a tan is
    synonymous with good health. The increasing popularity of this
    symbol of health status and the inability to suntan during
    non-summer months, has led to the growth of an artificial tanning
    industry using sunlamps or sunbeds (combination of fluorescent
    lamp-shaped sunlamps into a bed). The dangers of excessive exposure
    to UV have been described earlier in this document. They range from
    mild erythema to severe burns of the skin from acute exposure, to
    skin cancer and skin ageing from long-term exposure. When eyes are
    exposed, damage can occur to the cornea, lens and retina, depending
    on the UV wavelength.

         The first generation of sunlamps emitted primarily UVB and, if
    used correctly, were efficient tanning agents. Unfortunately, they
    also tended to cause a painful 'sunburn' and other undesirable
    side-effects. UVA lamps are now used and these, it is claimed, tan
    safely without burning. However, as the carcinogenesis action
    spectrum extends into the longer UVA spectral region, exposure to
    these lamps is not without risk.

         Recommendations regarding the use of sunbeds

         Following a thorough review of this topic, the IRPA/INIRC
    (1991a) issued recommendations on the use of sunlamps or sunbeds for
    cosmetic purposes as follows:

          General

         The use of sunbeds for cosmetic purposes is not recommended.

          Specific

    (1)  People with skin types I and II should not use sunbeds. They
         are likely to be disappointed with the results of the
         exposures, they have higher susceptibility to sunburn and have
         a higher risk of developing skin cancer.

    (2)  Any person with a large number of nevi (moles), a tendency to
         freckle, a history of severe sunburn especially in childhood,
         or a family history of malignant melanoma should not use a
         sunbed.

    (3)  Any person taking a medicine that is known to be photoactive
         should not use a sunbed. If in doubt, they should seek the
         advice of a physician.

    (4)  Any person who already has extensive skin "sunlight" damage, or
         who has had premalignant or malignant skin lesions, should not
         use sunbeds.

    (5)  Any person who has a skin disease should seek the advice of a
         physician before using a sunbed.

    (6)  Children should not use sunbeds.

    (7)  Sunbeds should not be used if perfumes, body lotions or sprays
         have been applied that day.

    (8)  Because the sensitivities of individuals vary greatly, it is
         advisable to limit the duration of the first session to about
         one-half of a regular session in order to establish the user's
         skin response. If following the first session any adverse
         reaction occurs, further use of the sunbed is not recommended.

    (9)  Regular exposure should not exceed two sessions per week with a
         maximum of 30 sessions per year or 30 minimum erythemal doses
         (MEDs) per year, whichever is the smaller erythemally effective
         exposure. An occasional break from the regularity of exposure
         is advisable.

    (10) With respect to recommendation (9), the manufacturer of the
         sunbed should supply a schedule of exposure and recommended
         maximum exposure durations based on the emission
         characteristics of the sunbed.

    (11) Appropriate protective eyewear should be provided by the
         manufacturer and should always be worn when using a sunbed.

    (12) When the sunbed is being provided for use by a commercial
         operator, it is the responsibility of the operator to provide
         the person intending the use the sunbed with the appropriate
         information as summarized in recommendations (1) to (11) above.

    14.7  Occupational Protection

         Occupational exposure to UV should be kept to a minimum. Some
    UV sources emit a considerable amount of visible radiation, and in
    this case, the natural aversion response is evoked, so there is
    little chance of accidental over-exposure of the eyes. On the other
    hand, artificial sources emitting short wavelength UV radiation
    exist where accidental exposure is quite likely. While working
    outdoors, skin and eye protection should be used. Exposure outdoors
    during the periods of 2 hours either side of noon should be avoided.

         Where UV levels, compared with the ELs and erythemal dose, are
    such as to constitute a hazard, protection against hazardous
    exposure may be achieved by a combination of engineering control

    measures; administrative control measures and personal protection.
    For artificial sources, wherever possible, priority should be given
    to engineering and administrative controls to reduce the requirement
    for personal protection.

          Engineering controls

         The principal and most effective engineering control measures
    are those intended to contain the radiation. Wherever possible UV
    from artificial sources should be contained within a sealed housing.
    If observation windows are required they should be made of suitably
    absorbent materials such as certain grades of acrylic and window
    glass. Where the exposure process is required to take place external
    to the source housing a screened area should be set aside where it
    may be carried out; an example of this is arc welding where screens
    must be provided to prevent exposure of people not involved with the
    welding process. Any such screened area should be subject to
    administrative control measures and persons working in the area
    should be adequately protected from UV as described below. Where a
    source of UV is normally enclosed during use but to which access is
    required, for example for maintenance, the housing should be fitted
    with safety interlocks. If anyone needs to gain access to the source
    while it is energized, the interlock should immediately switch off
    the power to the lamps and should not be able to be re-energized
    until the interlocks are re-engaged Many such sources are used for a
    variety of drying and curing purposes, particularly in the printing
    industry and all should be subject to strict engineering controls.

          Administrative control measures

         The principal administrative control measures are those that
    limit access to the source and provide information directed at
    making people aware of potential hazards associated with it. Access
    to an area where equipment emits UV should be limited to those
    persons directly concerned with its use. All persons concerned with
    the use of such equipment should be made aware of this and should be
    informed of the potential hazards. Appropriate hazard warning signs
    should be used to indicate the presence of UV and whenever possible
    warning lights may be used to show that equipment is energized. The
    user of a UV emitting source should keep as far from the source as
    possible. At large distances (greater than twice the greatest
    dimensions of the source) the irradiance (W m-2) falls off as the
    square of the distance from the source. Closer to the source
    irradiance falls off approximately linearly with distance. Exposures
    should be kept to a minimum and the EL's recommended by IRPA/INIRC
    (1991) given in chapter 13 should not be exceeded. Particular care
    should be taken to prevent exposure of persons taking
    photosensitising medications or concomitantly exposed to
    photosensitisers in the environment.

          Personal protection

         The most effective way to protect the skin from UV is to cover
    it. In indoor occupational situations, the areas of the body most at
    risk are the face (and eyes) and neck, the forearms and the backs of
    the hands. The face can be protected by a shield and this should
    also provide eye protection. The arms should be covered by clothing
    with a low UVB transmission; in general materials that are visibly
    opaque are suitable (see section 14.4). Hands can be protected by
    wearing gloves. Face shields, goggles or safety spectacles which
    absorb UV should be worn where there is a potential eye hazard.

         A range of suitable protective eyewear is commercially
    available (see figure 14.2). Welders should be protected by a helmet
    fitted with appropriate absorption filters. Some high pressure lamps
    are potential explosion hazards and the eyes and face should be
    protected against flying fragments of glass. Particular care should
    be taken to protect the eyes, face and hands when such lamps are
    being removed or replaced. The risk to outdoor workers such as
    agricultural workers, labourers, construction workers, fishermen etc
    from solar radiation exposure can be minimized by wearing
    appropriate tightly woven clothing, and most importantly a brimmed
    hat to reduce face and neck exposure. Sunscreens can be applied to
    exposed skin to reduce exposure further (see section 14.5).

          Hazards from ozone

         The absorption of short wavelength UV by oxygen in the air
    forms ozone, a powerful oxidising agent. The American Conference of
    Governmental Hygienists has published a threshold limit value for
    ozone exposure, 0.1 ppm, (ACGIH, 1993), and concentrations above
    this value should be avoided.

         Levels of ozone may be reduced by providing adequate
    ventilation in the area in which the source is located. Very intense
    sources emitting short wavelength UVB, for example, high pressure
    linear and compact mercury and xenon lamps will normally require an
    extraction system to remove ozone.

    14.8  Protection in Medicine and Dentistry

         Protection of both patients and staff must be considered. When
    a patient is being exposed to UV for clinical purposes, sites not
    intended to be treated should be covered and the eyes protected.

         In medical care, chlorpromazine and thioridazine are
    phenothiazine derivatives which are widely used as psychosedatives.
    As well as producing photosensitivity as a side-effect in some
    patients, these drugs may induce a similar type of reaction in
    hospital staff. Contact dermatitis may occur in staff handling
    phenothiazines or fouled laundry since metabolites of chlorpromazine

    are phototoxic (Moseley 1988). To avoid contamination, staff should
    wear protective clothing.

         Problems may also occur in dentistry, both for patients and
    staff. The mouth is lined with a relatively thin squamous epithelium
    and so there may be considerable penetration of UV to underlying
    cells. Patients and staff with systemic lupus erythematous (SLE)
    have been reported to be at risk from lights used in dentistry
    (Moseley 1988). Both operating lamps and sources used to polymerise
    resins have been found to cause damage. In one case, a dentist, who
    was a known SLE sufferer, experienced a facial eruption after using
    a visible-light resin-curing source on a patient. Evidently, special
    care is required when lights are used on patients and a known
    history of SLE would be contraindicated for UV or visible light
    polymerisation.

    14.9  Nutrition

         Nutrition can provide the body with essential antioxidants and
    these molecules are distributed throughout the body. At the cellular
    level, they enter a number of endogenous photoprotective systems to
    control photochemical processes (Roberts  et al., 1991). Quenchers
    which can negate specific reactive intermediates may be important as
    a defence mechanism against UV insult to the eye. Glutathione, due
    to the low energy of the SH (thiol) bond (65 kcal) is an efficient
    free-radical scavenger and singlet oxygen quencher. Ascorbic acid
    quenches free radical and superoxide reactions. alpha-Tocopherol
    quenches both singlet oxygen and free radicals. There are also
    various antioxidant enzymes present in the eye. Exogenous scavengers
    and quenchers may be able to prevent UV damage by interrupting
    transient intermediates which cause ocular damage. An approach is to
    increase the known endogenous quenchers, (antioxidants) ascorbic
    acid, alpha-tocopherol and ß-carotene in the diet (Roberts  et al.,
    1991).

    14.10  Additional Protective Agents

         Although it has been known that glutathione is a particularly
    effective quencher of excited state transients in the eye, until now
    there has been no success in increasing this endogenous sulphydryl
    compound in ocular tissues. There have recently been found (Roberts
     et al., 1991) a promising group of compounds, phosphorylated
    sulphydryls, which pass the blood retinal and blood ocular barrier,
    and appear to mimic the protective effect of the endogenous thiol
    glutathione. This offers a possible way of protecting the eye and
    other tissues from UV induced damage.

         Protection from UV damage was reported in mice maintained on a
    diet supplemented with either carnosine (ß alanylhistidine), an
    antioxidant known to have immunopotentiating properties (Reeve  et
     al., 1993) or retinal palmitate in combination with canthaxanthin,

    a carotenoid (Gensler, 1989). Some protection was also afforded by
    retinal palmitate alone. Protection has been shown in healthy humans
    receiving a daily supplement of ß-carotene (Fuller  et al., 1992).

    14.11  Eye Protection

         In industry there are many sources capable of causing acute eye
    injury within a short exposure time, while in the natural
    environment acute injury is likely to occur mostly in situations
    where solar UV is reflected onto the eye, such as from snow while
    skiing. A variety of eye protection is available with various
    degrees of protection appropriate to their intended use (see figure
    14.2). Those intended for industrial use include, welding helmets
    (additionally providing protection from intense visible and infrared
    radiation and face protection), face shields, goggles and UV
    absorbing spectacles. For use in the outdoor environment, they
    include (ski) goggles for extreme exposure conditions and
    sunglasses.

         The appropriateness and selection of protective eyewear is
    dependent on the:

    (1)  intensity and spectral emission characteristics of the UV
         source,

    (2)  behavioural pattern of people near UV sources (distance and
         time are important),

    (3)  transmission properties of the protective eyewear material, and

    (4)  design of the frame of the eyewear to prevent exposure of the
         eye from direct unabsorbed UV.

         In industrial exposure situations the degree of ocular hazard
    can be assessed by measurement and comparison with recommended
    limits for exposure (IRPA/INIRC, 1991) (see Chapter 13). Welders and
    nearby workers should routinely wear appropriate eye and face
    protection. For protection against other less intense sources of UV
    in the workplace tightly fitting goggles or spectacles with side
    shields may be appropriate, but consideration should also be given
    to the need for additional face protection. In general, protective
    eyewear provided for industrial use should fit snugly to the face,
    thus ensuring that there are no gaps through which UV can directly
    reach the eye and should have adequate mechanical construction to
    prevent physical injury.

         For outdoor workers and the general public, the most hazardous
    source of UV exposure is the sun. Adequately designed (ski) goggles
    afford protection against exposure to solar UV at high altitudes and
    on snow, but for most other exposure conditions, good UV absorbing
    sunglasses are an adequate means of eye protection.

         Transmission of UV through sunglasses varies considerably
    (Wester 1987, Gies  et al., 1992), yet consumers are provided with
    little information about the protection afforded by them. Some
    countries have drafted standards limiting UV transmission through
    sunglasses and Gies  et al. (1990a) have proposed a UV eye
    protection factor (EPF) for sunglasses similar to the one developed
    for fabrics and sunscreens.

    FIGURE 14.2

    15.  FUTURE RESEARCH

    15.1  Introduction

         At the United Nations Conference on the Environment and
    Development (UNCED) in 1992 it was declared under Agenda 21 that
    there should be activities on the effects of UV. Specifically:

    (i)  Undertake as a matter of urgency, research on the effects on
         human health of increasing UV reaching the earth's surface as a
         consequence of depletion of the stratospheric ozone layer;

    (ii) On the basis of the outcome of this research, consider taking
         appropriate remedial measures to mitigate the above mentioned
         effects on human beings.

         This report gives a thorough review of the health hazards of UV
    exposure. However, although it is known that the burden of
    UV-related diseases on human populations is high, the exact nature
    and extent of these diseases is still largely unknown.

         There is great uncertainty about future trends in atmospheric
    ozone. For example the Antarctic holes and large depletions of ozone
    that have occurred recently were not predicted in any of the ozone
    depletion models. While agreements have been reached to reduce
    releases of CFCs into the environment, and this will have future
    benefit on the ozone layer, there is uncertainty about the extent of
    ozone depletion caused by chemical pollutants. What is apparent is
    that decreased ozone levels will persist for many years to come and
    the corresponding increases in UV intensities will result in more
    significant adverse health effects on all populations of the world
    for many decades to come (WMO 1993).

         The health effects of UV are not restricted to fair skinned
    populations. UV exposure is thought to cause diseases of the eye and
    suppression of the immune system in all populations of the world. UV
    induced immune suppression may have adverse consequences on
    infectious disease immunization programmes, particularly in areas
    where the UV intensities are high. The possibility that UV will
    cause progression of various diseases such as for HIV positive
    patients still has to be elucidated. Many such important issues need
    to be resolved as a matter of urgency.

         The WHO Task Group reviewing this monograph strongly supported
    action and coordination of UV research at the international level.
    In particular the Task Group supported the concept of the
    International Research Programme on Health, Solar UV Radiation and
    Environmental Change (INTERSUN). INTERSUN is a collaborative effort
    between WHO, the United Nations Environment Programme (UNEP) and the
    International Agency on Cancer Research (IARC). The group recognised
    specific research needs in areas of exposure assessment, terrestrial

    plants, aquatic ecosystems, and human health effects related to the
    skin, immune system and eye. Some of these could be accomplished
    under the umbrella of INTERSUN, and some, at least initially, would
    require more basic laboratory research to be undertaken.

    15.2  INTERSUN

         The objectives and approach of INTERSUN are: to accurately
    evaluate the quantitative relationship between solar UV at the
    surface of the earth and human health effects; develop reliable
    predictions of health consequences of changes in UV; provide
    baseline estimates of the occurrence of health effects of UV in
    representative populations around the world; and develop practical
    ways of monitoring change in these effects over time in relation to
    environmental and behavioural change.

         It is intended to establish field research centres covering a
    range of latitudes in both hemispheres in which will be measured:
    ground level UV irradiance; incidence of non-melanocytic skin
    cancer; incidence and mortality of malignant melanoma; immune
    function; biological markers of UV exposure and its' early
    carcinogenic effects; constitutional sensitivity to sun exposure;
    and present and past sun-related behaviour. In addition, in these
    centres there will be undertaken case-control or cross-sectional
    studies of relationships between skin cancers, cataract and immune
    response and constitutional sensitivity to the sun, lifetime
    exposure to the sun, and biological markers of UV exposure and early
    carcinogenic effects. Surveys should be repeated at 5-yearly
    intervals to allow trends in population measurements to be
    established.

         Data collected at such centres would be used

    -    to describe quantitatively the relationship between ground
         level solar UV irradiance and the incidence of skin cancers and
         other health effects (particularly eye damage and effects on
         the immune system) of UV exposure in human populations.

    -    to estimate the change in occurrence of health effects of UV
         radiation that would result from change in ground level solar
         UV irradiance due to environmental change.

    -    to increase understanding of the relationship between personal
         risk of health effects of UV radiation and constitutional
         sensitivity to the sun and sun-related behaviour.

    -    to develop and validate appropriate ways of monitoring human
         exposure to UV and the occurrence of associated health effects.

    -    to develop a network of centres monitoring trends in ground
         level solar UV irradiance, sun exposure of populations, and the
         occurrence of health effects of UV radiation.

    -    to interpret these trends, as far as is possible, in relation
         to environmental change, changes in human behaviour, and the
         implementation of public policies aimed at ameliorating
         environmental change or human exposure to solar UV radiation.

    -    to provide a basis for development and evaluation of
         interventions to reduce the occurrence of adverse health
         effects of solar UV radiation.

         In line with the above objectives and the gaps in knowledge
    needed to be filled to make a satisfactory health risk assessment,
    the following recommendations for future research are provided.

    15.3  Solar and Personal UV Monitoring

         The quantification of health detriment to an individual or to a
    population requires the assessment of exposure of the individual or
    the average exposure of a member of the exposed population. In
    interpreting epidemiological data the exposures of the groups
    comprising the study populations must be known. The requirements
    here are precise measurement data of the environment occupied by the
    exposed individuals combined with data on their exposure habits,
    derived from representative personal exposure measurements, skin
    damage or from questionnaires.

    15.3.1  Solar monitoring

    1.   A comprehensive network of spectral and broad band monitoring
         stations is needed worldwide giving quality ground measurements
         of surface UV, to assess the effects of enhanced UVA and UVB
         radiation on exposed populations, and to document the impact of
         stratospheric ozone depletion on ambient UV. A long-term data
         record is important.

    2.   Quality and convenient monitoring data are needed as input to
         models and for model validation. Current models need to be
         further modified to account for cloud cover, gaseous pollutants
         and aerosols.

    3.   Common calibration and audit procedures should be developed and
         implemented for national and international monitoring
         programmes.

    4.   Monitoring activities should be coordinated with the efforts of
         other groups on a national and international scale. Close
         coordination is needed for inter-comparison studies and quality
         control/quality assurance efforts.

    5.   There is a need for global monitoring of UV radiation, with
         particular attention to long-term instrument stability and
         representative geographical deployment

    15.3.2  Personal monitoring

         Population studies using personal UV monitoring devices are
    needed to determine the fraction of the daily natural UV dose
    received by persons at risk. The daily amount of UV received by
    human skin varies greatly with occupation, behaviour, and local
    climatic and environmental conditions. Little is known about these
    factors and this seriously interferes with the interpretation of
    existing data on the relationship between UV and the development of
    skin cancer and of chronic skin and eye damage. Thus, the
    development of personal UV monitoring devices is important.

         Small passive personal UV dosimeters or monitoring devices
    should be further developed and used to provide data on personal
    exposures. Biomarkers for UV exposure and skin cancer risk should be
    further developed, e.g. UV-induced mutations in p53 tumour
    suppressor genes. Such biomarkers might constitute a person's own
    "built in" UV risk monitoring device.

         For some people, artificial sources of UV represents a
    significant component of their exposure. Occupational exposures and
    elective exposures such as from cosmetic sunbeds are important in
    this respect. Measurement studies are required in these areas.

         Standardized monitoring of exposure to the sun in at risk
    populations, particularly children, should be evaluated as a guide
    to public health action to control melanoma.

    15.4  Terrestrial Plants

         Particular attention must be given to the impact on food
    production in the developing world and to development of crop
    varieties resistant to higher levels of UV radiation.

         Substantial research is needed to describe and evaluate the
    effects of enhanced UV on plants, specifically in the following
    areas:

    1.   Field validations on crop plants should be extended to
         determine whether UV affects yield in other agriculturally
         important plants.

    2.   Studies should be initiated to determine the impacts on
         ecosystems. Little is known of the effects of UV in other
         ecosystems such as forests, where 80% of net primary production
         is currently stored.

    3.   It is important to study natural plants because they serve in
         supplying new drugs, medicines and other natural products. They
         also act as a reservoir of genetic diversity for modern crop
         breeding programmes.

    4.   Studies are needed to elucidate how UVB effects are modified by
         additional factors such as carbon dioxide, temperature, water
         and nutrient stress, heavy metals, diseases, and pests.

    5.   Of increasing interest are man-made air pollutants such as
         sulphur dioxide and nitrogen oxides, which are known to have
         damaging effects on crops and forests. Attention should be
         focused on whether the negative effects of these air pollutants
         may be aggravated by UV radiation.

    6.   The manner and magnitude of adaptation to increased UV,
         especially UVB, such as by increased screening pigments or
         enhanced DNA damage repair capacity, need to be investigated
         both as adaption capacity of individual plants and as
         genetically based changes in plant populations. The latter also
         provides a basis for plant breeding interventions for
         agricultural and tree species.

    15.5  Aquatic Ecosystems

         Current data suggest that predicted increases in UV, especially
    UVB radiation could have important negative effects in the marine
    environment. However, uncertainties regarding the magnitude of these
    effects remain large, including problems of extrapolating laboratory
    findings to the open sea, and the almost complete absence of data on
    long-term effects on ecosystems. Additional information is needed in
    several areas before more reliable assessments of risk are possible.
    Research is needed to:

    1.   Determine accurate and appropriate biological action spectra
         for selected endpoints of key marine species.

    2.   Produce dose-response data on a greater variety of ecologically
         important primary producers than is now available, as well as
         data for key higher organisms within the food web.

    3.   Determine long-term effects for embryos or larvae exposed to
         UVB radiation. Is the survival of the adult population (or
         their offspring) affected?

    4.   Determine effects of enhanced UV on major ecosystems, including
         the Antarctic ecosystems.

    5.   Obtain data on the mechanisms of damage and ranges of possible
         adaptation or genetic selection in response to increased UVB
         radiation.

    6.   Relate UVB penetration into the water column and through ice
         both in laboratory and field research to determine effects on
         phytoplankton and zooplankton.

    7.   Develop biomarkers to monitor current levels of UV damage in
         phytoplankton and zooplankton.

    15.6  Human Health

    15.6.1  Skin

         To a substantial degree, answers are still required to the
    following questions regarding the relationship between sun exposure
    and both melanoma and non-melanocytic skin cancer. Research is
    needed to determine:

    1.   The quantitative relationship between radiant exposure and
         incidence of these cancers; and in particular the shape of the
         exposure-response relationship.

    2.   How the effect of sun exposure is modified by: age at exposure,
         time since last exposure and pattern of exposure, particularly
         whether a particular dose is received intermittently or
         more-or-less continuously.

    3.   Biomarkers of UV exposure or early effects in the skin, e.g.
         UV-induced mutations for UV exposure and skin cancer risk.

    4.   The importance of UVA in causing skin cancers, both relative to
         UVB and comparatively between melanoma and non-melanocytic
         cancer.

    5.   The importance of indirect (e.g. oxidative) as opposed to
         direct UV damage in causing these cancers.

    6.   The contribution of sun exposure to the aetiology of these
         cancers in dark-skinned populations and how, in general,
         cutaneous sensitivity modifies both quantitatively and
         qualitatively the relationship between sun exposure and skin
         cancer.

    7.   If the associations of red hair and freckling with skin cancer
         are explained by effects of skin pigmentation or sensitivity to
         the sun or mediated by other susceptibility mechanisms.

    8.   If pigmentary characteristics or cutaneous sun sensitivity
         explain most of the ethnic variation in skin cancer incidence
         or are there other ethnically related susceptibility
         characteristics.

    9.   Whether the relationships between sun exposure and BCC and SCC
         are the same.

    10.  The role of UV-induced mutations of critical genes in causing
         melanoma.

    11.  The role of sun exposure in causing cancer of the lip. An
         epidemiological study is required with adequate control of
         potential confounding with alcohol and tobacco use.

    12.  The relationship between use of sunlamps and sunbeds
         (artificial tanning) and both melanoma and non-melanocytic skin
         cancer. Adequate control of likely confounding with sun
         exposure is required.

    13.  The effects of the interaction between UVB and the rest of the
         solar spectrum in relation to DNA repair, malignant
         transformation, and skin tumour development.

    14.  Additional animal models, particularly for the study of the
         experimental induction of malignant melanoma. Further, an
         action spectrum for induction of melanoma should be developed
         in  M. domestica to see if the apparent response of UVA
         exposure observed in hybrid fish can be confirmed.

    15.6.2  Immune System

         Suppression of immune functions results from UV exposure of
    humans. The following studies are needed to adequately assess the
    consequent risk to human health.

    1.   Studies to determine whether systemic effects occur in humans
         as a result of UV exposure.

    2.   Establish immunological biomarkers that could identify persons
         who are susceptible to specific long-term UV-induced immune
         suppression, e.g. to infectious agents.

    3.   Human and rodent studies are needed to determine the
         relationship between UV exposure and susceptibility to
         infection and vaccine effectiveness, and quantitative
         comparison of the relative sensitivities of mice and humans.

    4.   Studies of the effects of UVB on immune function in human skin
         of all types to determine if all skin types show the same
         quality and quantity of effects.

    5.   Studies to examine whether currently available sunscreens
         provide adequate protection and whether dietary supplements are
         protective. Specifically:

         a)   Protective effects for immunosuppression;

         b)   Systemic toxicity based on penetration of organic
              chemicals in sunscreens;

         c)   Carcinogenicity assays.

    6.   Studies of the spectral effectiveness of UV in producing immune
         function defects and enhanced susceptibility to disease.

    7.   Research to determine the role of UVB exposure in
         immediate-type hypersensitivity (allergic) reactions as well as
         specific antibody responses, particularly as it relates to
         impact on incidence and severity of asthma.

