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

    Published under the joint sponsorship of
    the United Nations Environment Programme
    and the World Health Organization

    World Health Organization Geneva, 1976

    ISBN 92 4 154061 3

    (c) World Health Organization 1976

        Publications of the World Health Organization enjoy copyright
    protection in accordance with the provisions of Protocol 2 of the
    Universal Copyright Convention. For rights of reproduction or
    translation of WHO publications, in part or  in toto, application
    should be made to the Office of Publications, World Health
    Organization, Geneva, Switzerland. The World Health Organization
    welcomes such applications.

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

        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.




         1.1. Some definitions
         1.2. Summary
               1.2.1. Analytical methods
               1.2.2. Sources of environmental pollution
               1.2.3. Environmental distribution and transport
               1.2.4. Environmental exposure levels
               1.2.5. Metabolism of mercury
               1.2.6. Experimental studies on the effects of mercury
               1.2.7. Epidemiological and clinical studies
               1.2.8. Evaluation of health risks to man and guidelines
                       for health protection
         1.3. Recommendations for further research
               1.3.1. Environmental sources and pathways of mercury
               1.3.2. Metabolic models in man
               1.3.3. Epidemiological studies
               1.3.4. Interaction of mercury with other environmental
               1.3.5. Biochemical and physiological mechanisms of

         2.1. Chemical and physical properties
         2.2. Purity of compounds
         2.3. Sampling and analysis
               2.3.1. Sample collection
               2.3.2. Analytical methods
               2.3.3. Analysis of alkyl mercury compounds in the presence
                       of inorganic mercury

         3.1. Natural occurrence
         3.2. Industrial production
         3.3. Uses of mercury
         3.4. Contamination by fossil fuels, waste disposal, and
               miscellaneous industries.

         4.1. Distribution between media -- the global mercury cycle
         4.2. Environmental transformation -- the local mercury cycle
         4.3. Interaction with physical or chemical factors
         4.4. Bioconcentration

         5.1. Levels in air, water, and food
         5.2. Occupational exposures
         5.3. Estimate of effective human exposure

         6.1. Uptake
               6.1.1. Uptake by inhalation
               6.1.2. Uptake by ingestion
               6.1.3. Absorption through skin
         6.2. Distribution in organisms
         6.3. Elimination in urine and faeces
         6.4. Transplacental transfer and secretion in milk
         6.5. Metabolic transformation and rate of elimination
         6.6. Accumulation of mercury and biological half-time (metabolic
         6.7. Individual variations -- strain and species comparisons

         7.1. Experimental animal studies
               7.1.1. Acute studies
               7.1.2. Subacute and chronic studies
                Reversible damage
                Irreversible damage
                Interactions with physical and chemical
               7.1.3. Biochemical and physiological mechanisms of

         8.1. Epidemiological studies
               8.1.1. Occupational exposure to mercury vapour,
                       alkylmercury vapour and other exposures
               8.1.2. General population
               8.1.3. Children and infants with in utero exposure
         8.2. Clinical studies of effects of mercury binding compounds
         8.3. Pathological findings and progression of disease
               8.3.1. Psychiatric and neurological disturbances
               8.3.2. Eye and visual effects
               8.3.3. Kidney damage
               8.3.4. Skin and mucous membrane changes

         9.1. General considerations
               9.1.1. Elemental mercury vapour
               9.1.2. Methylmercury compounds
               9.1.3. Ethylmercury compounds and other short-chain
               9.1.4. Inorganic mercury, aryl- and alkoxyalkylmercurials
         9.2. Summary and guidelines




        During the last two decades, evaluation of the health hazards from
    chemical and other environmental agents has received considerable
    attention in several WHO programmes. High priority was given to
    drinking water quality (1), food additives (2), and pesticide residues
    (3), to occupational exposure (4), air quality in urban areas (5),
    and, more recently, to the carcinogenic risk of chemicals to man (6).

        In most instances, man's  total exposure to a given agent, from
    different media or conditions (air, water, food, work, home), was not
    considered. The inadequacy of this approach is obvious for pollutants
    that may reach man by several pathways, as is the case with lead,
    cadmium, and some other metals, and certain persistent organic
    compounds. In response to a number of World Health Assembly
    resolutions (WHA23.60, WHA24.47, WHA25.58, WHA26.68) and taking into
    consideration the relevant recommendations of the United Nations
    Conference on the Human Environment (7) held at Stockholm in 1972, and
    of the Governing Council of the United Nations Environment Programme
    (UNEP) (8), an integrated and expanded programme on the assessment of
    health effects of environmental conditions was initiated in 1973 under
    the title of: WHO Environmental Health Criteria Programme, with the
    following objectives:

    (i)     to assess existing information on the relationship between
            exposure to environmental pollutants (or other physical and
            chemical factors) and man's health, and to provide guidelines
            for setting exposure limits consistent with health protection,
            i.e., to compile environmental health criteria documents;

    (ii)    to identify new or potential pollutants by preparing
            preliminary reviews on the health effects of agents likely to
            be increasingly used in industry, agriculture, in the home or

    (iii)   to identify gaps in knowledge concerning the health effects of
            recognized or potential pollutants or other environmental
            factors, to stimulate and promote research in areas where
            information is inadequate, and

    (iv)    to promote the harmonization of toxicological and
            epidemiological methods in order to obtain research results
            that are internationally comparable.


    a Prepared by the WHO Secretariat. References are listed on page 14.

        The general framework of the Environmental Health Criteria
    Programme was formulated by a WHO meeting held in November 1972 (9),
    and further elaborated by a WHO Scientific Group that met in April
    1973 (10).



        In the framework of the WHO Environmental Health Criteria
    Programme, it is understood that the term "criteria" designates the
    relationship between exposure to a pollutant or other factor and the
    risk or magnitude of undesirable effects under specified circumstances
    defined by environmental and target variables (9). This corresponds to
    the definition proposed by the Preparatory Committee for the United
    Nations Conference on the Human Environment (11). Other Preparatory
    Committee definitions of immediate interest to the criteria programme

    --  " exposure: the amount of a particular physical or chemical agent
        that reaches the target";

    --  " target (or receptor): the organism, population, or resource to
        be protected from specific risks";

    --  " risk: the expected frequency of undesirable effects arising
        from a given exposure to a pollutant".

    The WHO Scientific Group on Environmental Health Criteria (10)
    accepted these definitions for the purposes of its discussions, but
    felt that they were not altogether satisfactory, and recommended that
    WHO, in collaboration with other international organizations, should
    reconsider them, along with other necessary definitions, at an
    appropriate international meeting. In accordance with this
    recommendation, the WHO Secretariat is preparing a list of basic terms
    to be used in the Environmental Health Criteria Programme that will be
    submitted to the national institutions and other international
    organizations for discussion.

        The Scientific Group (10) found the definition of "exposure"
    particularly inadequate and considered that it should be expanded to
    include the concepts of concentration and length of exposure in
    addition to the amount of the agent.

        The WHO Secretariat considers it useful to attach specific
    meanings to the terms "effect", "response" and "dose" as was done by
    the Subcommittee on the Toxicology of Metals of the Permanent
    Commission and International Association on Occupational Health at the
    Tokyo meeting (12). These terms will be used in the following sense
    unless indicated differently in specific criteria documents:

    -- " effect: a biological change caused by (or associated with)a an

    -- " response: the proportion of a population that demonstrates a
    specific effect";

    -- " dose: the amount or concentration of a given chemical at the
    site of the effect".

        The concept of "response" as defined above is generally accepted
    but the terminology used to describe this concept varies widely. Many
    toxicologists use the terms "effect" and "response" interchangeably to
    denote a specific biological change associated with exposure, whereas
    different terms are used to indicate the proportion of a population
    affected (e.g., incidence, cumulative response frequency, response
    rate, etc.).

        There is no general agreement as to the use of the term "dose" for
    chemical agents. Its common usage is to express the amount of
    substance administered, for instance, to an experimental animal (e.g.,
    oral dose, injected dose, etc.). In most cases, the amount or
    concentration of a given agent at the site where its presence induces
    a given effect cannot be determined by direct measurement and has to
    be estimated from experimental, occupational, or general environmental
    exposure, or from measurements in biological indicator media such as
    blood, urine, faeces, sweat, or hair (12). To avoid misunderstanding,
    it is, therefore, necessary in each case to make as clear as possible
    the way in which the "dose" is measured or estimated, including the
    units used.

        Because of the existing differences in the use of terms, no
    attempt has been made at this stage to impose a uniform terminology in
    all criteria documents. Until an internationally agreed terminology
    becomes available, the task groups on specific criteria documents are
    given freedom to choose their terminology, provided the terms are
    defined and used consistently throughout the document under


    a Added by the WHO Secretariat.


        An attempt has been made to express all numerical values in a
    uniform fashion, for instance, the concentrations are always expressed
    as mass concentrations in units acceptable to the SI system (e.g.
    mg/litre or mg/kg) (13). Some departures from this are made where the
    introduction of new units would cause confusion, e.g., lead in blood
    is expressed in g/100 ml and not in g/litre.


        Considering the large number of environmental agents and factors
    that may adversely influence human health, a practical programme for
    the preparation of criteria documents must be based on clearly defined
    priorities. The list of priorities has been established by a WHO
    Scientific Group (10), and is based on the following considerations:

    --  " Severity and frequency of observed or suspected adverse effects
         on human health. Of importance are irreversible or chronic
        effects, such as genetic, neurotoxic, carcinogenic, and
        embryotoxic effects including teratogenicity. Continuous or
        repeated exposures generally merit a higher priority than isolated
        or accidental exposures.

    --   Ubiquity and abundance of the agent in man's environment. Of
        special concern are inadvertently produced agents, the levels of
        which may be expected to increase rapidly, and agents that add to
        a natural hazard.

    --   Persistence in the environment. Pollutants that resist
        environmental degradation and accumulate, in man, in the
        environment, or in food chains, deserve attention.

    --   Environmental transformations or metabolic alterations. Since
        these alterations may lead to the production of chemicals that
        have greater toxic potential, it may be more important to
        ascertain the distribution of the derivatives than that of the
        original pollutant.

    --   Population exposed. Attention should be paid to exposures
        involving a large portion of the general population, or
        occupational groups, and to selective exposures of highly
        vulnerable groups represented by pregnant women, the newborn,
        children, the infirm or the aged."

        The full list contains some 70 chemicals and physical hazards, and
    it will be periodically reviewed. In preparing this list, it was
    realized that each country must assess environmental health problems
    in the light of its own national situation and establish its own
    priorities, which may not have been covered by this list.



        As stated on page 5, the purpose of the criteria documents is to
    compile, review, and evaluate available information on the biological
    effects of pollutants and other environmental factors that may
    influence man's health, and to provide a scientific basis for
    decisions aimed at protecting man from the adverse consequences of
    exposure to such environmental factors, both in the occupational and
    general environment. Although attainment of this objective entails
    consideration of a wide range of data, no attempt is made to include
    in the documents an exhaustive review of all published information on
    the environmental and health aspects of specific agents. In the
    process of collecting the required information, the available
    literature has been carefully evaluated and selected as to its
    validity and its relevance to the assessment of human exposure, to the
    understanding of the mechanism of biological effects, and to the
    establishment of dose-effect and dose-response relationships.
    Environmental considerations are limited to information that can help
    in understanding the pathways leading from the natural and man-made
    sources of pollutants to man. Non-human targets (e.g., plants,
    animals) are not considered unless the effects of their contamination
    are judged to be of direct relevance to human health. For similar
    reasons much of the published information on the effects of chemicals
    on experimental animals has been omitted.


        The criteria documents consist of three parts:

    (i)     A summary, which highlights the major issues, followed by
            recommendations for research to fill existing gaps in

    (ii)    The bulk of the report, which contains the findings on which
            the evaluation of the health risks is based. This part has a
            similar structure in all the criteria documents on chemical
            agents and contains the following chapters: chemical and
            physical properties and analytical methods; sources of
            environmental pollution; environmental transport, distribution
            and transformation; metabolism; experimental studies of
            effects; and epidemiological and clinical studies of the
            effects. The subdivision of these chapters differs from
            document to document.

    (iii)   Evaluation of health risks to man from exposure to the
            specific agent. This part of the criteria document states the
            considered opinion of the task group, which examined the
            findings contained in the second part (see (ii) above), and
            typically contains the following sections: relative
            contributions to the total dose from air, food, water, and
            other exposures; dose-effect relationships; dose-response
            relationships and, whenever possible, guidelines on exposure
            or dose limits.

    Chemical and physical data

        The chemical and physical data included in the criteria documents
    are limited to the properties that are considered relevant to the
    assessment of exposure and to the understanding of the effects. Where
    applicable, the impurities that may occur in commercial products are
    examined. Analytical techniques are discussed only to the extent
    needed to understand and evaluate data on levels in the environment
    and biological samples. The methods described should not be considered
    as recommended procedures. Where feasible, information is included on
    the applicability of a given method for the analysis of different
    types of sample, on detection limits, precision, and accuracy. The
    detection limit represents the smallest total amount the method is
    able to determine. In most cases, the amount of sample is limited so
    that it is useful in practice to express the smallest concentration
    that can be determined by that method. Precision of a method is
    defined in terms of the standard deviation or the coefficient of
    variation of a number of analyses made on the sample. Accuracy denotes
    systematic deviation of the measured values from the true value. It is
    impossible to ascertain the accuracy with absolute certainty; the
    evidence for the accuracy of a method is often circumstantial and is
    based either on inter-laboratory data-quality control studies or on
    the agreement of results obtained with procedures using different
    approaches. The results of one "accurate" procedure should agree with
    those of another "accurate" procedure for a given set of samples.

    Production, use, and environmental levels

        Data on the production, use, and levels in the environment of
    pollutants are reported only to illustrate the magnitude and extent of
    the problem and are not meant to represent an exhaustive and critical
    review. It is hoped that, in the future, better data will be available
    and that closer collaboration will be established with other
    governmental and non-governmental organizations qualified to supply
    such information.

    Biological data

        Although every effort is made to review the whole literature, it
    is possible that some publications have been overlooked. Some studies
    have purposely been omitted because the information contained therein
    was not considered valid or relevant to the scope of the criteria
    documents, or because they only confirmed findings already described.
    In general, the information is summarized as given by the author;
    however, certain shortcomings of reporting or of experimental design
    are also pointed out. The data on carcinogenicity have been examined
    and evaluated in consultation with the International Agency for
    Research on Cancer.

        Whenever possible, the dose-effect and the dose-response
    relationships reported in the criteria documents are based on
    epidemiological and other human studies, and animal data are used, in
    general, as supporting evidence.


        In order to obtain balanced and unbiased information, the
    collection and evaluation of information is done in close
    collaboration with national scientific and health institutions. About
    20 Member States of WHO have designated national focal points for
    collaboration in the WHO Environmental Health Criteria Programme.
    Without this collaboration no progress could have been made in its

        In addition, a number of WHO collaborating centres on
    environmental health effects have been designated to extend and
    complement the expertise available in the WHO Secretariat.

        Two procedures have been used in preparing the criteria documents.
    One is based on the consolidation of national contributions and the
    other on a draft criteria document prepared by consultants or the
    collaborating centres in association with the Secretariat.

    Procedure based on national contributions

        Criteria documents are prepared in four stages: (1) the
    preparation of national contributions by focal points in the Member
    States reviewing all relevant research results obtained in these
    countries; (2) consolidation of the national contributions into a
    draft document, which is done on a contractual basis with individual
    experts or WHO collaborating centres; (3) the draft criteria documents
    are circulated to the national focal points for comments and
    additions, based on which a second draft is prepared, and (4) the
    second draft document is reviewed and the information assessed at a
    meeting of internationally recognized experts (the task group

        National contributions to the criteria documents consist of a
    review of data on health effects of environmental agents, as revealed
    by experimental, clinical, and epidemiological studies, and of other
    relevant information on research carried out in each country and
    published in scientific journals or official publications. In order to
    facilitate the integration of national contributions into draft
    criteria documents, detailed outlines are prepared for each
    environmental agent considered, and the national focal points are
    requested to follow these outlines as closely as possible and to
    attach all publications referred to in the review in the form of
    reprints or microfiches.

    Procedure for drafts prepared by the Secretariat

        With the exception of steps 1 and 2 (which are replaced by the
    preparation of a draft criteria document by individual experts or WHO
    collaborating centres), the procedure is the same as described above.
    This procedure is applied in cases where much preparatory work has
    been done in Member States and where criteria-like documents (WHO or
    national) already exist.

    Task group meetings

        The task group meetings that are convened to complete the criteria
    documents have the following terms of reference:

    (i)     to verify, as far as possible, that all available data have
            been collected and examined;

    (ii)    to select those data relevant to the criteria documents;

    (iii)   to determine whether the data, as summarized in the draft
            criteria document, will enable the reader to make his own
            judgement concerning the adequacy of an experimental,
            epidemiological, or clinical study;

    (iv)    to judge the health significance of the information contained
            in the draft criteria document, and

    (v)     to make an evaluation of the dose-effect, dose-response
            relationships and of the health risks from exposure to the
            environmental agents under examination.

        Members of task groups serve in a personal capacity, as experts
    and not as representatives of their governments or of any organization
    with which they are affiliated. In addition to the first and second
    draft criteria documents, the members of the task group are requested
    to refer to the original publications whenever they deem that
    necessary, and to review national and other comments on the first
    draft criteria document to make sure that no significant information
    is omitted and that the final document properly reflects the work done
    in different countries.

    Collaboration with the United Nations Environment Programme (UNEP) and
    other international organizations

        The WHO Environmental Health Criteria Programme has received
    substantial financial assistance from UNEP which is acknowledged with
    appreciation. In addition, the programme has been planned from the
    outset in consultation with the UNEP Secretariat. The UNEP Secretariat
    receives all the drafts of criteria documents and their comments are
    carefully considered in the preparation of the final documents. UNEP
    is regularly invited to be represented at the task group meetings.

        The United Nations, their subsidiary bodies and specialized
    agencies, and the IAEA are as a rule invited to provide comments on
    the draft criteria documents and to participate in the task group
    meetings. The same applies to selected nongovernmental organizations
    in official relationship with WHO.

    Note to readers of the criteria documents

        While every effort has been made to present information in the
    criteria documents as accurately as possible without unduly delaying
    their publication, mistakes might have occurred and are likely to
    occur in the future. In the interest of all users of the environmental
    health criteria documents, readers are kindly requested to communicate
    any errors found to the Division of Environmental Health, World Health
    Organization, Geneva, Switzerland, in order that they may be included
    in corrigenda which will appear in subsequent volumes.

        In addition, experts in any particular field dealt with in the
    criteria documents are kindly requested to make available to the WHO
    Secretariat any important published information that may have
    inadvertently been omitted and which may change the evaluation of
    health risks from exposure to the environmental agent under
    examination, so that the information may be considered in the event of
    updating and re-evaluation of the conclusions contained in the
    criteria documents.


    1.   International Standards for Drinking Water, third edition,
            Geneva, World Health Organization, 1971.

    2.  WHO Technical Report Series, Nos: 129 (1957), 228 (1962), 281
            (1964), 309 (1965), 339 (1966), 373 (1967), 383 (1968), 430
            (1969), 445 (1970), 462 (1971), 488 (1972), 505 (1972), 539

    3.  WHO Technical Report Series, Nos: 370 (1967), 391 (1968), 417
            (1969), 458 (1970), 474 (1971), 502 (1972), 525 (1973), 545
            (1974), 574 (1975), 592 (1976).

    4.  WHO Technical Report Series, No.: 415 (1969).

    5.  WHO Technical Report Series, No.: 506 (1972).

             the Evaluation of Carcinogenic Risk of Chemicals to Man,
            Vol. 1-11 (1972-76).

    7.  UNITED NATIONS GENERAL ASSEMBLY.  Report of the United Nations
             Conference on the Human Environment held at Stockholm, 5-16
            June 1972 A/CONF.48/14, 3 July 1972.

    8.  UNITED NATIONS ENVIRONMENT PROGRAMME.  Report of the Governing
             Council of the United Nations Environment Programme (First
             session) UNEP/GC/10, 3 July 1973.

    9.   The WHO Environmental Health Criteria Programme (unpublished
            WHO document EP/73.1).

    10.  Environmental Health Criteria. Report of a WHO Scientific Group
            (unpublished WHO document EP/73.2).

    11. UNITED NATIONS GENERAL ASSEMBLY.  Report of the Preparatory
             Committee for the United Nations Conference on the Human
             Environment on its Third Session. United Nations document
            A/CONF.48/PC/13, 30 September 1971.

    12. NORDBERG, G. F., ed.  Effects and dose-response relationships of
             toxic metals, Proceedings from an international meeting
             organized by the Sub-committee on the Toxicology of Metals
             of the Permanent Commission and International Associations
             on Occupational Health, Tokyo, 18-23 November 1974.
            Amsterdam, Oxford, New York, Elsevier Scientific Publishing
            Company, 1976.

    13. LOWE, D. A.  A guide to international recommendations on names and
             symbols for quantities and on units of measurement. Geneva,
            World Health Organization, 1975, 314pp.  (Progress in
             Standardization No. 2.)


     Geneva 4-10 February 1975



        Professor T. Beritic, Institute for Medical Research and
            Occupational Medicine, Zagreb, Yugoslavia

        Dr H. Blumenthal, Division of Toxicology, Bureau of Foods, Food
            and Drug Administration, Department of Health, Education and
            Welfare, Washington, DC, USA  (Rapporteur)

        Dr J. Bouquiaux, Department of the Environment, Institute of
            Hygiene and Epidemiology, Brussels, Belgium

        Dr G. J. van Esch, Laboratory for Toxicology, National Institute
            of Public Health, Bilthoven, Netherlands

        Professor L. Friberg, Department of Environmental Hygiene,
            Karolinska Institute, Stockholm, Sweden  (Chairman)

        Professor G. L. Gatti, Istituto Superio di Sanit, Rome, Italy

        Dr L. Magos, Toxicology Research Unit, Medical Research Council
            Laboratories, Carshalton, Surrey, England

        Dr J. Parizek, Institute of Physiology, Czechoslovak Academy of
            Sciences, Prague, Czechoslovakia

        Dr J. K. Piotrowski, Department of Biochemistry, Institute of
            Environmental Research, Medical Academy in Lodz, Lodz, Poland

        Dr E. Samuel, Health Protection Branch, Department of National
            Health and Welfare, Ottawa, Ontario, Canada

        Dr S. Skerfving, Department of Internal Medicine, University
            Hospital, Lund, Sweden

        Dr T. Tsubaki, Brain Research Insitiute, Niigata University,
            Niigata, Japan

        Professor H. Valentin, Institute for Occupational and Social
            Medicine, Erlangen, Federal Republic of Germany

     Representatives from other organizations

        Dr A. Berlin, Health Protection Directorate, Commission of the
            European Communities, Luxembourg

        Dr D. Djordjevic, Occupational Health and Safety Branch, ILO,
            Geneva, Switzerland

        Dr W. J. Hunter, Commission of the European Communities,

        G. D. Kapsiotis, Senior Officer, Food Policy and Nutrition
            Division, FAO, Rome, Italy

        Dr E. Mastromatteo, Chief, Occupational Health and Safety Branch,
            ILO, Geneva, Switzerland


        Dr T. Clarkson, University Center in Environmental Health
            Sciences, The University of Rochester, School of Medicine and
            Dentistry, Rochester NY, USA  (Temporary Adviser)

        Dr F. C. Lu, Chief, Food Additives, WHO, Geneva, Switzerland

        Dr B. Marschall, Medical Officer, Occupational Health, WHO,
            Geneva, Switzerland


        A WHO Task Group on Environmental Health Criteria for Mercury met
    in Geneva from 4-10 February 1975. Dr B. H. Dietrich, Director,
    Division of Environmental Health, opened the meeting on behalf of the
    Director-General. The Task Group reviewed and amended the second draft
    criteria document and made an evaluation of health risks from exposure
    to mercury and its compounds. The revised draft was sent for comments
    to all members of the Task Group.

        A group of WHO temporary advisers (Dr T. Clarkson, Dr L. Friberg,
    Dr A. Jernelv,a Dr L. Magos, and Dr G. Nordbergb) assisted the
    Secretariat in the final scientific editing of the document. They met
    in Geneva on 13 and 14 November 1975.

        The first and second draft criteria documents were prepared by
    Dr T. Clarkson, Environmental Health Sciences Centre, the University
    of Rochester School of Medicine and Dentistry, Rochester, New York,
    USA. The comments on which the second draft was based were received
    from the national focal points for the WHO Environmental Health
    Criteria Programme in Bulgaria, Czechoslovakia, the Federal Republic
    of Germany, Italy, Japan, the Netherlands, New Zealand, Poland,
    Sweden, the USA, and the USSR; and from the United Nations
    Industrial Development Organization (UNIDO), Vienna, and the United
    Nations Scientific, Educational and Cultural Organization (UNESCO),
    Paris. Comments from the International Labour Organisation, Geneva,
    the United Nations Food and Agriculture Organization, Rome, and the
    Commission of the European Communities Health Protection Directorate,
    Luxembourg, were submitted at the task group meeting.

        Comments were also received, at the request of the Secretariat,
    from Dr L. Amin-Zaki, Iraq, Dr G. J. van Esch, Netherlands, Dr K.
    Kojima, Japan, and Dr S. I. Shibko, USA.

        The collaboration of these national institutions, international
    organizations, WHO collaborating centres and individual experts is
    gratefully acknowledged. Without their assistance the document could
    not have been completed. The Secretariat wishes to thank in particular
    Dr T. Clarkson for his help in all phases of the preparation of the


    a Institute for Water and Air Pollution Research, Stockholm, Sweden.

    b Department of Environmental Hygiene, Karolinska Institute,
      Stockholm, Sweden.

        This document is based primarily on original publications listed
    in the reference section. However, several recent publications broadly
    reviewing health aspects of mercury and its compounds have also been
    used. These include reviews by the Swedish Expert Group (1971).,
    Hartung & Dinman (1972), IAEA (1972), and Wallace et al. (1971).
    Reviews devoted primarily to the biological effects of mercury have
    been published by Clarkson (1972a, 1972b) and Miller & Clarkson
    (1973). Furthermore, several recent symposia have provided extensive
    reviews of the environmental aspects of mercury (Bouquiaux, 1974;
    D'Itri, 1972; Krenkel, 1975). A systematic review of various
    environmental health aspects of mercury, including a broad review of
    the accessible literature up to 1971, has been presented by Friberg &
    Vostal (1972).