    8.   Research on the effects of UV on the development of autoimmune
         disease and on underlying mechanisms.

    9.   Studies on systemic effects caused by release of cytokines
         within the epidermis and blood stream.

    15.6.3  Eye

         Two issues underscore the need for undertaking studies of
    health risks of UV exposure of the eye on a priority basis. First,
    there is evidence of the depletion of the ozone layer and the
    reported consequent increase in ambient UV which will impact on many
    eye diseases. Second, the overall ageing of the population worldwide
    would further compound the magnitude of the cataract and other eye
    problems.

    1.   There is a need for well-conducted epidemiological studies to
         be undertaken to define the strength of the association between
         UVA and UVB exposure with cataract and other eye conditions,
         including climatic droplet keratopathy and pterygium, at both
         the individual and community levels. Ideally, this should be
         undertaken in a variety of locations. Careful attention needs
         to be given to quantifying, for each study participant, ocular
         exposure to UV and exposure to other possible risk factors that
         may confound observed associations.

    2.   Techniques need to be refined for measuring the proportion of
         UV of different wavelengths incident upon the human cornea and
         lens, and how these are affected by the behavioural response to
         avoid direct exposure to bright sunlight, and by different
         ground surfaces.

    3.   Studies should be conducted to provide data on the quantitative
         relationship between exposure to UV and cataract and other
         lesions of the anterior eye. In relation to this there is an

         urgent need to establish an internationally agreed system for
         grading the type and severity of cataract.

    4.   Investigations of the role of personal solar exposure in
         causing ocular melanoma, including the protective effect of
         ocular devices.

    5.   The relationship between sunlamp or sunbed exposure and
         malignant melanoma of the eye needs to be evaluated.

    6.   The relative efficacy of different protective devices, headwear
         and eye wear needs to be determined.

    7.   Techniques need to be defined for measuring the proportion of
         UV of different wavelengths penetrating the human cornea and
         lens in the intact eye, in view of possible retinal effects.

    8.   The action spectra for chronic exposure leading to ocular
         damage needs to be determined in primates and other animals.
         This includes action spectra for cataract formation, pterygium,
         climatic droplet keratopathy, and macular changes in aphakic
         animals.

    9.   Risks to aphakic people and those with intraocular lens
         implants for retinal changes, mainly from UVA and short
         wavelengths in the visible part of the spectrum, need to be
         determined.

    15.7  Laboratory Studies

         Little is known about the mechanisms of interaction of UV and
    environmental chemical agents on biological systems. Many widely
    distributed natural or artificial chemicals (pesticides,
    halocarbons, etc.) can be altered by UV, resulting in photoproducts
    that may be less or more biologically effective than the parent
    compound. Furthermore, many chemicals can be activated by UV  in
     situ in biological systems and this activation may elicit a
    biological effect which neither the chemical nor the radiation alone
    exhibits (Psoralens). An international registry of agents that
    interact with light to cause adverse health effects would speed
    identification of such agents.

         Much of the information on the chemical and biological effects
    of UV comes from experiments using UVC (particularly 254 nm)
    radiation not found in sunlight reaching the earth's surface. There
    are studies showing direct and indirect effects on cells and
    cellular constituents of UVB, UVA, and visible light that differ
    considerably from those of UVC. Thus, the chemical and biological
    effects of the wavelengths of UV found in sunlight should be
    studied. There is evidence that visible light can, under different
    conditions, either help cells to repair UV-induced damage or can

    potentiate the detrimental effects of UV. Thus, to better understand
    the effects of sunlight on humans and the environment, experiments
    should be performed using natural sunlight or artificial lamps with
    well-known continuous spectra.

         Cellular and molecular studies needed include:

    1.   Extension of relevant biological action spectra to the
         UVA/visible wavelengths, as well as interactions between UVA
         and UVB radiation. Crucial spectra required include:

         -    pre-mutagenic lesions

         -    gene activation

         -    activation of viruses specially HIV (activation of DNA
              binding, promoter activity, viral replication)

    2.   Determination of the free radical/oxidative component of UVB
         and UVA radiation effects. Determine the reactive intermediates
         involved.

    3.   Determination of adaptation responses in human cells.

    4.   Determination of endogenous cellular antioxidant defences in
         eye tissue and skin.

    5.   Elucidation of DNA repair pathways in humans using the advanced
         technology available for analyses of gene and function.

    6.   Identification of key endogenous and exogenous sensitizers.

    7.   Studies to examine the effect of UV on the balance between Th1
         and Th2 cells.

    15.8  Education

         It is essential to educate the general population and workers
    concerning the profound importance of sunlight and the possibilities
    of either UV deprivation or of acute and chronic UV injury. It is
    also important to overcome the lack of respect for the possible
    adverse effects on health from overexposure to sunlight, simply
    because sunlight is ubiquitous, and the concept that, if something
    is natural, it must be totally beneficial and safe. In this context
    there is a need to standardize a UV exposure index that can be used
    as part of health information/education campaigns.

    15.9  Administration

         An international panel composed of key national experts, with
    its secretariat at WHO, should be formed with the following terms of
    reference:

    (i)   to identify gaps in knowledge and set research priorities
          needed for a better health risk assessment of exposure to UV

    (ii)  to evaluate and monitor progress of research

    (iii) to establish mechanisms for funding research and initiate
          projects where crucial gaps in knowledge have been identified

    (iii) to facilitate international cooperation on UV monitoring
          efforts, including instrumentation intercomparisons,
          calibration standardization and uniform analysis of data, and

    (iv)  to facilitate international cooperation in measurement of
          human UV exposure, monitoring of trends in occurrence of UV
          health effects, and epidemiological studies of the
          relationship between UV exposure and health effects

    (v)   to provide a database of research projects to facilitate
          cooperation between researchers and institutions involved in
          research.

    16.  PREVIOUS EVALUATIONS BY INTERNATIONAL BODIES

    16.1  United Nations Environment Programme

         UNEP has undertaken a number of reviews on the environmental
    effects of increased UV intensities resulting from ozone depletion
    (1989, 1991, 1992). Their findings on each issue are summarized
    below:

    16.1.1  Ozone

         Significant global scale decreases in total ozone have occurred
    over the past ten years. All other factors being constant, there is
    no scientific doubt that decreases in total ozone will increase UVB
    radiation at ground level. Tropospheric ozone and aerosols may have
    masked the consequences of stratospheric ozone depletion for UVB in
    some industrialized regions. There are no reliable estimates of the
    direction or magnitude of effects of any cloud cover trends on UVB.
    Efforts to improve local and regional air quality may bring to light
    the increases in UVB associated with the depletion of stratospheric
    ozone.

    16.1.2  Human health

         The induction of immunosuppression by UVB has now been
    demonstrated in humans, not only those of light pigmentation, but
    also deeply pigmented individuals. This places all of the world's
    populations at risk of the potential adverse impacts of UVB on the
    immune system, including possible increases in the incidence or
    severity of infectious disease.

         An increased number of adverse ocular effects have been
    associated with exposure to UV. These include age-related
    nearsightedness, deformation of the lens capsule, and nuclear
    cataract (a form of cataract which previous information excluded
    from consideration). These effects appear to be independent of
    pigmentation. Estimates of risk would increase slightly if one were
    to include nuclear cataract among the forms of cataract increasing
    with ozone depletion. It is now predicted that, all other things
    being equal, a sustained 10% decrease in ozone will be associated
    with between 1.6 and 1.75 million additional cases of cataract per
    year world-wide.

         Recent information on the relationship of nonmelanoma skin
    cancer to UV exposures confirms previous findings and has allowed
    refinement of the carcinogenic action spectrum. Incorporation of
    this new information into the risk estimation process has led to
    slightly lower predictions. It is now predicted that a sustained 10%
    decrease in ozone will be associated with 26% increase in
    non-melanoma skin cancer. All other things remaining constant, this

    would mean an increase in excess of 300,000 cases per year
    world-wide.

    16.2  International Agency for Research on Cancer

         The carcinogenicity of ultraviolet has been evaluated by IARC
    (1992, 1993). IARC concluded that "There is sufficient evidence in
    humans for the carcinogenicity of solar radiation. Solar radiation
    causes cutaneous melanoma and nonmelanocytic skin cancer". With
    respect to other potential sources of UV, IARC has concluded that
    there is limited evidence for the carcinogenicity of exposure to
    fluorescent lighting and there is inadequate evidence for the
    carcinogenicity of other artificial sources of UV. Readers are
    referred to these excellent monographs for more details -much of
    which has been incorporated in this text.

    16.3  World Health Organization

         WHO and its Regional Office for Europe have completed reviews
    on the health effects of ultraviolet (WHO, 1979, 1989). They
    concluded that all people are exposed to UV from sunlight, and the
    risk to health varies with geographical, genetic and other factors.
    Similar risks are involved in the increasing exposure of people to
    UV from artificial sources, such as those used for suntanning, in
    phototherapy and in industrial processes. The biological effects of
    a single exposure differ significantly from the effects of repeated
    and cumulative exposures. Both types of risk increase markedly with
    excessive exposure.

         Most observed biological effects of UVB are extremely
    detrimental to living organisms. Much less is known about the
    biological effects of UVA. It can augment the biological effects of
    UVB, and doses of UVA, which alone do not show any biological
    effects, can in the presence of certain chemical agents, result in
    injury to tissues (phototoxicity, photoallergy, enhancement of
    photocarcinogenesis).

         All those who work out of doors are potentially at risk of
    overexposure, the consequences of which may be both acute and
    long-term effects. The fashion of exposing a large part of the body
    to sunlight has during recent years increased the exposure of the
    skin, resulting in quite high UV doses. This is true not only for
    outdoor work but is now also normal during leisure periods, as
    exemplified by the holiday exodus of a large part of the population
    of the northern European countries to the Mediterranean coast.

    16.4  International Commission on Non-Ionizing Radiation Protection

         Through its predecessor committee IRPA/INIRC the ICNIRP was
    involved in the publication of the original review of the health
    effects of UV (WHO/UNEP/IRPA, 1979). It has recommended

    international guidelines on limits (IRPA/INIRC, 1991) as shown in
    Chapter 13. It has also recommended against the use of sunlamps for
    cosmetic purposes (IRPA/INIRC, 1991a). A review of the studies on
    fluorescent lighting with reference to UV levels has suggested that
    they do not seem to be associated with an increased risk of
    malignant melanoma (IRPA/INIRC, 1991).


    REFERENCES

    Abrahams PJ, Huitema BA, & van der Eb AJ (1984) Enhanced
    reactivation and enhanced mutagenesis of herpes simplex virus in
    normal human and xeroderma pigmentosum cells. Mol Cell Biol, 4:
    2341-2346.

    ACGIH (1993) Threshold limit values for chemical substances and
    physical agents and biological exposure indices. Cincinnati, Ohio,
    The American Conference of Governmental Industrial Hygienists.

    Adam SA, Sheaves JK, Wright NH, Mosser G, Harris RW, & Vessey MP
    (1981) A case-control study of the possible association between oral
    contraceptives and malignant melanoma. Br J Cancer, 44: 45-50.

    Adams JS, Clemens TL, Parrish JA, (1982) Vitamin D synthesis and
    metabolism after UV irradiation of normal and vitamin D deficient
    subjects. N Engl J Med, 306: 722-725.

    Ainsleigh HG (1993) Beneficial effects of sun exposure on cancer
    mortality. Prev Med, 22: 132-140.

    Akslen LA & Morkve O (1992) Expression of p53 protein in cutaneous
    melanoma. Int J Cancer, 52: 13-16.

    Albino AP, Nanus DM, Davis ML, & McNutt NS (1991) Lack of evidence
    of Ki- ras codon 12 mutations in melanocytic lesions. J Cutan
    Pathol, 18: 273-278.

    Alcalay J, Freeman SE, Goldberg LH, & Wolf JE (1990) Excision repair
    of pyrimidine dimers induced by simulated solar radiation in the
    skin of patients with basal cell carcinoma. J Invest Dermatol, 95:
    506-509.

    Anderson RR & Parrish JA (1981) The optics of human skin. J Invest
    Dermatol, 77: 13-19.

    Andley UP (1987) Photodamage to the eye. Yearly review. Photochem
    Photobiol, 46: 1057-1066.

    Angel P, Poting A, Mallick U, Rahmsdorf HJ, Schorpp M, & Herrlich P
    (1986) Induction of metallothionein and other mRNA species by
    carcinogens and tumour promoters in primary human skin fibroblasts.
    Mol Cell Biol, 6: 1760-1766.

    Ansel JC, Mountz J, Steinberg AD, DeFabo E, & Green Ira (1985)
    Effects if UV radiation on autoimmune strains of mice: Increased
    mortality and accelerated autoimmunity in BXSB male mice. J Invest
    Dermatol, 85: 181-186.

    Araneo BA, Dowell T, Moon HB, & Daynes RA (1989) Regulation of
    murine lymphokine production  in vivo. Ultraviolet radiation
    exposure depresses Il-2 and enhances IL-4 production by T cells
    through an Il-1 dependent mechanism. J Immunol, 143: 1737-1744.

    Arlett CF & Harcourt SA (1983) Variation in response to mutagens
    amongst normal and repair-defective human cells. In: Lawrence CW ed.
    Induced mutagenesis: Molecular mechanisms and their implications for
    environmental protection. New York, London, Plenum Press, pp
    249-267.

    Arlett CF, Harcourt SA, Cole J, Green MH, & Anstey AV (1992) A
    comparison of the response of unstimulated and stimulated
    T-lymphocytes and fibroblasts from normal, Xeroderma pigmentosum and
    trichothiodystrophy donors to the lethal action of UVC. Mutat Res,
    273: 127-135.

    Armstrong BK (1984) Melanoma of the skin. Br Med Bull, 40: 346-350.

    Armstrong BK (1988) Epidemiology of malignant melanoma: intermittent
    or total accumulated exposure to the sun? J Dermatol Surg Oncol, 14:
    835-849.

    Armstrong BK (1993) Implications of increased solar UVB for cancer
    incidence. In: Chanin ML ed. The role of the stratosphere in global
    change. Berlin, Heidelberg, New York, Springer-Verlag, pp 517-540.

    Armstrong BK & English DR (1988) The epidemiology of acquired
    melanocytic naevi and their relationship to malignant melanoma.
    Pigment Cell, 9: 27-47.

    Armstrong BK, Woodings TL, Stenhouse NS, & McCall MG (1983)
    Mortality from cancer in migrants to Australia 1962 to 1971. Perth,
    University of Western Australia.

    Armstrong BK, de Klerk NH, & Holman CDJ (1986) Etiology of common
    acquired melanocytic nevi: constitutional variables, sun exposure
    and diet. J Natl Cancer Inst, 77: 329-335.

    Atkin M, Fenning J, Heady JA, Kennaway EL, & Kennaway NM (1949) The
    mortality from cancer of the skin and lip in certain occupations. Br
    J Cancer, 3: 1-15.

    Aubry F & MacGibbon B (1985) Risk factors of squamous cell carcinoma
    of the skin. A case-control study in the Montreal region. Cancer,
    55: 907-911.

    Auerbach H (1961) Geographic variation in incidence of skin cancer
    in the United States. Public Health Rep, 76: 345-348.

    Augustsson A, Stierner U, Rosdahl I, & Suurküla M (1990) Melanocytic
    naevi in sun-exposed and protected skin in melanoma patients and
    controls. Acta Dermato-Venereol, 71: 512-517.

    Azizi E, Lusky A, Kushelevsky AP, & Schewach-Millet N (1988) Skin
    type, hair colour and freckles are predictors of decreased minimal
    erythema ultraviolet radiation dose. J Am Acad Dermatol, 19: 32-38.

    Azizova AO, Islomov AI, Roshchupkin DI, Predvoditelev DA, Remizov
    AN, & Vladimir YA (1980) Free radicals formed on ultraviolet
    irradiation of the lipids of biological membrane. Biophysics, 24:
    407-414.

    Baadsgaard O, Wulf HC, Wantzin GL, & Cooper KD (1987) UVB and UVC,
    but not UVA, potently induce the appearance of T6-DR+
    antigen-presenting cells in human epidermis. J Invest Dermatol, 89:
    113-118.

    Baadsgaard O, Lisby S, Wulf HC, Wantzin GL, & Cooper KD (1989) Rapid
    recovery of Langerhans cell alloreactivity, without induction of
    autoreactivity, after  in vivo ultraviolet A, but not ultraviolet B
    exposure of human skin. J Immunol, 142: 4213-4218.

    Baadsgaard O, Salvo B, Mannie A, Dass B, Fox DA, & Cooper KD (1990)
     in vivo ultraviolet-exposed human epidermal cells activate T
    suppressor cell pathways that involve CD4+CD45RA+ suppressor-inducer
    T cells. J Immunol, 145: 2854-2861.

    Baasanhu J, Johnson GJ, Burendei G, & Minassian DC (in press)
    Prevalence and causes of blindness and visual impairment in
    Mongolia: A survey of populations aged 40 and older. Bull. World
    Health Organ.

    Bachem A (1956) Ophthalmic ultraviolet action spectra. Am J
    Ophthalmol, 41: 969-975.

    Baird EA, McHenry PM, & MacKie RM (1992) Effect of maintenance
    chemotherapy in childhood on numbers of melanocytic naevi. Br Med J,
    305: 799-801.

    Baker KS & Smith RC (1982) Spectral irradiance penetration in
    natural waters. In: Calkins J ed. The role of solar ultraviolet
    radiation in marine ecosystems. New York, London, Plenum Press, pp
    223-246.

    Barbareschi M, Girlando S, Cristofolini P, Cristofolini M, Togni R,
    & Boi S (1992) p53 Protein expression in basal cell carcinoma.
    Histopathology, 21: 579-581.

    Barker & Brainard (1993) Personal communication.

    Barker JNWN & MacDonald DM (1988) Eruptive dysplastic naevi
    following renal transplantation. Clin Exp Dermatol, 13: 123-125.

    Barnes PW, Jordan PW, Gold WG, Flint SD, & Caldwell MM (1988)
    Competition, morphology and canopy structure in wheat (Triticum
    aestivum L.) and wild oat (Avena fatua L.) exposed to enhanced
    ultraviolet-B radiation. Funct Ecol, 2: 319-330.

    Barnes PW, Flint SD, & Caldwell MM (1990) Morphological responses of
    crop and weed species of different growth forms to ultraviolet-B
    radiation. Am J Bot, 77: 1354-1360.

    Barrett SF, Robbins JH, Tarone RE, & Kraemer KH (1991) Evidence for
    defective repair of cyclobutane pyrimidine dimers with normal repair
    and other DNA photoproducts in a transcriptionally active gene
    transfected into Cockayne syndrome cells. Mutat Res, 255: 281-291.

    Basu-Modak S & Tyrrell RM (1993) Singlet oxygen. A primary effector
    in the UVA/near visible light induction of the human heme oxygenase
    gene. Cancer Res. 53: 4505-4510.

    Bech-Thomsen N, Wulf HC, Poulsen T, & Lundgren K (1988a)
    Pretreatment with long-wave ultraviolet light inhibits
    ultraviolet-induced skin tumor development in hairless mice. Arch
    Dermatol, 124: 1215-1218.

    Bech-Thomsen N, Wulf HC, & Lundgren K (1988b) Pre-treatment with UVA
    influences broad-spectrum UV photocarcinogenesis in hairless mice
    (Abstract). In: Rikles E ed. Proceedings of the 10th International
    Congress on Photobiology, Jerusalem. Jerusalem, Israel,
    International Association of Biology, p 34.

    Beer JZ & Smudzka BZ (1991) Activation of human immunodeficiency
    virus by radiation. In: Seymour CB & Murhersill C ed. New
    developments in fundamental and applied radiology. London, Taylor
    and Francis, pp 113-123.

    Beitner H, Norell SE, Ringborg U, Wennersten G, & Mattson B (1990)
    Malignant melanoma: aetiological importance of individual
    pigmentation and sun exposure. Br J Dermatol, 122: 43-51.

    Beral V & Robinson N (1981) The relationship of malignant melanoma,
    basal and squamous skin cancers to indoor and outdoor work. Br J
    Cancer, 44: 886-891.

    Beral V, Evans S, Shaw H, & Milton G (1982) Malignant melanoma and
    exposure to fluorescent lighting at work. Lancet, ii: 290-293.

    Berger DS & Urbach F (1982) A climatology of sunburning ultraviolet
    radiation. Photochem Photobiol, 35: 187-192.

    Bergmanson JP, Pitts DG, & Chu LW (1988) Protection from harmful UV
    radiation by contact lens. J Am Optom Assoc, 59:,178-182.

    Bergmanson JP, Pitts DG, & Chu LW (1987) The efficacy of a
    UV-blocking soft contact lens in protecting cornea against UV
    radiation. Acta Opthtalmol, 65: 279-286.

    Beukers R & Berends W (1960) Isolation and identification of the
    irradiation product of thymine. Biochim Biophys Acta, 41: 550-551.

    Bhatnagar R, West KP, Vitale S, Sommer A, Joshi S, & Venkataswamy G
    (1991) Risk of cataract and history of severe diarrhoeal disease in
    Southern India. Arch Ophthalmol, 109: 696-699.

    Bhuyan, K.C. and Bhuyan, D.K. (1983) Molecular mechanism of
    caractogenesis I reactive species of oxygen as triggering agents II.
    Evidence of lipid peroxidation and membrane damage. In Oxy Radicals,
    their scavenger syst. Proc. Int. Conf. Superoxide Dismutase 3rd,
    vol. 2 (Ed. G. Cohen and R.A. Greenwald), pp 343-356, Elsevier, New
    York.

    Bidigare RR (1989) Potential effects of UVB radiation on marine
    organisms of the Southern Ocean: distributions of phytoplankton and
    krill during Austral spring. Photochem Photobiol, 50: 469-477.

    Birch-Hirschfeld A (1914) The pathological effect of radiant energy
    on the eye. 1. Blinding by sunlight. Ergebn Allg Path des Menschen
    und der Tiere 16, 619-634.

    Bishop L (1992) Statistical considerations in network design and
    data analysis. Presentation to UVB Workshop: A review of the science
    and status of measuring and monitoring programs. Washington, DC,
    Science and Policy Associates, Inc.

    Bissett DL, Hannon DP, & Orr TV (1989) Wavelength dependence of
    histological, physical and visible changes in chronically
    UV-irradiated hairless mouse skin. Photochem Photobiol, 50: 763-769.

    Bittersmann E, Holzwarth AR, Agel G, & Nultsch W (1988) Picosecond
    time-resolved emission spectra of photoinhibited and photobleached
    Anabaena variabilis. Photochem Photobiol, 47: 101-105.

    Black HS (1987a) Photocarcinogenesis and diet. Fed Proc, 46:
    1901-1905.

    Black HS (1987b) Potential involvement of free radical reactions in
    ultraviolet light mediated cutaneous damage. Photochem Photobiol,
    46: 213-221.

    Black HS & Chan JT (1975) Suppression of ultraviolet light-induced
    tumor formation by dietary antioxidants. J Invest Dermatol, 65:
    412-412.

    Black HS, Thornby JI, Gerguis J, & Lenger, W (1992) Influence of
    dietary omega-6,-3 fatty acid sources on the initiation and
    promotion stages of photocarcinognesis. Photochem Photobiol, 56:
    195-199.

    Blum, HF (1948) Sunlight as a causal factor in cancer of the skin of
    man. J Natl Inst Cancer, 9: 247-258.

    Blum HF, Butler EG, Dailey TH, Daube JR, Mawe RC, & Soffen GA (1959)
    Irradiation of mouse skin with single doses of ultraviolet light. J
    Natl Cancer Inst, 22: 979-993.

    Blumthaler M, Ambach W, & Daxecker F (1987) On the threshold radiant
    exposure for keratitis solaris. Invest Opthalmol Vis Sci, 28:
    1713-1716.

    Blyth WA, Hill TJ, Field HJ, & Harbour DA (1976) Reactivation of
    herpes simples virus infection by ultraviolet light and possible
    involvement of prostaglandins. J Gen Virol, 33: 547-550.

    Bochow TW, West SK, Azar A, Munoz B, Sommer A, & Taylor HR (1989)
    Ultraviolet light exposure and risk of posterior subcapsular
    cataracts. Arch Ophthalmol, 107: 369-372.

    Boettner EA & Wolter JR (1962) Transmission of the ocular media.
    Invest Ophthalmol Vis Sci, 1: 776-783.

    Boiteux S, O'Connor TR, & Laval J (1987) Formamido pyrimidine-DNA
    glycosylase of Escherichia coli: cloning and sequencing of the fpg
    structural gene and overproduction of the protein. EMBO J, 6:
    3177-3183.

    Booth F (1985) Heredity in one hundred patients admitted for
    excision of pterygia. Aust N Z J Ophthalmol, 13: 59-61.

    Bornman JF & Vogelmann TC (1991) Effect of UV-B radiation on leaf
    optical properties measured with fibre optics. J Exp Vot, 42:
    547-554.

    Bose B, Agarwal S & Chafterjee SN (1989) UV-A induced lipid
    peroxidation in liposomal membrane. Radiat. Environ. Biophys. 28
    59-65.

    Bose SN & Davies RJH (1984) The photoreactivity of T-A sequences in
    oligodeoxyribonucleotides and DNA. Nucl Acids Res, 12: 7903-7913.

    Bouwes-Bavinck JN (1992) Risk of skin cancer in renal-transplant
    recipients. Leiden, The Netherlands (Thesis).

    Boyle J, Mackie RM, Briggs JD, Junor BJR, & Aitchison TC (1984)
    Cancer, warts, and sunshine in renal transplant patients: A
    case-control study. Lancet, 1: 702-705.

    Bradley MO, Hsu IC, & Harris CC (1979) Relationships between sister
    chromatid exchange and mutagenicity, toxicity and DNA damage. Nature
    (Lond), 282: 318-320.

    Brash DE & Haseltine WA (1982) UV-induced hotspots occur at DNA
    damage hotspots. Nature (Lond), 298: 189-192.

    Brash DE, Rudolph JA, Simon JA, Lin A, McKenna GJ, Baden HP,
    Halperin AJ, & Pontén J (1991) A role for sunlight in skin cancer:
    UV-induced p53 mutations in squamous cell carcinoma. Proc Natl Acad
    Sci (USA), 88: 10124-10128.