    1.1  Some definitions

        In order to clarify the meaning of certain terms used in the
    document, some definitions are given below. However, it should be
    noted that these definitions have not been formally adopted by WHO.

        The terms  critical effects, critical organ, and  critical organ
     concentration have recently been defined by the Sub-Committee on
    Toxicology of Metals of the Permanent Commission and International
    Association of Occupational Health (Nordberg, 1976). The term
    "critical" as defined by the Committee differs from its usual meaning
    in clinical medicine, where it refers to a situation in which the
    patient's condition may deteriorate suddenly and dramatically. It also
    differs in meaning from that used in the field of radiation
    protection, where the "critical" organ is defined as the organ of the
    body whose damage by radiation results in the greatest injury to the
    individual. In this document, the term "critical" does not refer to a
    life-threatening situation, but to a key decision point for taking
    preventive action. For example, at some point in the dose-effect
    relationship, a critical effect can be identified. The appearance of
    an effect in an individual signals the point at which measures should
    be taken to reduce or prevent further exposure.

    1.2  Summary

    1.2.1  Analytical methods

        The method of choice for determining total mercury in
    environmental and biological samples is flameless atomic absorption.
    The technique is rapid and sensitive and the procedure is technically
    simple. Neutron activation is now principally used as a reference
    method against which the accuracy of atomic absorption procedures may
    be checked. Gas-liquid chromatography combined with an
    electron-capture detector is the most widely used method for
    identifying methylmercury in the presence of other compounds of

        The methods of sampling require careful consideration of the type
    of exposure to be monitored and the material to be analysed. Errors
    arising in collection, storage, and transportation of samples may be
    as important as instrument errors in contributing to the total error
    in the measurement of mercury in the sample. These include
    contamination of the sample, and the loss of mercury by adsorption on
    the walls of the container, and by volatilization. In estimating human
    exposure, special care should be taken to see that the sample is truly
    representative, e.g. the mercury vapour concentration in the breathing
    zone and the concentration of methylmercury in the daily diet.

    1.2.2  Sources of environmental pollution

        The major source of mercury is the natural degassing of the
    earth's crust and amounts to between 25 000 and 125 000 tonnes per
    year. Anthropogenic sources are probably less than natural sources.
    World production of mercury by mining and smelting was estimated at
    10 000 tonnes per year in 1973 and has been increasing by an annual
    rate of about 2%. The chloralkali, electrical equipment, and paint
    industries are the largest consumers of mercury, accounting for about
    55% of the total consumption. Mercury has a wide variety of other uses
    in industry, agriculture, military applications, medicine, and

        Several of man's activities not directly related to mercury
    account for substantial releases into the environment. These include
    the burning of fossil fuel, the production of steel, cement, and
    phosphate, and the smelting of metals from their sulfide ores. It was
    extimated that the total anthropogenic release of mercury would amount
    to 20 000 tonnes per year in 1975.

    1.2.3  Environmental distribution and transport

        Two cycles are believed to be involved in the environmental
    transport and distribution of mercury. One is global in scope and
    involves the atmospheric circulation of elemental mercury vapour from
    sources on land to the oceans. However, the mercury content of the
    oceans is so large, at least seventy million tonnes, that the yearly
    increases in concentration due to deposition from the global cycle are
    not detectable.

        The other cycle is local in scope and depends upon the methylation
    of inorganic mercury mainly from anthropogenic sources. Many steps in
    this cycle are still poorly understood but it is believed to involve
    the atmospheric circulation of dimethylmercury formed by bacterial

        The methylation of inorganic mercury in the sediment of lakes,
    rivers, and other waterways and in the oceans is a key step in the
    transport of mercury in aquatic food chains leading eventually to
    human consumption. Methylmercury accumulates in aquatic organisms
    according to the trophic level, the highest concentrations being found
    in the large carnivorous fish.

        Alkylmercury fungicides used as seed dressings are important
    original sources of mercury in terrestrial food chains. Mercury is
    passed first to seed eating rodents and birds and subsequently to
    carnivorous birds.

        Accumulation of methylmercury in aquatic and terrestrial food
    chains represents a potential hazard to man by consumption of certain
    species of oceanic fish, of fish or shellfish from contaminated
    waters, and of game birds in areas where methylmercury fungicides are

    1.2.4  Environmental exposure levels

        The concentration of mercury in the atmosphere is usually below
    50 ng/m3 and averages approximately 20 ng/m3. A concentration of
    50 ng/m3 would lead to a daily intake of about 1 g. "Hot spots" near
    mines, smelting works, and refineries require further investigation
    but could lead to daily intakes as high as 30 g. Daily intakes would
    be higher for occupational exposures to mercury vapour. An average
    mercury concentration in air of 0.05 mg/m3 would lead to an average
    daily intake via inhalation of about 480 g. The highest occupational
    exposures usually occur in mining operations but over 50 specific
    occupations or trades involve frequent exposure to mercury vapour.

        Mercury in drinking water would contribute less than 0.4 g to the
    total daily intake. Bodies of fresh water for which there is no
    independent evidence of contamination contain mercury at less than
    200 ng/litre. Oceanic mercury is usually less than 300 ng/litre.

        Food is the main source of mercury in nonoccupationally exposed
    populations, and fish and fish products account for most of the
    methyl-mercury in food. Mercury in food other than fish is usually
    present at concentrations below 60 g/kg. Mercury is present in
    freshwater fish from uncontaminated waters at concentrations of
    between 100 and 200 g/kg wet weight. In contaminated areas of
    freshwater, mercury levels between 500 and 700 /kg wet weight are
    often described and in some cases, concentrations are even higher.
    Most species of oceanic fish have mercury levels of about 150 g/kg.
    However, the large carnivorous species (e.g. swordfish and tuna)
    usually fall in the range of 200-1500 g/kg. With few exceptions
    methylmercury accounts for virtually all the mercury in both
    freshwater and marine fish.

        Intake of mercury from food is difficult to estimate with
    precision. Daily intake from food other than fish is estimated as 5 g
    but the chemical form of mercury is not known. Most of the
    methylmercury in diet probably comes from fish and fish products. The
    median daily intake of methylmercury in Sweden has been estimated as
    5 g. In most countries the daily intake is less than 20 g but in
    subgroups in certain countries where there is an unusually high fish
    intake (dieters) the daily intake may rise to 75 g and may even be as
    high as 200-300 g (in coastal villages dependent on large oceanic
    fish as the main source of protein). In areas of high local pollution,
    daily intakes could be well in excess of 300 g and these levels have
    led to two recorded outbreaks of methylmercury poisoning.

    1.2.5  Metabolism of mercury

        Approximately 80% of inhaled mercury vapour is retained.
    Information on pulmonary retention of other forms of mercury in man is
    lacking. Absorption of inorganic mercury compounds from foods is about
    7% of the ingested dose. In contrast, gastrointestinal absorption of
    methylmercury is practically complete. Little information is available
    on skin absorption although it is suspected that most forms of mercury
    can penetrate the skin to some extent. In the case of methylmercury,
    poisoning has resulted from skin application.

        Animal data indicate that the kidneys accumulate the highest
    tissue concentrations no matter what form of mercury is administered.
    The distribution of mercury between red cells and plasma depends upon
    the form of mercury. The red cell to plasma ratio is highest for
    methylmercury (approximately 10) and lowest for inorganic mercury
    (approximately 1) in man.

        The hair is a useful indicator medium for people exposed to
    methylmercury. The concentration of mercury in hair is proportional to
    the concentration in the blood at the time of formation of the hair.
    The relationship between hair and blood concentrations is not known
    for other forms of mercury.

        Most forms of mercury are predominantly eliminated with urine and
    faeces. In workers exposed over a long period to mercury vapour,
    urinary excretion slightly exceeds faecal elimination. On a group
    basis, mercury excretion in urine is proportional to the time-weighted
    average air concentration. Large individual fluctuations are common in
    daily mercury excretion in urine in people under the same exposure

        Faecal elimination accounted for approximately 90% of total
    mercury elimination in volunteers given a single dose of
    methylmercury. Urinary concentrations of total mercury do not
    correlate with blood levels after exposure to methylmercury.

        Animal data indicate that elemental mercury vapour rapidly crosses
    the placenta. The transplacental transfer of methylmercury compounds
    is well documented in man. The mercury concentrations in plasma in the
    mother and the newborn infant are similar but the concentration in the
    fetal red blood cells is approximately 30% higher than in those of the

        Details on transmission into breast milk are available only for
    methylmercury. The concentration of mercury in breast milk is
    approximately 5% of the simultaneous mercury level in blood in the
    mother, and infants can accumulate dangerously high blood
    concentrations by suckling if their mothers are heavily exposed.

        Tracer studies in volunteers and in exposed populations have
    established the main features of the metabolic model for methylmercury
    in man. Clearance half-times from the whole body and from blood are
    about 70 days. Daily intakes of methylmercury will lead to a
    steady-state balance in about one year, when the body burden will be
    approximately one hundred times the daily intake. In steady-state, the
    numerical value of the concentration of mercury in whole blood in
    g/litre is virtually equal to the numerical value of the daily intake
    in g/day/70 kg body weight. Considerable individual variation around
    these average values has been noted, which must be taken into account
    in the estimation of risk in exposed populations.

        The metabolic models for other forms of mercury are less well

    1.2.6  Experimental studies on the effects of mercury

        Reversible and irreversible toxic effects may be caused by mercury
    and its compounds, depending upon the dose and duration of exposure.
    Reversible behavioural changes may be produced in animals by exposure
    to mercury vapour.

        Methylmercury compounds produce irreversible neurological damage
    in animals. Many of the neurological signs seen in man have been
    reproduced in animals. Methylmercury is equally toxic to animals
    whether it is given in the pure chemical state or in fish where it has
    accumulated naturally. A latent period lasting weeks or months is
    observed between cessation of exposure and onset of poisoning.
    Morphological changes have been seen in the brain before onset of
    signs. This phenomenon has been referred to as "silent damage". Animal
    data support epidemiological evidence from Japan, that the fetus is
    more sensitive than the adult.

        Little is known about the physical and chemical factors affecting
    the toxicity of mercury. Selenium is believed to be protective against
    inorganic and methylmercury compounds.

    1.2.7  Epidemiological and clinical studies

        The classic symptoms of poisoning by mercury vapour are erethism
    (irritability, excitability, loss of memory, insomnia), intentional
    tremor, and gingivitis. Most effects of mercury vapour are reversible
    on cessation of exposure, although complete recovery from the
    psychological effects is difficult to determine. Recovery may be
    accelerated by treatment with penicillamine and unithiol

        Studies of occupational exposure to mercury vapour reveal that the
    classic symptoms of mercurialism do not occur below a time-weighted
    average mercury concentration in air of 0.1 mg/m3. Symptoms such as
    loss of appetite and psychological disturbance have been reported to
    occur at mercury levels below 0.1 mg/m3.

        The most common signs and symptoms of methylmercury poisoning are
    paraesthesia, constriction of the visual fields, impairment of
    hearing, and ataxia. The effects are usually irreversible but some
    improvement in motor coordination may occur. Complexing and chelating
    agents may be useful in prevention if given early enough after
    exposure but BAL is contraindicated in cases of methylmercury
    poisoning as it leads to increased brain levels of mercury.

        Epidemiological investigations have been made on populations in
    whom the intensity and duration of exposure to methylmercury through
    diet differs, for example, a population in Iraq having-high daily
    mercury intakes (as high as 200 g/kg/day) for a brief period (about 2
    months), populations in Japan having lower daily intakes with exposure
    for several months or years, and several fish-eating populations
    having daily intakes of mercury usually below 5 g/kg but with
    exposure lasting for the lifetime of the individual. The results of
    these studies indicate that the effects of methylmercury in adults
    become detectable in the most sensitive individuals at blood levels of
    mercury of 20-50 g/100 ml, hair levels from 50-120 mg/kg, and body
    burdens between about 0.5 and 0.8 mg/kg body weight.

        Observations on the Minamata outbreak in Japan indicate that the
    fetus is more sensitive to methylmercury than the adult but the
    difference in degree of sensitivity has not yet been established.

    1.2.8  Evaluation of health risks to man from exposure to mercury
           and its compounds

        Adverse health effects have not yet been identified in workers
    occupationally exposed to a time-weighted average air concentration of
    mercuryof 0.05mg/m3. This air concentration is equivalent to an
    average mercury concentration in blood of 3.5 /100 ml and an average
    mercury concentration in urine of 150 g/litre on a group basis. The
    corresponding ambient air concentration of mercury for exposure of the
    general population would be 0.015 mg/m3.

        It is estimated that the first effects associated with long-term
    daily intake of methylmercury should occur at intake levels between 3
    and 7 g/kg/day. The probability of an effect (paraesthesia) at this
    intake level is about 5% or less in the general population. These
    figures apply only to adults. Prenatal life may be the most sensitive
    stage of the life cycle to methylmercury. Furthermore experiments on
    animals indicate a potential for genetic damage by methylmercury.

    1.3  Recommendations for Further Research

    1.3.1  Environmental sources and pathways of mercury intake

        More information is needed on the physical and chemical forms of
    mercury in air, food, and water. With the exception of fish tissue,
    little is known of the proportion of total mercury in the diet that is
    in the form of methylmercury.

        The concentration of mercury in the air in "hot spots" near points
    of industrial release is not yet adequately documented. The few
    reports reviewed in this criteria document indicate that people living
    near points of emission may receive substantial exposure to airborne
    mercury. Levels of mercury in the oceans are still inadequately
    documented. The pathways of methylation of mercury in the ocean and
    its uptake by fish of different trophic levels are poorly understood.

        Studies are needed to estimate quantitatively the dietary intake
    of methylmercury in populations dependent on fish for their main
    source of protein. Average dietary intakes for the populations of
    several industrialized countries have been reported. However, of much
    greater importance are the identification of those subgroups of the
    population having unusually high dietary intakes of methylmercury and
    the careful quantitative estimation of average daily intake in these

    1.3.2  Metabolic models in man

        The kinetic parameters of uptake, distribution, and excretion of
    methylmercury in man are documented in much more detail than for other
    forms of mercury. However, questions still remain on the linearity of
    this metabolic model at high toxic doses of methylmercury.
    Specifically, the applicability of the metabolic model derived from
    human tracer-dose studies should be verified at higher dose levels.
    Information on this point would greatly facilitate the interpretation
    of results of epidemiological studies on heavily exposed populations.

        Recent findings of large individual variations in clearance
    half-times of methylmercury from blood are of considerable importance
    in the estimation of risks from long-term dietary intake. Further
    studies are needed to establish the statistical parameters of the
    distribution of individual clearance half-times, and on the biological
    mechanisms underlying these differences.

        A more complete metabolic model for inhaled mercury vapour in man
    is urgently needed. Despite the continuous occupational exposure of
    thousands of workers annually and the long history of man's exposure
    to this form of mercury, we still do not have sufficient information
    to relate mercury concentrations in air to accumulated body burdens
    and to identify the most appropriate indicator media for levels of
    mercury vapour in the target organ (the brain). Animal experiments
    have indicated the ability of the inhaled vapour to cross the
    placenta; no information is available on human subjects concerning
    this important question.

    1.3.3  Epidemiological studies

        Several types of epidemiological study are needed. Long-term
    studies on adults should concentrate on those areas of the
    dose-response relationship where the effects of methylmercury become
    just detectable. There are still uncertainties concerning the
    concentrations of total mercury in indicator media and the equivalent
    long-term daily intake of mercury as methylmercury associated with the
    earliest effects in the most sensitive group in the adult population.

        So far, dose-response relationships in human populations have been
    based on outbreaks of poisoning in which daily exposure was high and
    limited to months or a few years at the most. To extrapolate these
    relationships to the general population, more information is needed on
    the potential influence of long-term exposure.

        In addition to continuing studies on mature adults, groups of the
    population specially sensitive to methylmercury should be identified.
    Special studies should be made on the relationship between the dose
    received by the expectant mother and the effect on her infant
    including the development and growth of the child.

        Further epidemiological studies are needed on groups
    occupationally exposed to mercury vapour. Whenever possible,
    collaborative studies should be carried out in which cohorts should be
    followed in time and different groups related to each other.

    1.3.4  Interaction of mercury with other environmental factors

        The extrapolation to the general population of epidemiological
    data from outbreaks of methylmercury poisoning that have occurred in
    certain parts of the world is fraught with uncertainties, unless the
    possible interaction of local environmental factors can be taken into
    account. For example, the conditions under which selenium exerts
    antagonistic and synergistic effects and its mode of action should be
    studied. Alcohol influences the metabolism of mercury and may affect
    the toxicity of inhaled vapour in man. Genetic factors should also be
    considered. Acatalasaemic individuals may metabolize inhaled mercury
    vapour differently from normal individuals.

        Mercury, along with other heavy metals, has the potential to alter
    the activity of drug metabolizing enzymes. Studies should be made on
    these potential effects with special emphasis on those individuals
    carrying high body burdens of mercury.

    1.3.5  Biochemical and physiological mechanisms of toxicity

        Long-term investigations of the mode of toxic action of mercury
    and its compounds are needed to give an insight into the causes of
    individual differences in sensitivity to mercury and into differences
    in metabolism such as clearance half-times. Methylmercury is known to
    produce "silent damage" in that morphological changes can be seen in
    the brains of experimental animals before functional disturbances are
    detectable. Biochemical disturbances such as inhibition of protein
    synthesis precede overt signs of damage. There is a great need to
    develop sensitive biochemical and physiological tests, especially in
    the case of methylmercury poisoning.

        A deeper understanding of the toxic action of mercury should lead
    to the development of more effective means of treatment. Present
    methods depend mainly on prevention, using complexing and chelating
    agents to remove the metal from the body before serious damage has


    2.1  Chemical and Physical Properties

        Mercury can exist in a wide variety of physical and chemical
    states. This property presents special problems to those interested in
    assessing the possible risk to public health. The different chemical
    and physical forms of this element all have their intrinsic toxic
    properties and different applications in industry, agriculture, and
    medicine, and require a separate assessment of risk.

        The chemistry of mercury and its compounds has been outlined in
    several standard chemistry texts (Rochow et al., 1957; Gould, 1962;
    Cotton & Wilkinson, 1972). Mercury, along with cadmium and zinc, falls
    into Group IIb of the Periodic Table. In addition to its elemental
    state, mercury exists in the + 1 (mercury(I)) and +2 (mercury(II))
    states in which the mercury atom has lost one and two electrons,
    respectively. The chemical compounds of mercury(II) are much more
    numerous than those of mercury(I).

        In addition to simple salts, such as chloride, nitrate, and
    sulfate, mercury(II) forms an important class of organometallic
    compounds. These are characterized by the attachment of mercury to
    either one or two carbon atoms to form compounds of the type RHgX and
    RHgR' where R and R' represent the organic moiety. The most numerous
    are those of the type RHgX. X may be one of a variety of anions. The
    carbon-mercury bond is chemically stable. It is not split in water nor
    by weak acids or bases. The stability is not due to the high strength
    of the carbon-mercury bond (only 15-20 cal/mol and actually weaker
    than zinc and cadmium bonds) but to the very low affinity of mercury
    for oxygen. The organic moiety, R, takes a variety of forms, some of
    the most common being the alkyl, the phenyl, and the methoxyethyl
    radicals. If the anion X is nitrate or sulfate, the compound tends to
    be "salt like" having appreciable solubility in water; however, the
    chlorides are covalent non-polar compounds that are more soluble in
    organic solvents than in water. From the toxicological standpoint, the
    most important of these organometallic compounds is the subclass of
    short-chain alkylmercurials in which mercury is attached to the carbon
    atom of a methyl, ethyl, or propyl group.

        An expert committee, considering occupational hazards of mercury
    compounds, distinguished two major classes of mercury compounds --
    "organic" and "inorganic" (MAC Committee, 1969). Inorganic mercury
    compounds included the metallic form, the salts of mercury(I) and
    mercury(II) ions, and those complexes in which mercury(II) was
    reversibly bound to such tissue ligands as thiol groups and protein.
    Compounds in which mercury was directly linked to a carbon atom by a

    covalent bond were classified as organic mercury compounds. This
    distinction is of limited value because the toxic properties of
    elemental mercury vapour differ from those of the inorganic salts and,
    furthermore, the short-chain alkylmercurials differ dramatically from
    other mercurials that fall within the definition of organic mercury.
    From the standpoint of risk to human health, the most important forms
    of mercury are elemental mercury vapour and the short-chain

        Mercury in its metallic form is a liquid at room temperature. Its
    vapour pressure is sufficiently high to yield hazardous concentrations
    of vapour at temperatures normally encountered both indoors and
    outdoors under most climatic conditions. For example, at 24C, a
    saturated atmosphere of mercury vapour would contain approximately
    18 mg/m3 -- a level of mercury 360 times greater than the average
    permissible concentration of 0.05 mg/m3 recommended for occupational
    exposure by the National Institutes of Safety and Health, USA (NIOSH,
    1973). Apart from the noble gases, mercury is the only element having
    a vapour which is monatomic at room temperature. However, little is
    known about the chemical and physical states of mercury found in the
    ambient air and in the air where occupational exposure occurs.

        Elemental mercury vapour is generally regarded as insoluble.
    Nevertheless, small amounts dissolved in water and other solvents are
    important from the toxicological point of view. At room temperatures,
    in air-free water, its solubility is approximately 20 g/litre. In the
    presence of oxygen, metallic mercury is rapidly oxidized to the ionic
    form -- mercury(II) -- and may attain concentrations in water as high
    as 40 g/litre.

        Calomel or mercury(I) chloride (Hg2Cl2) is the best known
    mercury(I) salt. Widely used in the first half of this century in
    teething powders and in anthelmintic preparations, the low toxicity of
    this compound is due principally to its very low solubility in water.
    Mercury(I) forms few complexes with biological molecules. However, in
    the presence of protein and other molecules containing SH groups, it
    gives one atom of metallic mercury and one mercury(II) ion. In
    general, an equilibrium is established between Hg0, Hg2++ and Hg++
    in aqueous solution. The distribution of mercury between the three
    oxidation states is determined by the redox (oxidation-reduction)
    potential of the solution and the concentration of halide, thiol, and
    other groups that form complexes with Hg++. The dissociation of
    mercury(I) chloride by thiol groups should be understood in this
    context. Extra halide and thiol compounds, added to solution, form
    complexes with mercury(II) ions and the mercury(I) chloride splits to
    restore the equilibrium between Hg0, Hg2++ and Hg++. The split
    results in the formation of one atom of mercury for every mercury(I)
    chloride molecule dissociated.

        The mercury(II) ion, Hg++, is able to form many stable complexes
    with biologically important molecules. Mercury(II) chloride (corrosive
    sublimate), a highly reactive compound, readily denatures proteins and
    was extensively used in the past century as a disinfectant. It is
    soluble in water and, in solution, forms four different complexes with
    chloride, HgCI+, HgCl2, HgCl3- and HgCl4=. It has been
    suggested that the negatively charged chlorine complexes are present
    in sea water (see section 5).

        Phenylmercury compounds have a low volatility. However, the halide
    salts of methyl-, ethyl-, and methoxyethylmercury can give rise, at
    20C, to saturated mercury vapour concentrations of the order of 90,
    8, and 26 mg/m3, respectively (Swensson & Ulfvarsson, 1968). In the
    case of methylmercury this saturated vapour concentration is several
    orders of magnitude greater than the maximum allowable concentration
    in the working atmosphere. This hazardous property of the halide salts
    of the short-chain alkylmercurials is not always fully appreciated in
    industrial and agricultural use and even in research laboratories
    (Klein & Hermen, 1971). In contrast, methylmercury dicyandiamide,
    previously widely used as a fungicide, has a much lower vapour
    pressure, being 340 times less volatile than the chloride salt.

        Although the carbon-mercury bond is chemically stable, in the
    living animal, the bond is subject to cleavage (for review, see
    Clarkson, 1972a). The nature of the R radical is all important. If R
    is a phenyl or methoxyalkyl group, rapid breakdown occurs in animal
    tissues so that most of the organic compound has disappeared within a
    few days. Enzymes that break the carbon mercury bond have been
    discovered and isolated (Tonomura et al., 1968a, 1968b, 1968c). The
    short-chain alkylmercurials undergo the slowest breakdown  in vivo
    with methylmercury being the most stable. Differences in the stability
    of the carbon-mercury bond play an important role in determining the
    toxicity and mode of action in man. The rapid breakdown of phenyl- and
    methoxymercury results in toxic effects similar to those of inorganic
    mercury salts. The relative stability of the alkylmercurials is one
    important factor in their unique position with regard to toxicity and
    risks to human health.

        The organic and inorganic cations of mercury, in common with other
    heavy metal cations, will react reversibly with a variety of organic
    ligandsa found in biologically important molecules. The chemical
    affinity of mercury(II) and of its monovalent alkylmercury cations for
    a variety of biologically occurring ligands is so great that free
    mercury would be present  in vivo at concentrations so low as to be
    undetectable by present methods.

    2.2  Purity of Compounds

        Impurities in mercury and its compounds are not important in
    assessing the hazards to man. Those compounds of mercury used in
    industry and agriculture have impurities of less than 10%. Bakir et
    al. (1973) reported that a methylmercury fungicide responsible for an
    epidemic of poisoning in Iraq contained 10% or less of ethylmercury as
    an impurity. Inorganic mercury usually amounts to no more than 1% of
    the total mercury in organomercurial preparations and rarely exceeds

        Impurities are of importance in the preparation of standard
    solutions for analytical procedures and in experimental research in
    animals where impurities in radioactive mercury may give misleading
    results. Preparations of methylmercury labelled with the isotope 203Hg
    are subject to radiolytic breakdown to inorganic compounds depending
    on the pH. This instability must be taken into account in the
    interpretation of some original reports in which the purity of the
    radioisotope was not checked properly.