    Braun J (1991) The protective function of phenolic compounds of rye-
    and oat seedlings against UV-B radiation and their biosynthetic
    regulation, (Thesis) pp. 1-237 in Karlsr Beitr Entw Okophysiol 9, M.
    Tevini (ed.), Bot Inst II, Karlsruhe.

    Braun J & Tevini M (1991) Regulation of UV-Protective pigment
    synthesis in the epidermal layer of rye seedlings ( Secale cereale
     L. cv. Kustro). Photochem Photobiol, 37: 318-321.

    Bridges BA (1990) Sunlight, DNA damage and skin cancer: a new
    perspective. Jpn J Cancer Res, 81: 105-107.

    Brilliant LB, Grasset NC, Pokhrel RP, Kolstad A, Lepkowski JM,
    Brilliant GE, Hawks WN, & Pararajasegaram R (1983) Associations
    among cataract prevalence, sunlight hours, and altitude in the
    Himalayas. Am J Epidemiol, 118: 250-264.

    Brodkin RH, Kopf AW, & Andrade R (1969) Basal-cell epithelioma and
    elastosis: a comparison of distribution. In: Urbach F ed. Biologic
    effects of ultraviolet radiation. Oxford, New York, Pergamon Press,
    pp 581-618.

    Brühl C & Crutzen PJ (1989) On the disproportionate role of
    tropospheric ozone as a filter against solar UV radiation. Geophys
    Res Lett, 16: 703-706.

    Bruls WAG, Slaper H, Van der Leun JC, & Berrens L (1984)
    Transmission of human epidermis and stratum corneum as a function of
    thickness in the ultraviolet and visible wavelengths. Photochem
    Photobiol, 40(4): 485-494.

    Bhuyan KC & Bhuyan DK (1983) Molecular mechanism of
    cataractogenesis, I. Reactive species of oxygen as triggering
    agents, II. Evidence of lipid peroxidation and membrane damage. In
    (ed. Cohen G & Greenwald RA) Oxy Radicals, their scavenger syst.
    Proc. Int. Conf. Superoxide Dismutase 3rd, vol 2, Elsevier, New
    York, pp 343-356.

    Cadet J, Ravanat JC, Buchko GW, Yeo HC, & Ames BN (in press) Singlet
    oxygen DNA damage: chromatographic and mass spectrometry analysis of
    damage products. Methods Enzymol.

    Caldwell MM (1976) The effect of solar UV-B radiation (280-315 nm)
    on higher plants: implications of stratosphere ozone reduction. In:
    Castellani A ed. Research in photobiology. New York, London, Plenum
    Press, pp 597-607.

    Caldwell MM, Termura AH, & Tevini M (1989) The changing solar
    ultraviolet climate and the ecological consequences for higher
    plants. Trends Ecol Evol, 4: 363-366.

    Cameron ME (1965) Pterygium throughout the world. Springfield,
    Illinois, Charles C. Thomas.

    Campbell CC, Quinn AG, & Rees JL (1993) Codon 12 Harvey-ras
    mutations are rare events in non-melanoma human skin cancer. Br J
    Dermatol, 128: 111-114.

    Carretto JJ, Carignana MO, Daleo G, & de Marco SG (1990) Occurrence
    of mycosporine-like amino acids in the red tide dinoflagellate
    Alexandrium excavarum, UV-photoprotective compounds. J Plankton Res,
    12: 909-921.

    Cartwright LE & Walter JF (1983) Psoralen-containing sunscreen is
    tumorigenic in hairless mice. J Am Acad Dermatol, 8: 830-836.

    Cerutti PA & Netrawali M (1979) Formation and repair of DNA damage
    induced by indirect action of ultraviolet light in normal and
    xeroderma pigmentosum skin fibroblasts. Radiat Res, Suppl: 423-432.

    Cervenka J, Witkop CJ, Okoro AN, & King RA (1979) Chromosome breaks
    and sister chromatid exchanges in albinos in Nigeria. Clin Genet,
    15: 17-21.

    Césarini JP & Muel B (1989) Erythema induced by quartz- halogen
    sources. Photodermatology, 6: 222-227.

    Challoner AVJ, Corless D, & Davis A (1976) Personnel monitoring of
    exposure to ultraviolet radiation. Clin Exp Dermatol, 1: 175-179.

    Chamberlain J & Moss SH (1987) Lipid peroxidation and other membrane
    damage produced in Escherichia coli K1060 by near-UV radiation and
    deuterium oxide. Photochem Photobiol, 45: 625-630.

    Chamberlin GJ & Chamberlin DG (1980) Colour: It's measurement,
    computation and application. London, Heyden & Son, p 46.

    Chan GL, Peak MJ, Peak JG, & Haseltine WA (1986) Action spectrum for
    the formation of endonuclease-sensitive sites and (6-4)
    photoproducts induced in a DNA fragment by ultraviolet radiation.
    Int J Radiat Biol, 50: 641-648.

    Chatterjee A, Milton RC, & Thyle S (1982) Prevalence and aetiology
    of cataract in Punjab. Br J Ophthalmol, 66: 35-42.

    Christman MF, Morgan RW, Jacobson FS, & Ames BN (1985) Positive
    control of a regulon for defences against oxidative stress and some
    heat shock proteins in Salmonella typhimurium. Cell, 41: 753-762.

    Chu G & Chang E (1988) Xeroderma pigmentosum group E cells lack a
    nuclear factor that binds to damaged DNA. Science, 242: 564-567.

    CIE (1987) International lighting vocabulary published jointly by
    the International Electrotechnical Commission (IEC) and the
    International Commission on Illumination (CIE). Geneva,
    International Electrotechnical Commission (ISBN 2-8273-0006-0).

    CIE (1990) Report on photosensitising chemicals - Updated in 1993.
    Vienna, International Commission on Illumination.

    CIE (1991) Sunscreen testing (UVB). Vienna, Commission on
    Illumination (CIE Technical report - Publication No. CIE 90).

    Cleaver JE (1973) Xeroderma pigmentosum - progress and regress. J
    Invest Dermatol, 60: 374-380.

    Clemmesen J (1965) Statistical studies in the aetiology of malignant
    neoplasms. I. Review and results. Copenhagen, Munksgaard, p 409.

    Coffman RL, Seymour WP, Lebman DA, Hirakkki DD, Christiansen JA,
    Shrader B, Cherwinski HM, Savelkoul HF, Finkelman FD, Bond MW, &
    Mosmann TR (1988) The role of helper T cell products in mouse B cell
    differentiation and isotype regulation. Immunol Rev, 102: 5-28.

    Cogan DG & Kinsey VE (1946) Action spectrum for keratitis produced
    by ultraviolet radiation. Arch Ophthalmol, 35: 670-677.

    Cole CA, Forbes PD, & Davies RE (1986) An action spectrum for UV
    photocarcinogenesis. Photobiology, 43: 275.

    Colin J, Bonissent JF, & Resnikoff S (1985) Epidemiology of the
    exfoliation syndrome. Proceedings of the 17th Congress of the
    European Society of Ophthalmolology, Helsinki, pp 230-231.

    Collmann GW, Shore DL, Shy CM, Checkoway H, & Luria AS (1988)
    Sunlight and other risk factors for cataract: an epidemiological
    study. Am J Public Health, 78: 1459-1462.

    Colston K, Berger U, & Coombes RC (1989) Possible role for vitamin D
    in controlling breast cancer cell proliferation. Lancet, 1: 188-191.

    Connor MJ & Wheeler LA (1987) Depletion off cutaneous glutathione by
    ultraviolet radiation. Photochem Photobiol, 46: 239-245.

    Cooke KR & Fraser J (1985) Migration and death from malignant
    melanoma. Int J Cancer, 36: 175-178.

    Coombs BD, Sharples KJ, Cooke KR, Skegg DCG, & Elwood JM (1992)
    Variation and covariates of the number of benign nevi in
    adolescents. Am J Epidemiol, 136: 344-355.

    Cooper KD (1993) Human studies in photo immunology. Photochem
    Photobiol, 57: 745.

    Cooper KD, Neises GR, & Katz SI (1986) Antigen-presenting OKM5+
    melanophages appear in human epidermis after ultraviolet radiation.
    J Invest Dermatol, 861: 363-370.

    Cooper KD, Oberhelman BS, Hamilton MS, Baadsgaard O, Terhune M,
    LeVee G, Anderson T, & Koren H (1992) UV exposure reduces
    immunization rates and promotes tolerance to epicutaneous antigens
    in humans; relationship to dose, CD1a-DR+ epidermal macrophage
    induction and Langerhans cell depletion. Proc Natl Acad Sci (USA),
    89: 8497-8501.

    Corominas M, Kamino H, Leon J & Pellicer A (1991) Oncogene
    activitation in human benign tumors of the skin (keratoacanthomas):
    Is H-ras involved in differentiation as well as proliferation? Proc
    Natl Acad Sci (USA), 86: 6372-6376.

    Coronel VP, Dai QJ, Vergara BS, & Teramura AH (1990) Preliminary
    study on response of rice seedlings to enhanced UV-B radiation. Int
    Rice Res Newsl, 15: 37.

    Coroneo MT (1993) Pterygium as an early indicator of ultraviolet
    insolation: a hypothesis. Br J Ophthalmol, 77: 734-739.

    Correll DL, Clark CO, Goldberg B, Goodrich VR, Hayes DR, Klein WH, &
    Schecher WD (1992) Spectral ultraviolet-B radiation fluxes at the
    earth's surface: long-term variations at 39°N, 77°W. J Geogr Res,
    97: 7579-7591.

    Cox CWJ (1987) Ultraviolet irradiance levels in welding processes.
    In: Passachier WF & Bosnjakovic BFM ed. Human exposure to
    ultraviolet radiation: Risks and regulations. Amsterdam, Oxford, New
    York, Elsevier Science Publishers, pp 383-386.

    Cristofolini M, Franceschi S, Tasin L, Zumiani G, Piscioli F,
    Talamini R, & La Vecchia C (1987) Risk factors for cutaneous
    malignant melanoma in a northern Italian population. Int J Cancer,
    39: 150-154.

    Crombie IK (1981) Distribution of malignant melanoma on the body
    surface. Br J Cancer, 43: 842-849.

    Cruickshanks KJ, Klein R, & Klein BE (1993) Sunlight and age-related
    macular degeneration: the Beaver Dam Eye Study. Arch Ophthalmol,
    111: 514-518.

    Cunningham ML, Johnson JS, Giovanazzi SM, & Peak MJ (1985)
    Photosensitized production of superoxide anion by monochromatic
    (290-405 nm) ultraviolet irradiation of NADH and NADPH coenzymes.
    Photochem Photobiol, 42: 125-128.

    Cyrlin MN, Pedris-Leftick A, & Sugar J (1980) Cataract formation in
    association with ultraviolet photosensitivity. Ann Ophthalmol, 12:
    786-790.

    Czochralska B, Bartosz W, & Shugar D (1984) Oxidation of
    excited-state NADH and NAD dimer in aqueous medium - involvement of
    O2- as a mediator in the presence of oxygen. Biochim Biophys Acta,
    801: 403-409.

    Danielle RP (1988) Pathophysiology of the asthmatic syndromes. In:
    Danielle RP ed. Immunologic diseases of the lung. Boston, Blackwell
    Scientific Publishers, pp 503-516.

    Dardanoni L, Gafá L, Paterno R, & Pavone G (1984) A case- control
    study on lip cancer risk factors in Ragusa (Sicily). Int J Cancer,
    34: 335-337.

    Dargent JL, Lespagnard L, Heenan M, & Verhest A (1992) Malignant
    melanoma occurring in a case of oculocutaneous albinism.
    Histopathology, 21: 74-76.

    Darrell RW & Bachrach CA (1963) Pterygium among veterans. Arch
    Ophthalmol, 70: 158-169.

    Davies DM (1985) Calcium metabolism in healthy men deprived of
    sunlight. Ann N Y Acad Sci, 453: 21-27.

    Daynes RA & Spellman CW (1977) Evidence for the generation of
    suppressor cells by UV radiation. Cell Immunol, 31: 182-187.

    Daynes RA, Spellman CW, Woodward JG, & Stewart DA (1977) Studies
    into the transplantation biology of ultraviolet light-induced
    tumours. Transplantation, 23: 343-348.

    De Fabo EC & Kripke ML (1979) Dose-response characteristics of
    immunologic unresponsiveness to UV-induced tumors produced by UV
    irradiation of mice. Photochem Photobiol, 30: 385-390.

    De Fabo EC & Noonan FP (1983) Mechanism of immune suppression by
    ultraviolet irradiation  in vivo. J Exp Med, 157: 84-98.

    de Gruijl FR & van der Leun JC (1980) A dose-response model for skin
    cancer induction by chronic UV exposure of a human population. J
    Theor Biol, 83: 487-504.

    de Gruijl FR & van der Leun JC (1982) Systemic influence of
    pre-irradiation of a limited skin area on UV-tumorigenesis.
    Photochem Photobiol, 35: 379-383.

    de Gruijl FR & van der Leun JC (1983) Follow up on systemic
    influence of partial pre-irradiation on UV-tumorigenesis. Photochem
    Photobiol, 38: 381-383.

    de Gruijl FR & van der Leun JC (1991) Action spectra for
    carcinogenesis. In: Urbach F ed. Biological responses to UVA.
    Overland Park, Kansas, Valdemar Publishing Company, p 91.

    de Gruijl FR, van der Meer JB, & van der Leun JC (1983) Dose-time
    dependency of tumor formation by chronic UV exposure. Photochem
    Photobiol, 37: 53-62.

    de Gruijl FR, Sterenborg HJCM, Forbes PD, Davies RE, Cole C,
    Kelfkens G, van Weelden H, Slaper H, & van der Leun JC (1993)
    Wavelength dependence of skin cancer induction by ultraviolet
    radiation of albino hairless mice. Cancer Res, 53: 53-60.

    Denkins Y, Fidler IJ, & Kripke ML (1989) Exposure of mice to UV-B
    radiation suppresses delayed hypersensitivity to Candida albicans.
    Photochem Photobiol, 49: 615-619.

    Desai ID, Sawant PL, & Tappel AL (1964) Peroxidative and radiation
    damage to isolated lysosomes. Biochim Biophys Acta, 86: 277-285.

    Detels R & Dhir SP (1967) Pterygium: a geographical study. Arch
    Ophthalmol, 78: 485-491.

    Devary Y, Gottlieb RA, Smeal T & Karin M (1992) The mammalian
    ultraviolet response is triggered by activation of Src tyrosine
    kinases. Cell 71: 1081-1091.

    Devary Y, Rosette C, Di Donato JA, & Karin M (1993) NFkB activation
    by ultraviolet light not dependent on a nuclear signal. Science,
    261: 1442-1445.

    Dhir SP, Detels R, & Alexander ER (1967) The role of environmental
    factors in cataract, pterygium and trachoma. Am J Ophthalmol, 64:
    128-135.

    Diffey BL (1977) The calculation of the spectral distribution of
    natural ultraviolet radiation under clear sky conditions, Phys Med
    Biol, 22: 309-316.

    Diffey BL (1986) Use of UVA sunbeds for cosmetic tanning. Br J
    Dermatol, 225: 67-76.

    Diffey BL (1987) Analysis of the risk of skin cancer from sunlight
    and solaria in subjects living in northern Europe. Photodermatology,
    4: 118-126.

    Diffey BL (1988) The risk of skin cancer from occupational exposure
    to ultraviolet radiation in hospitals. Phys Med Biol, 33(10):
    1187-1193.

    Diffey BL (1989) Ultraviolet radiation and skin cancer: Are
    physiotherapists at risk? Physiotherapy, 75(10): 615-616.

    Diffey BL (1989a) Ultraviolet radiation dosimetry with polysulphone
    film. In: Diffey BL ed. Radiation measurement in photobiology. New
    York, London, San Francisco, Academic Press, pp 135-159.

    Diffey BL (1990) Human exposure to ultraviolet radiation. Semin
    Dermatol, 9: 2-10.

    Diffey BL (1993) A photobiological evaluation of lamps used in the
    phototherapy of seasonal affective disorder. J Photochem Photobiol,
    B17: 203-207.

    Diffey BL (1993a) Personal communication.

    Diffey BL & Cheeseman J (1992) Sun protection with hats. Br J
    Dermatol, 127: 10-12.

    Diffey BL & McKinlay AF (1983) The UVB content of 'UVA fluorescent
    lamps' and its erythemal effectiveness in human skin. Phys Med Biol,
    28: 351-358.

    Diffey BL & Robson J (1992) The influence of pigmentation and
    illumination on the perception of erythema. Photodermatol
    Photoimmunol Photomed, 9: 45-47.

    Diffey BL, Tate TJ, & Davis A (1979) Solar dosimetry of the face:
    the relationship of natural ultraviolet exposure to basal cell
    carcinoma localisation. Phys Med Biol, 24: 931-939.

    Diffey BL, Challoner AVJ, & Key PJ (1980) A survey of the
    ultraviolet radiation emissions of photochemotherapy units. Br J
    Dermatol, 102: 301-306.

    Diffey BL, Larko O, & Swanbeck G (1982) UVB doses received during
    different outdoor activities and UVB treatment of psoriasis. Br J
    Dermatol, 106: 33-41.

    Dillon J, Wang RH, & Atherton S (1990) Photochemical and
    photophysical studies on human lens constituent. Photochem
    Photobiol, 52: 849-854.

    Dion M & Hammelin C (1987) Relationship between enhanced
    reactivation and mutagenesis of UV-irradiated human cytomegalovirus
    in normal human cells. EMBO J, 6: 397-399.

    Djavuheri-Mergny M, Maziére JC, Santus R, Mora L, Maziére C, Auclair
    M & Dubertet C (1993) Exposure to long wavelength ultraviolet
    radiation decreases processing of low density lipoprotein by
    cultured human fibroblasts. Photochem Photobiol 57: 302-305.

    Döhler G (1990) Effect of UVB (290-320nm) radiation on uptake of
    15N-nitrate by marine diatoms. Berlin, Heidelberg, New York,
    Springer-Verlag, pp 354-359.

    Döhler G & Alt MR (1989) Assimilation of 15N-ammonia during
    irradiance with ultraviolet-B and monochromatic light by
    Thalassiosira rotula, C R Acad Sci Paris, D308, 513-518.

    Dolezal JM, Perkins ES, & Wallace RB (1989) Sunlight, skin
    sensitivity, and senile cataract. Am J Epidemiol, 129: 559-568.

    Doll R (1991) Urban and rural factors in the aetiology of cancer.
    Int J Cancer, 47: 803-810.

    Doll R, Muir C, & Waterhouse J ed. (1970) Cancer incidence in five
    continents, Volume II. Berlin, Heidelberg, New York,
    Springer-Verlag.

    Doniger J, Jacobson ED, Krell K, & DiPaolo JA (1981) Ultraviolet
    light action spectra for neoplastic transformation and lethality of
    syrian hamster embryo cells correlate with spectrum for pyrimidine
    dimer formation in cellular DNA. Proc Natl Acad Sci (USA), 78:
    2378-2383.

    Dorn HF (1944b) Illness from cancer in the United States: IV.
    Illness from cancer of specific sites classed in broad groups.
    Public Health Rep, 59: 65-77.

    Dorn HF (1944a) Illness from cancer in the United States. Public
    Health Rep, 59: 33-48.

    Dorn CR, Taylor DON, & Schneider R (1971) Sunlight exposure and risk
    of developing cutaneous and oral squamous cell carcinomas in white
    cats. J Natl Cancer Inst, 46: 1073-1078.

    Doughty MJ & Cullen A (1990) Long-term effects of a single dose of
    ultraviolet B on albino rabbit cornea. II. Deturgescence and fluid
    pump assessed  in vitro. Photochem Photobiol, 54: 439-449.

    Driscoll CMH (1992) Solar UV trends and distributions. Natl Radiat
    Prot Board Bull, 137: 7-13.

    Dubin N, Moseson M, & Pasternack BS (1986) Epidemiology of malignant
    melanoma: pigmentary traits, ultraviolet radiation, and the
    identification of high-risk populations. Recent Results Cancer Res,
    102: 56-75.

    Duke-Elder WS (1926) The pathological action of light upon the eye.
    I. Action on the outer eye: Photophthalmia. Lancet, 1: 1137-1141.

    Duvic M, Lowe L, Rapini RP, Rodriguez S, & Levy ML (1989) Eruptive
    dysplastic nevi associated with human immunodeficiency virus
    infection. Arch Dermatol, 125: 397-401.

    Dyall-Smith D & Varigos G (1985) The malignant potential of
    papillomavirus. Aust J Dermatol, 26: 102-107.

    Eggersdorfer B & Häder DP (1991) Phototaxis, gravitaxis and vertical
    migration in the marine dinoflaggelates. Acta Protozool 30(2):
    63-71.

    Eisenstark A & Perrot G (1987) Catalase has only a minor role in
    protection against ultraviolet radiation damage in bacteria. Mol Gen
    Genet 207: 68-72.

    Eisman JA, MacIntyre I, Martin TJ, Frampton RJ, & King RJB (1980)
    Normal and malignant breast tissue is a target organ for 1,25-(OH)2
    vitamin D3. Clin Endocrinol, 13: 267-272.

    Elliott R (1961) The aetiology of pterygium. Trans Ophthalmol Soc N
    Z, 13: 22-41.

    Ellison MJ & Childs JD (1981) Pyrimidine dimers induced in
    Escherichia coli DNA by ultraviolet radiation present in sunlight.
    Photochem Photobiol, 34: 465-469.

    Elmets CA, Bergstresser PR, Tigelaar RE, Wood PJ, & Streilein JW
    (1983) Analysis of the mechanism of unresponsiveness produced by
    haptens painted on skin exposed to low dose ultraviolet radiation. J
    Exp Med, 158: 781-794.

    Elwood, JM & Gallagher RP (1983) Site distribution of malignant
    melanoma. Can Med Assoc J, 128: 1400-1404.

    Elwood JM, Gallagher RP, Hill GB, Spinelli JJ, Pearson JCG, &
    Threlfall W (1984) Pigmentation and skin reaction to sun as risk
    factors for cutaneous melanoma: Western Canada melanoma study. Br
    Med J, 288: 99-102.

    Elwood JM, Gallagher RP, Davison J, & Hill GB (1985a) Sunburn,
    suntan and the risk of cutaneous malignant melanoma: the Western
    Canada melanoma study. Br J Cancer, 51: 543-549.

    Elwood JM, Gallagher RP, Hill GB, & Pearson JCG (1985b) Cutaneous
    melanoma in relation to intermittent and constant sun exposure: the
    Western Canada melanoma study. Int J Cancer, 35: 427-433.

    Elwood JM, Williamson C, & Stapleton PJ (1986) Malignant melanoma in
    relation to moles, pigmentation, and exposure to fluorescent and
    other lighting sources. Br J Cancer, 53: 65-74.

    Elwood JM, Whitehead SM, Davison J, Stewart M, & Galt M (1990)
    Malignant melanoma in England: risk associated with naevi, freckles,
    social class, hair colour, and sunburn. Int J Epidemiol, 19:
    801-810.

    Emmett EA (1973) Ultraviolet radiation as a cause of skin tumors.
    Crit Rev Toxicol, 2, 211-255.

    Engel A, Johnson M, & Haynes S (1988) Health effects of sunlight
    exposure in the United States. Arch Dermatol, 124: 72-79.

    English DR & Armstrong BK (in press) Cutaneous malignant melanoma.
    In: Schottenfeld, D Fraumeni. J eds, Cancer epidemiology and
    prevention, New York, Oxford University Press.

    Enninga IC, Groenendijk RTL, Filon AR, van Zeeland AA, & Simons JWIM
    (1986) The wavelength dependence of UV-induced pyrimidine dimer
    formation, cell killing and mutation induction in human diploid skin
    fibroblasts. Carcinogenesis, 7: 1829-1836.

    Epe B, Pflaum M, & Boiteux S (1993) DNA damage induced by
    photosensitisers in cellular and cell-free systems. Mutat Res 299:
    135-145.

    Epstein JH (1965) Comparison of the carcinogenic and cocarcinogenic
    effects of ultraviolet light on hairless mice. J Natl Cancer Inst,
    34: 741-745.

    Epstein JH (1985) Animal models for studying photocarcinogenesis.
    In: Maibach H & Lowe N ed. Models in Dermatology. Basel, Karger, vol
    2, pp 303-312.

    Epstein JH (1988) Photocarcinogenesis promotion studies with benzoyl
    peroxide (BPO) and croton oil. J Invest Dermatol, 91: 114-116.

    Epstein JH & Epstein WL (1962) Cocarcinogenic effect of ultraviolet
    light on DMBA tumor initiation in albino mice. J Invest Dermatol,
    39: 455-460.

    Epstein JH & Roth HL (1968) Experimental ultraviolet light
    carcinogenesis. A study of croton oil promoting effects. J Invest
    Dermatol, 50: 387-389.

    Epstein JH, Tuffanelli DL, & Dubois EL (1965) Light sensitivity and
    lupus erythematosus. Arch Dermatol, 91: 483-485.

    Epstein JH, Epstein WL, & Nakai T (1967) Production of melanomas
    from DMBA-induced 'blue nevi' in hairless mice with ultraviolet
    light. J Natl Cancer Inst, 38: 19-30.

    Eriksen P (1987) Occupational applications of ultraviolet radiation:
    risk evaluation and protection techniques. In: Passchier WF &
    Bosnjakovic BFM ed. Human exposure to ultraviolet radiation: Risks
    and regulations. Amsterdam, Oxford, New York, Elsevier Science
    Publishers, pp 317-331.

    Everett MA, Olsen RL, & Sayer RM (1965) Ultraviolet erythema. Arch
    Dermatol, 92: 713-719.

    Farr PM, Marks JM, Diffey BL, & Ince P (1988) Skin fragility and
    blistering due to the use of sunbeds. Br Med J, 296: 1708-1709.

    FDA (1988) Quality control guide for sunlamp products, Washington,
    DC, US Dept. Health and Human Services, (HHS Publication FDA No.
    88-8234).

    FDA (1992) Medications that increase sensitivity to light (prepared
    by Levine JI). Rockville, Maryland, Food and Drug Administration,
    Center of Devices and Radiological Health.