    2.3  Sampling and Analysis

        Before reviewing various aspects of sample collection and analysis
    it may be worth taking an overview of the various sources of error in
    the determination of mercury content. Not only are there errors in the
    instrumental determination of mercury and in the laboratory
    procedures, but significant and often major errors occur during the
    collection, transportation, and storage of the samples. The accuracy
    of the determination of mercury in environmental samples should be
    assessed from this broad point of view. The error will be the sum of


    a Ligands are chemical groups within a molecule that are capable of
      donating electrons to a metal cation to form a chemical bond.
      Examples of biologically important ligands are the carboxyl, and
      especially with regard to heavy metals, the sulfhydryl (SH)

    the errors in collection, storage, transportation and, in the
    instrumental determination. It is of the greatest importance to
    determine the greatest source of error in each particular case. This,
    in itself, may lead to considerable improvement in the overall
    accuracy of the determination. For example, the introduction of a new
    and more sensitive instrumental technique may allow the collection of
    smaller samples and thus facilitate storage and transport. On the
    other hand, there is little value in proceeding further with
    improvements in instrumental measurements if major errors remain at
    the collection, storage, or transport stages.

    2.3.1  Sample collection

        Methods of sample collection for the determination of mercury in
    air have recently been reviewed (NIOSH, 1973). A recommended method
    for the determination of total mercury in air is presented.
    Essentially the method consists of using two bubblers in series,
    containing sulfuric acid and potassium permanganate. The mercury in
    these traps is subsequently determined by atomic absorption
    procedures. Problems of the determination of mercury in air are
    critically evaluated. Included in these problems is the fact that
    numerous chemical and physical forms of mercury may exist in air and
    that these are subject to interconversion. The volatility of mercury
    and its compounds is a special problem in the determination of mercury
    bound to particles. The separation of particulates from air, such as
    by filtration, may result in the loss of mercury by volatilization
    from the particulate. Published methods of sample collection consist
    of removal of mercury from the air by passing it through scrubbing
    devices, or direct collection of the air sample, for example in a
    plastic bag or syringe. The scrubbing device may take the form of
    bubblers, filters, absorbants, or amalgam collectors. Unfortunately
    many of the published procedures do not report collection efficiency.
    Attention is drawn to the importance of the use of standard dust
    chambers to check the efficiency of absorption.

        The procedure recommended by NIOSH (1973) has a collection
    efficiency for total mercury of more than 90%, when mercury is in the
    form of elemental vapour or inorganic salts. Organomercurials in air
    are collected with an efficiency of more than 80%, except in the case
    of the short-chain alkylmercurials. Bramen (1974) has described a
    procedure for separating and measuring different physical and chemical
    forms of mercury in air. Previous reports distinguishing between
    mercury vapour and particle-bound mercury have not reported the
    efficiency of collection.

        An early method (Polesajev, 1936) for the determination of mercury
    in air involved absorption in iodine and subsequent determination of
    the coloured complex in the sediment. This method is still widely used
    in the Soviet Union and some countries of eastern Europe.

        Commercially available portable monitoring devices are used to
    determine mercury directly in air. The air is pumped through an
    optical cell that measures the absorption of light emitted from a
    mercury vapour lamp. These units, although convenient, measure only
    elemental mercury vapour and are subject to a wide variety of
    interferences and interfering substances many of which are likely to
    be present in the working environment. These units should be
    calibrated each time before use. The commercial units also suffer from
    the deficiency that they sample only small volumes of air that may not
    give a representative picture of the working environment. Research
    should be directed towards the development of personal monitoring
    devices. These devices should be small and portable so that they can
    be carried by workmen throughout the working day and thereby give a
    cumulative picture of the exposure of each individual. In most cases
    it would be necessary only to devise systems for collecting total

        The method of Wolf et al. (1974) allows the direct detection of
    mercury using reactive tubes (Draeger tubes) providing a simple
    screening method for determining mercury in working places at sporadic

        The collection of samples for the determination of mercury in
    water must take into account the following factors; (a) the low
    concentration of mercury in water, normally of the order of
    10 ng/litre; (b) the tendency of mercury to adsorb on to the surface
    of the collection vessel at these low concentrations; (c) the
    possibility, if not likelihood, of volatilization of mercury from the
    sample (Toribara et al., 1970) and (d) the type of collection vessel.
    Greenwood & Clarkson (1970) have reported on the rates of loss of
    mercury from containers made from ten different materials and
    suggested that Pyrex, polycarbonate, and Teflon are the best materials
    for storing and handling mercury. Further studies of possible losses
    of organomercurials through the walls of some plastic containers
    should, however, be studied. Losses due to volatilization may be
    reituced by the addition of oxidizing substances such as potassium
    permanganate (Toribara et al., 1970). Lamm & Ruzika (1972) have
    recommended that radioactive-tracer mercury be added to the sample to
    check the losses discussed above. They note that this procedure has
    rarely been adopted to date.

        For the collection and storage of food samples, acceptable
    procedures are usually followed. The most important food items for
    determination of mercury are those containing fish and fish products.
    Mercury levels in other foodstuffs usually do not amount to a
    significant fraction of daily exposure unless the food has accidently
    been contaminated, such as by the use of pesticides. In the collection
    and storage of food samples prior to analysis, care should be taken to
    avoid bacterial growth leading either to the breakdown of organic
    mercury compounds or to the volatilization of mercury (Magos et al.,

        Samples of blood, hair, and urine have been used to monitor the
    exposure of human beings to mercury. The methods of collecting and
    storing these samples are of great importance. With respect to blood
    samples, care should be exercised to avoid any clot formation. If this
    does occur, the sample should be homogenized thoroughly before
    analysis. It is useful, in certain situations, to determine mercury in
    the red cells and plasma and it is thus important to avoid any
    haemolysis of the blood sample. The nature of the anticoagulants used
    does not affect the mercury determinations, of either the total
    mercury in whole blood or the distribution of mercury between plasma
    and red blood cells. "Vacutainers"a are convenient for blood
    collection and allow storage of the blood samples in Pyrex tubing
    under aseptic conditions. Blood samples that have been contaminated by
    microorganisms and stored in the refrigerator at 4C for a month or
    more may give misleading results due to the breakdown of methylmercury
    and other organic mercury compounds (Clarkson, personal communication,
    1974). The storage of blood samples in the frozen state or
    freeze-dried is suitable providing that mercury is determined only for
    whole blood. Significant losses of mercury do not occur during
    freeze-drying procedures (Albanus et al., 1972).

        Measurement of mercury in urine samples has been used as a measure
    of exposure to mercury under industrial conditions. The popularity of
    this approach in early studies was mainly due to the case of digestion
    of the urine sample. However, there are serious problems in the
    collection and storage of urine samples that may seriously influence
    the results. The following factors have been recognized;  (a) the
    time of day of urine collection (Piotrowski et al., 1975),
     (b) bacterial contamination, which might give rise to significant
    losses of mercury by volatilization (Magos et al., 1964),  (c) the
    nature of the container (Greenwood & Clarkson, 1970),
     (d) contamination from mercury in workers' clothing and from the
    collection of urine samples under working conditions. It should be
    noted that urine samples do not give a reliable indication of exposure
    to methylmercury (Bakir et al., 1973).

        Hair samples are becoming the samples of choice in determining
    exposure to methylmercury through diet. Depending upon the length of
    the hair sample, it is possible to recapitulate exposure to
    methylmercury for several yearsb. The concentration of mercury in
    hair when formed is directly proportional to the concentration of


    a Trade name of heparinized test-tube manufactured by Becton &
      Dickinson, USA, and used for collection of blood samples.

    b The average rate of growth of hair is approximately 1 cm
      per month (Giovanoli et al., 1974; Shahristani & Shihab,

    mercury in the blood, the concentration in hair being about 250 times
    the concentration in blood. The ratios are well established for
    exposure to methylmercury but only limited information is available
    for inorganic mercury. Attention has been drawn to the errors
    introduced during the collection and transportation of hair samples
    (Giovanoli & Berg, 1974). Usually 50-100 strands of hair are needed
    for analysis. Differential rates of growth for each strand and lateral
    displacement of the samples during cutting and transportation of the
    hair will affect the longitudinal profiles of mercury in the hair
    sample. Giovanoli & Berg (1974) have described a computerized
    procedure for the correction of these artifacts.

    2.3.2  Analytical methods

        Methods of analysis are usually classified according to the type
    of instrument used in the final measurement. This convenient
    classification will be used here. However this approach tends to
    belittle the role of the skill and experience of the analyst. In fact
    a poor method in the hands of a highly skilled analyst is more likely
    to yield accurate results than a good method in the hands of a poor
    analyst. In recent years it has become a practice to test methods by a
    "round robin" distribution of a standard sample. Comparison of results
    from the participating laboratories is more likely to give information
    on the competence of the analysts in the laboratory than it is to give
    a critical evaluation of the method itself.

        Measurement of the very low levels of mercury found in the
    non-contaminated environment makes special demands both on the skills
    of the analyst and the resources of the method employed. No matter how
    frequently used, a method for the determination of mercury in nanogram
    quantities cannot be regarded as a routine procedure. Continued
    vigilance over the results is an absolute requirement. Furthermore,
    where conditions allow, it is highly desirable that the results with
    one method and from one laboratory be checked against those with a
    different method from another laboratory. One useful combination of
    different procedures is the analysis of total and inorganic mercury by
    selective atomic absorption and the selective analysis of organic
    mercury compounds (usually methylmercury and other short-chain
    mercurials) by gas chromatography (Giovanoli et al., 1974).

        The literature is full of papers concerning methods of determining
    mercury. Several recent reviews have appeared (D'Itri, 1972; NIOSH,
    1973; Burrows, 1975, Swedish Expert Group, 1971; Wallace et al., 1971;
    CEC Working Group of Experts, 1974). The most frequently used methods
    for measurements of total mercury are colorimetric (dithizone),
    flameless atomic absorption, and neutron activation. The flameless
    atomic absorption method has become the "work-horse" for measurement
    of environmental samples. Difficulties might arise in the measurement
    of mercury owing to the fact that it is strongly bound to the organic
    materials in most samples. Many procedures require the destruction of

    organic materials by wet oxidation or by high temperatures. Loss of
    mercury by volatilization may occur. If the wet oxidation is too mild
    the result will be inadequate recovery. A high reagent blank may be
    introduced by the chemicals used for oxidation. In certain procedures
    involving atomic absorption or neutron activation the digestion of the
    sample or heating of the sample is not necessary. These procedures
    have the advantage of having a low blank but problems of variable
    recovery or interference may arise.

        The determination of mercury by colorimetric measurement of a
    mercury dithizonate complex has been the basis of most of the methods
    in the 1950s and in the 1960s. Other related methods using dithizone
    for measuring mercury in environmental samples have been described by
    Kudsk (1964) and Smart et al. (1969). The above procedures all make
    use of wet oxidation of the sample followed by extraction of mercury
    in an organic solvent as a dithizonate complex and finally the
    colorimetric determination of the complex itselfa. Selectivity for
    mercury is obtained by adjusting the conditions of extraction. Copper
    is the metal most likely to interfere with mercury measurement by

        The dithizone procedure has an absolute sensitivity of about
    0.5 g of mercury. A sample size of 10 g is suitable for most
    digestion procedures so that mercury can be determined at the
    0.05 mg/kg level in most foodstuffs and tissues.

        Kudsk (1964) has described a dithizone procedure for measuring
    mercury in air that will measure as little as 0.05 g of mercury. With
    the usual sample size of 0.1 m3, the detection limit would be
    0.5 g/m3. This is more than adequate sensitivity for monitoring air
    in the working environment with the MAC levels in force. The quoted
    recovery rates from foodstuffs and tissues are in the range of 85-99%
    and the reproducibility can yield a coefficient of variation of as low
    as 2%. On account of its long history of use, the dithizone procedure
    has been used to measure mercury in virtually all types of
    environmental samples including air, water, food, tissues, and soils.
    It suffers from the disadvantage that it is time consuming and its
    sensitivity is not high when compared with atomic absorption


    a The organic material may also be destroyed by combustion in an
      oxygen flask (Gutenmann & Lisk, 1960; White & Lisk, 1970; and
      Fujita et al., 1968). This allows all biological materials to be
      treated alike but has the disadvantage of requiring dried

        The latest developments in atomic absorption procedures have
    recently been reviewed by Burrows (1975). The most commonly used
    method in the USA is that of Hatch & Ott (1968) as modified by Uthe et
    al. (1970). The procedure involves oxidative digestion ("wet ashing"),
    followed by reduction, aeration, and measurement of mercury vapour
    absorption at 253.7 nm. The detection limit is approximately 1-5 ng of
    mercury. The wide popularity of cold vapour atomic absorption has
    resulted in a large number of publications dealing with various
    applications of this procedure to the measurement of mercury in
    sediments, soils, and biological samples (including foodstuffs). Of
    the 16 publications reviewed by Burrows (1975), 13 reported recoveries
    of 90% or more. The relative standard deviation was 10% or less in
    half of the published procedures, and was less than 20% in more than
    90% of these procedures.

        The measurement of very low levels of mercury in water samples
    requires some preconcentration. This may be achieved by dithizone
    extraction (Chau & Saiton, 1970; Thomson & McComas, 1973), by
    electrodeposition (Doherty & Dorsett, 1971) and by an amalgamation on
    silver wire (Hinkle & Learned, 1969; Fishman, 1970), in each case
    permitting detection limits of 1 ng/litre-10 ng/litre. Winter &
    Clements (1972) have described a procedure that will measure mercury
    in water in the range of 200 ng/litre and does not require

        Magos (1971) has described a reduction technique that selectively
    determines total and inorganic mercury in biological samples without
    digestion of the material. This technique has been modified by Magos &
    Clarkson (1972) to permit determination of mercury in blood samples at
    the low levels found in unexposed populations (0.1-1.0 g/100 ml). The
    technique has a sensitivity of approximately 0.5 ng of mercury.
    Recently it has been successfully applied to the measurement of total
    and inorganic mercury in hair samples (Giovanoli et al., 1974). The
    relative standard deviation was 2% and the recovery rates were quoted
    as being close to 100%. The technique has the advantage of high speed
    -- each determination taking less than 2 minutes -- high sensitivity,
    and the apparatus involved is light, portable, and suitable for field
    applications. Its widest application to date has been in the
    measurement of mercury in biological samples in the large Iraq
    outbreak (Bakir et al., 1973). Since the procedure does not require
    digestion of the biological sample, internal standards are used in
    each determination. The rates in this procedure must be checked for
    each new biological matrix.

        The atomic absorption techniques referred to above are subject to
    interference. The most common interfering substances are benzene and
    other aromatic hydrocarbons that absorb strongly in the 253.7 nm
    region. Interference from a variety of organic solvents has been
    reported by Kopp et al. (1972).

        The combustion-amalgamation method has undergone a series of
    developments to avoid difficulties due to interfering substances.
    Reference may be made to the work of Lidmus & Ulfvarson (1968), Okuno
    et al. (1972), and Willford (1973) who developed techniques for
    oxidation of the biological sample, and the trapping of mercury vapour
    on silver or gold followed by its release into an atomic absorption
    measuring device. All these methods have sensitivities down to the
    1 g/litre level and avoid the risk of interference from other
    substances. However, as pointed out by Burrows (1975), care must be
    taken in the design and operation of the combustion tube to avoid
    losses of volatile mercury derivatives.

        In summary, a wide variety of applications of atomic absorption
    procedures have now been published. The technique is rapid and
    sensitive and the procedure is technically simple. Procedures are
    available for avoiding difficulties due to interfering substances.
    Most procedures have a detection limit in the range of 0.5-5 ng of
    mercury and a relative standard deviation of about 10% or less.
    Recovery rates are usually of the order of 95-100% depending on the
    technique used in the preparation of the biological sample and the
    rate of release of mercury from it.

        Procedures for neutron activation analysis of total mercury have
    recently been reviewed by Wallace et al. (1971), Swedish Expert Group
    (1971), Westermark & Ljunggren (1972), and Burrows (1975). The method
    is based on the principle that when natural mercury (a mixture of
    stable isotopes) is exposed to a high flux of thermal (slow) neutrons,
    it is converted to a mixture of radioactive isotopes, principally
    197Hg and 203Hg, which have decay half-lives of 65 hours and 47 days,
    respectively. The Sjostrand (1964) technique has been used most in the
    measurement of environmental samples. After the sample has been
    irradiated with neutrons, a precise weight of carrier mercury is added
    and the sample subjected to digestion and organic destruction. On
    completion of digestion, mercury is isolated by electrodeposition on a
    gold foil and the radioactivity is determined with a gamma counter.
    The use of carrier mercury corrects for any losses of mercury during
    the digestion, extraction, and isolation procedures. The limit of
    detection is 0.1-0.3 ng of mercury. The sample size is 0.3 g giving a
    concentration limit of 0.3-1 g/kg in most biological samples. The
    relative standard deviation in samples of kale, fish, minerals, oil,
    blood, and water is less than 10%. Samuel (unpublished data)
    decomposed biological material irradiated with neutrons using fuming
    sulfuric acid and hydrogen peroxide and after the addition of hydrogen
    bromide, distilled the mercury as bromide together with other trace
    elements. This method, which is suitable for series analysis, is
    characterized by high recovery (96%) and good reproducibility. Trace
    mercury in biological and environmental materials can also be rapidly
    and satisfactorily determined through isolation as mercury(II) oxide
    or mercury(II) sulfide after digestion and clean-up procedures
    following neutron activation (Pillay et al., 1971; Samuel, unpublished

        In general, the analyst is faced with three major options in the
    use of neutron activation procedures;  (a) destruction or
    non-destruction of the sample, (destruction and isolation of the
    mercury is usually required in samples containing less than 1 g of
    mercury);  (b) the choice of isotope 197Hg (if the longer-lived
    isotope, 203Hg, is used the sample may be allowed to stand to avoid
    interference from short-lived elements activated along with the
    mercury -- however, 203Hg requires a more intense neutron flux or a
    longer irradiation time to achieve the same activity as the 197Hg);
     (c) the choice of detector (the sodium iodide (thallium) detector
    does not have as high a resolution as the germanium (lithium)
    detector, although its sensitivity is significantly higher).

        Interference may come from the following elements, produced at the
    same time as the radioactive mercury isotopes, 24Na, 82Br, 32P, and
    75Se. Interference from these isotopes may be avoided, as in the
    Sjostrand (1964) procedure, by chemical isolation of the radioactive
    isotope. However, 75Se may not be completely removed by the isolation
    procedures and might interfere if the sodium iodide (thallium)
    detector is used. The better resolution of the germanium (lithium)
    detector allows correction for 75Se interference through use of other
    lines in the 75Se spectrum. For samples containing more than 1 g of
    mercury, the required selectivity can be achieved without destruction
    of the sample, i.e., by instrumental analysis only. One procedure is
    to measure the 203Hg isotope, after allowing the sample to stand for
    approximately one month to eliminate interference due to sodium,
    phosphorous, and bromine. Another procedure is to make use of the
    discriminating germanium (lithium) detector when the gamma irradiation
    from the radioactive isotope may be determined to the exclusion of
    most of the interfering radioactivity.

        A recent non-destructive procedure for measuring mercury in coal
    makes use of a low-energy photon detector to estimate levels at the
    100 g/kg level with a precision of 10% (Weaver, 1973).

        Burrows (1975) has recently reviewed 11 publications describing
    the application of neutron activation to a variety of environmental
    samples. Non-destructive (instrumental) determination was used in only
    two of these publications. In 9 of these publications the 197Hg
    isotope was determined. Mercury levels were reported in lake water
    (4 g/litre, relative standard deviation 23%), in glacial ice
    (0.2 g/kg, relative standard deviation 90%), in coal (100 g/kg,
    relative standard deviation 10%), in whole blood (0.7 g/100 ml,a
    relative standard deviation 10%), in fish (1-3 mg/kg, relative
    standard deviation less than 10%). Many environmental samples were
    measured by neutron activation, especially in Sweden, before the
    introduction of the atomic absorption technique (Westermark &
    Ljunggren, 1972).

        Compared with other methods reviewed here, the neutron activation
    procedure has the following advantages; (1) high sensitivity
    (approximately 0.5 g/kg); (2) no reagent blank; (3) independence from
    the chemical form of the element; and (4) non-destructive instrumental
    methods applicable to samples containing 1 g of mercury or more. It
    has the disadvantages that it cannot be adapted to field use and, that
    if there are large numbers of samples, special radiation facilities
    and data processing are required. It is generally agreed that the
    neutron activation procedure finds its most important use as a
    reference method against which other procedures can be checked.

        A variety of other instrumental techniques, such as X-ray
    fluorescence, mass spectrometry, and atomic fluorescence, for the
    measurement of total mercury have been reviewed by Lamm & Ruzicka
    (1972) and by Burrows (1975). In general, some of these methods may
    have a potentially higher sensitivity or selectivity for mercury. The
    fact is that, at the time of writing, these procedures have not yet
    found useful application in the measurement of mercury in
    environmental samples.

        To summarize the present methods for the determination of total
    mercury in environmental samples, it would appear that the method of
    choice is that of flameless atomic absorption. No single procedure is
    appropriate, however, in all circumstances. The methods of sample
    handling depend upon the particular biological matrix to be analysed.
    Neutron activation is principally of use as a reference method against
    which atomic absorption methods may be checked.


    a In this document the concentration of mercury in blood is
      expressed in g/100 ml although in some original papers the values
      are given in g/100 g. For practical purposes the difference of
      about 5% can be neglected.

    2.3.3  Analysis of alkylmercury compounds in the presence of
           inorganic mercury

        Techniques for the identification and measurement of alkylmercury
    compounds in the presence of other compounds of mercury have been
    reviewed recently (Swedish Export Group, 1971; Tatton, 1972; Sumino,
    1975; West, 1973). In general, three methods are available for the
    identification of alkylmercury compounds. These include  (a) paper
    chromatography (Kanazawa & Sato, 1959; Sera et al., 1962),  (b) thin
    layer chromatography (Johnson & Vickery, 1970; West, 1966, 1967;
    Tatton & Wagstaffe, 1969),  (c) gas-liquid chromatography (West,
    1966, 1967; Sumino, 1968; Tatton & Wagstaff, 1969). The paper
    chromatographic techniques have given way to thin-layer chromatography
    (TLC) for qualitative identification of the organomercurial compounds.
    Most quantitative work is now carried out using TLC techniques, and
    also gas-liquid chromatography (West, 1966, 1967; Sumino, 1968;
    Tatton & Wagstaffe, 1969; Solomon & Uthe, 1971). However, the method
    of Magos & Clarkson (1972) that selectively determines organic mercury
    by cold vapour atomic absorption is frequently applicable to the
    determination of methylmercury at levels occurring in fish and blood.
    Methylmercury is the only organic form of mercury present in fish.
    Blood samples from people exposed to methylmercury contain only
    inorganic mercury and methylmercury compounds. Thus the determination
    of organic mercury by this procedure is an accurate measure of
    methylmercury in these situations.

        The basic procedures for samples of food, soil, and biological
    materials are first, homogenization of the sample, acidification by a
    hydrogen halide acid followed by extraction with an organic solvent,
    usually benzene, a clean-up step involving the conversion of the
    organomercurial compound to a water soluble compound usually the
    hydroxide or sulfate or a cysteine complex, and re-extraction with
    benzene. The benzene layer is now ready for analysis by thin-layer
    chromatography for qualitative purposes or by gas-liquid
    chromatography if quantitative measurements are required. A recent
    variant by Rivers et al. (1972) converts the organic into inorganic
    mercury and then makes use of cold vapour atomic absorption for final

        The gas-liquid chromatographic system is the one most commonly
    used. Problems may be encountered both in the pre-treatment of the
    sample and in the gas chromatographic determination itself. All these
    techniques involve non-destructive extraction of mercury from the
    sample. Thus recovery rates have to be checked for every different
    type of sample matrix. The efficiency of extraction of mercury is
    determined by both the nature of the sample matrix and the extraction
    procedures themselves. Von Burg et al. (1974) introduced the idea of

    adding a tracer amount of radioactively labelled methylmercury to the
    homogenate and counting the final benzene extract to check variations
    in the efficiency of extraction. This procedure is well worth
    consideration for routine use as it is most difficult to check
    extraction recovery rates.

        Acidification of the homogenate is usually achieved by the
    addition of a hydrogen halide acid (usually HCl). At this point
    mercury(II) chloride may be added to either the homogenate or the
    benzene to tie up excess sulfur compounds and prevent recombination of
    methylmercury with sulfur. West (1968) has shown that this approach
    may give high recovery rates but cannot be used with liver as there is
    a danger of methylation of the inorganic mercury. Clean-up of the
    first benzene extract is usually achieved by using solutions of
    cysteine. However, this complexing agent is subject to oxidation,
    particularly by substances in muds. A more suitable system in the
    presence of oxidizing agents is the ammonium hydroxide-sodium sulfate
    solution described by West. No problems are usually encountered in
    the reextraction ofmethylmercury from cysteine to benzene using
    3 mol/litre hydrochloric acid. However, in the extraction procedures,
    volumetric errors may arise especially when the concentration of
    hydrochloric acid is low (1 mol/litre) and when small amounts of
    methyl-mercury are extracted from large volumes (West, 1973).

        In gas chromatography, the main object is to produce sharp peaks
    and attain high sensitivity. Tatton (1972) has noted that most
    commercial preparations ofalkylmercury salts are not pure enough to
    use as standards. Sumino (1973) prepares pure methylmercury from the
    combination of inorganic mercury with tetramethyl lead salts. The peak
    is identified by electron-capture detectors using tritium or nickel as
    the source of beta particles. These detectors are subject to
    overloading and not more than 100 ng of mercury should be determined
    at one time (Tatton, 1972). Absolute confirmation of the identity of
    the peak should be made by mass fragmentation methods (Sumino, 1975).

        The detection limit in the West procedure is approximately
    1-5 g per kilogram of sample using a 10 g sample. The precision is 3%
    at the 0.05 mg/kg level for fish samples. Recovery rates are generally
    above 90% but do vary with the sample matrix. Solomon & Uthe (1971)
    developed a semimicro-method for the rapid determination of
    methylmercury in fish tissues. Samples of about 2 g were used. A
    precision of 2% was reported with recovery rates of about 99%. Samples
    such as blood, liver, and kidney are much more difficult to extract
    than fish tissues.