    Findlay GM (1928) Ultra-violet light and skin cancer. Lancet, ii:
    1070-1073.

    Finsen NR (1901) The treatment of lupus vulgaris by concentrated
    chemical rays. In: Phototherapy. London, Edward Armold, pp 73-75.

    Fisher GJ & Johns HE (1976) Pyrimidine photohydrates. In: (Ed. Wang,
    S.) Photochemistry and Photobiology of Nucleic Acids, vol. 1, New
    York, Academic Press, pp 169-224.

    Fisher MS & Kripke ML (1977) Systemic alteration induced in mice by
    ultraviolet light irradiation and its relationship to ultraviolet
    carcinogenesis. Proc Natl Acad Sci (USA), 74: 1688-1692.

    Fisher MS, Menter JM, & Willis I (1989) Ultraviolet
    radiation-induced suppression of contact hypersensitivity in
    relation to padimate O and oxybenzone. J Invest Dermatol, 92:
    337-341

    Fitzpatrick TB & Sober AJ (1985) Sunlight and skin cancer. N Engl J
    Med, 313: 818-820.

    Fitzpatrick TB, Pathak MA, Magnus IA, & Curwen WL (1963) Abnormal
    reactions of man to light. Annu Rev Med, 14: 195-214.

    Fitzpatrick TB, Pathak MA, Harber LC, Seiji M, & Kukita A (1974)
    Sunlight and man. Tokyo, University of Tokyo Press.

    Fleming ID, Barnawell JR, Burlison PE, (1975) Skin cancer in black
    patients. Cancer, 35: 600-605.

    Forbes PD, Davies RE, Sambuco CP, & Urbach F (1989) Topical urocanic
    acid enhances UV-induced skin tumors in hairless mice by topical
    application of the sunscreen 2-ethylhexyl-p-methoxycinnamate. J
    Toxicol Cutan Ocul Toxicol, 8: 209-226.

    Forbes PD, Blum HF, & Davies RE (1981) Photocarcinogenesis in
    hairless mice: dose-response and the influence of dose-delivery.
    Photochem Photobiol, 34: 361-365.

    Forbes PD, Davies RE, Urbach F, Berger D, & Cole C (1982) Simulated
    stratospheric ozone depletion and increased ultraviolet radiation:
    effects on photocarcinogenesis in hairless mice. Cancer Res, 42:
    2796-2803.

    Forbes PD, Davies RE, & Urbach F (1978) Experimental ultraviolet
    photocarcinogenesis: wavelength interactions and time-dose
    relationships. Natl Cancer Inst Monogr, 50: 31-38.

    Forsius H (1988) Exfoliation syndrome in various ethnic populations.
    Acta Ophthalmol, 68(Suppl 184): 71-85.

    Foster HM & Webb SJ (1988) Skin cancer in the North Solomons. Aust N
    Z J Surg, 58: 397-401.

    Fragu P, Lemarchand-Venencie F, Benhamou S, Francois P, Jeannel D,
    Benhamou E, Sezary-Lartigau I, & Avril MF (1991) Long-term effects

    in skin and thyroid after radiotherapy for skin angiomas - A French
    retrospective cohort study. Eur J Cancer, 27: 1215-1222.

    Franklin WA, Lo KM, & Haseltine WA (1982) Alkaline lability of
    fluorescent photoproducts produced in ultraviolet light-irradiated
    DNA. J Biol Chem, 157: 13535-13543.

    Frederick JE, Snell HE, & Haywoood EK (1989) Solar ultraviolet
    radiation at the earth's surface. Photochem Photobiol, 51: 443-450.

    Freeman RG (1975) Data on the action spectrum for ultraviolet
    carcinogenesis. J Natl Cancer Inst, 55: 1119-1122.

    Freeman SE, Hacham H, Gange RW, Maytum DJ, Sutherland JC, &
    Sutherland BM (1989) Wavelength dependence of pyrimidine dimer
    formation in DNA of human skin irradiated in situ with ultraviolet
    light. Proc Natl Acad Sci (USA), 86: 5605-5609.

    Friedberg EC (1984) DNA repair. New York, Freeman and Company.

    Friedmann PS, White SI, Parker S, Moss C, & Matthews JNS (1989)
    Antigenic stimulation during ultraviolet therapy in man does not
    result in immunological tolerance. Clin Exp Immunol, 76: 68-72.

    Fuchs J & Packer L (1991) Photooxidative stress in the skin. In:
    Sies H Ed. Oxidative stress: oxidants and antioxidants. New York,
    London, San Francisco, Academic Press, pp 554-584.

    Fuchs J, Huflejt M, Rothfuss L, Carcamero G, & Packer L (1989)
    Impairment of enzymic and nonenzymic antioxidants in skin by UVB
    irradiation. J Invest Dermatol 93, 769-773.

    Fuller CJ, Faulkner H, Bendich A, Parker RS, & Roe DA (1992) Effect
    of ß-carotene supplementation on photosuppression of delayed-type
    hypersensitivity in normal young men. Am J Clin Nutr, 56: 684-690.

    Funnell SGP & Keast D (1985) The effect of ultraviolet radiation on
    the generation of plaque-forming cells and on T-suppressor cell
    activity to sheep erythrocytes. Photodermatology, 3: 64-72.

    Gaboriau F, Morlière P, Marquis I, Maysan A, Gèze M, & Dubertret L
    (1993) Membrane damage induced in cultured human skin fibroblasts by
    UVA irradiation. Photochem Photobiol, 58: 515-520.

    Gafa L, Filippazzo MG, Tumino R, Dardanoni G, Lanzarone F, &
    Dardanoni L (1991) Risk factors of nonmelanoma skin cancer in
    Ragusa, Sicily: a case-control study. Cancer Causes Control, 2:
    395-399.

    Gallagher, RP (1988) Ocular melanoma in farmers (Letter to the
    Editor). Am J Ind Med, 13: 523-525.

    Gallagher PE & Duker NJ (1986) Detection of UV purine photoproduct
    in a defined sequence of human DNA. Mol Cell Biol, 6: 707-709.

    Gallagher RP, Elwood JM, & Hill GB (1986) Risk factors for cutaneous
    malignant melanoma: the Western Canadian melanoma study, Rec Res
    Cancer Res, 102: 38-55.

    Gallagher RP, Threlfall WJ, Jeffries E, Band PR, Spinelli J, &
    Coldman AJ (1984) Cancer and aplastic anemia in British Columbia
    farmers. J Natl Cancer Inst, 72: 1311-1315.

    Gallagher CH, Greenoak GE, Reeve VE, Canfield PJ, Baker RSU, & Bonin
    AM (1984a) Ultraviolet carcinogenesis in the hairless mouse skin -
    influence of the sunscreen 2-ethylhexyl-p-methoxycinnamate. Aust J
    Exp Biol Med Sci, 62: 577-588.

    Gallagher RP, Elwood JM, Rootman J, Spinelli JJ, Hill GB, Threlfall
    WJ, & Birdsell JM (1985) Risk factors for ocular melanoma: Western
    Canada melanoma study. J Natl Cancer Inst, 74: 775-778.

    Gallagher PE, Weiss RB, Brent TP, & Duker NJ (1989) Wavelength
    dependence of DNA incision by a human ultraviolet endonuclease.
    Photochem Photobiol, 49: 363-367.

    Gallagher RP, Ma B, McLean DI, Yang CP, Ho V, Carruthers JA, &
    Warshawski LM (1990a) Trends in basal cell carcinoma, squamous cell
    carcinoma, and melanoma of the skin from 1973 through 1987. J Am
    Acad Dermatol, 23: 413-421.

    Gallagher RP, McLean DI, Yang CP, Coldman AJ, Silver HKB, Spinelli
    JJ, & Beagrie M (1990b) Anatomic distribution of acquired
    melanocytic nevi in white children. A comparison with melanoma: the
    Vancouver mole study. Arch Dermatol, 126: 466-471

    Gallagher RP, McLean DI, Yang CP, Coldman AJ, Silver HKB, Spinelli
    JJ, & Beagrie M (1990c) Suntan, sunburn and pigmentation factors and
    the frequency of acquired melanocytic nevi in children. Similarities
    to melanoma: the Vancouver mole study. Arch Dermatol, 126: 770-776

    Gallagher RP, Rivers JK, Yang CP, (1991) Melanocytic nevus density
    in Asian, Indo-Pakistani, and white children: The Vancouver mole
    study. J Am Acad Dermatol, 25: 507-12.

    Gange RW, Blackett AD, Matzinger EA, Sutherland BM, & Kochevar IE
    (1985) Comparative protection efficiency of UVA-and UVB-induced tans
    against erythema and formation of endonuclease-sensitive sites in
    DNA by UVB in human skin. J Invest Dermatol, 85: 362-364.

    Garbe C, Krüger S, Stadler R, Guggenmoos-Holzmann I, & Orfanos CE
    (1989) Markers and relative risk in a German population for
    developing malignant melanoma. Int J Dermatol, 28: 517-523.

    Garcia-Pichel F & Castenholz RW (1991) Characterization and
    biological implications of scytonemin, a cyanobacterial sheath
    pigment, J Phycol, 27: 395-409.

    Garland C & Garland F (1980) Do sunlight and vitamin D reduce the
    likelihood of colon cancer? Int J Epidemiol, 9: 227-231.

    Garland CF, Comstock GW, Garland FC, Felsing K, Shaw EK, & Gorham ED
    (1989) Serum 25-hydroxyvitamin D and colon cancer: 8-year
    prospective study. Lancet, 2: 1176-1178.

    Garland FC, Garland CF, Gorham ED, & Young JF (1990) Geographic
    variation in breast cancer mortality in the United States: A
    hypothesis involving exposure to solar radiation. Prev Med, 19:
    614-622.

    Garland CF Garland FC, & Gorham ED (1992) Could sunscreens increase
    melanoma risk? (Letter) Am J Public Health, 82(4): 614-615.

    Garner A (1989) The pathology of tumours at the limbus. Eye, 3:
    210-217.

    Garsen J, Goettsch W, de Gruy F, & van Loveren H (1993) UVB
    suppresses immunity and resistance against systemic infections in
    the rat. Photochem Photobiol, 57: 755.

    Gasparro F & Fresco J (1986) Ultraviolet-induced 8,8-adenine
    dehydromimers in oligo- and polynucleotides. Nucl Acids Res, 14:
    4239-4251.

    Gellin GA, Kopf AW, & Garfinkel L (1965) Basal cell epithelioma. A
    controlled study of associated factors. Arch Dermatol, 91: 38-45.

    Gensler HL (1988) Enhancement of chemical carcinogenesis in mice by
    systemic effects of ultraviolet radiation. Cancer Res, 48: 620-623.

    Gensler HL (1989) Reduction of immunosuppression in UV-irradiated
    mice by dietary retinyl palmitate plus canthaxanthin.
    Carcinogenesis, 10: 203-207.

    Gensler HL & Bowden GT (1987) UVB-Induced modulation of mouse skin
    tumor induction by benzo[a]pyrene (Abstract no. 546). Proc Am Assoc
    Cancer Res, 28: 137.

    Gensler HL & Welch K (1992) Prevalence of tumor prevention rather
    than tumor enhancement when repetitive UV radiation treatments
    precede initiation and promotion. Carcinogenesis, 13: 9-13.

    Gerrish KE & Gensler HL (1993) Prevention of photocarcinogenesis by
    dietary vitamin E. Nutr Cancer, 19: 125-133.

    Giannini SH (1986a) Suppression of pathogenesis in cutaneous
    leishmaniasis by UV irradiation. Infect Immun, 51: 838-843.

    Giannini SH (1986b) Effects of UV-B on infectious disease. In: Titus
    JG ed. Effects of changes in stratospheric ozone and global climate.
    Washington, DC, US Environmental Protection Agency, vol 2, pp
    101-112.

    Giannini SH (1987) Abrogation of skin lesions in cutaneous
    leishmaniasis by ultraviolet irradiation. In: Hart DT ed.
    Leishmaniasis: The first centenary (1885-1985) new strategies for
    control. New York, London, Plenum Press, pp 677-684 (NATO ASI Series
    A: Life Sciences).

    Giannini SH (1992) Effects of ultraviolet B irradiation on cutaneous
    leishmaniasis. Parasitology Today, 8: 44-48.

    Gibbs NK (1993) Is ultraviolet immunosuppression initiated by
    photosomerisation of urocanic acid. In: de Gruijl FR ed. The dark
    side of sunlight. Utrecht, The Netherlands, Utrecht University.

    Gies HP & Roy CR (1990) Bilirubin phototherapy and potential UVR
    hazards. Health Phys, 58: 313-320.

    Gies HP, Roy CR, & Elliott G (1986) Artificial suntanning: Spectral
    irradiance and hazard evaluation of ultraviolet sources. Health
    Phys, 50(6): 691-703.

    Gies HP, Roy CR, & Elliot G (1990) A proposed UVR protection factor
    for sunglasses. Clin Exp Optom, 73: 183-189.

    Gies HP, Roy CR, & Elliott G (1992) Ultraviolet radiation protection
    factors for personal protection in both occupational and
    recreational situations. Radiat Prot Aust, 10(3): 59-66.

    Gies HP, Roy CR, Herlihy E, & Rivers J (1992a) Personal dosimetry of
    solar UVB using polysulphone film Congress Proceedings (IRPA 8) vol
    1, p 791, Montreal, May 1992, International Radiation Protection
    Association.

    Gilchrest BA (1990) Actinic injury. Annu Rev Med, 41: 199-210.

    Gilchrest BA, Soter NA, Stoff JS, & Mihm MC Jr (1981) The human
    sunburn reaction: histologic and biochemical studies. J Am Acad
    Dermatol, 5: 411-422.

    Giles GG, Marks R, & Foley P (1988) Incidence of non- melanocytic
    skin cancer treated in Australia. Br Med J, 296: 13-17.

    Glass AG & Hoover RN (1989) The emerging epidemic of melanoma and
    squamous cell skin cancer. J Am Med Assoc, 262: 2097-2100.

    Goldberg LH & Altman A (1984) Benign skin changes associated with
    chronic sunlight exposure. Cutis, 34: 33-39.

    Goodman GJ, Marks R, Selwood TS, Ponsford MW, & Pakes W (1984)
    Non-melanocytic skin cancer and solar keratoses in Victoria -
    clinical studies II. Aust J Dermatol, 25: 103-106.

    Gorham ED, Garland CF, & Garland FC (1989) Acid haze air pollution
    and breast and colon cancer mortality in 20 Canadian cities. Can J
    Public Health, 80: 96-100.

    Gorham ED, Garland FC, & Garland CF (1990) Sunlight and breast
    cancer incidence in the USSR. Int J Epidemiol, 19: 820-824.

    Graham S, Marshall J, Haughey B, Stoll H, Zielezny M, Brasure J, &
    West D (1985) An inquiry into the epidemiology of melanoma. Am J
    Epidemiol, 122: 606-619.

    Gray RH, Johnson GJ, & Freedman A (1992) Climatic droplet
    keratopathy. Surv Ophthalmol, 36: 241-253.

    Greaves MW, Hensby CN, Black AK, Plummer NA, Fincham N, Warin AP, &
    Camp R (1978) Inflammatory reactions induced by ultraviolet
    irradiation. Bull. Cancer, 65: 299-304.

    Green AC (1984) Sun exposure and the risk of melanoma. Aust J
    Dermatol, 25: 99-102.

    Green AC (1991) Premature ageing of the skin in a Queensland
    population. Med J Aust, 155: 473-478.

    Green AC & Battistutta D (1990) Incidence and determinants of skin
    cancer in a high-risk Australian population. Int J Cancer, 46:
    356-361.

    Green A & O'Rourke MGE (1985) Cutaneous malignant melanoma in
    association with other skin cancers. J Natl Cancer Inst, 74:
    977-980.

    Green A & Siskind V (1983) Geographical distribution of cutaneous
    melanoma in Queensland. Med J Aust, 1: 407-410.

    Green AES, Sawada T, & Shettle EP (1974) The middle ultraviolet
    reaching the ground. Photochem Photobiol, 19: 251.

    Green AC, Siskind V, Bain C, & Alexander J (1985) Sunburn and
    malignant melanoma. Br J Cancer, 51: 393-397.

    Green AC, Beardmore G, Hart V, Leslie D, Marks R, & Staines D
    (1988a) Skin cancer in a Queensland population. J Am Acad Dermatol,
    19: 1045-1052.

    Green AC, Sorahan T, Pope D, Siskind V, Hansen M, Hanson L, Leech P,
    Ball PM, & Grimley RP (1988b) Moles in Australian and British
    schoolchildren (Letter to the Editor). Lancet, ii: 1497.

    Green C, Diffey BL, & Hawk JLM (1992) Ultraviolet radiation in the
    treatment of skin disease. Phys Med Biol, 37(1): 1-20.

    Green A, Smith P, McWhirter W, Oregan P, Battistutta D, Yarker ME, &
    Lape K (1993a) Melanocytic naevi and melanoma in survivors of
    childhood cancer. Br J Cancer, 67: 1053-1057.

    Green A, MacLennan R, Youll P, & Martin N (1993b) Site distribution
    of cutaneous melanoma in Queensland. Int J Cancer, 53: 232-236.

    Greenberg JT, Monach PA, Chou JH, Josephy PD, & Demple B (1990)
    Positive control of a global antioxidant defence regulon activated
    by superoxide-generating agents in Escherichia coli. Proc Natl Acad
    Sci (USA), 87: 6181-6185.

    Greene MH & Wilson J (1985) Second cancer following lymphatic and
    hematopoietic cancers in Connecticut, 1935-82. Natl Cancer Inst
    Monogr, 68: 191-217.

    Greene MH, Young TI, & Clark WH Jr (1981) Malignant melanoma in
    renal transplant recipients. Lancet, May 30: 1196-1199.

    Grob JJ, Gouvernet J, Aymar D, Mostaque A, Romano MH, Collet AM, Noe
    MC, Diconstanzo MP, & Bonerandi JJ (1990) Count of benign
    melanocytic nevi as a major indicator of risk for nonfamilial
    nodular and superficial spreading melanoma. Cancer, 66: 387-395.

    Guerry RK, Ham WT, & Mueller HA (1985) Light toxicity in the
    posterior segment. In: Duane TD & Jaeger EA ed. Clinical
    ophthalmology. Philadelphia, Pennsylvania, Harper and Row, pp 1-17.

    Guex-Crosier Y & Herbort CP (1993) Presumed corneal intraepithelial
    neoplasia associated with contact lens wear and intense ultraviolet
    light exposure. Br J Ophthalmol, 77: 191-192.

    Gupta AK, Cardella CJ, & Haberman HF (1986) Cutaneous malignant
    neoplasms in patients with renal transplants. Arch Dermatol, 122:
    1288-1293.

    Gupta AK, Stern RS, Swanson NA, Anderson TF, & Arbor A (1988)
    Cutaneous melanomas in patients treated with psoralens plus
    ultraviolet A. J Am Acad Dermatol, 19: 67-76.

    Gutteridge JMC (1985) Superoxide dismutase inhibits the
    superoxide-driven Fenton reaction at two different levels. FEBS
    Lett, 185: 19-23.

    Häder DP & Worrest RC (1991) Effects of enhanced solar radiation on
    aquatic ecosystems. Photochem Photobiol, 53: 717-725.

    Häder DP & Lui SM (1991) Mobility and gravitactic orientation of the
    flagellate, Euglena gracilis, impaired by artificial and solar UVB
    radiation. Curr Microbiol, 21: 161-168.

    Haenszel W (1963) Variations in skin cancer incidence within the
    United States. Natl Cancer Inst Monogr, 10: 225-243.

    Halprin KM, Comerford M, Presser SE, & Taylor JR (1981) Ultraviolet
    light treatment delays contact sensitization to nitrogen mustard. Br
    J Dermatol, 105: 71-76.

    Ham WT, Mueller HA, Ruffolo JJ, Millen JE, Cleary SF, & Guerry RK
    (1984) Basic mechanisms underlying the production of photochemical
    lesions in the mammalian retina. Curr Eye Res, 3: 165-174.

    Hanchette CL & Schwartz GC (1992) Geographic patterns of
    prostatecancer mortality. Evidence for a protective effect of
    ultraviolet radiation. Cancer, 70: 2861-2869.

    Hardie IR, Strong RW, Hartley LCJ, Woodruff PWH, & Clunie GJA (1980)
    Skin cancer in Caucasian renal allograft recipients living in a
    subtropical climate. Surgery, 87: 177-183.

    Hartevelt MM, Bouwes Bavinck JN, Kootte AMM, Vermeer BJ, &
    Vandenbroucke JP (1990) Incidence of skin cancer after renal
    transplantation in the Netherlands. Transplantation, 49: 506-509.

    Hawk, JLM (1984) Photosensitising agents used in the United Kingdom.
    Clin Expl Dermatol, 9: 300-302.

    Hayashi Y & Aurelian L (1986) Immunity to herpes simplex virus type
    2: Viral antigen-presenting capacity of epidermal cells and its
    impairment by ultraviolet irradiation. J Immunol, 136: 1087-1092.

    Haynes RH (1966) The interpretation of microbiol inactivation and
    recovery phenomena. Radiat Res, Suppl 6: 1-29.

    Health Council of the Netherlands (1986) UV radiation: Human
    exposure to ultraviolet radiation. The Hague, Health Council of the
    Netherlands (Report 1986/93).

    Hedblom EE (1961) Snowscape eye protection. Arch Environ Health, 2;
    685-704.

    Hefferren JJ, Cooley RO, Hall JB, Olsen NH, & Lyon HW (1971) Use of
    ultraviolet illumination in diagnosis. J Am Dent Assoc, 82:
    1353-1360.

    Herity B, O'Loughlin G, Moriarty MJ, & Conroy R (1989) Risk factors
    for non-melanoma skin cancer. Ir Med J, 82: 151-152.

    Hersey P, Hasic E, Edwards A, Bradley M, Maran G, & McCarthy WH
    (1983a) Immunological effects of solarium exposure. Lancet, 1:
    545-548.

    Hersey P, Haran G, Hasic E, & Edwards A (1983b) Alteration of T cell
    subsets and induction of suppressor T cell activity in normal
    subjects after exposure to sunlight. J Immunol, 31: 171-174.

    Hersey P, MacDonald M, Burns C, Schibeci S, Matthews H, & Wilkinson
    FJ (1987) Analysis of the effect of sunscreen agent on the
    suppression of natural killer cell activity induced in human
    subjects by radiation from solarium lamps. J Invest Dermatol, 88:
    271-276.

    Hersey P, MacDonald M, Henderson C, Schibeci S, D'Alessandro G,
    Pryor M, & Wilkinson FJ (1988) Suppression of natural killer cell
    activity in humans by radiation from solarium lamps depleted of UVB.
    J Invest Dermatol, 90: 305-310.

    Hess C (1907) Experiments on the effect of ultraviolet light on the
    lens. Arch f Augenheilk 57 185-196.

    Higginson J & Oettlé AG (1959) Cancer incidence in the Bantu and
    "Cape Colored" races of South Africa: Report of a cancer survey in
    the Transvaal (1953-55). J Natl Cancer Inst, 24: 589-671.

    Hill D, White V, Marks R, Theobald T, Borland R, & Roy C (1992)
    Melanoma prevention: behavioural and nonbehavioural factors in
    sunburn among an Australian urban population. Prev Med, 21: 654-669.

    Hiller R, Sperduto RD, & Ederer F (1977a) Epidemiological
    associations with cataract in the 1971-72 national health and
    nutrition examination survey. Am J Epidemiol, 105: 450-459.

    Hiller R, Giacometti L, & Yuen K (1977b) Sunlight and cataract: an
    epidemiologic investigation. Am J Epidemiol, 105: 450-459.

    Hiller R, Sperduto RD, & Ederer F (1983) Epidemiologic associations
    with cataract in the 1971-1972 national health and nutrition
    examination survey. Am J Epidemiol, 118: 239-249.

    Hiller R, Sperduto RD, & Ederer F (1986) Epidemiologic associations
    with nuclear, cortical, and posterior subcapsular cataracts. Am J
    Epidemiol, 124: 916-925.

    Hirschfeld S, Levine AS, Ozato K, & Protic M (1990) A constitutive
    damage-specific DNA binding protein is synthesized at higher levels
    in UV-irradiated primate cells. Mol Cell Biol, 10: 2041-2048.

    Hoeijmakers JHJ & Bootsma D (1990) Molecular genetics of eucaryotic
    DNA excision repair. Cancer Cells, 2: 311-320.

    Hoffman JS (1987) Ultraviolet radiation and melanoma (with a special
    focus on assessing the risks of stratospheric ozone depletion).
    Washington, DC, US Environmental Protection Agency, Office of Air
    and Radiation (EPA 400/1-87-001D).

    Hogan DJ, To T, Gran L, Wong D, & Lane PR (1991) Risk factors for
    basal cell carcinoma. Int J Dermatol, 28: 591-594.

    Holick MF (1985) The photobiology of vitamin D and its consequences
    for human. Ann N Y Acad Sci, 453: 1-3.

    Hollows F & Moran D (1981) Cataract - the ultraviolet risk factor.
    Lancet, 2: 1249-1250.

    Holly EA, Kelly JW, Shpall SN, & Chiu S-H (1987) Number of
    melanocytic nevi as a major risk factor for malignant melanoma. J Am
    Acad Dermatol, 17: 459-468.

    Holly EA, Aston DA, Char DH, Kristiansen JJ, & Ahn DK (1990) Uveal
    melanoma in relation to ultraviolet light exposure and host factors.
    Cancer Res, 50: 5773-5777.

    Holly EA, Aston DA, Ahn DK, Kristiansen JJ, & Char DH (1991) No
    excess prior cancer in patients with uveal melanoma. Ophthalmology,
    98: 608-611.

    Holm-Hansen O (1990) UV radiation in Arctic waters: Effects on rates
    of primary production (Appendix G). La Jolla, CA, USA, Scripps
    Institution of Oceanography, pp 1-17.

    Holman CDJ & Armstrong BK (1984) Pigmentary traits, ethnic origin,
    benign nevi, and family history as risk factors for cutaneous
    malignant melanoma. J Natl Cancer Inst, 72: 257-266.