        Thin-layer chromatography usually requires, for optimum spot size,
    2 g of mercury for each type of compound.


        The sources of mercury leading to environmental pollution have
    been the subject of several recent reviews (Wallace et al., 1971;
    D'Itri et al., 1972; Joint FAO/WHO Expert Committee on Food Additives,
    1972; Heindryckx et al., 1974; Korringa & Hagel, 1974). Estimates of
    both natural and anthropogenic sources of mercury are subject to
    considerable error. In the first place the levels of mercury in
    environmental samples such as ice from Greenland are extremely low and
    close to the limit of sensitivity of the analytical methods. These low
    values are then converted by large multiplication factors (annual
    total global rainfall, 5.2 x 105 km3) so as to obtain values for the
    global sources and turnover of mercury. Enormous fluctuations may be
    seen in samples such as coal and oil, which are believed to be an
    important anthropogenic source of mercury. Values quoted by D'Itri
    (1972) indicate ranges of concentrations of mercury in crude oil
    varying by a factor of 1000 and ranges in coal even greater than this.
    Estimates of industrial production and consumption of mercury are
    subject to the vagaries of the economic market and in recent years to
    government regulation because of concern over mercury pollution.
    Nevertheless, despite all the assumptions and approximations in these
    procedures, the general picture that emerges from a variety of
    independent calculations is that the natural sources of mercury are at
    least as great as, and may substantially outweigh, the anthropogenic
    sources. However, man-made sources may be of considerable importance
    in terms of local contamination of the environment. For example,
    Korringa & Hagel (1974) have calculated that the man-made release of
    mercury in the Netherlands is 100 times greater than the release of
    mercury by natural degassing processes.

    3.1  Natural Occurrence

        A recent review by the Joint FAO/WHO Expert Committee on Food
    Additives (1972) quotes the major source of mercury as the natural
    degassing of the earth's crust and quotes figures in the range of
    25 000-150 000 tonnes of mercury per year. These figures originate
    from a paper by Weiss et al. (1971) on concentrations of mercury in
    Greenland ice that was deposited prior to 1900. The most recent
    calculations on natural sources of mercury have been published by
    Korringa & Hagel (1974). These authors also made use of the figures of
    Weiss et al. (1971) to calculate the annual amount of mercury reaching
    the earth's surface due to precipitation of rainfall and arrived at a
    figure of approximately 30 000 tonnes. It was admitted that the
    sources of this atmospheric mercury are not yet clearly established
    but that volcanic gases and evaporation from the oceans are probably

    significant sources. It was also calculated by these authors that the
    run-off of mercury from rivers having a "natural mercury" content of
    less than 200 ng/litre would account for approximately 5000 tonnes of
    mercury per year. Measurements of the concentrations of mercury in air
    attached to aerosols (Heindryckx et al., 1974) indicate that soil
    dispersion to the atmosphere is not an important source of mercury.

        Significant local contamination may result from natural sources of
    mercury. For example, Wershaw (1970) has shown that water sources
    located near mercury ore deposits may contain up to 80 g/litre as
    compared with the levels of 0.1 g/litre in non-contaminated sources.

    3.2  Industrial Production

        According to a recent review by Korringa & Hagel (1974), world
    production averaged about 4000 tonnes per year over the period
    1900-1940. Production in 1968 was 8000 tonnes per year and, in 1973,
    attained 10 000 tonnes per year. Although considerable yearly
    fluctuations were noted, the average rate of increase since 1950 has
    been about 2% per year. Recent concern over environmental problems
    related to the use of mercury seems to have stabilized production
    rates and to have led to a dramatic fall in the price of mercury. For
    example, according to figures quoted by Korringa & Hagel (1974), the
    1966 price was $452 per flask (a flask is 34.5 kg), the 1969 price had
    risen to $510.00 but by 1972 it had fallen dramatically to $202 per

        It is difficult to estimate the amount of mercury released into
    the environment as a result of the mining and smelting of this metal.
    High levels of mercury in lake and stream waters have been attributed
    to the dumping of materials and tailings (for review, see Wallace et
    al., 1971). It has been estimated that stack losses during smelting
    operations should not exceed 2-3%. Thus, based on a production figure
    for mercury of 10 000 tonnes in 1973, one might expect to find losses
    to the atmosphere of the order of 300 tonnes per year.

    3.3  Uses of Mercury

        Wallace et al. (1971) have attempted to give a picture of the use
    of mercury in the USA. They note that 26% of the mercury mined is not
    reusable. They point out, however, that at least from the theoretical
    point of view most of the remaining mercury (i.e. 74% of the mercury
    mined) is reusable. To what extent these theoretical possibilities are
    attained is debatable at the present moment.

        Rauhut & Wild (1973) reported on the consumption and fate of
    mercury in the Federal Republic of Germany in 1971. Flewelling (1975)
    noted that the chloralkali industry, one of the largest users of
    mercury, has been able to cut losses in water effluent by at least 99%
    in the last two or three years; consequently losses from chloralkali

    plants now occur predominantly by emission into the atmosphere. Losses
    by volatilization into the atmosphere have been reduced (approximately
    50%) by the introduction of cooling systems for effluent gases.
    Korringa & Hagel (1974) take a more pessimistic point of view and
    conclude that there is every reason to assume that by about 1975 all
    the 10 000 to 11 000 tonnes of mercury produced per year due to mining
    operations will finally find its way into the environment,
    predominantly via the atmosphere.

        Average consumption patterns for industrialized countries have
    been summarized by Korringa & Hagel (1974) as follows: chloralkali
    plants, 25%; electrical equipment, 20%; paints, 15%; measurements and
    control systems, such as thermometers and blood pressure meters, 10%;
    agriculture, 5%; dental, 3%; laboratory, 2%; and other uses including
    military uses as detonators, 20%. This pattern of consumption in
    industrialized countries is similar to that published by D'Itri (1972)
    for the consumption in the USA in 1968. Included in "other uses" are
    mercury compounds in catalysts, preservatives in paper pulp
    industries, pharmaceutical and cosmetic preparations, and in
    amalgamation processes. The use of mercury in the paper pulp
    industries is dramatically declining and it was banned in Sweden in
    1966 (Swedish Expert Group, 1971). Hasanen (1974) has reported that no
    mercury compounds have been used in the paper pulp industry in Sweden
    and Finland since 1968.

    3.4  Contamination by Fossil Fuels, Waste Disposal, and
         Miscellaneous Industries

        Industrial activities not directly related to mercury can give
    rise to substantial releases of this metal into the environment. The
    most significant source is probably the burning of fossil fuels.
    Heindryckx et al. (1974) calculated the following approximate figures
    based on reports published in 1971 and 1972 (Joensuu, 1971; Cardozo,
    1972): the combustion of coal and lignite, 3000 tonnes per year; the
    refining and combustion of petroleum and natural gas, 400 tonnes per
    year; the production of steel, cement, and phosphate, 500 tonnes per
    year. Korringa & Hagel (1974) made similar calculations from published
    material (Joensuu, 1971; Filby et al., 1970; Cardozo, 1972; Weiss et
    al., 1971). They estimated for the year 1970, an annual release of
    3000 tonnes of mercury from coal burning, 1250 tonnes from mineral
    oil, and 250 tonnes from the consumption of natural gas. They expected
    that, by 1975, a total of 5000 tonnes of mercury would be emitted from
    burning fossil fuels.

        Smelting of metals from their sulfate ores should contribute some
    2000 tonnes annually and the making of cement and phosphate and other
    processes involving heating should have contributed another 5000
    tonnes per year by 1975.

        D'Itri (1972) points out that the disposal of sewage might be an
    important source of environmental mercury. Calculations from data in
    the literature indicate that somewhere between 200 and 400 kg of
    mercury per million population may be released from sewage disposal
    units. This would amount to approximately 40-80 tonnes per year for
    the entire poptilation of the USA. He further points out that sewage
    sludge can retain high amounts of mercury according to published
    studies from Sweden (6-20 mg/kg). This sludge is sometimes used as a
    fertilizer resulting in widespread dispersal of mercury or is
    sometimes heated in multiple hearth furnaces when most of the mercury
    would probably be released into the atmosphere. If the United States
    production is taken as being roughly 30% of world consumption, one
    might extrapolate the sewage release figure for the United States to
    indicate that something of the order of 1000 tonnes of mercury may be
    released frow sewage systems on a global scale.

        The anthropogenic release of mercury has been well summarized in a
    recent article by Korringa & Hagel (1974) and will be briefly stated
    here. The total global release of mercury is taken as the sum of the
    global production (following their pessimistic view that all will be
    released into the environment) plus the release from fossil fuels and
    natural gas and release from non-mercury related industries.

        It was calculated that by 1975 the total anthropogenic release of
    mercury on a global scale would be about 20 000 tonnes per year. These
    figures should be compared with a minimum estimated release of 25 000
    to 30 000 tonnes per year from natural sources. The latter figure may,
    in fact, be as high as 150 000 tonnes per year, given the
    uncertainties in calculations on the natural global release of


        Jenson & Jernelov (1972) have suggested different types of cycle
    for the distribution of mercury. One cycle is global in scope and
    depends upon the atmospheric circulation of elemental mercury vapour.
    The other cycle is local and is based on an assumed circulation of
    volatile dimethylmercury compounds. In the global cycle most of the
    mercury is derived from natural sources whereas the local cycle is
    predominantly concerned with man-made release.

    4.1  Distribution between Media -- the Global Mercury Cycle

        Recent calculations on the global circulation of mercury have been
    reported by Korringa & Hagel (1974). Their calculations are based
    principally on data giving mercury levels in ice samples collected in
    Greenland and in the Antarctic as reported by Weiss et al. (1971). The
    circulation of mercury from natural sources was calculated using a
    figure of 0.06 g of mercury per kilogram of Greenland ice samples
    collected prior to the year 1900. Using a reported figure for the
    global precipitation of water as 5.2 x 105 km3 per year, they
    estimated that minimum transport from the atmosphere to the earth
    should have been about 30 000 tonnes annually, prior to 1900. The
    contribution by dust particles was regarded as insignificant, an
    assumption now supported by the findings of Heindryckx et al. (1974).
    Based on a published figure of 4.1 x 105 km3 for annual
    precipitation over the oceans, these authors estimated the annual
    delivery of mercury to the oceans as 25 000 tonnes.

        Korringa & Hagel (1974) also calculated the contribution of the
    man-made release of mercury to the atmospheric transport cycle. They
    assumed that 16 000 tonnes of mercury is now released per year to the
    atmosphere from man-made sources and that the mercury is returned to
    the continental land surfaces and would soon re-evaporate to the
    atmosphere. The 16 000 tonnes per year would eventually find its way
    into the oceans and thus the annual delivery to the oceans from both
    natural and man-made sources would be 25 000 plus 16 000 tonnes which
    on a proportional basis should increase the background level from the
    0.06 g/kg observed prior to the 1900s in Greenland ice to a predicted
    level of 0.1 g/kg. However, they point out that since most of the
    man-made release is probably in the northern hemisphere, the present
    level in Greenland ice should be somewhat higher than 0.1 g/kg. They
    note that this estimate agrees well with the observations of Weiss et
    al. (1971) who found present levels in Greenland ice to range from
    0.09 to 0.23 g/kg with an average of 0.125 g/kg. Thus, from these
    rough estimates, it would appear that present day "background" levels
    in rainwater, and presumably in the atmosphere, have a substantial
    component related to man-made release (approximately one-third).

        Observations on "background" mercury levels in the atmosphere tend
    to confirm the quantitative features of this global picture
    (Heindryckx et al., 1974). These authors assume that 50 000 tonnes are
    released each year from the continental land masses, that the mercury
    mixes up to a height of 1 km and that, in effect, the 50 000 tonnes
    are located over the continental land masses that account for 30% of
    the earth's surface.a The assumption of the location of this mercury
    over the land masses is not in contradiction with the calculations of
    Korringa & Hagel (1974). It assumes only that the atmosphere above the
    land masses is in steady state, and receives 50 000 tonnes of mercury
    a year as evaporation and loses 50 000 tonnes per year to the
    atmosphere over the oceans. Their figure of 50 000 tonnes per year
    comes from the publication of Bertini & Goldberg (1971) and agrees
    well with the figure of 41 000 tonnes per year as indicated above.
    With these assumptions, Heindryckx et al. (1974) concluded that the
    background continental levels of mercury vapour plus aerosols should
    be 10 ng/m3. The assumed mixing height of 1 km is probably the
    maximum level and they suggest that the actual level of mercury in air
    would lie between 1 and 10 ng/m3. These figures are in good agreement
    with the published air levels as indicated in section 5.1.

        Korringa & Hagel (1974) estimate the amount of mercury transported
    by rivers to the oceans to be 5000 tonnes per year based on quoted
    figures of 37 000 km3 of water flow via the rivers and a natural
    mercury content of less than 0.2 g/litre in river water. They note
    that this figure does not change substantially if one takes into
    account the fact that most of the mercury in river water is adsorbed
    to suspended matter with a mercury content of 200-500 g/kg and that
    some 1010-1011 tonnes of sediment are carried each year to the
    oceans. In fact river transport of mercury to the oceans may be less
    than 5000 tonnes per year. Heindryckx et al. (1974) noted that the
    concentrations of mercury in the North Sea and in the coastal areas
    around the North Sea were far less than would be predicted if all the
    mercury in the rivers entering this area were, in fact, delivered into
    the oceans. Presumably a considerable amount of mercury observed in
    river water is retained in sediments in the rivers and estuaries and
    does not reach the ocean by normal flow of the river. Thus it would
    appear that the major pathway of global transport of mercury is
    metallic mercury transported in the atmosphere.


    a Recent studies in Sweden cast some doubt on the validity of this

        An important conclusion from these calculations on the global
    cycle of mercury is that the concentration of mercury in the oceans
    should not change substantially in the foreseeable future, and that
    the mercury concentration in the oceans has not changed significantly
    since the beginning of the industrial era. The amount of mercury in
    the oceans has been calculated as 70 million tonnes using a figure for
    total ocean volume of 1.37 x 109 km3 and taking the average mercury
    content of ocean water as 50 ng/litre. Thus contrary to what has been
    observed for the mercury content of the atmosphere, it will be a long
    time before the mercury content in sea water is significantly
    increased. Since water is thought to remain in the surface layers of
    the ocean for 10-50 years, these authors concluded that the mercury
    resulting from man-made activities should be well distributed in the
    water of all the oceans and therefore should not lead to high local

        This conclusion is consistent with the findings reported in
    section 5.1 that mercury levels in swordfish and tuna fish caught at
    the beginning of the century fall within the same range as mercury
    levels reported in recent catches.

        The origin of mercury released by natural processes is not well
    established. Volcanic emissions are a possible source in view of the
    high concentrations of mercury vapour reported in the vicinity of
    volcanoes (for review, see Jonasson & Boyle, 1971). The general
    "degassing" of the earth's surface is probably a major source (Weiss
    et al., 1971). Levels of metallic mercury vapour in the atmosphere
    over soils rich in mercury (the humus layers of topsoil) have been
    reported in the range of 20-200 ng/m3 as compared to background
    levels of 5 ng/m3 according to a report by Barber et al. (quoted by
    Vostal, 1972). Korringa & Hagel (1974) have raised the possibility
    that evaporation from the oceans may make a contribution to the
    mercury present in the atmosphere in view of the substantial
    quantities of water vapour that evaporate (4.48 x 105 km3). However,
    it seems unlikely that mercury would evaporate at the same rate as
    water in view of the fact that it is believed to be in a complex form
    in the oceans (see section 5.1). Furthermore, the observations of
    Williston (1968) (referred to in section 5.1) indicate that the
    mercury content of the atmosphere over the oceans is considerably
    lower than that over land (industrialized and rural areas).

        The mechanisms of volatilization of mercury from the land masses
    are not well understood. Presumably release of mercury from volcanoes
    is due to the high temperatures associated with volcanic activity.
    Vostal (1972) has suggested two major mechanisms, firstly the

    reduction of mercury in soils by a chemical process depending on the
    local redox potential, and secondly reduction by the activity of
    microorganisms. The quantitative importance of these two processes is
    not known. Mercury-volatilizing microorganisms are known to exist and
    have been identified (Magos et al., 1964; Furukawa et al, 1969;
    Tonomura & Kanzaki, 1969).

    4.2  Environmental Transformationthe Local Mercury Cycle

        Mercury is present naturally in the environment and released from
    manmade sources in a variety of chemical and physical states. The
    principal mercury ore is cinnabar, which is mercury sulfide. Andersson
    (1967) has shown that mercury in soils is complexed to the organic
    (humus) content. Metallic mercury may be discharged into the
    environment from natural sources as discussed above and also from
    man-made sources such as chloralkali plants. A variety of
    organomercurial compounds are also discharged into the environment as
    a result of human activities (see section 3). Both the inorganic forms
    of mercury (such as metallic mercury vapour and cinnabar) and the
    organic forms of mercury are subject to conversion in the environment.

        Jensen & Jernelov (1972) have summarized the major pathways of
    transformation. The inorganic forms of mercury (Hg0 and HgS) undergo
    transformations in the environment mainly by oxidation-reduction
    reactions. Mercury vapour is oxidized to ionic divalent mercury
    (Hg++) in water in the presence of oxygen. Concentrations as high as
    40 g/litre have been attained when water saturated with oxygen was
    exposed to mercury vapour (Wallace et al., 1971). As pointed out by
    Jensen & Jernelov (1972) the oxidation of metallic mercury to
    inorganic divalent mercury is greatly favoured when organic substances
    are present in the aquatic environment.

        Ionic mercury, once present in water, is capable of forming a wide
    variety of complexes and chelates with organic materials. Of
    considerable importance is its reaction with the sulfide (S--) ion to
    form highly insoluble mercury(II) sulfide. This reaction is likely to
    occur in anaerobic aquatic environments owing to the presence of
    hydrogen sulfide gas. This sulfide complex of mercury is highly stable
    and will not normally become involved in transformation under
    anaerobic conditions. However, in the presence of oxygen, the
    insoluble mercury(II) sulfide can become oxidized to the soluble
    sulfite and sulfate salts of mercury which allow the metal to ionize
    and enter subsequent chemical reactions.

        In addition to the oxidation of metallic vapour, inorganic mercury
    (Hg++) can be formed by the breakdown of a variety of organic
    mercury compounds. The alkoxyalkylmercury compounds are very unstable
    in acid conditions and it has been reported (see Jensen & Jernelov,
    1972) that, in humid soil (pH = 5), methoxyethylmercury has a
    half-life of only 3 days. Aryl- and alkylmercury compounds can all be
    degraded in the environment by chemical and physical processes and by
    biologically mediated processes.

        Divalent inorganic mercury (Hg++) can undergo two important
    reactions in the environment. The first is the reduction to metallic
    mercury vapour, a reaction that will occur in nature under appropriate
    reducing conditions. As mentioned above, certain bacteria,
    particularly of the genus  Pseudomonas, can convert divalent mercury
    into metallic mercury (Magos et al., 1964; Furukawa et aI., 1969). The
    formation of inorganic divalent mercury in nature and its reduction to
    metallic mercury vapour are probably key processes in the global cycle
    of mercury. The reduction to metallic mercury vapour must be the key
    step in the release of mercury because of degassing of the earth's
    surface. The oxidation of metallic mercury vapour to divalent ionic
    mercury must be the critical step in the uptake of mercury vapour in
    rainwater and in the oceans. Unfortunately, other than these crude
    generalizations, little is known of the details of the kinetics of
    these processes in nature.

        The second important reaction that ionic divalent mercury (Hg++)
    undergoes in nature is its conversion to methylmercury and
    dimethylmercury compounds and the interconversions between these
    compounds. These reactions play a critical role in the so called
    "local cycle" of mercury and are worth further discussion. Some
    countries, particularly those in Scandinavia, that used methylmercury
    fungicides extensively, experienced a general rise in the mercury
    content of their agricultural products. High levels were also noted in
    some species of birds. The increase corresponded with the onset of the
    use of methylmercury fungicides. However, it was discovered that
    mercury levels in fish were also high and that these fish were
    obtained in areas where methylmercury compounds were not used (Jensen
    & Jernelov, 1969). It was subsequently discovered that methylmercury
    was the predominant form of mercury in fish regardless of the nature
    of the mercury pollutant. This was the first evidence that
    transformations of mercury compounds must occur in the environment and
    that, indeed, they must be of great significance. It has now been
    demonstrated that biological methylation of mercury occurs in the
    organic sediments of aquaria and in sediments from freshwater and
    coastal waters of Sweden (Jensen & Jernelov, 1967, 1969; Jernelov,

        Two biochemical pathways of methylation of mercury have been
    identified, one anaerobic the other aerobic. The anaerobic pathway
    involves the methylation of inorganic mercury by methylcobalamine
    compounds produced by methanogenic bacteria in a mildly reducing
    environment (Wood et al., 1968). The process is non-enzymic and is
    strictly anaerobic. The aerobic pathway has been described by Landnet
    (1971) in studies of  Neurospora crassa. His findings indicate that
    methylmercury bound to homocysteine becomes methylated by those
    processes in the cell normally responsible for the formation of
    methionine. In other words, the methylmercury-homocysteine complex is
    methylated by "mistake".

        Despite the fact that an anaerobic pathway for methylmercury
    production is well known, it seems unlikely that significant amounts
    of methylmercury are formed in the aquatic environment under anaerobic
    conditions. The chief reason for this, as pointed out by Jensen &
    Jernelov (1972), is that, in natural water when oxygen is exhausted,
    hydrogen sulfide is formed and divalent mercury becomes bound up as
    mercury(II) sulfide. In this sulfide form, mercury is not available
    for methylation under anaerobic conditions (Jernelov, 1968; Rissanen,
    quoted by Jensen & Jernelov, 1972), and methylation is slow even under
    aerobic conditions (Fagerstrom & Jernelov, 1971).

        In an aquatic environment under aerobic conditions, it must be
    borne in mind that the upper sedimentary layers and sedimentary
    particles suspended in the water may be both aerobic and anaerobic,
    the exterior being well oxygenated and the interior deficient in
    oxygen. Thus both pathways, aerobic and anaerobic, are possible routes
    of methylation in water that is oxygenated.

        The ability to methylate mercury is not confined to a limited
    number of species of microorganism. Thus, conditions that promote
    bacterial growth in general, will lead to enhanced methylation of
    mercury. The highest rates of methylation in the aquatic environment
    are, therefore, seen in the uppermost part of the organic sediments
    and on suspended organic material in water (Jernelov, 1973).

        The formation of dimethylmercury from monomethylmercury compounds
    has been shown to occur in decomposing fish (Jensen & Jernelov, 1968),
    and from (originally) inorganic mercury in sediments. The anaerobic
    pathway using methylcobalamines is one means by which dimethylmercury
    can be synthesized. The reaction is greatly favoured by high pH
    whereas the formation of monomethylmercury is favoured by a low pH

        The ability to methylate mercury at a high rate correlates with
    the resistance of the microorganism to concentrations of inorganic
    mercury (for review, see Jernelov, 1973).

        The observations, reviewed above, of the interconversion of the
    various mercury compounds in nature have led to a hypothesis for a
    local cycle (Jensen & Jernelov, 1972). Inorganic divalent mercury is
    formed either by the oxidation of metallic mercury vapour by
    physico-chemical processes or by the cleavage of the carbon-mercury
    bond in organomercurial compounds either chemically or enzymatically.
    The divalent ionic mercury becomes attached to sediments either
    suspended in the water or in the sedimentary layers. The upper
    sedimentary layers are biologically active but it is postulated that,
    with the passage of time, large quantities of inorganic mercury will
    penetrate down to the inorganic mineral layers of the sediments where
    the mercury should remain inactive. In the surface layers of the
    sediment, part of the inorganic mercury becomes methylated.
    Methylation significantly increases the ability of mercury to cross
    biological membranes. This is why aquatic organisms contain mainly

        If conditions of pH are appropriate, dimethylmercury will be
    formed. Dimethylmercury is water insoluble, possesses a very high
    volatility, and is postulated to diffuse from the aquatic environment
    into the atmosphere. Once in the atmosphere, it is subject to removal
    by rainfall. If the rainwater is acidic, the dimethylmercury is
    converted to monomethylmercury compounds and is thereby returned to
    the aquatic environment completing the cycle. In the presence of
    mercury(II), dimethylmercury is converted to two methylmercury
    molecules (Jensen & Jernelov, 1969).

        Key parts of this local cycle remain conjectural. It is known that
    dimethylmercury compounds can be formed and that the conditions for
    their formation can exist in an aquatic environment. Unfortunately
    analytical data are sparse but Bramen & Johnson (1974) have identified
    both mono-and dimethyl compounds in the atmosphere both outdoors and
    indoors in the USA. Evidence is still lacking for methylmercury
    compounds in rainwater. The analytical difficulties are considerable.
    Nevertheless the present weight of evidence supports the existence of
    a local cycle for the transport of mercury involving dimethylmercury
    as the key intermediary for the atmospheric turnover in this cycle.
    The observations available today on this cycle refer to local bodies
    of water such as lakes and rivers and the cycle itself would represent
    the best available explanation for the presence of methylmercury
    compounds in freshwater fish.

        The origin of methylmercury compounds in oceanic fish has not been
    well described. Inorganic mercury is available in unlimited quantities
    in the oceans, as has been indicated in the calculations reported in
    section 4.1. The site of methylation of this mercury is not known.
    Sediment suspended in oceanic water would seem to be a prime suspect.
    Methylation of mercury is also known to occur in the slime covering

    fish but it does not occur in the fish tissues themselves (Jensen &
    Jernelov, 1972). It would seem an important research priority to
    describe the methylation pathways in ocean waters. Only then will it
    be possible to state whether the rate of formation of methylmercury in
    ocean waters and uptake in oceanic fish is related to the total
    deposit of mercury in the oceans (70 million tonnes) or whether it is
    related to a very small sub-fraction of the mercury in the oceans that
    may respond to man's activities more dramatically than the total ocean

    4.3  Interaction with Physical or Chemical Factors

        The interaction of mercury with physical or chemical factors has
    been referred to frequently in the previous section, so that only a
    brief summary will be given here. In terms of the global distribution
    of mercury, such physicochemical factors as temperature, pH, redox
    potential, and chemical affinities for the organic materials in soil
    will interact to determine the degree of volatility of mercury under
    specific local conditions and the rate of release of mercury from the
    earth's crust as elemental mercury vapour. The interplay between these
    factors is so complex that studies of mercury volatilization from soil
    and from the earth's crust, in general, do not lend themselves easily
    to experimental work. Once in the atmosphere, metallic mercury is
    liable to both physical and chemical interactions. Physically it may
    be adsorbed on to particulate materials in air but evidence reviewed
    in section 5.1 indicates that the aerosol fraction of mercury is 5% or
    less of the total mercury in air. Metallic mercury vapour should
    distribute more or less evenly between air and water providing it
    remains in the unoxidized metallic state (Hughes, 1957). However, the
    reported levels in rainwater (see section 5.1) are higher than the
    background level by a factor of at least 2 or 3. This is no doubt a
    consequence of the oxidation of metallic mercury to ionic mercury in
    the water in the presence of oxygen. Once deposited in the ocean from
    rainwater, any remaining metallic mercury should be liable to
    oxidation to ionic mercury whereupon it will undergo rapid chemical
    combination with various chemical compounds in ocean water. Sillen
    (1963) has estimated that the mercury may be present as negative
    chloride complexes (section 5.1). However, it seems probable that,
    because of its affinity for sulfhydryl groups, mercury will also bind
    strongly to living organisms in ocean waters.

        Another aspect that should be considered is the relationship
    between mercury and selenium. Recent data indicate that selenium
    compounds known to detoxify mercury, increase mercury retention in
    some organisms changing the tissue distribution (Parizek et al.,
    1971). High mercury concentrations were accompanied by high selenium
    concentrations in tissues of several animal species (Ganther et al.,
    1972; Koeman et al., 1972, 1973) and also in man (Kosta et al., 1975;
    Byrne & Kosta, 1974). This relationship is further discussed in
    section 7 of this document.

        In the local cycle of mercury, the same physico-chemical factors
    will be operative. Oxygen tension in the aquatic environment will
    determine the degree of formation of insoluble mercury(II) sulfide
    that will limit the rate of methylation. The pH of the aquatic
    environment and also of the rainwater will determine the distribution
    of the methylated forms of mercury between dimethyl and monomethyl

    4.4  Bioconcentration

        The short-chain alkylmercurials, especially methylmercury
    compounds, have a strong tendency to bioaccumulation since they
    possess a group of properties that makes them unique among the mercury
    compounds. Methylmercury is very efficiently absorbed through
    biological membranes. In mammals, absorption of methylmercury from
    food is virtually complete. Methylmercury is degraded much more slowly
    into inorganic mercury than are the other classes of organomercurial
    compounds. It is excreted from living organisms much more slowly than
    other mercury compounds. It possesses a very high chemical affinity
    for the sulfhydryl group. Since this group occurs mainly in proteins
    in living organisms, methylmercury, once it has entered the organism,
    is soon convened to a non-diffusible protein-bound form. However, even
    though most of the methylmercury is bound to protein, a small fraction
    remains in a diffusible form. Methylmercury rapidly equilibrates
    between diffusible and non-diffusible binding sites and thus retains
    its mobility within animal tissues.

        In view of its ability to accumulate in living organisms, one
    would, in general, expect to see higher concentrations of
    methylmercury at higher trophic levels in natural food chains.
    Qualitatively, this generalization appears to be true but quantitative
    predictions are not possible because of the complex interplay of a
    host of factors that influence the accumulation and movement of
    mercury in food chains. For example, remarkably large species
    differences exist in biological half-times which vary from
    approximately 7 days in the mouse, to 70 days in the monkey and man,
    500 days in seals, and over 1000 days in some species of fish (for
    review, see Clarkson, 1972a).

        The origin of methylmercury in terrestrial food chains is
    predominantly the use of mercury fungicides in the treatment of seed
    grain (D'Itri, 1972). The seeds are consumed by grain-eating birds or
    rodents and the rodents themselves become victims of the large
    carnivorous birds. The dramatic increase in the concentration of
    mercury in feathers of carnivorous birds in Sweden was associated with
    the introduction of methylmercury fungicides in 1940 (for review, see
    Swedish Expert Group, 1971). High concentrations of mercury in
    pheasants and other game birds are also a result of this terrestrial

    food chain and have led to restrictions on hunting in certain areas of
    North America. The replacement of methylmercury by the
    alkoxyalkylmercury compounds in Sweden led to a diminished level in
    this terrestrial food chain. Generally speaking, alkoxyalkyl- and
    phenylmercury compounds are either less well absorbed or more easily
    degraded to inorganic mercury and more rapidly excreted.

        The accumulation of methylmercury compounds in aquatic food chains
    has been the subject of a recent review (Fagerstrom & Larsson,
    unpublished report). This chain or group of chains is considerably
    more complex than the terrestrial ones. Nevertheless, several
    tentative generalizations seem plausible at this time. Once
    methylmercury is formed in the upper sedimentary layers or in
    suspended sediments in water, it readily leaves the sedimentary
    particle (Gavis & Ferguson, 1973). The reason for this is not fully
    established but Fagerstrom & Larsson suggest that it may be due to the
    pathway of synthesis of methylmercury compounds. For example, if
    methylmercury is formed by the pathway proposed by Landner (1971), it
    will be in the form of a diffusible complex with homo-cysteine. In
    contrast, inorganic mercury in the sediment is probably bound to large
    macromolecules. Once methylmercury has diffused from the sedimentary
    particle into the water, it must be rapidly accumulated by living
    organisms. This accumulation is so efficient that methylmercury has
    never been detected in filtered water. Fagerstrom & Larsson, in
    reviewing recent experimental work on methylmercury accumulation,
    noted that this form of mercury accumulates in all species, whether
    plant or animal, that possess membranes for gas exchange with their
    aquatic environment.

        The accumulation of methylmercury in food chains in freshwater
    systems has been proposed as a three-step process by Fagerstrom &
    Larsson. The first step is an accumulation by bottom fauna that are in
    closest proximity to the active sedimentary layers where the
    methylmercury is formed. Accumulation in the bottom fauna, including
    plankton, would be followed by accumulation in species such as the
    roach and finally in the large carnivorous fish such as the northern
    pike. The authors point out that the relative importance of uptake of
    methylmercury directly from water through the gill membranes, as
    opposed to intake from food, should depend upon the trophic level of
    the fish. The higher the trophic level the more important the intake
    from food. However, for the overall food chain, uptake through the
    gills is the key process. If for some reason there is a dramatic
    change in the environmental layers of methylmercury, the authors
    predict that it would take from 10-15 years for the levels in the top
    predators to readjust to the new environment.

        These generalizations on freshwater species should be expected to
    apply to oceanic fish. The remarkably high levels of methylmercury
    seen in swordfish and tuna fish are due to a variety of factors. First
    these species are large carnivorous fish at the end of a food chain.
    They live for a relatively long time compared with other species of
    fish and it is well established that methylmercury levels show a
    positive correlation with age (and or weight) of the fish. They are
    highly active fish having insatiable appetites. Because of their
    activity, large quantities of oceanic water pass through the gill
    membranes each day. Thus it is possible that tuna fish, swordfish and
    related species have a high intake of methylmercury both from their
    food supply and from the surrounding water.

        Accumulation of mercury in the terrestrial and aquatic food chains
    (Fagerstrom & Larsson) results in risks for man mainly through the
    consumption of: game birds in areas where methylmercury fungicides are
    in use; fish from contaminated waters, especially predator species,
    tuna fish, swordfish and other large oceanic fish even if caught
    considerably off shore; other seafoods including muscles and crayfish;
    fish-eating birds and mammals; and eggs of fish-eating birds.

        Space does not permit a full discussion of the important questions
    concerning the chain of mercury transport from soil to plant to
    domestic animals and ultimately to man. Important parameters in this
    transport include absorption and availability in the soil, intake and
    distribution in the plant, toxic effects on the plant, and intake by
    domestic animals and by man. The maximum amounts tolerated in the soil
    may be key factors in determining the possible enrichment in food
    chains and the ultimate hazards to man (Koronowski, 1973; Kloka,


        The levels of mercury in the environment have been reviewed either
    partially or completely by: Swedish Expert Group (1971), Joint FAO/WHO
    Expert Committee on Food Additives (1972), Holden (1972), D'Itri
    (1972), Petersen et al. (1973), Bouquiaux (1974), and CEC (1974). The
    principal findings may be summarized as follows. The concentration of
    mercury vapour in the atmosphere is so low that it does not contribute
    significantly to human intake of mercury. A few "hot spots" may exist
    but these require further investigation. Concentrations of mercury in
    water, particularly drinking water, are also sufficiently low as not
    to contribute significantly to human exposure. The industrial release
    of methylmercury compounds into a sheltered ocean bay (Minamata Bay)
    and into a river (the Agano River) in Japan have led to extremely high
    concentrations of methylmercury in fish (up to 20 000 g/kg wet
    weight) and resulted in human poisonings and fatalities. The
    industrial release of a variety of chemical and physical forms of
    mercury into inland waters has led to local pollution, to mercury
    levels in fish occasionally over 10 000 g/kg but usually less than
    5000 g/kg, and to the restriction of fishing for sport and commercial
    fishing in these areas. The mercury level in most freshwater and
    oceanic fish is below 200 g/kg. However, in large carnivorous fish
    such as tuna, swordfish, halibut, and shark, levels are usually above
    200 g/kg and can be as high as 5000 g/kg wet weight. The general
    population face no significant hazards from the consumption of
    methylmercury in the diet. However, certain sub-populations, either
    those eating locally contaminated fish or those with an unusually high
    consumption of large carnivorous oceanic fish eventually develop blood
    levels of mercury in the range of the lowest levels associated with
    signs and symptoms of poisoning in the Japanese outbreak. It is
    estimated that the average daily intake of the general population is
    less than 20 g of mercury per day in the diet. An appreciable amount
    of this would be methylmemury. However, individuals in certain
    sub-populations having unusually high exposure may ingest daily
    amounts of mercury of up to 200 g, mainly as methylmercury compounds.

    5.1  Levels in Air, Water, and Food


        The average concentration of mercury in the general atmosphere was
    reported by Stock & Cucuel (1934) to be 20 ng/m3. These results were
    confirmed by Eriksson (1967) in Sweden. Sergeev (1967) noted
    concentrations of 10 ng/m3 in the USSR. Fujimura (1964) reported
    concentrations of 0-14 ng/m3 in non-industrialized regions of Japan.
    The lowest reported levels are those reported by McCarthy (1968) in
    Denver, USA, of 2-5 ng/m3. Williston (1968) reported mercury levels
    in the vicinity of San Francisco, USA, of 0.5-50 ng/m3, the level
    depending greatly on the direction of the wind. Williston's method
    would have detected only mercury vapour.

        Levels of particle-bound mercury have also been reported.
    Goldwater (1964) noted that airborne dust in New York City contained
    from 1 to 41 ng/m3 and that outdoors the concentration was from 0 to
    14 ng/m3. Brar et al. (1969) noted that particle-bound mercury in air
    above Chicago ranged from 3 to 39 ng/m3. Heindryckx et al. (1974), in
    the most recent study, found that aerosol mercury levels corresponding
    to remote background levels in Norway and Switzerland were as low as
    0.02 ng/m3. In a heavily industrialized area of Belgium, near Liege,
    the aerosol mercury levels noted were as high as 7.9 ng/m3. Other
    sampling stations in Belgium reported values roughly an order of
    magnitude below this. Unfortunately it is not known to what extent
    particle-bound mercury contributes to total mercury levels in the
    atmosphere. An indirect reference to Jervis by Heindryckx et al.
    (1974) indicates that aerosol mercury accounts for only 5% of total
    mercury in the atmosphere. All the particle-bound mercury reported by
    Heindryckx et al. (1974) had a particle size of less than 0.4 m.

        "Hot spots" of mercury concentration have been reported in
    atmospheres close to industrial emissions or above areas where mercury
    fungicides have been used extensively. Fujimura (1964) reported air
    levels up to 10 000 ng/m3 near rice fields where mercury fungicides
    had been used and values of up to 18 000 ng/m3 near a busy super
    highway in Japan. McCarthy et al. (1970) noted air values of up to 600
    and 1500 ng/m3 near mercury mines and refineries. Fernandez et al.
    (1966) reported maximum values of 800 000 ng/m3 in a village close to
    a large mercury mine in Spain. The remarkably high mercury vapour
    levels reported by these authors indicate the need for further studies
    into localized high concentrations of mercury in the atmosphere.


        Limited data are available for concentrations of mercury in
    rainwater and snow. First reported values were 50-500 ng/litre (Stock
    & Cucuel, 1934). Eriksson (1967) found values from 0 to 200 ng/litre.
    Brune (1969) noted values of approximately 300 ng/litre in rainwater
    in Sweden. Values for mercury in snow have been reported by Johnels et
    al. (1967) as 70 ng/kg and by Byrne & Kosta (quoted by Holden, 1972)
    as 1000-3000 ng/kg in centrifuged melted snow. It is probable that
    mercury levels in snow depend greatly on the collection conditions and
    upon how long the snow has laid on the ground. For example, Strabya
    noted values of 80 ng/kg in fresh snow but 400-500 ng/kg in snow that


    a STRABY, A. (1968)  Analysis of snow and water. In: Westermark, T.
      & Ljunggeren, K., ed.  Development of analytical methods for
       mercury and studies of its dissemination from industrial sources.
      Stockholm, Swedish Technical Research Council, mimeographed

    may have partly melted or evaporated over the winter. Analysis of ice
    deposited in Greenland prior to the 1900s (Weiss et al., 1971)
    indicates values of 60 ng/kg.

        Bodies of freshwater for which there is not independent evidence
    for mercury contamination, contain levels of mercury of less than
    200 ng/litre. Stock & Cucuel (1934) reported 10-50 ng/litre in
    well-water and 100 ng/litre in the River Rhine. Dall'Aglio (1968) in
    measurements of 300 samples from natural water in Italy found values
    in the range of 10-50 ng/litre. Voege (1971) reported levels up to
    40 ng/litre for uncontaminated Canadian waters. Durum et al. (1971)
    have reported data on the concentration of mercury in surface waters
    of the USA. In areas where mercury mineralization was present, values
    of up to 200 ng/litre were seen. The results of the CEC International
    Symposium, reviewed by Bouquiaux (1974), indicate that the purest
    surface water (drinking quality) contains less than 30 ng/litre based
    on over 700 samples collected from drinking reservoirs in the Federal
    Republic of Germany. Rivers believed to have low contamination, such
    as the Danube, and bodies of water such as the Boden See, have values
    close to 150 ng/litre based on the analysis of 152 samples. The rivers
    in the lowland countries of Western Europe that flow into the North
    Sea have mercury values in the range of 400-700 ng/litre no doubt
    reflecting the high industrialization of this area (Schramel et al.,
    1973). Reports by Hasselrota, Fonds (1971), and Smith et al. (1971a),
    indicate that mercury is predominantly particle-bound in contaminated
    water-ways. In the Federal Republic of Germany the mercury
    concentration measured was around 400 ng/litre in inland waters,
    between 100 ng and 1800 ng/litre in rivers, and 600 ng/litre in a
    sample of potable water. (Reichert, 1973; Schramel et al., 1973.)

        Data for mercury concentrations in ocean waters are not as
    extensive as those reported for freshwater. Findings of Stock & Cucuel
    (1934) giving a mean value of 30 ng/litre were confirmed by Sillen
    (1963). Sillen, on the basis of physico-chemical arguments, suggested
    that most of the mercury in seawater would be present as negatively
    charged halide complexes. Hosohara (1961) noted the following levels
    in the Pacific Ocean: at the surface, 80-150 ng/litre; at a depth of
    500 metres, 60-240 ng/litre and at a depth of 3000 metres,
    150-270 ng/litre. Levels reported at the CEC International Symposium
    (reviewed by Bouquiaux, 1974) were 20 ng/litre in 14 samples from the


    a HASSELROT, T. (1971) Mercury in fish, water, and bottomless
      sediments. Investigations at the research laboratories of the
      National Swedish Environment Protection Board (mimeographed

    English Channel but were as high as 150 ng/litre in samples taken from
    the Belgian shoreline and the Waddenzee in the Netherlands. Other
    references such as Burton & Leatherland (1971) and Leatherland et al.
    (1971) also support the general rule that oceanic levels are below
    300 ng/litre. Higher concentrations have been produced as a result of
    local contamination such as in Minamata Bay where Hosohara et al.
    (1961) have reported values up to 600 ng/litre.

        In view of questions, discussed earlier, on the total mercury
    content of the ocean, the stability of mercury levels in the ocean
    over the past 50 years, and on the high mercury levels in species of
    oceanic fish, the paucity of data on oceanic levels of mercury is
    remarkable. This would seem to be one area for future studies of
    environmental levels of mercury. These efforts should include attempts
    to analyse the different physical (particulate, or soluble) and
    chemical (inorganic, or methyl) forms of mercury.

     Food (except fish)

        Smart (1968) has reviewed data concerning mercury concentrations
    in foods and the most recent data from Europe have been summarized by
    Bouquiaux (1974). Mercury levels in milk products (81 samples from the
    Federal Republic of Germany and the United Kingdom) ranged from 0 to
    40 g/kg with a median value of 6 g/kg. Levels in eggs (440 samples,
    taken from Denmark, the Federal Republic of Germany and the United
    Kingdom, ranged from 0 to 100 g/kg with most of the values between 10
    and 20 g/kg. Levels in meat, meat products, and prepared meat
    products (318 samples from the United Kingdom) ranged from 0 to
    50 g/kg with most values lying between 10 and 20 g/kg. Various kinds
    of cereal and flour (2133 samples, taken from the Federal Republic of
    Germany and the United Kingdom) ranged from 0 to 20 g/kg with most
    values being close to 3 g/kg. Mercury levels in cereal products from
    the same countries (52 samples) ranged up to 50 g/kg with most values
    close to 20 g/kg. Vegetables and fruits (288 samples) from Belgium,
    the Federal Republic of Germany, and the United Kingdom had mercury
    levels up to 50 g/kg with most values close to 7 g/kg. The analysis
    of nearly 1400 foods, excluding fish, in Canada during 1970 showed
    mercury residues to be less than 60 g/ kg in bread, flour, grains,
    and eggs and less than 40 g/kg in meats and vegetables (Somers,

        A Swedish Expert Group (1971) has reviewed Swedish experience on
    the effects of widespread use of methylmercury fungicides on food
    levels of mercury. As a result of a ban on the use of methylmercury
    fungicides, food levels fell by a factor of three. For example, the
    mercury levels in Swedish hen eggs (whole) averaged 29 g/kg prior to
    April 1966. Between October 1967 and September 1969, following the ban

    on methylmercury fungicides instituted in 1966, the level in Swedish
    hen eggs fell to 9 g/kg.

        The chemical form of mercury in foodstuffs other than fish has not
    been well identified. The reason is that the levels are, in general,
    so low as to preclude gas chromatographic identification. However,
    West (quoted by a Swedish Expert Group, 1971) has noted that
    methylmercury accounts for over half the total mercury in samples of
    pork chop and liver, filet of beef, and egg white. Inorganic mercury
    can account for more than half the total mercury in pig kidney, pig
    brain, ox liver, and egg yolk.


        The earliest reported mercury levels for freshwater fish are those
    of Stock & Cucuel (1934) and Raeder & Snekvik (1949) and range from 30
    to 180 g/kg wet weight. Upper limits for mercury levels have been
    quoted as, 200 g/kg wet weight (Lofroth, 1969) for Sweden, 150 g/kg
    (Sprague & Carson, 1970) for Canada, and 100 g/kg (Ui, 1967) for
    Japan. These are probably to be regarded as normal levels, i.e. for
    fish in uncontaminated water. The WHO Regional Office for Europe
    (1973) has summarized references indicating that fish caught in
    contaminated freshwater areas may have values of 200-5000 g/kg and,
    where the water is heavily polluted, values may be as high as
    20 000 g/kg.

        The CEC International Symposium (Bouquiaux, 1974) quote levels in
    freshwater fish caught in Western Europe as ranging from 0 to
    1000 g/kg with most values being between 200 and 400 g/kg wet
    weight. Canned fish, excluding tuna taken from several Western
    European countries (597 samples), had values up to 500 g/kg with an
    average close to 50 g/kg wet weight. Canned tuna from the same areas
    (1798 samples) had values ranging up to 4000 g/kg with most values
    falling into the range of 200-500 g/kg. Salmon appears to have
    remarkably low levels of mercury. Measurements of some 260 samples of
    Atlantic Ocean, Canadian, and Baltic Sea salmon had mercury levels
    ranging up to 150 g/kg with most values being close to 50 g/kg. On
    the other hand, pike caught in contaminated rivers in Denmark had
    average mercury values of 5000 g/kg, results which are in agreement
    with experiences summarized by a Swedish Expert Group (1971) in
    contaminated freshwater areas in Sweden and Finland. The concentration
    of mercury in marine fish showed marked variations. Not all the
    factors responsible for these variations are understood but it is
    generally realized that the species of fish, the geographical
    location, and the age and/or weight of the fish are important. The
    highest values of mercury are usually seen in those fish at the end of
    a long food-chain such as the large carnivorous species.

        The concentration of mercury in marine fish has been the subject
    of intense study in recent years. The first measurements reported by
    Stock & Cucuel (1934) and Raeder & Snekvik (1941) are in agreement
    indicating levels from 44 to 150 g/kg wet weight. The most recent
    reports (Peterson et al., 1973; Bouquiaux, 1974) indicate that mercury
    levels in most species of oceanic fish fall in the range of
    0-500 g/kg wet weight with most values close to 150 g/kg wet weight
    (more than 1600 samples). The most important exceptions to this rule
    are swordfish, tuna fish, and halibut, whose values usually range from
    200 to 1500 g/kg (reviewed by the Joint FAO/WHO Expert Committee on
    Food Additives, 1972). Skipjack, white tuna, and yellow fin tuna (911
    samples) ranged from 0 to 1000 g/kg with most values ranging from 200
    to 300 g/kg. These samples were caught in the Atlantic, Pacific, and
    Indian Oceans. Bluefin tuna from the Bay of Biscay (285 samples)
    ranged from 200 to 800 g/kg with most values close to 500 g/kg. The
    same species caught in the Mediterranean Sea (136 samples) ranged from
    500 to 2500 g/kg with most values close to 1100 g/kg. Big-eye tuna
    (20 samples from various origins) had mercury values ranging from 400
    to 1000 g/kg. Over 5200 samples of tuna, variety not specified but
    originating from Italy, had levels in the range of 0-1750 g/kg with
    most values ranging from 300 to 500 g/kg wet weight.

        Swordfish caught in the western Atlantic (210 samples) had mercury
    values ranging from 50 to 4900 g/kg with a mean value of 1150 g/kg.
    40 samples of swordfish, originating near Italy, had values ranging
    from 650 to 1750 g/kg with most values close to 1100 g/kg wet

        The geographical location appears to be important. This is
    illustrated by mercury analysis of cod (Dalgaard-Mikkelsen, 1969).
    Samples recovered from the strait between Denmark and Sweden, which is
    heavily contaminated, had values up to 1290 g/kg; cod caught in the
    area of Greenland had values of 12-36 g/kg whereas North Sea cod had
    values in the range of 150-195 g/kg wet weight. Peterson et al.
    (1973) quote evidence that halibut caught in the southern areas of the
    Northern Pacific had higher mercury levels than those caught in the
    North. Beckett & Freeman (quoted by Peterson et al., 1973) in a study
    of 210 swordfish from six areas extending from the Caribbean Sea to
    the Grand Banks noted significant variations from one area to another
    in average mercury levels.

        Metabolic differences may also affect mercury levels. For example,
    Barber et al. (1972) noted differences in mercury content in different
    species of benthopelagic fish despite the fact that they had identical
    feeding habits and ecological requirements and were exposed to mercury
    in the same area for the same length of time.

        The age (or weight) of the fish appears to be an important
    determinant of mercury levels. A positive correlation between mercury
    concentrations and the weight of the fish has been demonstrated by
    Beckett & Freeman (quoted by Peterson et al., 1973) for swordfish,
    halibut, benthopelagic morid (Barber et al., 1972), spiny dogfish
    (Forrester et al., 1972), blue marlin (Rivers et al., 1972), and tuna
    (quoted by Peterson et al., 1973). In the last study, mercury levels
    were measured in 88 yellowfin tuna whose sizes ranged up to 100 kg.
    Tuna having weights below 25 kg had mercury levels not exceeding
    250 g/kg; tuna having body weights below 50 kg had mercury levels not
    exceeding 500 g/kg. Tuna with body weights above 60 kg had values
    ranging up to 1000 g/kg. However, large variations in mercury content
    were noted in tuna with body weights in the range of 60-100 kg. A
    relationship between mercury content and body weight has previously
    been noted for freshwater fish (Johnels, 1967; Kleinart, 1972; Bache
    et al., 1971).

        Mercury content may also differ with the sex of the fish. For
    example, Forrester et al. (1972), in studies of spiny dogfish on the
    coast of British Columbia, noted that males had a higher mercury
    content than females for a given body weight. These authors suggested
    that this difference may be due to the fact that the males grow more
    slowly than the females.

        Mercury in fish appears to be predominantly in the form of
    methylmercury. Swedish measurements of freshwater fish, summarized by
    a Swedish Expert Group (1971), indicated that virtually all of the
    mercury is present in the form of methylmercury compounds. Smith et
    al. (1971b) confirmed these findings for fish on the North American
    continent and for swordfish and tuna fish. Exceptions to this rule are
    Pacific marlin caught off the coast of Hawaii where methylmercury
    accounts for only a small fraction of the total mercury (Rivers et
    al., 1972) and also lake trout where methylmercury seems to account
    for only 21-35% of total mercury (Bache et al., 1971).