    Holman CDJ, Gibson IM, Stephenson M, & Armstrong BK (1983)
    Ultraviolet irradiation of human body sites in relation to
    occupation and outdoor activity: field studies using personal UVR
    dosimeters. Clin Exp Dermatol, 8: 269-277.

    Holman CDJ, Armstrong BK, Evans PR, Lumsden GJ, Dallimore KJ, Meehan
    CJ, Beagley J, & Gibson IM (1984a) Relationship of solar keratosis
    and history of skin cancer to objective measures of actinic skin
    damage. Br J Dermatol, 110: 129-138.

    Holman CD, Evans PR, Lumsden GJ, & Armstrong BK (1984b) The
    determinants of actinic skin damage: problems of confounding among
    environmental and constitutional variables. Am J Epidemiol, 120:
    414-422.

    Holman CDJ, Armstrong BK, & Heenan PJ (1986) Relationship of
    cutaneous malignant melanoma to individual sunlight-exposure habits.
    J Natl Cancer Inst, 76: 403-414.

    Hoover R (1977) Effects of Rugs - Immunosuppression. In: Hiatt HH,
    Watson JD, & Winsten JA ed. Origins of human cancer. Book A:
    Incidence of cancer in humans. Proceedings of the Cold Spring Harbor
    Conferences on Cell Proliferation. Cold Spring Harbor, New York,
    Cold Spring Harbor Laboratory, vol 4.

    Hoppeler Th, Hendrickson Ph, Dietrich C, & Reme (1988) Morphology
    and time-course of defined photochemical lesions in the rabbit
    retina. Curr Eye Res, 7(9): 849-860.

    Howie S, Norval M, & Maingay J (1986) Exposure to low-dose
    ultraviolet radiation suppresses delayed-type hypersensitivity to
    herpes simplex virus in mice. J Invest Dermatol, 86: 125-128.

    Howie SEM, Ross JA, Norval M, & Maingay JP (1986)  In vivo
    modulation of antigen presentation generates Ts rather than TDH in
    HSV-1 infection. Immunology, 60: 419-423.

    Hsu J, Forbes PD, Harber LC, & Lakow E (1975) Induction of skin
    tumors in hairless mice by a single exposure to UV radiation.
    Photochem Photobiol, 21: 185-188.

    Hughes BR, Cunliffe WJ, & Bailey CC (1989) Excess benign melanocytic
    naevi after chemotherapy for malignancy in childhood. Br Med J, 299:
    88-91.

    Hunter DJ, Colditz GA, Stampfer MJ, Rosner B, Willett WC, & Speizer
    FE (1990) Risk factors for basal cell carcinoma in a prospective
    cohort of women. Ann Epidemiol, 1: 13-23.

    Hunter DJ, Colditz GA, Stampfer MJ, Rosner B, Willett WC, & Speizer
    FE (1992) Diet and risk of basal cell carcinoma of the skin in a
    prospective cohort of women. Ann Epidemiol, 2: 231-239.

    Husain Z, Yang Q, & Biswas DK (1990) cHa-ras Proto oncogene.
    Amplification and overexpression in UVB-induced mouse skin
    papillomas and carcinomas. Arch Dermatol, 126: 324-330.

    Husain Z, Pathak MA, Flotte T, & Wick MM (1991) Role of ultraviolet
    radiation in the induction of melanocytic tumors in hairless mice
    following 7,12-dimethylbenz(a)anthracene application and ultraviolet
    irradiation. Cancer Res, 51: 4964-4970.

    Hyman LG, Lilienfeld AM, Ferris FL, & Fine SL (1983) Senile macular
    degeneration: a case-control study. Am J Epidemiol, 118: 213-227.

    IARC (1993) In: Kricker A, Armstrong BK, Jones ME & Burton RD ed.
    Health, solar UV radiation and environmental change. Lyon,
    International Agency for Research on Cancer (IARC Technical Report
    No. 13).

    IARC (1992) Solar and ultraviolet radiation. Lyon, International
    Agency for Research on Cancer (Monographs on the Evaluation of
    Carcinogenic Risks to Humans, Volume 55).

    IARC (1980) International Agency for Research on Cancer, Monographs
    on the Evaluation of the Carcinogenic Risk of Chemicals to Humans,
    Vol 24, Some Pharmaceutical Drugs, Lyon, 101-124.

    Ignatiades L (1990) Photosynthetic capacity of the surface
    microlayer during the mixing period. J Planktonic Res, 12: 851-860.

    Imlay JA, Chir SM, & Linn S (1988) Science, 240: 640-642.

    IRPA/INIRC (1985) International Radiation Protection Association/
    International Non-Ionizing Radiation Committee. Concepts, units and
    terminology for NIR protection. Health Phys, 49(6): 1329-1362.

    IRPA/INIRC (1991a) International Radiation Protection Association/
    International Non-Ionizing Radiation Committee. Health issues of
    ultraviolet A sunbeds used for cosmetic purposes. Health Phys,
    61(2): 285-288.

    IRPA/INIRC (1991b) International Radiation Protection Association/
    International Non-Ionizing Radiation Committee. In: Duchêne AS,
    Lakey JRA, & Repacholi MH ed. IRPA guidelines on protection against
    non-ionizing radiation., New York, McGraw-Hill.

    Isaacson C, Selzer G, Kaye V, Greenberg M, Woodruff JD, Davies J,
    Ninin D, Vetten D, & Andrew, M (1978) Cancer in the urban blacks of
    South Africa. S Afr Cancer Bull, 22: 49-84.

    Ishizaki K, Tsujimura T, Nakai M, Nishigori C, Sato K, Katayama S,
    Kurimura O, Yoshikawa K, Imamura S, & Ikenaga M (1992) Infrequent
    mutation of the  ras genes in skin tumors of Xeroderma Pigmentosum
    patients in Japan. Int J Cancer, 50: 382-385.

    Italian-American Cataract Study Group (1991) Risk factors for
    age-related cortical, nuclear, and posterior subcapsular cataracts.
    Am J Epidemiol, 133: 541-553.

    Ito A & Ito T (1983) Possible involvement of membrane damage in the
    inactivation by broad-band near-UV radiation in Saccharomyces
    cerevisiae cells. Photochem Photobiol 37: 395-401.

    Iversen OH (1988) Skin tumorigenesis and carcinogenesis studies with
    7,12-dimethylbenz[a]anthracene, ultraviolet light, benzoyl peroxide
    (Panoxyl gel 5%) and ointment gel. Carcinogenesis, 9: 803-809.

    Jagger J (1985) Solar-UV actions on living cells. New York, Praeger.

    Jeevan A & Kripke ML (1990) Alteration of the immune response to
    Mycobacterium bovis BCG in mice exposed chonically to low doses of
    UV radiation. Cell Immunol, 130: 32-41.

    Jeevan A, Evans R, Brown EL, & Kripke ML (1992a) Effect of local
    ultraviolet irradiation on infections of mice with Candida albicans,
    Mycobacterium bovis BCG, and Schistosoma mansoni. J Invest Dermatol,
    99: 59-64.

    Jeevan A, Gilliam K, Heard H, & Kripke ML (1992b) Effects of
    ultraviolet radiation on the pathogenesis of Mycobacterium
    lepraemurium infection in mice. Exp Dermatol, 1: 152-160.

    Jeevan A, Ullrich SE, Dizon V, & Kripke ML (1992c) Supernantants
    from ultraviolet-irradiated keratinocytes decrease the resistance
    and delayed-type hypersensitivity response to Mycobacterium bovis
    bacillus Calmette-Guerin in mice and impair the phagocytic ability
    of macrophages. Photodermatol Photoimmunol Photomed, 9: 255-263.

    Jeffery SW & Humphrey GH (1975) New spectrophotometric equations for
    determining chlorophylls a, b, c1 and C2 in higher plants,
    algae, and natural phytoplankton. Biochem Physiol Pflanz, 167:
    191-194.

    Jensen OM & Bolander (1980) Trends in malignant melanoma of the
    skin, WHO Stat Q, 33: 2-26.

    Johnson GJ (1981) Aetiology of spheroidal degeneration of the cornea
    in Labrador. Br J Ophthalmol, 65: 270-283.

    Johnson GJ & Overall M (1978) Histology of spheroidal degeneration
    of the cornea in Labrador. Br J Ophthalmol, 62: 53-61.

    Johnson GJ, Paterson GD, Green JS, & Perkins ES (1981) Ocular
    conditions in a labrador community In: Harvald B & Hansen JP ed.
    Circumpolar health 81. Copenhagen, Nordic Council for Arctic Medical
    Research.

    Johnson GJ, Minassian DC, & Franken S (1989) Alterations of the
    anterior lens capsule associated with climatic keratopathy. Br J
    Ophthalmol, 73: 229-234

    Jones CA, Huberman E, Cunningham ML, & Peak MJ (1987a) Mutagenesis
    and cytotoxicity in human epithelial cells by far- and

    near-ultraviolet radiations : action spectra. Radiat Res, 110:
    244-254.

    Jones SK, Moseley H, & Mackie RM (1987b) UVA-induced melanocytic
    lesions. Br J Dermatol, 117: 111-115.

    Jones ME, Shugg D, Dwyer T, Young B, & Bonett A (1992) Interstate
    differences in incidence and mortality from melanoma: a
    re-examination of the latitudinal gradient. Med J Aust, 157:
    373-377.

    Jose JG (1986) Posterior cataract induction by UVB radiation in
    albino mice. Exp Eye Res, 42: 11-20.

    Jose JG & Pitts DG (1985) Wavelength dependency of cataracts in
    albino mice following chronic exposure. Exp Eye Res, 41: 545-563.

    Kalimo K, Loulu L, & Jansen CT (1983) Effect of a single UVB or PUVA
    exposure on immediate and delayed skin hypersensitivity reactions in
    humans. Correlation to erythemal response and Langerhans cell
    depletion. Arch Dermatol Res, 275: 374-378.

    Karai I & Horiguchi S (1984) Pterygium in welders. Br J Ophthalmol,
    68: 347-349.

    Karentz D & Lutze LH (1990) Evaluation of biologically harmful
    ultraviolet radiation in Antarctica with a biological dosimeter
    designed for aquatic environments. Limnol Oceanogr, 35: 549-561.

    Karentz D, Mc Euen FS, Land MC, & Dunlap WC (1991) Survey of
    microsporine-like amino acid compounds in Antarctic marine
    organisms: potential protection from ultraviolet exposure. Mar Biol,
    108: 157-166.

    Karjalainen S, Salo H, & Teppo L (1989) Basal cell and squamous cell
    carcinoma of the skin in Finland Int J Dermatol, 28: 445-450.

    Kataoka H & Fujiwara Y (1991) UV damage specific protein in
    xeroderma pigmentosum complementation group E. Biochem Biophys Res
    Commun, 175: 1139-1143.

    Katz L, Ben-Tuvia S, & Steinitz R (1982) Malignant melanoma of the
    skin in Israel: effect of migration. In: Magnus K ed Trends in
    cancer incidence: Causes and practical implication. Washington, New
    York, Hemisphere, pp 419-426.

    Kelfkens G, van Weelden H, de Gruijl FR, & van der Leun JC (1991)
    Influence of dose rate in ultraviolet tumorigenesis. J Photochem
    Photobiol, B10: 41-50.

    Kelfkens G & van der Leun JC (1989) Skin temperature changes after
    irradiation with UVB or UVC: Implications for the mechanism
    underlying ultraviolet erythema. Phys Med Biol, 34: 599-608.

    Keller, AZ (1970) Cellular types, survival, race, nativity,
    occupations, habits and associated diseases in the pathogenesis of
    lip cancers. Am J Epidemiol, 91, 486-499.

    Kelly, GE, Meikle, WD, & Moore, DE (1989) Enhancement of UV-induced
    skin carcinogenesis by azathioprine: role of photochemical
    sensitisation. Photochem Photobiol, 49, 59-65.

    Kelly GE, Meikle WD, & Sheil AGR (1987) Effects of immunosuppressive
    therapy on the induction of skin tumors by ultraviolet irradiation
    in hairless mice. Transplantation, 44, 429-434.

    Kennedy JC, Poltier RH, & Pross DC (1990) Photodynamic therapy with
    endogenous protoporphyrin. IX. Basic principles and present clinical
    experience. J Photochem Photobiol, 6: 143-148.

    Keyse SM (1993) The induction of gene expression in mammalian cells
    by radiation. Sem Cancer Biol, 4: 119-128.

    Keyse SM & Tyrrell RM (1989) Heme oxygenase is the major 32-kDa
    stress protein induced in human skin fibroblasts by UVA radiation,
    hydrogen peroxide and sodium arsenite. Proc Natl Acad Sci (USA), 86:
    99-103.

    Khlat M, Vail A, Parkin M, & Green A (1992) Mortality from melanoma
    in migrants to Australia: variation by age at arrival and duration
    of stay. Am J Epidemiol, 135: 1103-1113.

    Kim TY, Kripke ML, & Ullrich SE (1990) Immunosuppression by factors
    released from UV-irradiated epidermal cells: Selective effects on
    the generation of contact and delayed hypersensitivity after
    exposure to UVA or UVB radiation. J Invest Dermatol, 94: 26-32.

    Kinlen LJ, Sheil AGR, Peto J, & Doll R (1979) Collaborative United
    Kingdom-Australasian study of cancer in patients treated with
    immunosuppressive drugs. Br Med J, 2: 1461-1466.

    Klamen DK & Tuveson RW (1982) The effect of membrane fatty acid
    compositions on the near-UV (300-400 nm) sensitivity of Escherichia
    coli K1060. Photochem Photobiol, 35: 167-173.

    Klepp O & Magnus K (1979) Some environmental and bodily
    characteristics of melanoma patients. A case-control study. Int J
    Cancer, 23: 482-486.

    Kligman LH, Kaidbey KH, Hitchens VM, & Miller SA (1987) Long
    wavelength (>340 nm) ultraviolet-A induced damage in hairless mice

    is dose dependent. In: Passchier W & Bosnjakovi, BFM ed., Human
    exposure to ultraviolet radiation: Risks and regulations. Amsterdam,
    Oxford, New York, Elsevier Science Publishers, pp 77-81.

    Knekt P, Aromaa A, Maatela J, Alfthan G, Aaran RK, Nikkari T, Hakama
    M, Hakulinen T, & Teppo, L (1991) Serum micronutrients and risk of
    cancers of low incidence in Finland. Am J Epidemiol, 134: 356-361.

    Kollias, N, Sayre RM, Zeise L, & Cehedekel MR (1991) Photoprotection
    by melanin, J Photochem Photobiol, 9: 135-160

    Kopecky KE, Pugh GW Jr, Hughes DE, Booth GD, & Cheville, NF (1979)
    Biological effect of ultraviolet radiation on cattle: bovine ocular
    squamous cell carcinoma. Am J Vet Res, 40: 1783-1788.

    Kopf AW, Lazar M, Bart RS, Dubin N, & Bromberg J (1978) Prevalence
    of nevocytic nevi on lateral and medial aspects of arms. J Dermatol
    Surg Oncol, 4: 153-158.

    Kopf AW, Lindsay AC, Rogers GS, Friedman RJ, Rigel DS, & Levenstein
    M (1985) Relationship of nevocytic nevi to sun exposure in
    dysplastic nevus syndrome. J Am Acad Dermatol, 12: 656-662.

    Kraemer KH, Lee MM, & Scotto J (1987) Xeroderma Pigmentosum Arch
    Dermatol, 123: 241-50.

    Kraemer KH, Seetharam S, Protic-Sabljic M, Brash DE, Bredberg A, &
    Seidman MM (1988) Defective DNA repair and mutagenesis by dimer and
    non-dimer photoproducts in xeroderma pigmentosum measured with
    plasmid vectors. In: Freidberg EC & Hanawalt PC ed. Mechanisms and
    consequences of DNA damage processing. New York, Alan R. Liss, pp
    325-335.

    Kramer GF & Ames BN (1987) Oxidative mechanisms of toxicity of
    low-intensity near-UV light in Salmonella typhimurium. J Bacteriol,
    164: 2259-2266.

    Kremers JJM & van Norren D (1988) Two classes of photochemical
    damage to the retina. In. Lasers and light in ophthalmology.
    Amsterdam, Berkeley, Milano, Klugler Publications, vpl 2, pp 41-52.

    Kress S, Sutter C, Strickland PT, Mukhtar H, Schweizer J, & Schwarz
    M (1992) Carcinogen-specific mutational pattern in the p53 gene in
    ultraviolet B radiation-induced squamous cell carcinomas of mouse
    skin. Cancer Res, 52: 6400-6403.

    Kricker A, Armstrong BK, English DR, & Heenan PJ (1991a) Pigmentary
    and cutaneous risk factors for non-melanocytic skin cancer - a
    case-control study. Int J Cancer, 48: 650-662.

    Kricker A, Armstrong BK, English D, Heenan PJ, & Randell PL (1991b)
    A case-control study of non-melanocytic skin cancer and sun exposure
    in Western Australia (Abstract No. III, P2). Cancer Res Clin Oncol,
    117(Suppl II): S75.

    Kricker A, English DR, Randell PL, Heenan PJ, Clay CD, Delaney TA, &
    Armstrong BK (1990) Skin cancer in Geraldton, Western Australia: a
    survey of incidence and prevalence. Med J Aust, 152: 399-407.

    Kricker A, Armstrong BK, Jones ME, & Burton RC (1993) Health, solar
    UV radiation and environmental change. IARC Technical Report No 13.
    Lyon, International Agency for Research on Cancer.

    Kripke ML (1974) Antigenicity of murine skin tumors induced by
    ultraviolet light. J Natl Cancer Inst, 53: 1333-1336.

    Kripke Ml & Fidler IJ (1980) Enhanced experimental metastasis of
    ultraviolet light-induced fibrosarcomas in ultraviolet
    light-iradiated syngeneic mice. Cancer Res, 40: 625-629.

    Kripke ML & Fisher MS (1976) Immunologic parameters of ultraviolet
    carcinogenesis. J Natl Cancer Inst, 57: 211-215.

    Kripke ML & Sass ER ed. (1978) International Conference on
    Ultraviolet Carcinogenesis. Natl Cancer Inst Monogr, 50: 16-17.

    Kripke ML, Lofgreen JS, Beard J, Jessup JM, & Fisher MS (1977)  in
     vivo immune responses of mice during carcinogenesis by ultraviolet
    irradiation. J Natl Cancer Inst, 59: 1227-1230.

    Kromberg JG & Jenkins T (1982) Prevalence of albinism in the South
    African Negro. S Afr Med J, 61: 383-386.

    Kromberg JGR, Castle D, Zwane EM, & Jenkins T (1989) Albinism and
    skin cancer in southern Africa. Clin Genet, 36: 43-52.

    Kulandaivelu G, Neducheshian N & Annamalainathan K (1993)
    Ultraviolet-B (280-320) radiation induced changes in photochemical
    activities and polypeptide components of C3 and C4 chloroplasts.
    Photosynthetica, 25:12-14.

    Kune GA, Bannerman S, Field B, Watson LF, Clelan H, Merenstein D, &
    Vitetta L (1992) Diet, alcohol, smoking, serum ß-carotene, and
    vitamin A in male nonmelanocytic skin cancer patients and controls.
    Nutr Cancer, 18: 237-244.

    Lancaster HO & Nelson J (1957) Sunlight as a cause of melanoma: a
    clinical survey. Med J Aust, 1: 452-456.

    Langer B & Wellmann E (1990) Phytochrome induction of
    photoreactivation in Phaseolus vulgaris L. seedlings. Photochem
    Photobiol, 52: 861-864.

    Larko O & Diffey BL (1983) Natural UVB radiation received by people
    with outdoor, indoor and mixed occupations and UVB treatment of
    psoriasis. Clin Exp Dermatol, 8: 279-285.

    Lassam NJ, From L, & Kahn HJ (1993) Overexpression of p53 is a late
    event in the development of malignant melanoma. Cancer Res, 53:
    2235-2238.

    Laycock KA, Lee SF, Brady RH, & Pepose JS (1991) Characterization of
    a murine model of recurrent herpes simples viral keratitis induced
    by ultraviolet B radiation. Invest Ophthalmol Vis Sci, 32:
    2741-2746.

    Lê, MG, Cabanes, PA, Desvignes, V, Chanteau, MF, Mlika, N, & Avril,
    MF (1992) Oral contraceptive use and risk of cutatneous malignant
    melanoma in a case-control study in French women. Cancer Causes
    Control, 3, 199-205.

    Leach, JF, McLeod, VE, Pingstone, AR, Davis, A and Deane, GHW (1978)
    Measurement of the ultraviolet doses received by office workers.
    Clinical and Experimental Dermatology, 3,77-79.

    Leclerc, JE, Borden, A and Lawrence, CW (1991) The thymine-thymine
    pyrimidine-pyrimidine (6-4) ultraviolet light photoproduct is highly
    mutagenic and specifically induces 3'-thymine-to-cytosine transition
    in Escherichia coli Proc Natl Acad Sci USA, 88: 9685-9689.

    Lee JAM (1989) The relationship between malignant melanoma of the
    skin and exposure to sunlight. Photochem Photobiol, 50: 493-496.

    Lee GA & Hirst LW (1992) Incidence of ocular surface epithelial
    dysplasia in metropolitan Brisbane: a 10-year survey. Arch
    Ophthalmol, 110: 525-527.

    Lerman S (1980) Human ultraviolet radiation cataracts. Ophthalmic
    Res, 12: 303-314.

    Leske MC, Chylack LT, Wu S, & The Lens Opacities Case-Control Study
    Group (1991) The lens opacities case-control study: risk factors for
    cataract. Arch Ophthalmol, 109: 244-251.

    Levi F, La Vecchia C, Te V-C, & Mezzanotte G (1988) Descriptive
    epidemiology of skin cancer in the Swiss canton of Vaud. Int J
    Cancer, 42: 811-816.

    Levine EA, Ronan SG, Shirali SS, & Das Gupta TK (1992) Malignant
    melanoma in a child with oculocutaneous ablinism. J Surg Oncol, 51:
    138-142.

    Lew RA, Sober AJ, Cook N, Marvell R, & Fitzpatrick TB (1983) Sun
    exposure habits in patients with cutaneous melanoma: a case control
    study. J Dermatol Surg Oncol, 9: 981-986.

    Lewis MG (1967) Malignant melanoma in Uganda (The relationship
    between pigmentation and malignant melanoma on the soles of the
    feet). Br J Cancer, 21: 483-495.

    Ley RD (1993) Photoreactivation in humans. Proc Natl Acad Sci (USA),
    90: 4337.

    Ley RD Peak MJ, & Lyon LL (1983) Induction of pyrimidine dimers in
    epidermal DNA of hairless mice by UVB: an action spectrum. J Invest
    Dermatol, 80: 188-191.

    Ley RD (1985) Photoreactivation of UV-Induced pyrimidine dimers and
    erythema in the marsupial Monodelphis domestica. Proc Natl Acad Sci
    (USA), 82: 2409-2411.

    Ley RD, Applegate LA, Stuart TD, & Fry RJM (1987) UV
    radiation-induced skin tumors in Monodelphis domestica.
    Photodermatology, 4: 144-147.

    Ley RD, Applegate LA, Padilla RS, & Stuart TD (1989) Ultraviolet
    radiation-induced malignant melanoma in Monodelphis domestica.
    Photochem Photobiol, 50: 1-5.

    Ley RD, Applegate LA, Fry RJM, & Sanchez AB (1991) Photoreactivation
    of ultraviolet radiation-induced skin and eye tumors of Monodelphis
    domestica. Cancer Res, 51: 6539-6542.

    Li YF, Kim ST, & Sancer A (1993) Evidence for lack of DNA
    photoreactivity enzyme in humans. Proc Natl Acad Sci (USA), 90:
    4389-4393.

    Lieu F-M, Yamanishi K, Jonishi K, Kishimoto S, & Yasuno, H (1991)
    Low incidence of Ha-ras oncogene mutations in human epidermal
    tumors. Cancer Lett, 59: 231-235.

    Lill PH (1983) Latent period and antigenicity of murine tumors
    induced in C3H mice by short-wavelength ultraviolet radiation. J
    Invest Dermatol, 81: 342-346.

    Lindberg JG (1917) Clinical study on iris depigmentation and on
    transparency of iris in cataract and in normal eyes among elderly
    persons. Dissertation, Helsinki, Finland (in Swedish).

    Lindelöf B, Sigurgeirsson B, Tegner E, Larkö O, Johannesson A, Berne
    B, Christensen OB, Andersson T, Törngren M, Molin L,
    Nylander-Lundqvist E, & Emtestam L (1991) PUVA and cancer; a
    large-scale epidemiological study. Lancet, 338: 91-93.

    Lindqvist C (1979) Risk factors in lip cancer: a questionnaire
    survey. Am J Epidemiol, 109: 521-530.

    Lippke JA, Gordon LK, Brash D, & Haseltine WA (1981) Distribution of
    UV light-induced damage in a defined sequence of human DNA:
    Detection of alkaline-sensitive lesions at pyrimidine
    nucleoside-cytidine sequences. Proc Natl Acad Sci (USA), 78:
    3388-3392.

    Lischko AM, Seddon JM, Gragoudas ES, Egan KM, & Glynn RJ (1989)
    Evaluation of prior primary malignancy as a determinant of uveal
    melanoma. A case-control study. Ophtalmology, 96: 1716-1721.

    Logani MK, Sambuco CP, Forbes PD, & Davies RE (1984) Skin-tumour
    promoting activity of methyl ethyl ketone peroxide - a potent
    lipid-peroxidizing agent. Food chem Toxicol, 22: 879-882.

    Luande J, Henschke CI, & Mohammed N (1985) The Tanzanian human
    albino skin. Cancer, 55: 1823-1828.

    Lubin D, Frederick JE, Booth CR, Lucas T, & Neuschuler D (1989)
    Measurements of enhanced springtime ultraviolet radiation at Palmer
    Station, Antarctica. Geophys Res Lett, 16: 783-785.

    Lynch DH & Daynes RA (1983) Evaluation of naturally occurring
    cell-mediated cytotoxic activity in normal and UV-irradiated mice.
    Transplantation, 35: 216-223.

    Lynge E & Thygesen L (1990) Occupational cancer in Denmark. Cancer
    incidence in the 1970 census population. Scand J Work Environ
    Health, 16(Suppl 2): 1-35.

    Lytle CD, Miller SA, Jacobs ME, Cyr WH, James RH, Kaczmarek RG, Beer
    JZ, Landry RJ, Sharkness CM, Gaylor D, & De Gruijl FR (1993) An
    estimation of squamous cell carcinoma risk from ultraviolet
    radiation emitted by fluorescent lamps. In : Proceedings of the 1993
    Meeting of the American Society for Photobiology, Chicago, 26-30
    June 1993.

    McCormick JP, Fisher JR, Pachlatko JP, & Eisenstark A (1976)
    Characterization of a cell lethal product from the photooxidation of
    tryptophan: hydrogen peroxide. Science, 198: 468-469.

    McCredie M & Coates MS (1989) Cancer incidence in migrants to New
    South Wales; 1972 to 1984, New South Wales Central Cancer Registry.
    Woolloomooloo, New South Wales Cancer Council, pp 22-23, 62-63.

    MacDonald EJ (1948) Malignant melanoma in Connecticut. In: The
    biology of melanomas. New York, Academy of Sciences, pp 71-81
    (Special publication of the Academy of Sciences No. 4).

    McGregor JM, Barker JNWN, & MacDonald DM (1992) The development of
    excess numbers of melanocytic naevi in an immunosuppressed identical
    twin. Clin Exp Dermatol, 16: 131-132.

    Mack TM, & Floderus B (1991) Malignant melanoma risk by nativity,
    place of residence at diagnosis, and age at migration. Cancer Causes
    Control, 2: 401-411.

    Mackenzie FD, Hirst LW, Battistutta D, & Green A (1992) Risk
    analysis in the development of pterygia. Ophthalmol, 99: 1056-1061.

    MacKie RM & Aitchison T (1982) Severe sunburn and subsequent risk of
    primary cutaneous malignant melanoma in Scotland. Br J Cancer, 40:
    55-960.

    MacKie RM, Freudenberger T, & Aitchison TC (1989) Personal
    risk-factor chart for cutaneous melanoma. Lancet, ii: 487-490.

    McKinlay AF (1992) Artificial sources of UVA radiation. Loss and
    emission characteristics. In: Urbach F ed Biological responses to
    UVA radiation. Vandermor Pub Corp Kansas.

    McKinlay AF & Diffey BL (1987) A reference action spectrum for
    ultra-violet induced erythema in human skin. CIE J, 6: 17-22

    McKinlay AF & Whillock MJ (1987) Measurements of ultraviolet
    radiation from fluorescent lamps used for general lighting and other
    purposes in the UK. In: Passchhier WF & Bosnjakovic BFM ed. Human
    exposure to ultraviolet radiation: Risks and regulations.,
    Amsterdam, Oxford, New York, Elsevier Science Pubishers, pp 253-258.

    McKinlay AF, Whillock MJ, & Meulemans CCE (1989) Ultraviolet
    radiation and blue-light emissions from spotlights incorporating
    tungsten halogen lamps. London, National Radiological Protection
    Board (Report NRPB No. R228).

    McKinlay AF, Harlen F, & Whillock MJ (1988) Hazards of optical
    radiation: A guide to sources, uses and safety. Bristol,
    Philadelphia, Adam Hilger.

    Madewell BR, Conroy JD, & Hodgkins EM (1981) Sunlight-skin cancer
    association in the dog: a report of three cases. J Cutan Pathol, 8:
    434-443.

    Magnus K (1977) Incidence of malignant melanoma of the skin in 5
    Nordic countries: significance of solar radiation. Int J Cancer, 20:
    477.

    Magnus K (1986) Malignant melanoma in Norway. Tidsskr Nor
    Lacgeforen, 106: 2309-2313.

    Magnus K (1991) The Nordic profile of skin cancer incidence. A
    comparative epidemiological study of the three main types of skin
    cancer. Int J Cancer, 47: 12-19.

    Mandal TK & Chatterjee SN (1980) Ultraviolet- and sunlight-induced
    lipid peroxidation in liposomal membrane. Radiat Res, 83: 290-302.

    Mao W & Hu T (1982) An epidemiologic survey of senile cataract in
    China. Chin Med J, 95: 813-818.

    Marks R & Hill D ed. (1992) The public health approach to melanoma
    control: Prevention and early detection. Geneva, International Union
    Against Cancer (UICC Publication).

    Marks R, Ponsford MW, Selwood TS, Goodman G, & Mason G (1983)
    Non-melanocytic skin cancer and solar keratoses in Victoria. Med J
    Aust, 2: 619-622.

    Marks R, Jolley D, Dorevitch AP, & Selwood TS (1989) The incidence
    of non-melanocytic skin cancers in an Australian population: results
    of a five-year prospective study. Med J Aust, 150: 475-478.

    Marks R, Staples M, & Giles GG (1993) Trends in non melanocytic skin
    cancer treated in Australia: the second national survey. Int J
    Cancer, 53: 585-590.

    Martin EK (1912) The effects of ultra-violet rays upon the eye. Proc
    of the Soc Ser B, 85: 319-334.

    Mathews-Roth MM & Krinsky NI (1987) Carotenoids affect development
    of UV-B induced skin cancer. Photochem Photobiol, 46: 507-509.

    Matsunaga T, Hieda K, & Nikaido O (1991) Wavelength dependent
    formation of thymine dimers and (6-4) photoproducts in DNA by
    monochromatic ultraviolet light ranging from 150 to 365 nm.
    Photochem Photobiol, 54: 403-410.

    Matsuoka LY, Wortsman J, Hanifan N, & Holick MF (1988) Chronic
    sunscreen use decreases circulating concentration of 25-hydroxy
    vitamin D. A preliminary study. Arch Dermatol, 124: 1802-1804.

    Mayne LV, Mullenders LHF, & vasn Zeeland AA (1988) Cockayne
    syndrome: a UV sensitive disorder with a defect in the repair of
    tanscribing DNA but normal overall excision repair. In: Friedberg EC
    & Hanawalt PC ed. Mechanisms and consequences of DNA damage
    processing. New York, Alan R. Liss, pp 349-353.

    Menzies, SW, Greenoak, GE, Reeve, VE, & Gallagher, CH (1991)
    Ultraviolet radiation-induced murine tumors produced in the absence
    of ultraviolet radiation-induced systemic tumor immunosuppression.
    Cancer Res, 51: 2773-2779.

    Miguel AG & Tyrrell RM (1983) Induction of oxygen-dependent lethal
    damage by monochromatic UVB (313 nm) radiation: strand breakage,
    repair and cell death. Carcinogenesis 4, 375-380.

    Milham S Jr (1983) Occupational mortality in Washington State
    1950-1979. Cincinnati, Ohio, National Institute for Occupational
    Safety and Health (DHSS (NIOSH) Publication No. 83-116).

    Minassian DC, Mehra V, & Johnson GJ (1992) Mortality and cataract:
    findings from a population-based longitudinal study. Bull World
    Health Organ, 70: 219-223.

    Miskin R & Ben-Ishai R (1981) Induction of placmenogen activator by
    UV light in normal and Xeroderma pigmentosum fibroblasts. Proc Natl
    Acad Sci (USA), 78: 6236-6240.

    Mitchell DL, Pfeifer GP, Taylor JS, Zdienicka MZ, & Nikaido O (1993)
    Biological role of (6-4) photoproducts and cyclobutane dimers. In:
    Shima A, Ichihashi M, Fujiwara Y, & Takebe H ed. Frontiers of
    photobiology. Amsterdam, Oxford, New York, Elsevier Science
    Publishers, pp 337-344.

    Mitchell DL & Rosenstein BS (1987) The use of specific
    radioimmunoassays to determine action spectra for the photolysis of
    (6-4) photoproducts. Photochem Photobiol, 45: 781-786.

    Mitchell RL & Andersen IC (1965) Catalase Photoinactivation. Science
    150: 74.

    Moan J & Dahlback A (1992) The relationship between skin cancers,
    solar radiation and ozone depletion. Br J Cancer, 65: 916-921.

    Modan B, Alfandary E, Shapiro D, Lusky A, Chetrit A, Shewach-Millet
    M, & Moushovitz M (1993) Factors affecting the development of skin
    cancer after scalp irradiation. Radiat Res, 134: 125-128.

    Mohan M, Sperduto RD, Angra SK, Milton RC, Mathur RL, Underwood BA,
    Jaffery N, Pandya CB, Chhabra VK, Vajpayee RB, Kalra VK, Sharma YR,
    & The Indian-US Case-Control Study Group (1989) India-US
    case-control study of age-related cataracts. Arch Ophthalmol, 107:
    670-676.

    Moan J, Dahlbach A, Henriksen T, & Magnus K (1989) Biological
    amplification factor for sunlight-induced nonmelanoma skin cancer at
    high latitudes. Cancer Res, 49: 5207-5212.

    Molès J-P, Moyret C, Guillot B, Jeanteur P, Guilhou J-J, Theillet C,
    & Basset-Sèguin N (1993) p53 Mutations in human epithelial skin
    cancers. Oncogene, 8: 583-588.

    Molesworth EH (1927) Rodent ulcer. Med J Aust, 1: 878-899.

    Moller H, Mellemgaard A, Jacobsen GK, Pedersen D, & Storm HH (1993)
    Incidence of second primary cancer following testicular cancer. Eur
    J Cancer, 29A: 672-676.

    Mooy CM, Van der Helm MJ, Van der Kwast ThH, De Jong PTVM, Ruiter
    DJ, & Zwarthoff EC (1991) No N- ras mutations in human uveal
    melanoma: the role of ultraviolet light revisited. Br J Cancer, 64:
    411-413.

    Moran DJ & Hollows FC (1984) Pterygium and ultraviolet radiation: a
    positive correlation. Br J Ophthalmol, 68: 343-346.

    Morimoto S & Kunihiko Y (1989) Psoriasis and vitamin D: A review of
    our experience. Arch Dermatol, 125: 231-234.

    Morin RJ, Hu B, Peng S-K, & Sevanian A (1991) Cholesterol oxides and
    carcinogenesis. J Clin Lab Anal, 5: 219-225.

    Morison WL (1984) The effect of a sunscreen containing
    para-aminobenzoic acid on the systemic immunologic alterations
    induced in mice by exposure to UVB radiation. J Invest Dermatol, 83:
    405-408.

    Morlière P, Mayson A, Santus R, Huppe G, Mazière JP, & Dubertet L
    (1991) UVA-induced lipid peroxidation in cultured human fibroblasts.
    Biochem Biophys Acta, 1084(3): 261-268.

    Morlière P, Moysan A, Gaboriou F, Santus R, Mazière JC, & Dubertret
    L (1992) Ultraviolet A et la peau: implications d'espèces activées
    de l'oxygène, tendances actuelles et résultats récents. Pathol Biol,
    40: 160-168.

    Moseley H (1988) Hospital Physicists' Association - Non-ionizing
    radiation, medical physics handbook 18. Bristol, Philadelphia, Adam
    Hilger Publishers.

    Mountin JW & Dor H(1939) Some peculiarities in the geography of
    cancer. J Am Med Assoc, 113: 2405-2410.

    Muir C, Waterhouse J, Mack T, Powell J, & Whelan S ed. (1987) Cancer
    incidence in five continents, Vol V (IARC Scientific Publications No
    88), Lyon, International Agency for Research on Cancer.

    Munkata N (1993) Biological effective dose of solar ultraviolet
    radiation estimated by spore dosimetry in Tokyo since 1989.
    Photochem Photobiol, 58(3): 386-392.

    Munch-Petersen B, Frentz G, Squire B, Wallevik K, Horn CC, Reymann
    F, & Faber M (1985) Abnormal lymphocyte response to uv radiation in
    multiple skin cancer. Carcinogenesis, 6: 843-845.

    Murali NS & Teramura AH (1987) Insensitivity of soyabean
    photosynthesis to UVB radiation under phosphorus deficiency. J Plant
    Nutr, 10: 501-515.

    Murphy GF, Walsh Lj, Kaidby K, & Lavker RM (1991) UV and sunbeds
    (Abstract) Clin. Res. 39,195.

    Murphy GM, Wright J, Nicholls DSH, McKee PH, Messenger AG, Hawk JLM,
    & Levene GM (1989) Sunbed induced pseudoporphyria. Br J Dermatol,
    120: 555-562.

    NSF (1993) National Science Foundation, Div Polar Programs, Report
    to P Penhale (NSF) & CR Booth (Biospherical Instruments), Antarctic
    Support Assoc US.

    Nachtwey DB & Rundel RD (1981) A photobiological evaluation of
    tanning booths. Science, 211: 405-407.

    Nakazawa H, English D, Randell PL, Nakazawa K, Martel N, Armstrong
    BK, & Yamasaki H (in press) UV and skin cancer; specific p53 gene
    mutation in normal skin as a biologically relevant exposure
    measurement. Proc Natl Acad Sci (USA).

    Naumann GOH & Apple D (1986) Pathology of the eye. Berlin,
    Heidelberg, New York, Springer-Verlag.

    Nelemans PJ, Groenendal H, Kiemeney LALM, Rampen FHJ, Ruiter DJ, &
    Verbeek ALM (1993) Effect of intermittent exposure to sunlight on
    melanoma risk among indoor workers and sun sensitive individuals.
    Environ Health Perspect, 101(3): 252-255.

    Nelson EW, Eichwald EJ, & Shelby J (1987) Increased ultraviolet
    radiation-induced skin cancers in cyclosporine-treated mice. Transpl
    Proc, 19: 526-527.

    NIH (1991) Executive summary: Guidelines for the diagnosis and
    management of asthma. Bethesda, Maryland, National Institutes of
    Health, US Department of Health and Human Services (Publication No.
    91-3042A).

    Noell WK, Walker VS, & Kang BS (1966) Retinal damage by light in
    rats. Invest Ophthalmol Vis Sci, 5: 450-473.

    Noonan FP & De Fabo EC (1990) Ultraviolet-B-dose-response curves for
    local and systemic immunosuppression are identical. Photochem
    Photobiol, 52, 801-810.

    Noonan FP, Kripke ML, Pedersen GM, & Green MI (1981a) Suppression of
    contact hypersensitivity in mice by ultraviolet irradiation is
    associated with defective antigen presentation. Immunology, 43:
    527-533.

    Noonan FP, De Fabo EC, & Kripke ML (1981b) Suppression of contact
    hypersensitivity by UV radiation and its relationship to UV-induced
    suppression of tumor immunity. Photochem Photobiol, 34: 683-689.

    Noonan FP, De Fabo EC, & Morrison H (1988) Cis-urocanic acid, a
    product formed by ultraviolet B irradiation of the skin, initiates
    an antigen presentation defect in splenic dendritic cells  in vivo.
    J Invest Dermatol, 90: 92-99.

    Norbury KC, Kripke ML, & Budmen M (1977)  in vitro reactvity of
    macrophages and lymphocytes from ultraviolet-irradiated mice. J Natl
    Cancer Inst, 59: 1231-1235.

    Norn MS (1982) Spheroid degeneration, pinguecula, and pterygium
    among Arabs in the Red Sea Territory, Jordan. Acta Ophthalmol, 60:
    949-954.

    Norval M, Howie SE, Ross JA, & Maingay JP (1987) A murine model of
    herpes simplex virus recrudescence. J Gen Virol, 68: 2693-2698.

    Nultsch W, Pfan J, & Huppertz K (1990) Photoinhibition of
    photosynthetic oxygen production and its recovery in the subtidal
    red data Polyneura hilliae, Bot. Acta, 103: 62-67.

    O'Beirn SF, Judge P, Urbach F, MacCon CF, & Martin F (1970) Skin
    cancer in County Galway, Ireland. Proc Natl Cancer Conf, 6: 489-500.

    Oberhelman L, Koren H, & Cooper KD (1992) Depressed capacity to
    contact sensitiza women at the onset of menses; elevated contact
    sensitization capacity during mid cycle can be inhibited by UV. Clin
    Res, 40: 542A.

    O'Dell BL, Jessen RT, Becker LE, Jackson RT, & Smith EB (1980)
    Diminished immune response in sun-damaged skin. Arch Dermatol, 116:
    559-561.

    Oettlé,AG (1963) Skin cancer in Africa. Natl Cancer Inst Monogr, 10:
    197-214.

    Office of Population, Censuses and Surveys (1986) Occupational
    mortality: the registrar general's decennial, supplement for Great

    Britain 1979-80, 1982-83. London, Office of Population Censuses and
    Surveys, Her Majesty's Stationary Office (Report Series DS No. 6).

    Oldenburg JB, Gritz DC, & McDonnell PJ (1990) Topical ultraviolet
    light-absorbing chromophore protects against experimental
    photokeratitis. Arch Ophthalmol, 108: 1142-1144.

    O'Loughlin C, Moriarty MJ, Herity B, & Daly, L (1985) A re-
    appraisal of risk factors for skin carcinoma in Ireland: a case
    control study. Ir J Med Sci, 154: 61-65.

    Olsen JH & Jensen OM (1987) Occupation and risk of cancer in
    Denmark. An analysis of 93 810 cancer cases, 1970-1979. Scand J Work
    Environ Health, 13(Suppl 1): 1-91.

    Oluwasanmi JO, Williams AO, & Alli AF (1969) Superficial cancer in
    Nigeria. Br J Cancer, 23: 714-728.

    Omdahl JR, Garry TJ, & Hunsaker A (1982) Nutritional status in the
    healthy elderly population: vitamin D. Am J Clin Nutr, 6: 1225-1233.

    Organisciak DT & Winkler BS (1993) Retinal light damage: Practical
    and theoretical considerations. In: Osborne N & Chader G ed.
    Progress in retinal research. Oxford, New York, Seoul, Tokyo,
    Pergamon Press.

    Orth AB, Teramura AH, & Sisler HD (1990) Effects of UV-B radiation
    on fungal disease development in Cucumis sativus. Am J Bot, 77:
    1188-1192.

    Ostendfeld-Åkerblom A (1988) Pseudoexfoliation in Eskimos (Inuit) in
    Greenland. Acta Ophthalmol, 66: 467-468.

    Osterlind A (1987) Trends in incidence of ocular malignant melanoma
    in Denmark 1943-1982. Int J Cancer, 40: 161-164.

    Osterlind A, Olsen JH, Lynge E, & Ewertz M (1985) Second cancer
    following cutaneous melanoma and cancers of the brain, thyroid,
    connective tissue, bone, and eye in Denmark, 1943-80. Natl Cancer
    Inst Monogr, 68: 361-388.

    Osterlind A, Hou-Jensen K, & Jensen OM (1988a) Incidence of
    cutaneous malignant melanoma in Denmark 1978-1982. Anatomic site
    distribution, histologic types, and comparison with non-melanoma
    skin cancer. Br J Cancer, 58: 385-391.

    Osterlind A, Tucker MA, Stone BJ, & Jensen OM (1988b) The Danish
    case-control study of cutaneous malignant melanoma. II. Importance
    of UV-light exposure. Int J Cancer, 42: 319-324.

    Otani T & Mori R (1987) The effects of UV irradiation of the skin on
    herpes simplex virus infection: Alteration in immune function
    mediated by epidermal cells and in the course of infection. Arch
    Virol, 96(1-2): 1-15.

    Oxholm A, Oxholm P, Staberg B, & Bendtzen K (1988)
    Immunohistological detection of interleukin-like molecule and tumor
    necrosis factor in human epidermis before and after UVB-irradiation
     in vivo. Br J Dermatol, 118: 369-376.

    Paffenbarger RS Jr, Wing AL, & Hyde RT (1978) Characteristics in
    youth predictive of adult-onset malignant lymphomas, melanomas, and
    leukemias: brief communication. J Natl Cancer Inst, 60: 89-92.

    Paksoy N, Bouchardy C, & Parkin DM (1991) Cancer incidence in
    Western Samoa. Int J Epidemiol, 20: 634-641.

    Pang Q & Hays JB (1991) UV-B inducible and temperature- sensitive
    photoreactivation of cyclobutane pyrimidine dimers in Arabidopsis
    thaliana. Plant Physiol, 95: 536-543.

    Parkin DM, Muir CS, Whelan SL, Gao Y-T, Ferlay J, & Powell J ed.
    (1992) Cancer incidence in five continents, Vol VI (IARC Scientific
    Publications No. 120), Lyon, International Agency for Research on
    Cancer.

    Parrish JA, Jaenicke KF, & Anderson RR (1982) Erythema and
    melanogenesis action spectra of normal human skin. Photochem
    Photobiol, 36: 187-191.

    Patrick MH (1977) Studies on thymine-derived UV photoproducts in DNA
    - I. Formation and biological role of pyrimidine adducts in DNA.
    Photochem Photobiol, 25: 357-372.

    Peak MJ & Peak JG (1982) Single-strand breaks induced in  Bacillus
     subtilis DNA by ultraviolet light: action spectrum and properties.
    Photochem Photobiol, 35: 675-680.

    Peak JG & Peak MJ (1990) Ultraviolet light induces double- strand
    breaks in DNA of cultured human P3 cells as measured by neutral
    filter elution. Photochem Photobiol, 52: 387-393.

    Peak JG & Peak MJ (1991) Comparison of initial yields of
    DNA-to-protein crosslinks and single-strand breaks induced in
    cultured human cells by far-and near-ultraviolet light, blue light
    and x-rays, Mutat Res, 246: 187-191.

    Peak MJ, Peak JG, Moehring MP, & Webb RB (1984) Ultrviolet action
    spectrum for DNA dimer induction, lethality, and mutagenesis in
     Escherichia coli with emphasis on the UVB region. Photochem
    Photobiol, 40: 613-620.

    Peak MJ, Peak JG, Sikorski RS, & Jones CA (1985a) Induction of
    DNA-protein crosslinks in human cells by ultraviolet and visible
    radiations: action spectrum. Photochem Photobiol, 41: 295-302.

    Peak MJ, Peak JG, & Jones CA (1985b) Different (direct and indirect)
    mechanisms for induction of DNA-protein crosslinks in human cells by
    far- and near-ultraviolet radiation (290 and 405 nm). Photochem
    Photobiol, 42: 141-146.

    Peak MJ, Peak JG, & Carnes BA (1987) Induction of direct and
    indirect single-strand breaks in human cell DNA by far and
    near-ultraviolet radiations: action spectrum and mechanisms.
    Photochem Photobiol, 45: 381-387.

    Perna JJ, Mannix ML, Ronney JF, Notkins AL, & Straus SE (1987)
    Reactivation of latent herpes simplex virus infection by ultraviolet
    light: A human model. J Am Acad Dermatol, 17(3): 473-478.

    Philips (1983) Lighting, comprehensive handbook including technical
    section. Croydon, Philips Lighting.

    Pierceall WE, Mukhopadhyay T, Goldberg LH, & Ananthasswamy HN (1991)
    Mutations in the P53 tumor suppressor gene in human cutaneous
    squamous cell carcinomas. Mol Carcinog, 4: 445-449.

    Piltingsrud HV, Fong CW, & Odland LT (1978) An evaluation of
    ultraviolet radiation personnel hazards from selected 400-watt high
    intensity discharge lamps. Am Ind Hyg Assoc J, 39: 406-413.

    Pincus MW, Rollings PK, Craft AB, & Green A (1991) Sunscreen use on
    Queensland beaches, Australia. J Dermatol, 32: 21.

    Pitcher H & Longtreth J (1991) Melanoma mortality and exposure to
    ultraviolet radiation: An empirical relationship. Environ Int, 17:
    7-21.

    Pitts DG (1974) The human ultraviolet action spectrum. Am J Optom
    Arch Am Acad Optom, 51: 946-960.

    Pitts DG (1978) The ocular effects of ultraviolet radiation. Am J
    Optom Phys Optics, 55: 19-35.

    Pitts DG, Cullen AP, & Hacker PD (1977) Ocular effects of
    ultraviolet radiation from 295 to 365nm. Invest Ophthalmol Vis Sci,
    16: 932-939.

    Pound AW (1970) Induced cell proliferation and the initiation of
    skin tumour formation in mice by ultraviolet light. Pathology, 2:
    269-275.

    Prystowsky JH (1988) Photoprotection and the vitamin D status of the
    elderly. Arch Dermatol, 124: 1844-1848.

    Punnonen K, Puntela A, & Ahotupa M (1991) Effects of ultraviolet A
    and B irradiation on lipid peroxidation and activity of the
    antioxidant enzymes in keratinocytes in culture. Photochem
    Photoimmunol Photomed, 8: 3-6.

    Putvinsky AV, Sokolov AI, Roshchupkin DI, & Vladimirov Ya (1979)
    Electric breakdown of bilayer phospholipid membranes under
    ultraviolet irradiation-induced lipid peroxidation. FEBS Lett, 106:
    53-55.

    Quan MB & Moy RL (1991) The role of human papillomavirus in
    carcinoma. J Am Acad Dermatol, 25: 698-705.

    Quintern LE, Horneck G, Eischweiler V, & Büker H (1992) A biofilm
    used as ultraviolet dosimeter. Photochem Photobiol, 55: 389-395.

    Rady P, Scinicariello F, Wagner RF, & Tyring SK (1992) p53 Mutations
    in basal cell carcinomas. Cancer Res, 52: 3804-3806.

    Rapp LM & Smith SC (1992) Morphorlogical comparisons between
    rhodopsin-mediated and short wavelength classes of retinal light
    damage. Invest Ophthalmol Vis Sci, 33: 3367-3377.

    Rasanen L, Reunala T, Lehto M, Jansen C, Rantala I, & Leinikki P
    (1989) Immediate decrease in antigen-presenting function and delayed
    enhancement of interleukin-1 production in human epidermal cells
    after  in vivo UVB irradiation. Br J Dermatol, 120: 589-596.

    Rathbun WB (1989) In: Dolphin D, Poulson R, & Avramoric O ed.
    Coenzymes and cofactors - Volume III, Glutathione. New York, John
    Wiley and Sons, Inc., pp 467-509.

    Raven JA, (1991) Responses of aquatic photosynthetic organisms to
    increased solar UVB. J Photochem Photobiol, B9: 239-244.

    Reeve VE, Greenoak GE, Gallagher CH, Canfield PJ, & Wilkinson FJ
    (1985) Effect of immunosuppressive agents and sunscreens on UV
    carcinogenesis in the hairless mouse. Aust J Exp Biol Med Sci, 63:
    655-665.