        Interpretation of the results of observations on museum specimens
    of tuna fish and swordfish caught at the turn of the century (Miller
    et al., 1972) indicates that mercury levels in these species of fish
    have not changed significantly throughout the twentieth century.
    Specimens from preserved fish of this age are necessarily limited. In
    seven samples of tuna reported by Miller et al. (1972), the mercury
    concentrations ranged from 180 to 640 g/kg. These compare with
    present values in tuna ranging roughly from 200 g/kg to over
    1000 g/kg wet weight. Given this variation, it is true to say that
    there is no statistically significant difference between samples
    caught in 1900 and those caught in 1970. However, because of the wide
    range of values, the data at present available do not preclude the
    possibility that some change may have taken place and that the change
    might be quite substantial.

    5.2  Occupational Exposures (See also section 8.1.1)

        Occupational exposure to elemental mercury vapour is still the
    principal hazard to human health when mercury is considered. More than
    50 specific occupations or trades involving frequent exposure to
    mercury have been described by Gafafer (1966). Diseases caused by
    mercury or its toxic compounds are classical occupational diseases and
    in most countries are notifiable and qualify for compensation.
    Reporting of occupational poisoning by mercury has been inadequate, as
    is the case with all other occupational diseases, particularly in
    developing countries where there is evidence that large numbers of
    workers are exposed to high concentrations of mercury leading to
    poisoning. Occurrence of occupational mercury poisoning in a wide
    variety of industries in different parts of the world has been
    reported. In accordance with the information available, most people
    exposed to elemental mercury vapour appear to be employed in the
    mining industry, or in chloralkali plants (McGill et al., 1964; Ladd
    et al., 1966; West & Lim, 1968; Smith et al., 1970) and in the
    manufacturing of instruments where mercury finds application. These
    publications, all appearing within the last ten years, indicate that
    mercury levels in air may attain values as high as 5 mg/m3. The
    highest mercury concentrations in air are reported in papers on
    exposure in mining operations. The concentration of mercury in urine
    may attain levels as high as 2175 g/litre.

        In mining for metals other than mercury (e.g. copper), mercury ore
    may be present in the mine and give rise to occupational exposure.
    Donovan (1974) has reported levels of mercury in urine samples (91
    samples, number of workers not stated) ranging from 30 to 700 g/litre
    in a non-mercury related mining operation. In the two years (1972-73),
    seven urine samples were found with mercury levels in excess of
    250 g/litre and some of the miners were admitted to hospital.

        Ladd et al. (1964) have reported on occupational exposure to
    phenylmercury compounds. Air mercury concentrations ranged up to
    0.1 mg/m3 and urinary mercury levels ranged from 1 to 788 g/litre. A
    total of 67 workers were involved in these studies. Phenylmercury
    compounds continue to be used as fungicides in the paint industry (for
    review, see Goldwater, 1973) so that occupational exposure to
    phenylmercury compounds is still significant.

        The Swedish Expert Group (1971) have summarized reports on
    occupational exposures to methyl- and ethylmercury compounds. All
    these reports were published within the period 1940-60 except for the
    reports on laboratory personnel published by Edwards in 1865 and 1866.
    Restrictions on the agricultural application of ethyl- and
    methylmercury compounds by various industrialized countries probably
    accounts for the lack of recent reports on occupational exposure.

    5.3  Estimate of Effective Human Exposure

        The daily intake of elemental mercury vapour by the general
    population may be calculated from the published data on ambient air
    levels discussed above and on the assumption that 80% of the inhaled
    mercury vapour is retained and that the daily ventilation in the
    average person is 20 m3 of air. The ambient air level, except in
    polluted areas, appears to be of the order of 20 ng/m3 and appears
    not to exceed 50 ng/m3 (see section 5.1). Assuming an ambient air
    level of 50 ng/m3, the average daily intake of metallic mercury
    vapour would amount to 1 g/day due to inhalation. The average daily
    intake of those sub-groups of the general population living in
    specially polluted areas is difficult to estimate with any accuracy.
    If we use the figures of McCarthy (1970), it is possible to find
    mercury levels as high as 0.0015 mg/m3 close to points of emission.
    Individuals living continuously in these areas would have intakes of
    30 g/day. Daily intake from occupational exposure is almost
    impossible to estimate because of the wide variation in exposure
    conditions in industry (see section 5.2). Assuming that, generally,
    the time-weighted average threshold limit value of 0.05 mg/m3 (ACGOH,
    1976) is being followed, average occupational exposure would lead to
    an average daily intake of 300 g of mercury or less, assuming a
    ventilation of 10 m3/day at work and 225 working days per year. The
    published reports are insufficient to estimate occupational daily
    intake from other forms of mercury. The proposed guideline of
    0.1 mg/m3 for phenylmercury (MAC Committee, 1969) should lead to an
    intake in workers exposed to phenylmercury compounds of 500 g/day or

        The intake of mercury from drinking water by the general
    population is more difficult to estimate but it is probably very low
    in comparison with intake from diet. The major problem is that the
    chemical form of mercury in water has not always been identified and
    the efficiency of absorption from the gastrointestinal tract depends
    greatly on the form of mercury. Methylmercury compounds are absorbed
    almost completely whereas absorption of inorganic mercury may be 15%
    or less. In making the following calculations the worst case will be
    assumed, namely that all mercury in drinking water is methylmercury.
    it will also be assumed that the daily intake of water in adults is
    2 litres/day (Joint FAO/WHO Expert Committee on Food Additives, 1972).
    Published reports indicate that pure well-water and drinking water
    from reservoirs have mercury levels not exceeding 50 ng/litre (see
    section 5.1). Thus the daily intake of mercury from drinking water
    would not normally exceed 0.1 g/day. However, drinking water in
    certain areas may derive either from natural waters such as those
    reported in Italy that have levels as high as 300 ng/litre because of
    exposure to mineralized mercury deposits, or from rivers in heavily
    industrialized areas reported to have values up to 700 ng/litre (see
    section 5.1). Taking the highest reported figure and assuming that

    mercury is not removed during purification of the water, the highest
    daily intake would be close to 1.4 g/day. The advised upper limit for
    mercury in drinking water is 1 g/litre (World Health Organization,
    1971) which would allow intakes of up to 2 g/day from this source.

        The intake of mercury from food is the most difficult of all to
    estimate because of the different levels of mercury in different
    classes of foodstuffs and different dietary habits of individuals in
    the general population. The one important generalization that emerges
    is that the intake of mercury as methylmercury is related to fish
    intake. Thus normal levels for intake of mercury cannot be stated in
    general without some reference to the fish intake of the population in

        Over the past forty years, various estimates have been made on the
    intake of mercury by the general population assuming that fish intake
    is close to the average values for that population. These reports have
    been reviewed by a Joint FAO/WHO Expert Committee on Food Additives
    (1972) and indicate that the range of daily intake of mercury in the
    general population is from 1 to 20 g/day. The most complete reviews
    of dietary intake published to date are those of a Swedish Expert
    Group (1971) and Jonsson et al. (1972). The reports refer specifically
    to the Swedish population. It was noted that the intake of mercury in
    the diet from sources other than fish in Sweden is about 5 g/day and
    that the methylmercury content is not known precisely. The median
    supply of methylmercury from fish is stated to be 5 g/day or less. As
    fish consumption exceeds the median value for Sweden, the daily intake
    of methylmercury will increase in proportion. It was noted that the
    average daily intake of fish flesh was 30 g, that 10% of the adult men
    might consume between 80 and 100 g and that a few individuals may
    consume as much as 500 g/day.

        Epidemiological studies summarized by a Swedish Expert Group
    (1971) indicate that in fishermen and their families, daily intakes of
    methylmercury can rise to values of 200 g/day and that one individual
    had an unusually high intake of 800 g/day. Another example of a
    Swedish fish eater with very heavy methylmercury exposure has now been
    published (Skerfving, 1974b).

        Dietary intake of mercury in other countries is not as well
    documented as that in Sweden. Recent studies reported in a CEC
    Symposium (Bouquiaux, 1974) indicate that average dietary intake in
    the United Kingdom, based on total diet samples, is less than
    20 g/day. Observations on fish eating groups, such as fishermen based
    in American Samoa, indicate that blood mercury levels of up to
    20 g/100 ml can be obtained through fish intake (Clarkson et al.,
    1975). Such blood levels would be equivalent to a daily intake of
    between 200 and 300 g/day of methylmercury in fish. McDuffie (1973)

    has reported on intakes of mercury in dieters in the United States who
    consume substantial amounts of tuna and swordfish. He estimated that
    in the 40 dieters, who had the highest daily intake of fish, 25%
    consumed 9-16 g/day, that the second quartile consumed 17-26 g/day,
    the third quartile consumed 27-38 g/day and that the highest quartile
    consumed 40-75 g/day. On the basis of radio-chemical measurements,
    Diehl & Schellenz (1974) estimate the total intake of mercury with
    food in the Federal Republic of Germany to be between 57 and 192 g
    per person per week.

        Some industrial countries appear to have an average daily intake
    of less than 20 g/day but sub-groups in these countries with
    unusually high fish intakes (dieters, fishermen's families) may have
    intakes rising to 75 g/day (dieters) and even to 800 g/day (an
    extremely heavy fish-eater in Sweden).

        In countries depending greatly on fish as the major source of
    dietary protein, there is a great need for dietary studies including
    the measurement of mercury in the diet of these populations. Initial
    studies from a South American country indicate that coastal villages
    have populations that are comparable to the Swedish fishermens'
    families in terms of daily intake of methylmercury (Turner et al.,


    6.1  Uptake

    6.1.1  Uptake by Inhalation

        Inhalation is the most important route of uptake for elemental
    mercury vapour. From what is known of the general principles governing
    pulmonary retention of vapours, the high diffusibility and appreciable
    lipid solubility of metallic mercury vapour should ensure a high rate
    of absorption in the alveolar regions of the lung (Task Group on Metal
    Accumulation, 1973). Calculations made by Nordberg & Skerfving (1972)
    indicate that mercury vapour should be distributed between air and
    body tissues in the proportion of 20 to 1 in fayour of tissue
    deposition. Experiments on animals confirm that the major site of
    absorption is alveolar tissue where virtually complete absorption of
    the vapour takes place (Magos, 1967; Berlin et al., 1969; Hayes &
    Rothstein, 1962). If mercury vapour is completely absorbed across the
    alveolar membranes, one would expect that, owing to the physiological
    dead space, 80% of the inhaled vapour would be retained. This has been
    confirmed by observations in man where retention of the inhaled vapour
    was in the range of 75-85%, at mercury concentrations between 50 and
    350 g/m3. (Teisinger & Fierova-Bergerova, 1965; Kudsk, 1965a). The
    retention of mercury vapour in man can be reduced by moderate amounts
    of alcohol (Kudsk, 1965b). Magos et al. (1973) have shown that the
    action of alcohol is due to the inhibition or, oxidation of the vapour
    in the red blood cells and other tissues. More recently Magos et al.
    (1974) have shown that the herbicide, aminotriazole, has a similar
    action to that of alcohol.

        No specific data are available on the monoalkylmercury compounds.
    However, it is generally believed that absorption is high, of the
    order of 80% of the inhaled amount (Task Group on Metal Accumulation,
    1973). Ostlund (1969a, 1969b) reported a high retention of inhaled
    dimethylmercury in mice. The inorganic and organic compounds of
    mercury may also exist in the atmosphere in particulate form (see
    section 5). No detailed studies have been reported on pulmonary
    retention and clearance of mercury aerosols. In general, one would
    expect that aerosols of mercury should follow the general physical
    laws governing deposition in the respiratory system.

        Particulates with a high probability of deposition in the upper
    respiratory tract should be cleared quickly. For particulates
    deposited in the lower respiratory tree, longer retention will be
    expected, the length of which will depend on solubility, among other
    factors (Task Group on Lung Dynamics, 1966). Approximately 45% of a
    mercury(II) oxide aerosol having a mean diameter of 0.16 m was
    cleared in less than 24 hours and the remainder cleared with a
    half-time of 33 days according to experiments on dogs by Morrow et al.

    (1964). Information on pulmonary retention of aerosols of the
    organomercurials is lacking. Pulmonary absorption of monoalkylmercury
    must be significant to judge from the incidents of poisoning resulting
    from occupational exposures to dusts and vapours of the alkylmercury
    fungicides. It should be noted that the gastrointestinal route may
    include those particulates of mercury compounds that have been cleared
    from the lung in the bronchociliary tract.

    6.1.2  Uptake by ingestion

        The general principles underlying the gastrointestinal absorption
    of mercury and its compounds are not clearly understood. Probably the
    formation of soluble salts and complexes is a prerequisite for
    absorption of metals ingested from food.

        Liquid metallic mercury has long been considered to be poorly
    absorbed from the gastrointestinal tract. Based on the data of
    Bornmann et al. (1970), in animals given gram quantities by mouth,
    Friberg & Nordberg (1973) have calculated that less than 0.01% of an
    administered dose of metallic mercury was absorbed. Persons who had
    accidently ingested several grams of metallic mercury showed increased
    blood levels of mercury (Suzuki & Tanaka, 1971).

        The efficiency of absorption from food depends greatly upon the
    type of mercury compound (Clarkson, 1972a). Studies on mice revealed
    that the absorption of inorganic salts of mercury from food was 15% or
    less in contrast with 80% or more in the case of phenyl- or
    methylmercury compounds. Observations on volunteers given tracer doses
    of inorganic mercury revealed that the efficiency of absorption was
    the same with both free and protein-bound mercury. The absorption from
    food in these volunteers was an average of about 7%, (Rahola et al.,

        Aberg et al. (1969) and Miettinen (1973) have reported on the
    absorption of radioactive methylmercury compounds in volunteers given
    oral doses. The absorption of the administered dose was 95%
    irrespective of whether the methylmercury was administered as a salt
    dissolved in water or in a protein-bound form. Information on the
    absorption in humans of other organic compounds of mercury including
    the other short-chain alkylmercurials is not available. As episodes of
    accidental poisoning due to ingestion of food contaminated with
    ethylmercury compounds have occurred, absorption must be significant.

        The Task Group on Metal Accumulation (1973) considered the
    possibility that the gastrointestinal absorption of one metal may be
    influenced by the presence of another. Studies on animals and animal
    tissues (Sahagain et al., 1966, 1967) suggest the possibility that
    some interaction may occur between zinc, manganese, cadmium, and
    inorganic mercury.

    6.1.3  Absorption through skin

        Debate has persisted throughout most of the present century about
    the importance of skin as a route for entry of metallic mercury into
    the body. Early studies on man (Juliusberg, 1901) and animals
    (Schamberg et al., 1918), where inhalation of mercury vapour was
    prevented, indicated that appreciable skin absorption of metallic
    mercury took place. It would appear that metallic mercury can cross
    the skin barrier but to what extent is not known.

        Studies on experimental animals reveal that inorganic salts of
    mercury, principally mercury(II) chloride, may be absorbed in
    significant amounts through skin. For example, Friberg et al. (1961)
    and Skog & Wahlberg (1964) indicate that 5% of mercury in a 2% water
    solution of mercury(II) chloride was absorbed through intact skin of
    guinea-pig over a 5-hour period. Such a penetration rate, if
    applicable to man, could result in absorption of substantial amounts
    of mercury under conditions of high exposure.

        Friberg et al. (1961) and Wahlberg (1965) have demonstrated in
    guinea-pigs that methylmercury dicyandiamide was absorbed from a water
    solution through intact skin, the rate was more or less the same as
    that for mercury(II) chloride reported above. No information is
    available on animals with respect to ethyl- or other alkylmercury

        No quantitative data are available for skin absorption of the
    short-chain alkylmercurials in man. People have been poisoned by
    administration of methylmercury compounds locally to the skin such as
    methyl-mercury thioacetamide (Tsuda et al., 1963; Ukita et al, 1963;
    Okinaka et al., 1964; Suzuki & Yoshino, 1969; Suzuki et al., 1970).
    The methylmercury compound was absorbed in sufficient amounts to cause
    severe poisoning although the possibility of some inhalation exposure
    cannot be excluded.

    6.2   Distribution in the Organism

        Details on the organ distribution of mercury have been recently
    reviewed (Clarkson, 1972a; Nordberg & Skerfving, 1972). New
    publications since that time have not substantially changed the
    general picture. Methylmercury and its homologous short-chain
    alkylmercurials, which are much more uniformly distributed throughout
    the body than are the other organomercurials, and inhaled elemental
    mercury vapour are distinguished from other types of mercury compound
    in their ability to cross the blood-brain barrier and placenta

        Organ distribution is not only affected by the type of mercury
    compound ingested or inhaled but also changes with time after
    exposure. For example, the phenylmercurials are subject to rapid
    conversion in the body to inorganic mercury so that the distribution

    of mercury following administration of these compounds and related
    organomercurials approaches that of inorganic mercury with increasing
    time after exposure (for details, see Clarkson, 1972b).

        The distribution between cells and plasma (the red cell/plasma
    ratio) depends upon the form of mercury to which the subject is
    exposed. Studies on fish-eating populations reported by Birke et al.
    (1972) and on a heavily exposed population in Iraq (Bakir et al. 1973)
    indicate that the cell to plasma ratio for methylmercury is
    approximately 10, as was found in human volunteers given tracer doses
    of radioactive methylmercury (Aberg et al., 1969; Miettinen, 1973).
    The red cell to plasma ratio in human volunteers given radioactive
    inorganic mercury salts was 0.4 (Miettinen, 1973).

        The distribution of mercury between hair and blood tends to follow
    a constant ratio in people exposed to methylmercury (Table 1). In
    various populations having a broad range of dietary methylmercury
    intake from fish, the concentration of total mercury in hair is
    proportional to the concentration in whole blood. The ratio of hair to
    blood concentration is about 250 as determined by linear regression
    analysis. The data in Table 1 are from populations of individuals,
    most of whom probably have a steady concentration of methylmercury in
    hair and blood. In the Iraq epidemic, hair and blood concentrations
    underwent rapid changes. Two cases have been reported in Iraq in which
    blood and hair concentrations were measured when both were declining
    following cessation of heavy exposure. (Amin-Zaki et al., in press.)
    The ratios of hair to blood concentrations were constant and the value
    of the ratio was close to 250. However it should be noted that, when
    hair and blood concentrations are changing, it is important to choose
    the segment of hair for analysis that corresponds to the blood sample.
    Depending on the length of hair segment used for analysis and the rate
    of growth of hair, there is a delay of about 2-4 weeks between the
    time of sampling the blood, and the emergence of the appropriate
    segment of hair above the scalp (Amin-Zaki et al., in press).

        In people occupationally exposed to metallic mercury vapour, the
    red cell to plasma ratio may be as high as 2 (Lundgren et al., 1967;
    Suzuki et al., 1970; Einarsson et al., 1974). Work on experimental
    animals has shown that the ratio was higher in animals given
    radioactive vapour compared with those given salts of inorganic

        Table 1.  Relationship between concentrations of mercury in samples of blood and hair
              in people having long-term exposure to methylmercury from fish

    No. of      Whole blood     Hair (y)
    subjects    (x) (mg/kg)     (mg/kg)     Linear regression    References

    12          0.004-0.65      1-180       y = 280x - 1.3       Birke et al. (1972)
    51          0.004-0.11      1-30        y = 230x + 0.6       Swedish Expert Group (1971)
    50          0.005-0.27      1-56        y = 140x + 1.5       Swedish Expert Group (1971)
    45          0.002-0.80      20-325      y = 260x + 0         Tsubaki (1971)
    60          0.044-5.5       1-142       y = 230x - 3.6       Skerfving (1974b)
        Studies on a variety of experimental animals indicate that the
    kidney is the chief depository of mercury after the administration of
    inorganic salts and exposure to elemental mercury vapour. Over 50% of
    the body burden of mercury can be found in the kidneys of rats exposed
    to mercuric salts and metallic mercury vapour a few days after
    receiving the dose. This percentage may rise to 90% or more as the
    length of time after exposure increases (Rothstein & Hayes, 1960;
    Hayes & Rothstein, 1962; Trojanovska, 1966). However, it should be
    noted that in experimental animals, the brain levels of mercury
    following exposure to elemental mercury vapour were ten times higher
    than brain levels after equal doses of inorganic salts (Berlin et al.,
    1966; Magos, 1967; Nordberg & Serenius, 1969). A more uniform
    distribution of methylmercury throughout the body also results in much
    higher brain levels for a given body burden of mercury as compared
    with inorganic salts.

        Little information is available on the distribution of mercury in
    human organs following exposure to elemental mercury vapour. Takahata
    et al. (1970) and Watanabe (1971) have reported mercury levels in the
    brain several times higher than those in the liver and other organs
    (except the kidney) of miners with long-term exposure to high
    concentrations of mercury vapour. These concentration ratios were
    maintained even several years after cessation of exposure. High
    mercury concentrations in the thyroid and pituitary glands in persons
    connected with mercury mining have been reported (Kosta et al., 1975).
    It should be noted that organ distribution of mercury after inhalation
    of elemental mercury vapour can be dramatically affected by moderate
    intakes of alcohol and small doses of the herbicide aminotriazole, as
    shown in animals (Magos et al., 1973, 1974). These agents reduce
    levels in the lung and increase levels in the liver several-fold.

        The percentage of the body burden of methylmercury found in the
    brain is much higher in primates than in other animal species (Swedish
    Expert Group, 1971). Observations on human volunteers given tracer
    doses of radioactive methylmercury (Aberg et al., 1969) indicate that
    10% of the radioactivity in the whole body is located in the posterior
    part of the head. Probably not all of this represents methylmercury in
    the brain but would include methylmercury attached to the hair.
    Studies by Miettinen's group (quoted by a Swedish Expert Group, 1971),
    on volunteers given tracer doses of radioactive methylmercury,
    indicate that an initial rapid distribution throughout the body is
    followed by a further slow redistribution of methylmercury to the

    6.3  Elimination in Urine and Faeces

        Urine and faeces are the principal routes of elimination of
    mercury from the body. The contribution of each pathway to total
    elimination depends upon the type of mercury compound and the time
    that elapses after exposure. Experiments in animals indicate that
    elimination of inorganic mercury by the gastrointestinal tract depends
    on the size of the dose and the time after exposure. The faecal route
    is dominant soon after exposure. The urinary route is favoured when
    high doses are given (Prickett et al., 1950; Friberg, 1956; Rothstein
    & Hayes, 1960; Ulfvarson, 1962; Cember, 1962; Phillips & Cember, 1969;
    Nordberg & Skerfving, 1972).

        Data obtained on rats subjected to a single exposure of labelled
    203Hg vapour indicated that about 4 times more mercury was eliminated
    in the faeces than in the urine (Hayes & Rothstein, 1962). In
    prolonged exposure of rats, the proportion changed in favour of
    urinary excretion (Gage, 1961). In workers exposed to mercury vapour,
    the output of mercury in urine slightly exceeded that in the faeces
    (Tejning & Ohman, 1966), High individual variation and great
    fluctuation from day to day were the principal features of urinary
    excretion in workers under similar exposure conditions (Goldwater et
    al., 1963; Jacobs et al., 1964). There is evidence that, on a group
    basis, urinary excretion is roughly proportional to exposure (air
    concentration) to elemental vapour (MAC Committee, 1969). Occupational
    exposure of at least 6 months, 5 days per week at average air
    concentrations of mercury of 0.05 mg/m3, should lead to average
    urinary concentrations of mercury of about 150 g/litre.

        Piotrowski et al. (1975) have reported changes in urinary rates of
    excretion in workmen following exposure to elemental mercury vapour.
    They noted that urinary excretion could be described by a two-term
    exponential equation with rate constants equivalent to half-times of 2

    and 70 days. The short half-time compartment accounted for about
    20-30% of the excretion rate under conditions of steady-state
    excretion. Piotrowski et al. (1975) suggested that there is variation
    in urinary mercury excretion in individuals and that this can be
    greatly reduced by collecting the urine samples at the same time in
    the morning.

        Mercury exhalation found in animals after exposure to the
    elemental vapour (Clarkson & Rothstein, 1964) has also been confirmed
    in man (Hursh et al., 1975). This pathway of excretion accounted for
    about 7% of the total excretion of mercury in volunteers following
    inhalation of a tracer dose. Recent observations indicate that the
    concentration of mercury in sweat may be sufficiently high to be taken
    into account in the overall mercury balance in workers exposed to
    elemental mercury vapour (Lovejoy et al., 1974).

        The faecal route is most important in the elimination of mercury
    after acute or chronic dosing with methylmercury. Studies on human
    volunteers (Aberg et al., 1969; Miettinen, 1973) indicate that
    approximately 90% of the elimination takes place via the faeces. This
    proportion does not change with time after exposure. Concentrations of
    total mercury in urine showed no correlation with blood mercury in
    people heavily exposed to methylmercury (Bakir et al., 1973).

    6.4 Transplacental Transfer and Secretion in Milk

        The transplacental movement of mercury in women exposed to
    elemental mercury vapour has not been studied thoroughly.

        Experiments on animals reveal that after brief (approximately 20
    minutes) exposure to radioactive elemental mercury vapour, the
    radioactive mercury easily penetrates the placental barrier (Clarkson
    et al., 1972). These authors report that, after equal exposure of
    pregnant rats, the fetal uptake was 10-40 times higher after exposure
    to elemental mercury vapour than to inorganic salts. In contrast, the
    placental content of mercury after exposure to elemental mercury
    vapour was only about 40% of that after exposure to inorganic salts of

        The alkylmercuric compounds have been known for some time to
    penetrate the placenta readily as indicated from studies on
    experimental animals (for review, see Swedish Expert Group, 1971). In
    a recent study, Childs (1973) noted that the level of methylmercury in
    the fetus may be twice that in the maternal tissues when low levels of
    methylmercury are fed to rats in a tuna fish matrix. At higher dose
    levels, the ratio between fetal and maternal tissues becomes close to
    unity. Transplacental movement of methylmercury in women has been
    sufficient to cause several cases of prenatal poisonings in various

    countries (Engleson & Herner, 1952; Harada, 1968; Bakulina, 1968;
    Snyder, 1971; Bakir et al., 1973; Amin-Zaki et al., 1974a). Tejning
    (1970) has reported methylmercury levels in fetal blood cells to be
    30% higher than in maternal cells in studies on women having normal
    pregnancies and a low to moderate fish intake. The relatively higher
    concentrations in fetal blood have been confirmed in a study by Suzuki
    et al. (1971). It was noted that the plasma levels in both types of
    blood were similar and that differences arose only in terms of
    concentrations in the red blood cells.