    Reeve VE, Matheson M, Greenoak GE, Canfield PJ, Boehm- Wilcox C, &
    Gallagher CH (1988) Effect of dietary lipid on UV light
    carcinogenesis in the hairless mouse. Photochem Photobiol, 48:
    689-696.

    Reeve VE, Greenoak GE, Canfield PJ, Boehm-Wilcox C, & Gallagher CH
    (1989) Topical urocanic acid enhances UV- induced tumour yield and
    malignancy in hairless mouse. Photochem Photobiol, 49: 459-464.

    Reeve VE, Bosnic M, & Boehm-Wilcox C (1990) Effect of ultraviolet
    (UV) radiation and UVB-absorbing sunscreen ingredients on
    7,12-dimethyllbenz(a)anthracene-initiated skin tumorigenesis in
    hairless mice. Photodermatol Photoimmunol Photomed, 7: 222-227.

    Reeve VE, Bosnic M, Boehm-Wilcos C, & Ley RD (1991) Differential
    protection by two sunscreens from UV radiation-induced
    immunosuppression. J Invest Dermatol, 97: 624-628.

    Reeve VE, Bosnic M, & Rozinova E (1993) Carnosine
    (beta-alanylhistidine) protects from the suppression of contact
    hypersensitivity by ultraviolet B (280-320 nm) radiation or by cis
    urocanic acid. Immunology, 78: 99-104.

    Remé CH, Braschler U, Roberts J, & Dillon J (1991) Light damage in
    the rat retina: effect of a radioprotective agent (WR-77913) on
    acute rod outer segment disk disruption. Photochem Photobiol, 54:
    137-142.

    Reynolds P, Saunders LD, Layefsky ME, & Lemp GF (1993) The spectrum
    of acquired immunodeficiency syndrome (AIDS)- associated
    malignancies in San Francisco, 1980-1987. Am J Epidemiol, 137:
    19-30.

    Rhodes AR, Albert LS, Barnhill RL, & Weinstock MA (1991) Sun-induced
    freckles in children and young adults: A correlation of clinical and
    histopathologic features. Cancer, 67: 1990-2001.

    Rigel DS, Friedman RJ, Levenstein MJ, & Greenwald DI (1983)
    Relationship of fluorescent lights to malignant melanoma: another
    view. J Dermatol Surg Oncol, 9: 836-838.

    Ringvold A & Davangar M (1985) Changes in the rabbit corneal stroma
    caused by UV irradiation. Acta Ophthalmol, 63: 601-606.

    Ringvold A, Davander M, & Olsen EG (1982) Changes of the cornea
    epithelium after ultraviolet radiation. Acta Ophthalmol, 60: 41-52.

    Rippey J & Schmaman A (1972) Skin tumours of Africans. In: Marshall
    J ed. Essays on tropical dermatology. Amsterdam, Excerpta Medica,
    vol 2, pp 98-115.

    Rivas JM & Ullrich SE (1992) Systemic suppression of delayed-type
    hypersensitivity by supenatants from UV-irradiated keratinocytes. An
    essential role for keratinocyte-derived Il-10. J Immunol, 149:
    3865-3871.

    Rivers JK, Norris PG, Murphy GM, Chu AC, Midgley G, Morris J, Morris
    RW, Young AR, & Hawk JLM (1989) Tanning, photoprotection, acute
    adverse effects and immunological changes. Br J Dermatol, 120:
    767-777.

    Ro YS, Cooper PN, Lee JA, Quinn AG, Harrison D, Lane D, Horne CHW
    Rees JL, & Angus B (1993) p53 Protein expression in benign and
    malignant skin tumours. Br J Dermatol, 128: 237-241.

    Roberts DL (1990) Incidence of non-melanoma skin cancer in West
    Glamorgan, South Wales. Br J Dermatol, 122: 399-403.

    Roberts LK & Daynes RA (1980) Modification of the immunogenic
    properties of chemically induced tumors arising in hosts treated
    concomitantly with ultraviolet light. J Immunol, 125: 438-447.

    Roberts JE, Kinley J, Young A, Jenkins G, Atherton S, & Dillon J
    (1991)  in vivo and photophysical studies on photooxidative damage
    to lens proteins and their protection by radioprotectors. Photochem
    Photobiol, 53: 33-38.

    Robinson JK & Rademaker AW (1992) Relative importance of prior basal
    cell carcinomas, continuing sun exposure, and circulating
    T-Lymphocytes on the development of basal cell carcinoma. J Invest
    Dermatol, 99: 227-231.

    Roffo AH (1934) Cancer et soleil: carcinomes et sarcomes provoqués
    par l'action du soleil in toto. Bull Assoc Fr Étude Cancer, 23:
    590-616.

    Roff AH (1939) Physico-chemical etiology of cancer (with special
    emphasis on the association with solar radiation). Strahlentherapie,
    66: 328-350 (in German).

    Ros J (1990) On the effect of UV-radiation on elongation growth of
    sunflower seedlings (Helianthus annuus L.)(Thesis), pp. 1-157 in
    Karlst. Beitr. Entw Okophysiol 8, M Tevini (ed), Bot Inst II,
    Karlsruhe.

    Roschupkin DJ, Pelenitsyn AB, Potapenko AY, Talitsky VV & Vladimirov
    YA (1975) Study of the effects of ultraviolet light on biomembranes
    - IV. The effect of oxygen on UV-induced hemolysis and lipid
    photoperoxidation in rat erythrocytes and liposomes. Photochem.
    Photobiol. 21: 63-69.

    Rose EF (1973) Pigment variation in relation to protection and
    susceptibility to cancer. Pigment Cell, 1:, 236-45.

    Rosen ES (1986) Filtration of non-ionizing radiation by the ocular
    media. In: Cronley-Dillon J, Rosen ES, & Marshall J ed. Hazards of
    light: myths and realities of eye and skin. Oxford, New York,
    Pergamon Press, pp 145-152.

    Rosenstein BS & Ducore JM (1983) Induction of DNA strand breaks in
    normal human fibroblasts exposed to monochromatic ultraviolet and

    visible wavelengths in the 240-546 nm range. Photochem Photobiol,
    38: 51-55.

    Rosenthal F, Phoon C, Bakalian A, & Taylor H (1988) The ocular dose
    of ultraviolet radiation to outdoor workers. Invest Ophthalmol Vis
    Sci, 29: 649-656.

    Rosenthal F, Safran M, & Taylor H (1985) The ocular dose of
    ultraviolet radiation from sunlight exposure. Photochem Photobiol,
    42: 163-171.

    Roshchupkin DI, Pelenitsyn AB, Potapenko AY, Talitsky VV, &
    Vladimirov YA (1975) Study of the effects of ultraviolet light on
    biomembranes. IV. The effect of oxygen on UV-induced hemolysis and
    lipid photoperoxidation in rat erythrocytes and liposomes. Photochem
    Photobiol, 21: 63-69.

    Ross JA, Howie SE, Norval M, Maingay J, & Simpson TJ (1986)
    Ultraviolet-irradiated urocanic acid suppresses delayed-type
    hypersensitivity to herpes simplex virus in mice. J Invest Dermatol,
    87: 630-633.

    Roth M, Müller J, & Boyle JM (1987) Immunochemical determination of
    an initial step in thymine dimer excision repair in xeroderma
    pigmentosum variant fibroblasts and biopsy material from the normal
    population and patients with basal cell carcinoma and melanoma.
    Carcinogenesis, 8: 1301-1307.

    Roy CR & Gies HP (1993) Personal protection against solar
    ultraviolet radiation. In: Proceedings of the International
    Symposium on UV, Munich 4-6 May 1993. German Ministry of the
    Environment, Munich, Germany, pp 47-52.

    Roy CR, Gies HP, & Elliot G (1988) Solar ultraviolet radiation:
    personel exposure and protection. J Occup Health Safety (Aust N Z),
    4: 133.

    Roza L, Baan RA, Van Der Leun JC, Kligman L, & Young AR (1989) UVA
    hazards in skin associated with the use of tanning equipment. J
    Photochem Photobiol, B3: 281-287.

    Russell WO, Wynne ES, Loquvam GS, & Mehl DA (1956) Studies on bovine
    ocular squamous carcinoma ('cancer eye'). I. Pathological anatomy
    and historical review. Cancer, 9: 1-52.

    Ryel RJ, Barnes PW, Beyschlag W, Caldwell MM, & Flint SD (1990)
    Plant competition for light analyzed with a multispecies canopy
    model. I Model development and influence of enhanced UV-B conditions
    on photosynthesis in mixed wheat and wild oat canopies. Oecologia,
    82: 304-310.

    Sadamori N, Mine M, & Honda T (1991) Incidence of skin cancer among
    Nagasaki atomic bomb survivors. J Radiation Res, 32, Suppl 2:
    217-225.

    Saftlas AF, Blair A, Cantor KP, Hanrahan L, & Anderson HA (1987)
    Cancer and other causes of death among Wisconsin farmers. Am J Ind
    Med, 11: 119-129.

    Schaeffer L, Ray R, Humbert S, Moncollin V, Vermeulen W, Hoeijmakers
    JHJ, Chambon P, & Egly JM (1993) The basic transcription factor
    BTF2/TFIIH contains a helicase involved in both transcription and
    DNA repair. Science, Apr 2; 206(5104):37-38.

    Scharffetter K, Wlaschek M, Hogg A, Bolsen K, Schothorst A, Goerz G,
    Krieg T, & Plewig G (1991) Arch Dermatol Res, 283: 506-511.

    Schmitt C, Schmitt J, Wegener A, & Hockwin O (1988) Effect of an
    aldose reductase inhibitor, AL-1576, on the development of UV-B and
    X-Ray cataract. Graefe's Arch Clin Exp Ophthlamol, 226: 455-460.

    Schothorst AA, Slaper H, Schouten R, & Suurmond D (1985) UVB dose in
    maintenance psoriasis phototherapy versus solar UVB exposure.
    Photodermatology, 3: 213-220.

    Schwartz GG & Hulka BS (1990) Is vitamin D deficiency a risk factor
    for prostate cancer? Anticancer Res, 10: 1307-1311.

    Schwartz SM & Weiss NS (1988) Place of birth and incidence of ocular
    melanoma in the United States. Int J Cancer, 41: 174-177.

    SCOPE/UNEP (1993) Effects of increased ultraviolet radiation on
    global ecosystems (Workshop proceedings). Paris, Scientific
    Committee on Problems of the Environment, United Nations Environment
    Programme.

    Scotto J, Kopf AW, & Urbach F (1974) Non-melanoma skin cancer among
    Caucasians in four areas of the United States. Cancer, 34:
    1333-1338.

    Scotto J & Fears TR (1987) The association of solar ultraviolet and
    skin melanoma incidence among Caucasians in the United States.
    Cancer Invest, 5: 275-283.

    Scotto J, Fears TR, & Fraumeni JF Jr (1982) Solar radiation. In:
    Schottenfeld D & Fraumeni JF Jr ed. Cancer epidemiology and
    prevention. Philadelphia, Pennsylvania, W.B. Saunders Company, pp
    254-276.

    Scotto J, Fears TR, & Fraumeni JF Jr (1983) Incidence of nonmelanoma
    skin cancer in the United States. Bethesda, Maryland, National
    Cancer Institute (NIH Publication No. 83-2433).

    Scotto J, Cotton G, Urbach F, Berger D, & Fears T (1988)
    Biologically effective ultraviolet radiation: surface measurements
    in the US 1974-1985. Science, 235: 762-764.

    Seddon JM, Gragoudas ES, Glynn RJ, Egan KM, Albert DM, & Blitzer PH
    (1990) Host factors, UV radiation and risk of uveal melanoma. A
    case-control study. Arch Ophthalmol, 108: 1274-1280.

    Serrano H, Scotto J, Shornick G, Fears TR, & Greenberg ER (1991)
    Incidence of nonmelanoma skin cancer in New Hampshire and Vermont. J
    Am Acad Dermatol, 24: 574-579.

    Servilla KS, Burnham DK, & Daynes RA (1987) Ability of cyclosporine
    to promote the growth of transplanted ultraviolet radiation-induced
    tumors in mice. Transplantation, 44; 291-295.

    Setlow RB, Grist E, Thompson K, & Woodhead AD (1993) Wavelengths
    effective in induction of malignant melanoma. Proc Natl Acad Sci
    (USA), 90: 6666-6670.

    Setlow RB, Woodhead AD, & Grist E (1989) Animal model for
    ultraviolet radiation-induced melanoma: platyfish-swordtail hybrid.
    Proc Natl Acad Sci (USA), 86: 8922-8926.

    Shea CR, McNutt NS, Volkenandt M, Lugo J, Prioleau PG, & Albino AP
    (1992) Overexpression of p53 protein in basal cell carcinoma of
    human skin. Am J Pathol, 141: 25-20.

    Shearer GM & Clerici M (1992) T-helper cell immune dysfunction in
    asymptomatic HIV-1 seropositive individuals: The role of TH1-TH2
    cross regulation. In: Coffman RL ed. Regulation and functional
    significance of T-cell subsets. Prog Chem Immunol, 54, 21-43

    Shetlar MD (1980) Cross-linking of proteins to nucleic acids by
    ultraviolet light. Photochem Photobiol Rev, 5: 105-197.

    Shibata T, Katoh N, Hatano T, & Sasaki K. Population based
    case-control study of cortical cataract in the Noto area, Japan,
    Ophthalmic Res. (1993, in press).

    Shore RE (1990) Overview of radiation-induced skin cancer in humans.
    Int J Radiat Biol, 57: 809-827.

    Shore RE, Albert RE, Reed M, Harley N, & Pasternack BS (1984) Skin
    cancer incidence among children irradiated for ringworm of the
    scalp. Radiat Res, 100: 192-204.

    Shukla VK, Hughes DC, Hughes LE, McCormick F, & Padua RA (1989)  ras
    mutations in human melanotic lesions: K- ras activation is a
    frequent and early event in melanoma development. Oncogene Res, 5:
    121-127.

    Siemiatycki J (1991) Risk factors for cancer in the workplace. Boca
    Raton, Florida, CRC Press.

    Simon JC, Cruz PD, Bergstresser PR, & Tigelaar RE (1990) Low dose
    ultraviolet B-irradiated Langerhans cells preferentially activate
    CD4+ cells of the T helper 2 subset. J Immunol, 145: 2087-2091.

    Simon JC, Tigelaar RE, Bergstresser PR, Edelbaum D, & Cruz PD (1991)
    Ultraviolet B radiation converts Langerhans cells from immunogenic
    to tolerogenic antigen-presenting cells. J Immunol, 146: 485-491.

    Sjovall P, Christensen OB, & Moller H (1985) Single exposure to
    ultraviolet irradiation and elicitation of human allergic contact
    dermatitis. Acta Derm Venereol, 65: 93-96.

    Slaper H (1987) Skin cancer and UV exposure: Investigations on the
    estimation of risks. University of Utrecht, The Netherlands (PhD
    Thesis).

    Slaper H, Schothorst AA, & Van der Leun JC (1986) Risk evaluation of
    UVB-therapy for psoriasis: comparison of calculated risk for
    UVB-therapy and observed risk in PUVA- treated patients.
    Photodermatology, 3: 271-283.

    Sliney, DH (1983) Eye protective techniques for bright light.
    Ophthalmology 90(8), 937-944.

    Sliney D (1986) Physical factors in cataractogenesis: ambient
    ultraviolet radiation and temperature. Invest Ophthalmol Vis Sci,
    27(5):, 781-790.

    Sliney DH (1987) Estimating the solar ultraviolet radiation exposure
    to an intraocular lens implant, J Cataract Refract Surg,
    13(5),296-301.

    Sliney DH & Wolbarsht ML (1980) Safety with lasers and other optical
    sources -A comprehensive handbook. New York, London, Plenum Press.

    Smith KC (1976) The radiation-induced addition of proteins and other
    molecules to nucleic acids. In: Wang SY ed. Photochemistry and
    photobiology of nucleic acids. New York, London, San Francisco,
    Academic Press, vol 2, pp 187-218.

    Smith GJ & Ryan KG (1993) The effect of changes on differences in
    Robertson-Berger radiometer responsitivity on solar ultraviolet-B
    measurement. Photochem Photobiol, 58(4): 512-514.

    Smith RC, Baker KS, Holm-Hansen O, & Olson R (1980) Photoinhibition
    of photosynthesis in natural waters. Photochem Photobiol, 31:
    585-592.

    Smith EL, Walworth NC, & Holick MF (1986) Effects of 2
    alpha-25-dihydrxy vitamin D3 on the morphologic and biological
    differentiation of cultured human epidermal keratinocytes grown in
    serum-free conditions. J Invest Dermatol, 86: 709-714.

    Söderberg PG (1989) Mass alteration in the lens after exposure to
    ultraviolet radiation, 300nm. Acta Ophthalmol, 67: 633-644.

    Söderberg PG (1991) Na and K in the lens after exposure to radiation
    in the 300 nm wavelength region. J Photochem Photobiol, 8: 279-294.

    Soparker CN, O'Brien JM, & Albert DM (1993) Investigation of the
    role of the  ras protooncogene point mutation in human uveal
    melanomas. Invest Ophthalmol Vis Sci, 34: 2203-2209.

    Sorahan T & Grimley RP (1985) The aetiological significance of
    sunlight and fluorescent lighting in malignant melanoma: a
    case-control study. Br J Cancer, 52; 765-769.

    Spector A (1991) In: Sies H ed. The lens and oxidative stress.
    Endogenous antioxidant defences in human blood plasma in oxidative
    stress: Oxidants and antioxidants. New York, London, San Francisco,
    Academic Press.

    Spellman CW & Daynes RA (1978) Properties of ultraviolet
    light-induced suppressor lymphocytes within a syngeneic tunour
    system. Cell Immunol, 36: 383-387.

    Spellman CW, Woodward JG, & Daynes RA (1977) Modification of
    immunological potential by ultrviolet radiation. I Immune status of
    short-term UV-irradiated mice. Transplantation, 24: 112-119.

    Spitzer WO, Hill GB, Chambers LW, Helliwell BE, & Murphy HB (1975)
    The occupation of fishing as a risk factor in cancer of the lip. N
    Engl J Med, 293: 419-424.

    Spruance SL (1985) Pathogenesis of herpes simplex labialis:
    Experimental induction of lesions with UV light. J Clin Microbiol,
    22: 366-368.

    Staberg B, Wulf HC, Poulsen T, Klemp P, & Brodthagen H (1983)
    Carcinogenic effect of sequential artificial sunlight and UVA
    irradiation in hairless mice. Consequences for solarium 'therapy'.
    Arch Dermatol, 119: 641-643.

    Stadtman ER (1990) Metal ion catalysed oxidation of
    protein-biochemical mechanism and biological consequences. Free
    Radic Biol Med, 9: 315-325

    Stamnes K, Tsay SC, Wiscombe WJ, & Jayaweera K (1988) Numerically
    stable algorithm for discreet-ordinate method radiative transfer in
    multiple scattering and emitting media. Appl Optics, 27: 2502-2509.

    Stein B, Rahmsdorf HJ, Steffen A, Liffin M, & Herrlich P (1989)
    UV-induced damage is an intermediate step in UV-induced expression
    of human immunodeficiency virus type 1, collagenase, c-fos and
    metallothionein. Mol Cell Biol, 9: 5169-5181.

    Steinitz R, Parkin DM, Young JL, Bieber CA, & Katz L ed. (1989)
    Cancer incidence in Jewish migrants to Israel, 1961-1981 (IARC
    Scientific Publications No 98), Lyon, International Agency for
    Research on Cancer, 114-115, 134-135.

    Stenbäck F (1975a) Species-specific neoplastic progression by
    ultraviolet light on the skin of rats, guinea pigs, hamsters and
    mice. Oncology, 31: 209-225.

    Stenbäck F (1975b) Ultraviolet light irradiation as initiating agent
    in skin tumor formation by the two-stage method. Eur J Cancer, 11:
    241-246.

    Stephenson TJ, Royds J, Silcocks PB, & Bleehen SS (1992) Mutant p53
    oncogene expression in keratoacanthoma and squamous cell carcinoma.
    Br J Dermatol, 127: 566-570.

    Sterenborg HJCM & van der Leun JC (1990) Tumorigenesis by a long
    wavelength UVA source. Photochem Photobiol, 51: 325-330.

    Sterenborg HJCM, van der Putte SCJ, & van der Leun JC (1988) The
    dose-response relationship of tumorigenesis by ultraviolet radiation
    of 254 nm. Photochem Photobiol, 47: 245-253.

    Stern RS & Docken W (1986) An exacerbation of SLE after visiting a
    tanning salon. J Am Med Assoc, 255: 3120.

    Stern RS & Lange R (1988) Non-melanoma skin cancer occurring in
    patients treated with PUVA five to ten years after the first
    treatment. J Invest Dermatol, 91: 120-124.

    Stern RS, Thibodeau LA, & Kleinerman RA (1979) Risk of cutaneous
    carcinoma in patients treated with oral methoxsalen
    photochemotherapy. N Engl J Med, 300: 809-813.

    Stocker R & Frei B (1991) In Sies H ed. Endogenous antioxidant
    defences in human blood plasma in oxidative stress: Oxidants and
    antioxidants.New York, London, San Francisco, Adacemic Press.

    Stretch JR, Gatter KC, Ralfkiaer E, Lane DP, & Harris, AL (1991)
    Expression of mutant p53 in melanoma. Cancer Res, 51: 5976-5979.

    Strickland PT, Burns FJ, & Albert, RE (1979) Induction of skin
    tumors in the rat by single exposure to ultraviolet radiation.
    Photochem Photobiol, 30: 683-688.

    Strickland PT, Creasia D, & Kripke ML (1985) Enhancement of
    two-stage skin carcinogenesis by exposure of distant skin to UV
    radiation. J Natl Cancer Inst, 74: 1129-1134.

    Strickland PT, Vitasa BC, West SK, Rosenthal FS, Emmett EA, & Taylor
    HR (1989) Quantitative carcinogenesis in man: solar ultraviolet B
    dose dependence of skin cancer in Maryland watermen. J Natl Cancer
    Inst, 81: 1910-1913.

    Stryker WS, Stampfer MJ, Stein EA, Kaplan L, Louis TA, Sober A, &
    Willett WC (1990) Diet, plasma levels of beta-carotene and
    alpha-tocopherol, and risk of malignant melanoma. Am J Epidemiol,
    131: 597-611.

    Sullivan JH & Teramura AH (1991) The effects of UV-B radiation on
    loblolly pine. 2 Growth of field-grown seedlings, Trees (Berl)
    6(3):115-120.

    Sutherland JC & Griffin K (1981) Absorption spectrum of DNA for
    wavelengths greater than 300nm. Radiat Res, 86: 399-410.

    Swerdlow AJ, English JSC, MacKie RM, O'Doherty CJ, Hunter JAA, Clark
    J, & Hole DJ (1988) Fluorescent lights, ultraviolet lamps, and risk
    of cutaneous melanoma. Br Med J, 297: 647-650.

    Sydenham MM, Wong CF, Hirst LW, & Collins MJ (1991) Cular UVB
    dosimetry made possible for the first time using a CR-39 contact
    lens. In: Proceedings of 22nd Session of CIE, Melbourne. Vienna,
    International Commission of illumination, vol 1, part 2, pp 27-28.

    Talbot G (1948) Pterygium. Trans Ophthalmol Soc N Z, 2: 42-45.

    Taylor HR (1979) Pseudoexfoliation, an environmental disease? Trans
    Ophthalmol Soc UK, 99: 302-307.

    Taylor HR (1980a) Aetiology of climatic droplet keratopathy and
    pterygium. Br J Ophthalmol, 64: 154-163.

    Taylor HR (1980b) The environment and the lens. Br J Ophthalmol, 64:
    303-310.

    Taylor HR, West SK, Rosenthal FS, Munoz B, Newland HS, Abbey H, &
    Emmett EA (1988) Effect of ultraviolet radiation on cataract
    formation. N Engl J Med, 319: 1429-1433.

    Taylor HR, West SK, Rosenthal FS, Munoz B, Newland HS, & Emmett EA
    (1989) Corneal changes associated with chronic UV irradiation. Arch
    Ophthalmol, 107: 1481-1484.

    Taylor HR, West S, Munoz B, Rosenthal FS, Bressler SB, & Bressler NM
    (1992) The long-term effects of visible light on the eye. Arch
    Ophthalmol, 110: 99-104.

    Taylor JR, Schmieder GF, Shimizu T, & Streilein JW (1993)
    UVB-susceptibility is a risk factor for recurrent herpes labialis.
    Photochem Photobiol, 57: 135.

    Templeton AC (1967) Tumours of the eye and adnexa in Africans of
    Uganda. Cancer, 20: 1689-1698.

    Teppo L, Pukkala E, Hakama M, Hakulinen T, Herva A, & Saxén E (1980)
    Way of life and cancer incidence in Finland. A municipality-based
    ecological analysis. Scand J Soc Med, Suppl 19: 5-84.

    Terman M, Reme CE, Rafferty B, Gallin PF, & Terman JS (1990) Bright
    light therapy for winter depression: Potential ocular effects and
    theoretical implications. Yearly Review. Photochem Photobiol, 51:
    781-792.

    Tevini M, Braun J, & Fieser G (1991a) The protective function of the
    epidermal layer of rye seedlings against ultraviolet-B radiation,
    Photochem Photobiol, 53: 329-333.

    Tevini M, Mark U, & Saile-Mark M (1991b) Effects of enhanced solar
    UV-B radiation on growth and function of crop plant seedlings. In:
    Randall D & Blevins D ed. Current topics in plant biochemistry and
    physiology. Columbia, University of Missouri.

    Tevini M, Mark U, Fieser G, & Saile M (1991c) Effects of enhanced
    solar UV-B radiation on growth and function of selected crop plan
    seedlings. In: Riklis, D ed. Photobiology. New York, London, Plenum
    Press, pp 635-649.

    Tobal K, Warren W, Cooper CS, McCartney A, Hungerford J, & Lightman
    S (1992) Increased expression and mutation of p53 in choroidal
    melanoma. Br J Cancer, 66: 900-904.

    Toews GB, Bergstresser PR, & Streilein JW (1980) Epidermal
    langerhans cell density determines whether contact hypersensitivity
    or unresponsiveness follows skin painting with DNFB. J Immunol, 124:
    445-453.

    Tomatis, L (1990) ed, Cancer: Causes, Occurrence and Control. Lyon,
    International Agency for Research on Cancer (IARC Scientific
    Publications No 100).

    Travis, LB, Curtis, RE, Boice, JD, Hankey, BF, & Fraumeni Jr, JF
    (1991) Second cancers following non-Hodgkin's lymphoma. Cancer, 67:
    2002-2009.

    Travis, LB, Curtis, RE, Hankey, BF, & Fraumeni Jr, JF (1992) Second
    cancers in patients with chronic lymphocytic leukemia. J Natl Cancer
    Inst, 84: 1422-1427.

    Tucker, MA, Boice, JD, Jr & Hoffman, DA (1985a) Second cancer
    following cutaneous melanoma and cancers of the brain, thyroid,
    connective tissue, bone and eye in Connecticut, 1935-82. Natl Cancer
    Inst Monogr, 68: 161-189.

    Tucker, MA, Misfeldt, D, Coleman, N, Clark, WH, & Rosenberg, SA
    (1985b) Cutaneous malignant melanoma after Hodgkin's Disease. Ann
    Int Med, 102: 37-41.

    Tucker, MA, Shields, JA, Hartge, P, Augsberger, J, Hoover, RN, &
    Fraumeni, JF, Jr (1985c) Sunlight exposure as risk factor for
    intraocular malignant melanoma. New Engl J Med, 313: 789-792.

    Turner, BJ, Siatkowski, RM, Augsburger, JJ, Shields, JA, Lustbader,
    E, & Mastrangelo, MJ (1989) Other cancers in uveal melanoma patients
    and their families. Am J Ophthalmol, 107: 601-608.

    Tuveson RW & Sammartano, LJ (1986) Sensitivity of HemA mutant
    Escherichia coli cells to inactivation by near-UV light depends on
    the level of supplementation with L-aminolevulinic acid. Photochem
    Photobiol 43: 621-626.

    Tyrrell RM (1973) Induction of pyrimidine dimers in bacterial DNA by
    365 nm radiation, Photochem Photobiol 17: 69-73.

    Tyrrell RM, Ley RD & Webb RB (1974) Induction of single strand
    breaks (Alkali-labile bonds) in bacterial and phage DNA by near-UV
    (365 nm) radiation. Photochem. Photobiol. 20: 395-398.

    Tyrrell RM (1982) Cell inactivation and mutagenesis by solar
    ultraviolet radiation In: Hélène C, Charlier M, Montenay-Garestier
    Th & Laustriat G ed. Trends in Photobiology, Plenum Press, New York,
    pp 155-172.

    Tyrrell, RM (1984) Exposure of nondividing populations of primary
    human fibroblasts to UV (254 nm) radiation induces a transient
    enhancement in capacity to repair potentially lethal cellular
    damage. Proc Natl Acad Sci USA 81: 781-784.

    Tyrrell, RM (1991) UVA (320-380 nm) radiation as an oxidative stress
    In: Sies H ed. Oxidative Stress : oxidants and antioxidants, London,
    Academic Press, pp 57-83.

    Tyrrell, RM (1992) Inducible responses to UVA exposure. In: (ed. F.
    Urbach) Biological responses to ultraviolet A radiation, Valdenmar
    Press, Overland Park, pp. 59-64.

    Tyrrell, RM and Pidoux, M (1986) Endogenous glutathione protects
    human skin fibroblasts against the cytotoxic action of UVB, UVA and
    near-visible radiations. Photochem Photobiol 44: 561-564.

    Tyrrell RM & Pidoux M (1987) Action spectra for human skin cells.
    Estimates of the relative cytotoxicity of the middle ultraviolet,
    near ultraviolet and violet regions of sunlight on epidermal
    keratinocytes. Cancer Res 47: 1825-1829.

    Tyrrell, RM and Pidoux, M (1988) Correlation between endogenous
    glutathione content and sensitivity of cultured human skin cells to
    radiation at defined wavelengths in the solar UV range. Photochem
    Photobiol 47: 405-412.

    Tyrrell, RM & Pidoux M (1989) Singlet oxygen involvement in the
    inactivation of cultured human fibroblasts by UVA (334 nm, 365 nm)
    and near-visible radiations. Photochem Photobiol 49: 407-412.

    Ullrich SE (1985) Suppression of lymphoproliferation by
    hapten-specific suppressor T lymphocytes from mice exposed to
    ultraviolet radiation. Immunology, 54: 343-352.

    Ullrich SE (1986) Suppression of the immune response to allogeneic
    histocompatibility antigens by a single exposure to ultraviolet
    radiation. Transplantation, 42: 287-291.

    Ullrich SE & Kripke ML (1984) Mechanisms in the suppression of tumor
    rejection produced in mice by repeated UV irradiation. J Immunol,
    133: 2786-2790.

    Ullrich SE, Azizi E, Kripke ML (1986a) Suppression of the induction
    of delayed-type hypersensitivity reactions in mice by a single
    exposure to ultraviolet radiation. Photochem Photobiol, 43: 633-638.

    Ullrich SE, Yee GK, & Kripke ML (1986b) Suppressor lymphocytes
    induced by epicutaneous sensitization of UV- irradiated mice control
    multiple immunological pathways. Immunology, 58: 185-190.

    UNEP (1987) The ozone layer, United Nations Environment Programme,
    UNEP/GEMS Environmental Library No 2, UNEP, Nairobi.

    UNEP (1989) Environmental Effects Panel Report, Van der Leun, JC,
    Tevini, M eds. United Nations Environment Programme , Nairobi,
    Kenya.

    UNEP (1991) Environmental effects of ozone depletion: 1991 Update.
    United Nations Environment Programme, Nairobi, Kenya.

    UNEP (1992) Effects of increased ultraviolet radiation on biological
    systems. Proceedings of meeting, Scientific Committee on Problems of
    the Environment (SCOPE). Budapest, United Nations Environment
    Programme, Nairobi, Kenya.

    UNEP (1993) United Nations Environment Programme.  Report of the
    ninth meeting of the Open-Ended Working Group of the parties to the
    Montreal Protocol, Geneva 30 August-1 September 1993.

    UNEP-WMO (1989) Scientific assessment of stratospheric ozone: 1989
    WMO Global ozone research and monitoring project, United Nations
    Environment Programme, World Meteorological Organization, Geneva
    (Report No 29).

    Urbach F (1987) Man and ultraviolet radiation. In: Passchier WF &
    Bosnjakovic BFM eds. Human exposure to ultraviolet radiation - risks
    and regulations. New York, Excerpta Medica.

    Urbach F (1987a) Cutaneous photocarcinogenesis.  J Environ Sci
    Health-Environ Carcinogen Rev, C5, 211-34.

    Urbach F (1989) Testing the efficacy of sunscreens: effect of choice
    of source and spectral distribution of ultraviolet radiation, and
    choice of endpoint. Photodermatology, 6: 177-181.

    Urbach F, Davies RE, & Forbes PD (1966) Ultraviolet radiation and
    skin cancer in man In: Montagna W & Dobson RL eds. Advances in
    Biology of Skin, Oxford, Pergamon Press, Carcinogenesis, Vol VII, pp
    195-214.

    Urbach F, Epstein JH, & Forbes PD (1974) Ultraviolet carcinogenesis:
    experimental, global and genetic aspects. In: Pathak MA, Harber LC,
    Seiji M, & Kukita A, eds. Sunlight and Man - Normal and Abnormal
    Photobiological Responses, Tokyo, University of Tokyo Press, pp
    259-283.

    US Food and Drug Administration (1978) Sunscreen drug products for
    over-the-counter use. Fed Regis, 43: 38206-28269.

    US National Cancer Institute (1989) Sunscreens, Rockville, MD,
    Tracor Technological Resources Inc (Class Study Report; Contract No
    NO1-CP-71082 (7/89)).

    Vågerö D, Ringbäck G, & Kiviranta H (1986) Melanoma and other
    tumours of the skin among office, other indoor and outdoor workers
    in Sweden 1961-1979. Br J Cancer, 53: 507-512.

    Vågerö D, Swerdlow AJ, & Beral V (1990) Occupation and malignant
    melanoma: a study based on cancer registration data in England and
    Wales and in Sweden. Br J Ind Med, 47: 317-324.

    Valerie K, Delers A, Bruck C, Thiriatt C, Rosenberg H, Debouck C &
    Rosenberg M (1988) Activation of human immmunodefficiency virus type
    1 by DNA damage in human cells. Nature 333: 78-81.

    Van der Leun JC (1984) Yearly review: UV-carcinogenesis. Photochem
    Photobiol, 39: 861-868.

    Van der Leun JC (1987) Principles of risk reduction and protection.
    In: Passchier WF & Bosnjakovic BFM, eds. Human Exposure to
    Ultraviolet Radiation: Risks and Regulations, Amsterdam, Elsevier,
    pp 293-303.

    Van der Leun JC (1992) Interactions of UVA and UVB in
    photodermatology: what was photoaugmentation? In: Urbach F ed. The
    Biological Responses to Ultraviolet A Radiation, Overland Park, KS
    Valdenmar, pp 309-319.

    Van der Leun JC & de Gruijl FR (1993) Influences of ozone depletion
    on human and animal health. In: Tevini M ed. UV-B radiation and
    ozone depletion. Effects on humans, animals, plants, microorganisms,
    and materials. Boc Raton, Lewis Publishers, pp 95-123.

    Van der Schroeff JG, Evers LM, Boot AJM & Bos JL (1990) ras oncogene
    mutations in basal cell carcinomas and squamous cell carcinomas of
    human skin. J Invest Dermatol, 94: 423-425.

    Van Weelden H, de Gruijl FR & van der Leun JC (1986) Carcinogenesis
    by UVA, with an attempt to assess the carcinogenic risks of tanning
    with UVA and UVB. In: Urbach F & Gange RW eds. The Biological
    Effects of UVA Radiation, New York, Praeger, pp 137-146.

    Van Weelden H and van der Leun J, (1987) IVA induced tumours in
    pigmented and albino hairless mice in Human exposure to ultraviolet
    radiation - risks and regulations eds. Passchier WF and Bosnjakovic
    BFM (New York: Excerpta Medica).

    Van Weelden H, de Gruijl FR, van der Putte SCJ, Toonstra J & van der
    Leun JC (1988) The carcinogenic risks of modern tanning equipment:
    is UVA safer than UVB? Arch Dermatol Res, 280: 300-307.

    Van't Veer LJ, Burgering BMT, Versteeg R, Boot AJM, Ruiter DJ,
    Osanto S, Schrier PI & Bos JL (1989) N- ras Mutations in human
    cutaneous melanoma from sun-exposed body sites. Mol Cell Biol, 9:
    3114-3116.

    Varghese AJ & Wang SY (1967) Ultraviolet irradiation of DNA  in
     vitro and  in vivo produces a third thymine-derived product.
    Science 156: 955-957.

    Verhoeff FH, Bell L & Walker CB (1916) The pathological effects of
    radiant energy on the eye. Proc Am Acad Arts Sci, 510: 1-810.

    Vermeer M & Streilein JW (1990) Ultraviolet B light-induced
    alterations in epidermal Langerhans cells are mediated in part by
    tumor necrosis factor-alpha. Photodermatol Photoimmunol Photomed, 7:
    258-265.

    Vermeer M, Schmieder GJ, Yoshikawa T, Van den Berg J-W, Metzman MS,
    Taylor JR, & Streilein JW (1991) Effects of ultraviolet B light on
    cutaneous immune responses of humans with deeply pigmented skin. J
    Invest Dermatol, 97: 729-734.

    Vile GF, Basu-Moda S, Waltner C & Tyrrell RM (1994) Haem oxygenase 1
    mediates an adaptive response to oxidative stress in human skin
    fibroblasts. Proc Natl Acad Sci USA 91: 2607-2610.

    Vincek V, Jurimoto I, Medema JP, Prieto E, & Streilein JW (1993)
    Tumor necrosis factor alpha polymorphism correlates with deleterious
    effects of ultraviolet B light on cutaneous immunity. Cancer Res,
    53: 728-732.

    Vitale S, West S, Munoz B, Schein OD, Maguire M, Bressler, N &
    Taylor HR (1992) Watermen Study II: mortality and baseline
    prevalence of nuclear opacity. Invest Ophthalmol Vi Sci, 33: 1097.

    Vitasa BC, Taylor HR, Strickland PT, Rosenthal FS, West S, Abbey H,
    Ng SK, Munoz B, & Emmett EA (1990) Association of nonmelanoma skin
    cancer and actinic keratosis with cumulative solar ultraviolet
    exposure in Maryland watermen. Cancer, 65: 2811-2817.

    Volkenandt M, Schlegel U, Nanus DM, & Albino AP (1991) Mutational
    analysis of the human p53 gene in malignant melanoma. Pigment Cell
    Res, 4: 35-40.

    Vosjan, JH, Dohler, G, & Nieuwland G (1990) Effect of UVB irradiance
    on the ATP content of microorganisms of the Weddell Sea Antarctica,
    Neth J Sea Res 25, 391-394.

    Vuillaume M, Daya-Grosjean L, Vincens P, Pennetier JL, Tarroux P,
    Baret A, Calvayrac R, Taieb A & Satorin, A (1992) Striking
    differences in cellular catalase activity between two DNA
    repair-deficient diseases. Xeroderma pigmentosum and
    trichothiodystophy. Carcinogenesis 13: 321-328.

    Walker GC (1987) The SOS response of Escherichia coli In: Niedhardt
    FC, Ingraham JL, Low KB, Magasanik B, Shaechter M, Umbarger HE eds.
    Escherichia coli and Salmonella typhimurium. Cellular and Molecular
    Biology, American Society of Microbiology, Washington DC pp
    1346-1357.

    Walter SD, Marrett LD, From L, Hertzman C, Shannon HS, & Roy P
    (1990) The association of cutaneous malignant melanoma with the use
    of sunbeds and sunlamps. Am J Epidemiol, 131: 232-243.

    Warfel AH, Moy JA, Meola T, Sanchez M, Soter NA, & Belsito DV (1993)
    Effect of ultraviolet B (UVB) on the expression of human
    immunodeficiency virus (HIV) in mice and humans, Photochem
    Photobiol, 57: 755.

    Waring GO, Roth AM, & Ekins MB (1984) Clinical and pathological
    description of 17 cases of corneal intraepithelial neoplasia. Am J
    Ophthalmol, 97: 547-559.

    Waterhouse J, Muir C, Correa P, & Powell J eds (1976) Cancer
    Incidence in Five Continents, Lyon, International Agency for
    Research on Cancer, Vol III (IARC Scientific Publications No 15).

    Waterhouse J, Muir C, Shanmugaratnam K, & Powell J eds (1982) Cancer
    Incidence in Five Continents, Lyon, International Agency for
    Research on Cancer, Vol IV (IARC Scientific Publications No 42).

    Webb RB (1977) Lethal and mutagenic effects of near-ultraviolet
    radiation Photochem Photobiol Rev 2: 169-261.

    Wei Q, Matanoski GM, Farmer ER, Hedayati MA, & Grossman L (1993) DNA
    repair and aging in basal cell carcinoma: A molecular epidemiology
    study. Proc Natl Acad Sci USA, 90: 1614-1618.

    Weinstock MA, Colditz GA, Willett WC, Stampfer MJ, Bronstein BR,
    Mihm MC Jr, & Speizer FE (1989) Nonfamilial cutaneous melanoma
    incidence in women associated with sun exposure before 20 years of
    age. Pediatrics, 84: 199-204.

    Wellmann E, (1971) Phytochrome mediated flavone glycoside synthesis
    in cell suspension cultures of Petroselinum hortense after
    preirradiation with ultraviolet light, Planta, 101: 283-286.

    Wellmann E, (1991) Specific ultraviolet effects in plant
    development, J Exp Bot, (Suppl), 32: 42.

    Welsh D & Diffey B (1981) Protection against solar actinic radiation
    afforded by common clothing fabrics. Clinical and Experimental
    Dermatology 6: 577-582.

    Werner JS (1991) The damaging effects of light on the eye and
    implications for understanding changes in vision across the life
    span, in (P Bagnoli and W Hodos) The Changing Visual System, New
    York, Plenum Press.

    West SK, Rosenthal FS, Bressler NM, Bressler SB, Munoz B, Fine SL, &
    Taylor HR (1989) Exposure to sunlight and other risk factors for
    age-related macular degeneration. Arch Ophthalmol, 107: 875-879.

    Wester A (1987) Ultraviolet transmission properties of sunscreens
    and sunglasses. In: Passchier W & Bosnjakovic B eds. Human exposure

    to ultraviolet radiation: risks and regulations, Elsevier,
    Amsterdam.

    Westerveld A, Hoeijmakers JHJ, van Duin M, de Wit J, Odijk H,
    Pavlink A, Wood RD & Bootsma D (1984) Molecular cloning of a human
    DNA repair gene. Nature 310: 425-529.

    Whillock, MJ, Clark, IE, McKinlay, AF, Todd, CD & Mundy, SJ (1988)
    Ultraviolet radiation levels associated with the use of fluorescent
    general lighting, UVA and UVB lamps in th eworkplace and home. 
    National radiological Protection Board, Research Report R221,
    London, HMSO.

    Whillock MJ, McKinlay AF, Kemmlert J & Forsgren PG (1990)
    Ultraviolet radiation emissions from miniature [compact] fluorescent
    lamps. Lighting Res Technol 22: 125-128.

    Whitaker CJ, Lee WR, & Downes JE (1979) Squamous cell skin cancer in
    the north-west of England, 1967-69, and its relation to occupation.
    Br J Ind Med, 36: 43-51.

    WHO (1948) International Statistical Classification of Diseases,
    Injuries and Causes of Death, Sixth Revision, Geneva.

    WHO (1977) Manual of the International Statistical Classification of
    Diseases, Injuries, and Causes of Death. International
    Classification of Diseases, 1975 rev, Vol 1, Geneva, p 102.

    WHO (1989) Non-ionizing radiation protection. MJ Suess & DA
    Benwell-Morison eds. World Health Organization Regional Office for
    Europe, European series No 25, WHO, Copenhagen.

    WHO/UNEP/IRPA (1979) Ultraviolet Radiation, Environmental Health
    Criteria 14, World Health Organization, United Nations Environment
    Programme, WHO, Geneva.

    Widmark (1889) Ueber den Einfluss des Lichtes auf die vorderen
    medien des auges. Skand Arch f physiol 1, 264-330.

    Widmark J (1901) Ueber den Einfluss des Lichtes auf die Linse.
    Mitteil aus d Augenklin d Carol med chir Inst zu Stockholm 3:
    135-149.

    Wilesmith JW, Wells GA, Cranwell MP and Ryan JB (1988) Bovine
    spongiform encephalopathy : epidemiological studies. Vet Rec, 123,
    638-644.

    Willis I, Menter JM, & Whyte HJ (1981) The rapid induction of
    cancers in the hairless mouse utilizing the principle of
    photoaugmentation. J Invest Dermatol, 76: 404-408.

    Wittenberg S (1986) Solar radiation and the eye: a review of
    knowledge relevant to eye care. Am J Optom Physiol Optics, 63:
    676-689.

    WMO (1993) World Meteorological Organization Report of the second
    meeting of the ozone research managers of the parties to the Vienna
    Convention for the protection of the ozone layer. Geneva 10-12
    March.

    Wolf P, Donawho CK, & Kripke ML (1993a) Analysis of the protective
    effect of different sunscreens on ultraviolet radiation-induced
    local and systemic suppression of contact hypersensitivity and
    inflammatory responses in mice. J Invest Dermatol 100: 254-259.

    Wolf P, Yarosh DB, & Kripke ML (1993b) Effects of sunscreens and a
    DNA excision repair enzyme on ultraviolet radiation - induced
    inflammation, immune suppression and cyclobutane pyrimidine dimer
    formation in mice. J invest Dermatol 101: 523-527.

    Wong CF, Fleming R, & Carter SJ (1989) A new dosimeter for
    ultraviolet B radiation. Photochem Photobiol 50: 611-615.

    Wong L, Ho SC, Coggon D, Cruddas AM, Hwang CH, Ho CP, Robertshaw AM,
    & MacDonald DM (1993) Sunlight exposure, antioxidant status, and
    cataract in Hong Kong fishermen. J Epidemiol Comm Health, 47: 46-49.

    Wood RD, Robbins P & Lindahl T (1988) Complementation of the
    Xeroderma pigmentosum DNA repair defect in cell-free extracts. Cell
    53: 91-106.

    Wucherpfenning V (1931) Biologie und prackische verwenbarkeit der
    erythemschwelle des UV. Strahlentherapie 40: 201-244.

    Yannuzzi L, Fisher Y, Slakter J, Krueger A (1989) Solar retinipathy.
    A photobiologic and geophysical analysis, Retina, 9, 28-43.

    Yarosh D, Alas LG, Yee V, Oberyszyn A, Kibitel JT, Mitchell D,
    Rosenstein R, Spinowitz A, & Citron M (1992) Pyrimidine dimer
    removal enhanced by DNA repair liposomes reduces incidence of UV
    skin cancer in mice. Cancer Res, 52: 4227-4231.

    Yasumoto S, Yoshinobu H, & Aurelian L (1987) Immunity to herpes
    simplex virus type 2: Suppression of virus-induced responses in
    ultraviolet B-irradiated mice. J Immunol, 139: 2788-2793.

    Yoshida A, Ishiguro S, & Tamai M (1993) Expression of glial
    fibrillary acidic protein in muoller cells after
    lensectomy-vitrectomy. Invest. Ophthalmol. Vis. Sci. 34: 3154-3160.

    Yoshikawa T & Streilein JW (1990) On the genetic basis of the
    effects of ultraviolet B on cutaneous immunity. Evidence that

    polymorphisms at the Tnfý and Lps loci govern susceptibility.
    Immunogenetics, 32: 398-405.

    Yoshikawa T, Rae V, Bruins-Slot W, Van den Berg J-W, Taylor Jr, &
    Streilein JW (1990) Susceptibility to effects of UVB radiation on
    induction of contact hypersensitivity as a risk factor for skin
    cancer in humans. J Invest Dermatol, 95: 530-536.

    Young AR, Magnus IA, Davies AC, & Smith NP (1983) A comparison of
    the phototumorigenic potential of 8-MOP and 5-MOP in hairless albino
    mice exposed to solar simulated radiation. Br J Dermatol, 108:
    507-518.

    Young AR, Walker SL, Kinley JS, Plastow SR, Averbeck D, Morlière P,
    & Dubertret L (1990) Phototumorigenesis studies of 5-methoxypsoralen
    in bergamot oil: evaluation and modification of risk of human use in
    an albino mouse skin model. J Photochem Photobiol B Biol, 7:
    231-250.

    Young AR, Potten CS, Chadwick CA, Murphy GM, Hawk JLM & Cohen AJ
    (1991) Photo-protection and 5-MOP photochemoprotection from
    UVR-induced DNA damage in humans: the role of skin type. J Invest
    Dermatol, 97: 942-948.

    Young RW (1991) Age-related cataract. New York, Oxford University
    Press.

    Yu CC, MacGregor JM, Dublin EA, Barnes DM, MacDonald DM, & Levison
    DA (1992) Patterns of immunostaining for p53 in benign and malignant
    melanocytic lesions (meeting abstract). J Pathol, 167: (Suppl),
    130A.

    Zadjela E & Bisagni E (1981) 5-Methoxypsoralen, the melanogenic
    additive in sun-tan preparations, is tumorigenic in mice exposed to
    365 nm uv radiation. Carcinogenesis, 2: 121-127.

    Zamansky G & Chou, I-N (1987) Environmental wavelength of
    ultraviolet light induce cytoskeletal damage. J Inv Dermatol 89:
    603-606.

    Zanetti R, Rosso S, Faggiano F, Roffino R, Colonna S, & Martina G
    (1988) A case-control study on cutaneous malignant melanoma in the
    province of Torino, Italy (Fr). Rev Epidemiol Santé publ, 36:
    309-317.

    Zanetti R, Franceschi S, Rosso S, Colonna S, & Bidoli E (1992)
    Cutaneous melanoma and sunburns in childhood in a southern European
    population. Eur J Cancer, 28A: 1172-1176.

    Zaridze D, Mukeria A, & Duffy SW (1992) Risk factors for skin
    melanoma in Moscow. Int J Cancer, 52: 159-161.

    Zaunuddin D & Sasaki K (1991) Risk factor analysis in a cataract
    epidemiology survey in West Samatara, Indonesia. Dev Ophthalmol, 21:
    78-86.

    Ziegler A-M, Leffell DJ, Kunala S, Sharma HW, Gailani M, Simon JA,
    Halperin AJ, Baden HP, Shapiro PE, Bale AE, & Brash DE (1993)
    Mutation hotspots due to sunlight in the p53 gene of nonmelanoma
    skin cancers. Proc Natl Acad Sci USA, 90: 4216-4220.

    Zigman S (1993) Yearly review: ocular light damage. Photochem.
    Photobiol. 57: 1060-1068.

    Zigman S, Yulo T, & Schultz J (1974) Cataract induction in mice
    exposed to near UV light. Ophthalmol Res, 6: 259-270.

    Zigman S, Graff J, Yulo T, & Vaughen T (1975) The response of mouse
    ocular tissue to continuous near-UV light exposure. Invest
    Ophthalmol, 14: 710-713.

    Zmudzka BZ & Beer JZ (1990) Activation of human immunodeficiency
    virus by ultraviolet radiation. Photochem Photobiol, 52: 1153-1162.

    Zuclich JA (1989) Ultraviolet induced photochemical damage in ocular
    tissues. Health Physics 56: 671-682.

    Zuclich J & Kurtin W (1977) Oxygen dependence of near-ultraviolet
    induced corneal damage. Photochem Photobiol 25: 133-135.

    Zundorf, I, & Hader, DP (1991) Biochemical and spectroscopic
    analysis of UV effects on the marine flagellate, cryoptomonas
    maculata, Arch Microbiol.


    See Also:
       Toxicological Abbreviations
       Ultraviolet radiation (EHC 14, 1979)