        Information on the transplacental movement of other compounds of
    mercury in women is lacking. Animal experiments indicate that those
    compounds rapidly converted to inorganic mercury in the body, such as
    phenylmercury compounds, behave in this respect like inorganic mercury
    (for review, see Clarkson, 1972b).

        Mercury has been reported in breast milk in women exposed to
    methylmercury from fish (Harada, 1968; Skerfving, 1974a) and from
    bread contaminated with methylmercury fungicides in the 1971-72
    outbreak in Iraq (Bakir et al., 1973). In Iraq, it was noted that
    levels of total mercury in milk correlated closely with levels in
    whole blood and averaged 5% of simultaneous concentrations in.
    maternal blood. The total mercury in milk consisted of two fractions
    identified as inorganic mercury (40%) and methylmercury (60%).
    Skerfving's (1973) observations on 15 lactating mothers exposed to
    methylmercury in fish are in general agreement with the findings in
    Iraq except that methylmercury accounted for only 20% of the total
    mercury in milk.

        Despite the relatively low concentration in milk as compared with
    maternal blood, the suckling infants accumulated high concentrations
    of mercury in their blood if their mothers were heavily exposed
    (Amin-Zaki et al., 1974b). Some Iraqi infants, exposed only through
    maternal milk, had blood levels in excess of 100 g/100 ml. In
    prenatally exposed infants, intake of methylmercury by suckling is one
    factor responsible for the slower decline in blood levels as compared
    with the mother (Amin-Zaki et al., 1974a).

    6.5  Metabolic Transformation and Rate of Elimination

        The most dramatic example of metabolic transformation is the
    conversion of metallic mercury to divalent ionic mercury in the body.
    This oxidation reaction has been shown to take place  in vitro in the
    red cells (Clarkson et al., 1961). More recent studies indicate that
    it probably takes place in most other tissues (for details, see Kudsk,
    1973). The process is enzyme mediated and the catalase complex is the
    most likely site of biochemical oxidation (Kudsk, 1973; Magos et al.,

        Studies on the biotransformation of elemental mercury make it
    possible to develop a picture of the role of the oxidation process in
    the accumulation of mercury vapour in the body and its transport to
    the site of action (for details, see Clarkson, 1972a). Elemental
    mercury vapour, after inhalation, is absorbed into the blood stream.
    Despite the rapid oxidation that has been shown to take place in the
    red blood cells, some elemental mercury remains dissolved in the blood
    long enough for it to be carried to the blood-brain barrier and to the
    placenta. Its lipid solubility and high diffusibility allow rapid
    transit across these barriers. Tissue oxidation of the mercury vapour
    in brain and fetal tissues converts it to the ionic form which is much
    less likely to cross the blood-brain and placental barriers. Thus
    oxidation in these tissues serves as a trap to hold the mercury and
    leads to accumulation in brain and fetal tissues.

        Most studies on the metabolic transformation of organomercury
    compounds have concentrated on measurements of the rate of cleavage of
    the carbon-mercury bond. There is no evidence in the literature
    supporting the possibility of the synthesis of organomercury compounds
    in human or mammalian tissues.

        The absolute rates of cleavage of the carbon-mercury bond in man
    or experimental animals is not known. The relative rates of cleavage
    of different mercury compounds have been estimated by measurements of
    the amounts of inorganic mercury deposited in tissues following single
    doses of organomercury compounds. In general, these studies reveal
    that the phenyl-(aryl) and the methoxyethyl compounds are converted
    rapidly to inorganic mercury in the body (for reviews, see Gage, 1974;
    Clarkson, 1972a). The short-chain alkylmercurials are converted more
    slowly to inorganic mercury with the methylmercury compounds being
    converted the most slowly of all. The phenyl- and
    methoxyethylmercurials are probably converted to inorganic mercury
    more or less completely within a few days whereas methylmercury can be
    detected in human tissue months after exposure has stopped (Bakir et
    al., 1973). Suzuki et al. (1973) have reported the only case in which
    the metabolic conversion of ethylmercury has been studied in man.
    Proportional values of inorganic mercury to total mercury ranging from
    12 to 69% were detected in red cells, plasma, brain, spleen, liver,
    and kidney in a patient exposed for about 3 months to

        The role of biotransformation in determining the toxicity of
    organomercurials is not well understood (for discussion, see Clarkson,
    1972b). The rapid conversion of phenylmercury to inorganic mercury
    probably accounts for the fact that, in chronic studies on animals,
    the effects of this organomercury compound on kidneys were similar to
    those of inorganic mercury (Fitzhugh et al., 1950).

        The conversion of organic to inorganic mercury may increase or
    decrease the total rate of excretion of mercury from the body. If the
    intact molecule of an organomercurial is more rapidly excreted than
    inorganic mercury, biotransformation will decrease the overall
    excretion rate. This has been demonstrated in the case of the
    diuretic, chlormerodrin, where the intact molecule is almost
    completely excreted within 24 hours, but inorganic mercury remains in
    the animal for much longer periods (Clarkson et al., 1965). The
    phenyl- and methoxyethylmercury compounds are excreted at a rate
    similar to that of inorganic mercury according to studies on
    experimental animals. In the case of methylmercury, biotransformation
    may play an important part in determining the rate of excretion of
    total mercury from the body (Swensson & Ulvarson, 1968, 1969).
    Inorganic mercury accounts for approximately 50% of the total mercury
    in faeces, the principal pathway of excretion following single or
    chronic doses of methylmercury compounds. Methylmercury undergoes
    extensive enterohepatic recirculation in rats but inorganic mercury
    does not (Norseth & Clarkson, 1971). Thus a small rate of metabolic
    transformation in the liver leading to biliary excretion of inorganic
    mercury could make an important contribution to the faecal elimination
    of mercury.

    6.6  Accumulation of Mercury and Biological Half-time
         ("Metabolic Model")

        The body accumulates a metal when uptake exceeds elimination. At a
    certain stage a steady state may be reached when uptake and
    elimination are equal. A common way to express the accumulation is in
    terms of biological half-time. The biological half-time for mercury
    would be the time taken for the amount of mercury in the body to fall
    by one-half. The concept of biological half-time is meaningful,
    however, only if the elimination can be approximated to a single
    exponential first-order function. This will be true if the
    distribution and turnover of a metal in different tissues of the body
    are faster than the elimination from the body as a whole. If
    elimination from one organ is slow compared with that from other
    organs then the calculation of a biological half-time for the whole
    body may be completely misleading from the toxicological point of view
    (Task Group on Metal Accumulation, 1973; Nordberg, 1976).

        Studies on experimental animals and volunteers indicate that, for
    methylmercury compounds, the elimination can be approximated to a
    single exponential first-order function (Miettinen, 1973; Aberg et
    al., 1969; for reviews, see Clarkson, 1972b; Swedish Expert Group,
    1971; Task Group on Metal Accumulation, 1973). Observations on
    experimental animals indicate that the elimination of mercury after

    exposure to mercury vapour, inorganic mercury salts, and the phenyl
    and methoxyethyl compounds does not follow such a pattern and thus the
    accumulation and elimination of mercury ("the metabolic model") is
    much more complex. The pattern of elimination of these mercury
    compounds, when administered to animals, is dose- and time-dependent
    (Rothstein & Hayes, 1960; Ulfvarson, 1962; Piotrowski et al., 1969).

        In cases where the elimination of a metal such as methylmercury
    follows a single exponential first order function, the concentration
    in an organ at any time can be expressed by the following equation;

                    C =  Co. e-b.t                                  (1)

               C  = concentration in the organ at time  t
               Co = concentration in the organ at time  o
               b  = elimination constant
               t  = time

    The relation between the elimination constant and the biological
    half-time is the following:

                    T = 1n 2/ b

               T = biological half-time
              1n 2 (natural logarithm of 2) = 0.693

        If data on exposure and absorption of the metal are known, then it
    will be possible to predict the body burden of the metal at constant
    exposure over different time periods. If a constant fraction of the
    intake is taken up by a certain organ, the accumulated amount in that
    organ can also be calculated. The following expression gives the
    accumulated amount of metal in the total body (or organ):

                    A = ( a/b)(1 - exp ( -b.t))                     (2)

               A = accumulated amount
               a = amount taken up by the body (or organ) daily

    At steady state the following applies:

                    A =  a/b                                        (3)

        In other words, the steady state amount in the body (or organ)
     A is proportional to the average daily intake and inversely
    proportional to the elimination rate. The latter point will be taken
    up later (section 9) in discussing hazards to man, as large individual
    variations in elimination rates imply large individual variation in
    steady state body burden, even in people having the same average daily

        Equations (1), (2), and (3) are illustrated graphically in Fig. 1.
    During the period of steady daily intake (assumed to be 10 g/70 kg
    body weight), the amount in the body rises rapidly at first, reaching
    half its maximum (steady state) value in a time equivalent to one
    elimination half-time (assumed to be 69 days for methylmercury in
    man). After an exposure period equivalent to 5 elimination half-times
    (approximately one year for methylmercury), the body burden is within
    3% of its final steady state value. The steady state value is one
    hundred times the average daily intake assuming an elimination
    half-time of 69 days. On cessation of exposure, the body burden will
    immediately begin to fall following an exponential curve that is an
    inverse image of the accumulation curve. Thus the body burden will
    have returned to within 3% of pre-exposure values in 5 half-times.

        In this example, it is assumed that the hair to blood ratio is
    constant and equal to 250 and that 1% of the body burden is found in
    1 litre of blood in a 70-kg man.

        That this model provides a reasonable approximation to the
    accumulation of methylmercury in man over a wide range of daily
    intakes is indicated by the data in Tables 2 and 3. Data on
    elimination rates for the whole body reported by Aberg et al. (1969)
    on 5, and by Miettinen (1973) on 15 volunteers were in good agreement
    indicating average values close to the value of 69 days used in Fig.
    1. An average value of 50  7 days for clearance half-time from blood
    was reported by Miettinen (1973) in volunteers receiving a single
    tracer dose. Blood clearance values are difficult to measure
    accurately with tracer doses owing to the low counting rates in the
    blood samples. Skerfving (1974b) reported clearance from whole blood
    ranging from 58 to 87 days in 4 people having high intake (up to
    5 g/kg body weight) of methylmercury from fish, one individual had a
    clearance half-time of 164 days. Bakir et al. (1973) reported that
    patients having very high blood levels (over 100 g/100 ml) in Iraq,
    had clearance half-times in the same range (45-105 days, mean 65
    days). "Clearance" from hair is estimated by analysis of consecutive
    short (0.2-1 cm) segments of hair samples and plotting the mercury
    concentrations against the distance from the scalp on semilogarithmic
    paper (for details, see Birke et al., 1972). A straight line is
    usually obtained, the slope of which is equivalent to a biological
    half-time if the growth rate of the hair is known. "Clearance"
    half-times from hair are assumed to reflect clearance half-times for
    blood. Data from a fish eating population (daily intake up to 5 g/kg)
    and on a highly exposed population in Iraq (daily intake up to
    50 g/kg) are compatible with this assumption given the wide range of
    individual variations (Table 2).

    FIGURE 1

        The relationship between steady state body burden  (A) and
    average daily intake is given by equation (3); using data derived from
    tracer observations on volunteers (Aberg et al., 1969; Miettinen,
    1973), one would predict that the steady state blood level ( y ng/ml)
    is numerically equal to the average daily intake ( x g/day/70 kg
    body weight) as indicated in Table 3. This calculation assumes a
    69-day elimination half-time from the whole body, that 1% of the body
    burden is found in 1 litre of blood in a 70-kg "standard man".
    Observed steady state relationships between blood concentration
     (y) and daily intake  (x) are given in Table 3 for several
    populations. These populations consist of fishermen and their families
    who had had a high dietary intake of fish for many years. The range of
    intake between different individuals is high -- up to 800 g/day. The
    relationship between blood concentrations and average daily intake was
    found to be linear for each population studied. Linear regression
    analysis reveals that the observed relationship between  y and  x is
    less than that predicted by tracer studies. The coefficient of
     x lies between 0.5 and 0.8 in the fish eating populations as
    compared with the predicted value of unity from tracer studies.

        Table 2.  Mercury intake and clearance

                                Clearance half-times (days)
    No. of      Hg intake
    subjects    (g/kg/day)    Body        Blood         Hair        References

    5           tracer         70          --            --          Aberg et al. (1969)

    15          tracer         76          50            --          Miettinen (1973)

    5           up to 5        --          --            (33-120)    Birke et al. (1972)

    5           up to 5        --          seea          --          Skerfving (1974)

    16          up to 50       --          65            --          Bakir et al. (1973)

    48          up to 50       --          --            72b         Shahristani & Shihab
                                                         (35-189)    (1974)

    a One person had a biological half-time of 164 days. The other four were in the
      range of 58-87 days.

    b The data were distributed bimodally. One group accounting for 89% of the samples
      had a mean value of 65 days and the other group had a mean value of 119 days.

    Table 3.  Relationship of steady state blood concentrations to daily intake of
                              Ave. Hg intake    Steady blood       
    No. of      Time of       (g/day/          concentration
    subjects    exposure      70 kg B.W.)       (ng/ml)         References

                              (x)               (y)
    6 + 26b     years         0-800             y = 0.7x + 1    Birke et al. (1967)

    139 + 26b   years         0-400             y = 0.3x + 5    Tejning (1967, 1969a,
                                                                1969b, 1969c)

    6 + 14b     years         0-800             y = 0.8x + 1    Birke et al. (1972)

    725c        years         0-800             y = 0.5x + 4    Estimated from Kojima &
                                                                Araki (unpublished data)

    22          years         0-800             y = 0.5x + 10   Skerfving (1974b)

    30d         1-2 months    0-2340            r = 0.8x        Estimated from Shahristani
                                                                & Shihab (1974) and
                                                                Shahristani et al. (1976)

    15          single                          y = 1.0x        Estimated from Miettinen
                tracer dose                                     (1973)

    a For details of these calculations, see text.

    b None or low fish consumers.

    c Estimated from data on hair concentrations and daily intake. The hair to blood
      concentration was assumed to be 250 and the average body weight of the population
      under study to be 60 kg.

    d Estimated from data on hair concentrations and daily intake. The hair to blood
      concentration ratio was assumed to be 250.
        Given the difficulties in the accurate measurement of dietary
    intake and the uncertainty in tracer studies based on counting blood
    samples, it is likely that differences between the observed and the
    tracer values are not real. This conclusion is supported by the fact
    that the Iraqi populations (Table 3, Shahristani et al., 1974), having
    an extremely high dietary intake yielded a factor of 0.8 suggesting
    that the relationship between  y and  x is not substantially changed
    at high doses.

        In summary, a considerable body of evidence exists to support the
    linearity of the metabolic model for methylmercury in man. No
    definitive evidence is yet available that refutes this conclusion.
    However, we cannot exclude the possibility that the mean values of the
    parameters of the metabolic model could change by about a factor of
    two over a wide dose range. We have, however, taken the predicted
    value from the tracer data, since this approach would offer a greater
    margin of safety in estimates of hazards to human health.

        Biological half-times for other mercury compounds are not well
    established and this is particularly true for those organs that are of
    toxicological importance. It seems, however, that the biological
    half-time for the greatest part of retained salts of inorganic mercury
    has an average value of about 40 days (Miettinen, 1973). In five
    female volunteers the average half-time was 37 days and in a similar
    number of males the average half-time was 48 days.

        The biological half-time of both methyl and inorganic salts of
    mercury does not appear to be affected whether the compound is
    administered in an ionic form or bound to protein (Miettinen, 1973).

        Limited information is available on biological half-times of
    mercury in the body following exposure to elemental vapour. Five
    volunteers inhaled radioactive mercury vapour for 10-15 minutes and
    were subjected to whole body counting for up to 43 days after exposure
    (Hursh et al., 1975). Elimination from the body followed a single
    exponential process having a biological half-time of 58 days with a
    range of individual values of 35-90 days.

        As noted above, whole body elimination half-times may not be a
    reliable guide to accumulation in specific organs. For example, the
    fact that Takahata et al. (1970) and Watanabe (1971) found high
    mercury concentrations in the brain in an individual 10 years after
    cessation of exposure indicates that accumulation in the brain does
    not follow the same kinetics as seen by whole-body counting.
    Observations on experimental animals also indicate that the half-time
    in brain is longer than in other organs (Task Group on Metal
    Accumulation, 1973).

    6.7  Individual Variations -- Strain and Species Comparisons

        As a general rule, the processes of absorption by inhalation or
    ingestion do not appear to be subject to large species and strain
    differences (for detailed review, see Clarkson, 1972b). Elemental
    mercury vapour and methylmercury compounds are well absorbed across
    the pulmonary epithelium and the gastrointestinal tract, respectively,
    in a variety of animal species. Distribution in the body tissues is
    subject to species differences. A very pronounced example is the case
    of the red cell to plasma ratios of methylmercury where the ratio can
    be as high as 300 in the rat and as low as 10 in primates. Differences
    in red cell to plasma ratio may account for species differences in
    blood to brain ratios (Vostal, 1972). The blood to brain ratio has
    been reported to be approximately 10-20 for the rat, approximately
    unity for the cat, 0.5 for the dog and pig, and 0.1 for primates (from
    data reviewed by a Swedish Expert Group, 1971). Careful and accurate
    quantitative comparison of species differences and distribution is not
    possible because of the different experimental conditions in these

        The oxidation of elemental mercury vapour to ionic mercury and the
    cleavage of the carbon-mercury bond in a variety of mercurials has
    been described for several different species of animal. However the
    observations in this field are not adequate to allow quantitative
    comparison of metabolic rates of breakdown in different species.

        The rate of elimination of mercury from the body is subject to
    wide species variation (for detailed review, see Clarkson, 1972a). In
    general, animals of small body weight tend to excrete mercury more
    rapidly than larger animals and the cold blooded species, particularly
    fish, appear to retain mercury for an extremely long time.

        Species differences in the elimination of methylmercury have been
    reported. The mouse and the rat have half-times between 8 and 16 days
    as compared with 70 days in primates. The seal is reported to have a
    half-time of 500 days and fish and crustaceans appear to have
    half-times ranging from 400 to over 1000 days. These species
    differences indicate that we cannot extrapolate parameters describing
    the metabolic fate of mercury in animals to that in man. Furthermore,
    since these parameters determine the amount of mercury accumulated in
    the body, it would appear that quantitative information on toxicities
    cannot be directly extrapolated from animals to man.

        The half-time of clearance of mercury from blood and hair varies
    considerably between individuals exposed to methylmercury (Birke et
    al., 1972; Skerfving, 1974). In the outbreak in Iraq the half-time of
    clearance from blood ranged from 40-105 days in 16 subjects (Bakir et
    al., 1973) and from hair the range was from 35-189 days (Shahristani &
    Shihab, 1974). The Iraqi data on blood samples (Bakir et al., 1973)
    were obtained about 1-2 months after the termination of exposure.

        Average biological half-times for groups of at least 15
    individuals seem to be remarkably constant. The average half-time in
    blood was 65 days (Bakir et al., 1973), in hair 72 days (Shahristani &
    Shihab, 1974), and in the whole body in 15 subjects given a tracer
    dose, 76 days (Miettinen, 1973).


    7.1  Experimental Animal Studies

    7.1.1  Acute studies

        Little information is available on the acute toxicity of elemental
    mercury vapour to animals. Ashe et al. (1953) reported evidence of
    damage to brain, kidney, heart, and lungs in rabbits exposed to
    mercury vapour at a mercury concentration of 29 mg/m3 of air. This
    concentration under the circumstances of their experiments would
    represent an atmosphere saturated with mercury vapour. The first
    effects were seen within 1 hour of exposure and subsequent severe
    changes resulted after longer exposures.

        Information on the LD50 (the dose of mercury that kills half the
    test population) has been reviewed by a Swedish Expert Group (1971).
    The results reported for different mercury compounds are not easily
    comparable since different animal species and different routes of
    administration have been used for the test. Nevertheless, despite all
    these differences, the LD50 lies between approximately 10 and 40
    mg/kg body weight for all compounds tested to date including inorganic
    mercury, arylmercury, alkoxyalkyl- and alkylmercury compounds. The
    remarkable similarity in LD50s of these various types of mercury
    compound is probably due to the fact that when given in acute massive
    doses, mercury in whatever chemical form will denature proteins,
    inactivate enzymes, and cause severe disruption of any tissue with
    which it comes into contact in sufficient concentration. The symptoms
    of acute toxicity usually consist of shock, cardiovascular collapse,
    acute renal failure and severe gastrointestinal damage. A variety of
    complexing and chelating agents, all of which contain sulfhydryl
    groups can modify the LD50s of mercury and its compounds (for review,
    see Clarkson, 1972a). These agents are most effective when given
    either prior to the mercury dose or in the few hours following a
    single dose of mercury. The importance of time of administration is to
    be expected since the effects of these agents are to reduce the
    reactivity of mercury in the body and to do so before irreversible
    damage has been inflicted on the tissue (see also, section 9.2).

        Irreversible damage is used in this document to define cellular or
    organ damage that is not repaired even after the cessation of
    exposure; some improvement in function or in the condition of the
    poisoned tissue may occur but recovery is never complete. In the case
    of reversible damage, regeneration of cells and restoration of
    function takes place after the cessation of exposure.

    7.1.2  Subacute and chronic studies  Reversible damage

        This section deals with toxic effects known to be reversible, at
    least up to a certain dose and/or duration of exposure. It should be
    noted, however, that at higher doses or longer duration of exposure,
    the damage can surpass the stage of reversibility.

        Studies by Trahtenberg (1969) (reviewed by Friberg & Nordberg,
    1973) indicate that exposure of rats to concentrations of elemental
    mercury vapour in the range of 0.1-0.3 mg/m3 for over 100 days
    increased uptake of radioactive iodine by the thyroid. Friberg &
    Nordberg (1973) also refer to unpublished observations by Aveckaja
    which, under different conditions including prolonged exposure,
    indicate the opposite effect. Kournossov reported in 1962 (reviewed by
    Friberg & Nordberg, 1973) effects on the behaviour of rats at mercury
    concentrations in air as low as 0.005 mg/m3. Studies by Armstrong et
    al. (1963) on pigeons showed irreversible behavioural changes only at
    vapour levels well in excess of recommended maximum allowable
    concentrations. It is clear that many more studies need to be carried
    out on behavioural and other subtle changes resulting from exposure to
    mercury vapour at these low concentrations.

        Fitzhugh et al. (1950) studied the toxicity of mercury(II)
    chloride and phenylmercury acetate when added to the diet of rats for
    periods of up to 2 years. Morphological changes were induced in kidney
    tissue at approximately the same mercury levels for both compounds.
    The similarity in the effects of the two compounds on the kidney is
    probably due to the fact that phenylmercury compounds are rapidly
    converted to inorganic mercury in animal tissues.

        The main effect of alkylmercurials is the irreversible action on
    the central nervous system. However, Lucier et al. (1971, 1972, 1973)
    have reported that subacute doses (with no neurological signs) caused
    a marked decrease in rat liver mixed-function oxidase activity. This
    effect was shown to be due to an increased degradation rate of
    cytochrome P-450  in vivo. Methylmercury also depressed the activity
    of enzymes dependent upon cytochrome P-450. Ultrastructural changes
    involving the endoplasmic reticulum in liver have been reported by
    Chang & Yamaguchi (1974) and these effects were reversible.

        In rats, administration of methylmercury can produce kidney damage
    manifested by tubular degeneration in the distal convoluted tubules
    after daily doses of 10 mg/kg for 7 days (Klein et al., 1972). With
    lower doses of methylmercury, morphological and functional damage is
    produced in kidney tissue in the absence of any signs of neurological

    dysfunction (Fowler, 1972a, 1972b; Magos & Butler, 1972; Klein et al.,
    1973). Ultrastructural studies by Fowler (1972a) showed that female
    rats given 2 mg/kg methylmercury in their diet were more sensitive to
    methylmercury than males. The primary lesion was characterized by
    extrusion of cyto-plasmic masses from proximal tubular cells. It has
    been suggested that the nephrotoxic effect is due to inorganic mercury
    split from methyl-mercury  in vivo (Klein et al, 1973).  Irreversible damage

        With the exception of massive doses of inorganic compounds or
    prolonged exposure to extremely high concentrations of elemental
    mercury vapour, as in the experiments of Ashe et al. referred to
    below, the effects of inorganic mercury on tissues are generally

        Ashe et al. (1953) reported microscopically detectable changes in
    the organs of dogs, rabbits, and rats exposed to concentrations of
    elemental mercury vapour ranging from 0.1 to 30 mg/m3 for different
    periods of time. Severe damage was noted in kidneys and brains at
    mercury levels in air of 0.9 mg/m3 after an exposure period of about
    12 weeks. After longer periods of exposure to 0.1 mg/m3, no
    microscopically detectable effects could be seen.

        The short-chain alkylmercurials are primarily neurotoxic in man.
    After a single dose there is a latency period of days or weeks before
    signs of poisoning occur. Many of the signs of methylmercury poisoning
    observed in man can be reproduced in animals under appropriate
    conditions. For example, Berlin et al. (1973) noted that a sudden
    visual disturbance occurred in monkeys given subacute doses of
    methylmercury. Prolonged exposure to methylmercury resulted in a
    gradual constriction of the visual field and impaired motor
    coordination and possibly sensory disturbances. Neurological signs of
    damage have also been produced in the mouse, rat, ferret, cat, and dog
    by feeding them methylmercury compounds (Chang et al., 1974; for
    review, see Swedish Expert Group, 1971). The toxicity of methylmercury
    to animals does not appear to be affected whether it is given to them
    as a pure chemical, e.g. methylmercury chloride, or whether it has
    accumulated naturally as in fish such as the Northern Pike (Swedish
    Expert Group, 1971; Albanus et al., 1972).

        Grant (1973), in his studies of primates experimentally poisoned
    with methylmercury, reported findings confirming those of Hunter &
    Russell (1954). Neuronal damage and destruction was observed in the
    visual cortex, and the granular layer of the cerebellum. The dose-rate
    of methylmercury, the period of dosing, and the animal species all
    influence the pattern of pathological damage. For example, in monkeys

    after short periods of exposure to high doses, there is an abrupt
    visual change over two days leading to blindness. This is accompanied
    by damage to the neurons in the visual cortex. Longer exposure to
    lower daily doses of methylmercury leads to more generalized damage to
    the cortex and is accompanied by gradual onset of visual changes and
    other signs of central nervous involvement such as ataxia.

        Recent studies on rats reviewed by Somjen et al. (1973a) have
    confirmed the findings of Hunter et al. (1940) that the earliest
    neurological effects in these animals is damage to the peripheral
    sensory nerves. Later the disease affects other parts of the central
    and peripheral nervous systems. There is now evidence that the primary
    site of the disease is the cell bodies in the dorsal root ganglia with
    secondary deterioration in their fibres (Chang & Hartman, 1972a,
    1972b). Consistent with this interpretation, Somjen et al. (1973b)
    found that the spinal dorsal root ganglia contained the highest
    concentrations of mercury. Electrophysiological investigations
    confirmed the findings drawn from morphological evidence that the cell
    bodies in the spinal ganglia are the primary sites of action (Somjen
    et al., 1973a).

        Morphological, electrophysiological and biochemical changes have
    been demonstrated in animals prior to the onset of overt signs of
    poisoning. These phenomena, especially with respect to morphological
    changes, have been referred to as "silent damage". For example,
    Nordberg et al. (1971) and Grant (1973) noted that morphological
    damage was present in certain of the test monkeys before signs of
    visual impairment could be detected. Somjen et al. (1973a) reported
    electrophysiological manifestations of methylmercury intoxication in
    rats preceding clinical signs. Yoshino et al. (1966) noted a decreased
    uptake of amino acids in brain slices taken from animals given high
    doses of methylmercury at a time before signs of poisoning had
    appeared. Cavanagh & Chen (1971) reported that incorporation of amino
    acids into protein was impeded in spinal root ganglia of rats, treated
    with methylmercury before signs of poisoning were present. Chang &
    Hartman (1972c) noted damage to the blood-brain barrier as early as 12
    hours after a dose of 1 mg/kg of methylmercury chloride to rats. If
    these observations on experimental animals may be extrapolated to man,
    the possibility must be considered that significant damage to the
    central and peripheral nervous systems may take place prior to the
    onset of clinical signs and symptoms.

        Methylmercury compounds have been demonstrated to disturb mitosis
    in the plant cell, in human leucocytes treated  in vivo, and in human
    cells in tissue culture. The short-chain alkylmercurials cause
    chromosome breakage in plant cells and point mutations in  Drosophila
    (for detailed reviews, see Swedish Expert Group, 1971; Ramel, 1972).

        Clegg (1971) has given a detailed review of the embryotoxicity of
    the short-chain alkylmercurials. In general, animal experiments
    confirm the idea derived from epidemiological observations in the
    Minamata epidemic that much more damage was inflicted on the fetus
    than on the mother. Spyker et al. (1972) have recently reported on
    performance deficits in mice treated prenatally with methylmercury.
    The alkylmercury compounds may also damage the gametes prior to
    fertilization (Khera, 1973). Virtually no information is available on
    the morphological and biochemical factors related to prenatal damage
    in experimental animals. The results, however, do point to incipient
    hazards to human fetuses exposed before birth.

        In discussing the biological effects of methylmercury compounds,
    species differences should be considered. Although man, monkeys, and
    pigs may become blind at high exposures, similar visual disturbances
    in cats were not detected (Albanus et al., 1972; Charbonneau et al.,
    1974). Pronounced morphological changes are seen in the peripheral
    nervous system of rats (Somjen et al., 1973a), but again, were not
    detected in cats (Albanus et al., 1972; Charbonneau, 1974). However,
    one cannot exclude the possibility that qualitative differences,
    reported in studies of different species including man, may reflect
    differences in degree and intensity of exposure in man and in
    experimental conditions in animal studies.  Interactions with physical and chemical factors

        Parizek & Ostadalova (1967) reported that selenite salts could
    protect experimental animals against the toxic effects of inorganic
    mercury. Selenium also depressed the passage of inorganic mercury into
    fetuses and its secretion into milk (for review, see Parizek et al.,
    1969, 1971, 1974). Parallelism between tissue concentrations of
    mercury and selenium has been reported in human subjects exposed to
    elemental mercury vapour (Kosta et al., 1975). Several recent
    publications have claimed that selenite added to the diet protects
    experimental animals against methylmercury compounds (Ganther et al.,
    1972; El Bergerami et al., 1973; Potter & Metrone, 1973). Ganther &
    Sunde (1974) have reviewed evidence indicatting that the content of
    selenium in tuna fish is sufficiently high to provide substantial
    protection against methylmercury. However, the protective factor in
    tuna fish, whether selenium or some other substance, has yet to be

        Toxicological interactions have been reported between
    methylmercury and the chlorinated hydrocarbon pesticide, dieldrin.
    Rats dosed with both dieldrin and methylmercury, showed less
    morphological damage of the pars recta tubule than animals given only
    methylmercury. However, there was degeneration in proximal tubular
    cells (Fowler, 1972b). It has been demonstrated recently that
    phenobarbital administration increases the biliary excretion of
    methylmercury compounds (Magos & Clarkson, 1973).

        Estrogenic hormones (Lehotzky, 1972) and spironolactone (Selye,
    1970) protect the kidney from methoxyethylmercury salts and
    mercury(II) chloride, respectively. The mechanism of these actions is

        The question of the interaction of the physical and chemical
    factors on the toxicity of methylmercury is important and should be a
    major priority in future research studies. Extrapolation of
    epidemiological and toxicological data from populations in Japan and
    in Iraq suffering from methylmercury poisoning is fraught with
    difficulties when the possible interactions of local factors are not
    taken into account.

    7.1.3  Biochemical and physiological mechanisms of toxicity

        A physiological basis for the action of mercury and other heavy
    metals had already been propounded prior to 1967 (for reviews, see
    Hughes, 1957; Passow et al., 1961; Peters, 1963; Webb, 1966). Two
    general concepts on the mechanisms of action of mercury and other
    heavy metals are discussed in these reviews. The first dates back to
    the 1940s and is attributed to Peters, 1963. The toxic sequelae of
    heavy metal action on tissues result from a primary "biochemical
    lesion" whereby a critical enzyme or metabolic process is inhibited.
    Unfortunately, despite a considerable amount of research work (see
    Webb, 1966), it has not been possible to locate the biochemical lesion
    associated with the toxic actions of mercury.

        An alternative general concept mainly proposed by Passow et al.
    (1961) is that the cell membrane is the first site of attack by heavy
    metals. Topographically, this would seem reasonable. Furthermore, the
    membrane is known to contain sulfhydryl groups that are essential to
    the normal permeability and transport properties of the cell membrane.
    These same sulfhydryl groups are known to have a very high affinity
    for mercury and its compounds. Passow et al. (1961) summarized a great
    many experimental studies that support this general idea. However, it
    must be admitted that most experimental work testing this idea is
    based on  in vitro studies on isolated cells and tissues so that the
    role of membrane damage in the pathogenesis of heavy metal poisoning
    remains to be established. The effect of mercury in intercellular
    membranes is also of some interest.

        The affinity of mercury for thiol groups in proteins and other
    biological molecules is far in excess of its affinity for other
    biologically occurring ligands (Clarkson, 1972b). As pointed out by
    Rothstein (1973), the affinity of mercury(II) cations for the
    sulfhydryl groups of proteins creates a "severe logistics problem" for
    those interested in elucidating the mechanisms of action of
    mercurials. "Although mercurials are highly specific for sulfhydryl

    groups, they are highly unspecific in terms of proteins. Almost all
    proteins contain sulfhydryl groups that are metal-reactive.
    Furthermore, because most sulfhydryl groups are important in most
    protein functions, mercurials can disturb almost all functions in
    which proteins are involved. Thus almost every protein in the body is
    a potential target". In other words, the mercurials are potent but
    non-specific enzyme poisons. Mercury will inflict cellular damage
    wherever it accumulates in sufficient concentrations. This reasoning
    has given rise to the idea that the selective toxicity of mercury is
    related to its selective distribution. In general, there seems to be
    some truth in this. Inorganic mercury compounds are avidly accumulated
    by the kidney which is the target organ for this compound.

        Studies by Somjen et al. (1973b) on the microdistribution of
    mercury in the nervous system also lend credence to the importance of
    distribution of methylmercury in that the spinal root ganglia, the
    site of peripheral nerve damage, are also the area of highest
    accumulation of methylmercury in rats.

        However, it seems that distribution factors alone cannot give a
    complete explanation for the toxicity of methylmercury. The kidney is
    always the site of the highest accumulation of mercury irrespective of
    the form of the mercury compound involved. For example, kidney levels
    of methylmercury are much higher than brain levels and yet kidney
    damage, except in the rat, is much less than that seen in the central
    nervous system.

        In recent years, interest has arisen in the biocomplexes of
    mercury in the body. The toxicity of any mercury compound will be
    determined by its chemical activity close to its site of action. For
    example the chloride salts of mercury compounds when added  in vitro
    to tissue preparations are highly toxic, whereas when mercury is added
    in the presence of sulfhydryl compounds the toxicity is very much less
    (for review of this subject, see Clarkson & Vostal, 1973). Thus the
    chemical state of combination of mercury in plasma and other body
    fluids may be of primary importance in determining the particular site
    of action of the mercurial.

        Mercury accumulated in the kidney is contained there partly in
    form of a metallothioneine-like complex (Jakubowski et al., 1970;
    Wisniewska, et al., 1970). In the rat, binding by this protein is
    especially effective in repeated exposure to mercury(II) chloride,
    owing to the induction of higher levels of the metallothioneine-like
    protein by mercury (Piotrowski et al., 1974a, 1974b). This form of
    binding probably also occurs in the case of exposure to elemental
    mercury vapour since this exposure results in enhancement of the
    metallothioneine level in the kidneys of rats (Sapota et al., 1974).
    Binding of inorganic mercury in other organs may also involve storage
    of a similar form, as found in the liver (Wisniewska et al., 1972) and
    brain of the rat (Sapota et al., 1974).

        The binding of mercury by metallothioneine-like protein of the
    kidneys is enhanced by the presence of cadmium (Shaikh et al., 1973)
    and therefore may play an important role in man whose kidneys
    accumulate, in normal conditions, considerable amounts of cadmium
    (Piscator & Lind, 1972). The above applies also to organic mercurials,
    which are rapidly converted into inorganic mercury, as in the case of
    phenylmercury acetate (Piotrowski & Bolanowska, 1970).

        The primary biochemical lesions associated with mercury poisoning
    have not yet been established. Virtually nothing is known of the
    biochemical disturbances associated with exposure to metallic mercury
    vapour. Studies referred to in section by Cavanagh & Chen
    (1971) and Yoshino et al. (1966) suggest that protein synthesis may
    undergo an early biochemical change preceding clinical signs and
    symptoms of methylmercury poisoning. This may give an explanation for
    the latent period associated with this form of mercury poisoning.


    8.1  Epidemiological Studies

    8.1.1  Occupational exposure to mercury vapour, alkylmercury
           vapour and other compounds

        Occupational exposures to elemental mercury vapour have been the
    subject of recent reviews by Friberg & Nordberg (1972, 1973) and NIOSH
    (1973). Many studies dating back to the 1930s have related the
    frequency of signs and symptoms of mercury poisoning to exposure.
    These studies, involving observations of more than one thousand
    individuals, indicate that the classical signs and symptoms of
    elemental mercury vapour poisoning (objective tremors, mental
    disturbances, and gingivitis) may be expected to appear after chronic
    exposure of workers to air concentrations of mercury above 0.1 mg/m3
    (Neal et al., 1937; Smith & Moskowitz, 1948; Smith et al., 1949;
    Friberg, 1951; Bidstrup et al., 1951; Vouk et al., 1950; Kesic &
    Heusler, 1951; Baldi et al., 1953; Seifert & Neudert, 1954; McGill et
    al., 1964; Ladd et al., 1966; Copplestone & McArthur, 1967; Smith et
    al., 1970). The industries involved included the chloralkali industry,
    the manufacture of thermometers and graduated scientific glassware,
    the repair of DC electrical meters, the mining and milling of mercury,
    the manufacture of artificial jewellery, the felt hat industry and
    others (NIOSH, 1973). Most of the publications referred to above do
    not report time-weighted average exposures and few give information as
    to the physical and chemical forms of mercury in the atmosphere.
    Different methods of measurement of mercury in air were employed some
    of which measured only mercury vapour, while others attempted to
    include particulate forms of mercury. Most of the studies, if not all,
    assumed that exposure occurred only during the working day. However,
    evidence has now come to light that, in certain industries, metallic
    mercury may be entrapped in the clothing and contaminate the home,
    particularly in those industries actually handling liquid metallic
    mercury (West & Lim, 1968; Danzinger & Possick, 1973).

        Effects of elemental mercury vapour, other than those designated
    as classical mercurialism, have been reported (Smith et al., 1970;
    Trahtenberg, reviewed by Friberg & Nordberg, 1973). The study of Smith
    and co-workers involved observations on 567 workers exposed to mercury
    in chloralkali plants. The air concentrations of mercury (measured by
    a mercury vapour meter) ranged from less than 0.01 to 0.27 mg/m3 and
    time-weighted averages were calculated for each worker. A significant
    increase in the frequency of objective tremors was noted at mercury
    levels in air above 0.1 mg/m3 in agreement with previous reports on
    occupational exposure. However, a significant increase was observed at
    mercury concentrations in air of 0.06-0.1 mg/m3 in such non-specific
    signs and symptoms as loss of appetite, weight loss, and shyness.

        Studies related to assessment of the occurrence of a so-called
    "asthenic-vegetative syndrome" or "micromercurialism" have been
    reported by Trahtenberg (1969). This syndrome may occur in persons
    with or without mercury exposure. For a diagnosis of mercury-induced
    asthenic vegetative syndrome Trahtenberg (1969) (reviewed by Friberg &
    Nordberg, 1972) required that not only neurasthenic symptoms should be
    present but as supporting evidence three or more of the following
    clinical findings; tremor, enlargement of the thyroid, increased
    uptake of radio-iodine in the thyroid, labile pulse, tachycardia,
    dermographism, gingivitis, haemotological changes, and excretion of
    mercury in the urine which was above normal or increased 8-fold after
    medication with unithiol.

        Trahtenberg, in her monograph (1969) considered that this syndrome
    would be more frequently found in persons exposed to mercury
    concentrations between 0.004-12 mg/m3 but, upon detailed scrutiny of
    her data (see review by Friberg & Nordberg, 1972), there does not seem
    to be any difference between exposed and control groups that can be
    related to mercury exposure.

        Studies on the prevalence of a similar syndrome defined as a
    combination of insomnia, sweating, and emotional lability have been
    reported by Trahtenberg, et al., quoted by Trahtenberg (1969), in
    workers exposed to mercury levels of 0.006-0.01 mg/m3 and
    temperatures of 40-42C in the summer and 28-38C in the winter. They
    found 28-50% prevalence of this syndrome in this exposed group and
    only 13% of the same syndrome in a control group exposed to only
    38-42C temperatures. Details about the selection and evaluation of
    these workers are not known.

        The studies of Bidstrup et al. (1951) and Turrian et al. (1956)
    also indicate that psychological disturbances may be seen at air
    concentrations of mercury below 0.10 mg/m3. Thus it is impossible, at
    this time, to establish a lower exposure limit at which no effects
    occur. There is a continuous need for research studies on the effects
    of exposure of people to mercury vapour concentrations below
    0.1 mg/m3.

        Short-chain alkyl compounds have been the subject of recent
    reviews by a Swedish Expert Group (1971) and Kurland (1973), but no
    new information has appeared in the literature on occupational
    exposures to methylmercury or other short-chain alkylmercury
    compounds. Following the description of the first two cases of
    occupational exposure to diethylmercury compounds in 1865 by Edwards
    (1866), occupational exposures have been infrequent and usually

    limited to a few individuals. For example, exposure has occurred in
    laboratory personnel, workers, and farmers involved in either the
    production of alkylmercury fungicides or their application to cereal
    seeds, in people in seed testing institutes, and in workers in pulp
    mills and saw mills. Exposure is presumed to be mainly by inhalation
    of the vapour or dust but it is possible that, in some cases,
    absorption of the liquid preparation of the fungicide may have
    occurred through the skin.

        Occupational exposures to alkylmercury compounds, although not
    important numerically or epidemiologically, have been the occasion for
    the description of the signs and symptoms of poisoning. For example,
    four workmen exposed to methylmercury fungicide were the basis of the
    now classic reports of Hunter et al. (1940) and Hunter & Russell
    (1954) which gave the detailed pathology of methylmercury poisoning in

        Occupational exposures to the phenyl- and methoxyethylmercury
    compounds have been the subject of a recent review (Goldwater, 1973).
    The hazards from industrial exposure to these compounds appear to be
    very low. Ladd et al. (1964), in a study of 67 workers occupationally
    exposed to phenylmercury compounds, found no evidence of adverse
    health effects. Mercury levels in air in this study were mainly below
    0.1 mg/m3. Comparison of readings with a mercury vapour detector and
    estimates of total mercury in the atmosphere revealed that elemental
    mercury vapour was the principal form of mercury. Goldwater (1964,
    1973) makes reference to seven workers who had spent approximately six
    weeks preparing and packaging a batch of material containing
    methoxyethylmercury chloride. Four weeks after they had completed this
    task their whole-blood mercury levels were in the range of
    34-109 g/100 ml with an average of 65 g/100 ml. At no time did any
    toxic sign or symptom appear. These workers may also have had a
    limited exposure to phenylmercury compounds.

        At this point it is worth while to quote from the Task Group on
    Metal Accumulation (1973) indicating the need for carefully controlled
    studies of occupationally exposed groups. "There is a need in
    industrially exposed populations for standardized, wherever possible
    collaborative, epidemiological studies, where cohorts can be followed
    in time and where groups can be related to each other. With some
    occupational exposures to the less common metals, only small groups
    may be available for study in any one country, so that international
    collaboration in epidemiological studies would again be of value".

    8.1.2  General population

        Epidemics of poisoning in the general population due to exposure
    to phenyl- and methoxyethylmercury compounds have not been reported.
    Two outbreaks of poisoning due to elemental mercury vapour occurred in
    the 19th century, one due to a fire in the mercury mines in Idria, the
    other being caused by spillage of metallic mercury in a British
    warship in the early 1800s (Bidstrup, 1964). Fernandez et al. (1966)
    have reported that, in the village of Almaden, the site of the large
    mercury mines in Spain, air levels exceeded 0.1 mg/m3. However, there
    are no reports as yet about the health status of the population in the

        Methyl- and ethylmercury compounds have been the cause of several
    major epidemics of poisoning in the general population due either to
    the consumption of contaminated fish or to eating bread prepared from
    cereals treated with alkylmercury fungicide. The two major epidemics
    of methylmercury poisoning in Japan in Minamata Bay (Katsuna, 1968)
    and in Niigata (Niigata Report, 1967) were caused by the industrial
    release of methyl- and other mercury compounds into Minamata Bay and
    into the Agano River followed by accumulation of the mercury by edible
    fish. The median level of total mercury in fish caught in Minamata Bay
    at the time of the epidemic has been estimated as 11 mg/kg fresh
    weight and in the Agano river in Niigata as less than 10 mg/kg fresh
    weight (Swedish Expert Group, 1971).

        A recent report by Tsubaki (1971) indicates that follow-up
    observations on exposed people in Niigata revealed a much larger
    number having mild signs and symptoms than the original 46 that had
    been reported. These milder cases may only have had paraesthesia. By
    1971 a total of 269 cases of methylmercury poisoning had been reported
    in Minamata and Niigata, of which 55 proved fatal. By 1974, more than
    700 cases of methylmercury poisoning had been identified in Minamata
    and more than 500 cases had been identified in Niigata (personal
    communication, Tsubaki, 1975).a The two Japanese epidemics have been
    the subject of intensive studies on the effects of methylmercury on
    man and have resulted in important conclusions concerning
    dose-response relationships (Swedish Expert Group, 1971).


    a It has not been possible for this group to review data on the new
      cases reported since the publication of the Niigata and Minamata

        Epidemics resulting in the largest number of cases of poisoning
    and of fatalities have been caused by the ingestion of contaminated
    bread prepared from wheat and other cereals treated with alkyl-
    (methyl- or ethyl-) mercury fungicides. The largest recorded epidemic
    took place in the winter of 1971-72 in Iraq resulting in the admission
    of over 6000 patients to hospital and over 500 deaths in hospital
    (Bakir et al., 1973). Previous epidemics have occurred in Iraq (Jalili
    & Abbasi, 1961), in Pakistan (Haq, 1963), in Guatemala (Ordonez et
    al., 1966), and on a limited scale in other countries (Snyder, 1971).
    Reports on these epidemics have resulted in interesting clinical
    findings but quantitative studies relating exposure to effects have
    been reported only on the recent epidemic in Iraq (Bakir et al., 1973;
    Kazantzis et al., 1976; Mufti et al., 1976; Shahristani et al., 1976).

        In the Iraqi outbreak, the mean methylmercury content of the wheat
    was 7.9 mg/kg with most samples falling between 3.7 and 14.9 mg/kg.
    The mean methylmercury content of wheat flour samples was 9.1 mg/kg
    with a range of 4.8-14.6 mg/kg in 19 samples (Bakir et al., 1973). The
    average weight of the home-made loaves was about 200 g with a moisture
    content of about 30% of the fresh weight (Damluji, 1962). The range of
    daily intake of bread varied widely. In an epidemiological survey of a
    heavily affected village, Mufti et al. (1976) reported that the
    average total ingested dose of a group of 426 people was about 150 mg
    of mercury but some people may have consumed as much as 600 mg. The
    average daily intake of contaminated loaves was 3.2 loaves although
    individual variation was large, some people eating up to 10 loaves per
    day. The daily intake of methylmercury would vary greatly. The average
    daily intake of mercury in this village would be 80 g/kg assuming a
    body weight 50 kg for the population, with extremes of daily intake
    attaining 250 g/kg. In the most severely affected group, reported by
    Bakir et al. (1974), the highest daily intake of mercury was about
    130 g/kg body weight. The average period of consumption for groups of
    patients reported by Bakir et al. (1973) ranged from 43-68 days. Mufti
    et al. (1976) reported mean consumption periods in villages to be
    about 32 days but some people continued for up to 3 months. Birke et
    al. (1972) and Skerfving (1974b) have reported on families in Sweden
    consuming fish containing mercury levels of 0.3-7 mg/kg. Daily intake
    ranged up to approximately 5 g/kg body weight. In two cases, intake
    was as high as 10-20 g/kg. The highest recorded blood level of
    mercury was 1.2 g/g of red cells or approximately 60 g/100 ml of
    whole blood. A total of 188 people were referred to in these studies.
    No signs or symptoms of poisoning attributable to methylmercury were

    FIGURE 2

        Clarkson et al. (1975) and Marsh et al. (1974) have offered
    preliminary information on 163 fishermen based in American Samoa who
    ingested unusually high amounts of fish containing methylmercury. Data
    on daily intake of mercury were not reported but the mercury levels in
    blood in this population ranged as high as 28 g/100 ml. No signs or
    symptoms of poisoning could be ascribed to methylmercury.

        Turner et al. (1974) have reported on neurological examinations of
    186 persons living in two fishing villages in northern Peru.
    Concentrations of total mercury, methyl- and inorganic mercury were
    measured in blood samples from 141 of these villagers. The
    concentration range of methylmercury in blood was from 1.1 to
    27.5 g/100 ml with a mean of 8.9 g/ 100 ml. The mean intake of fish
    was approximately 10 kg per family per week and the average family
    size was six.

        Fifty-one persons from a "control" village were also examined. The
    mean intake of fish was 1.0 kg per family per week and the mean family
    size was 6.4. Methylmercury concentrations in blood averaged
    0.99 g/100 ml with a range of 0.33-2.5 g/100 ml. No correlation was
    observed between blood levels of methylmercury and the frequency of
    signs and symptoms usually associated with methylmercury poisoning
    (paraesthesia, ataxia of gait and limbs, impaired vision, and

        Paccagnella et al. (1974) have reported blood concentrations of
    mercury in a community in the Mediterranean island of S. Peitro
    (Cagliari). The average dietary weekly intake of fish was 300 g. Tuna
    fish, with an average total content of mercury of 1.23 mg/kg and a
    methylmercury content of 0.92 mg/kg was an important dietary item.
    Other fish in their diet had an average total mercury concentration of
    0.33 mg/kg. The concentration of total mercury in blood was measured
    in 115 people aged from 10 years upwards. The mean mercury
    concentration in blood in the males was approximately 8 g/100 ml and
    the maximum recorded concentration was 23 g/100 ml. The females had
    average mercury levels in blood of 6 g/100 ml with a maximum level of
    24 g/100 ml. Three individuals in the sample population of 115 had
    blood levels higher than 20 g/100 ml.

    FIGURE 3

    FIGURE 4

        Epidemiological studies show that these exposed populations may be
    classified into three categories distinguished by intensity and
    duration of exposure to the short-chain alkylmercurials. Populations
    consuming contaminated grain had a high daily intake of mercury
    (reaching over 200 g/kg for brief periods averaging 1-2 months). The
    outbreaks in Japan fall into the second category where daily mercury
    intakes ranged from 5 to 100 g/kg with a median of 30 g/kg/day with
    exposure times lasting from several months to years, in Niigata,
    although doubts on the accuracy of those figures have been expressed
    by a Swedish Expert Group (1971). The third category includes
    populations having unusually high fish intakes for years if not, for
    most of their lives, such as in parts of Sweden (Skerfving, 1974) the
    Samoan fishermen (Marsh et al., 1974), and in Peru (Turner et al.,
    1974). A small proportion of those in this category may attain mercury
    levels up to 5 g/kg/day or even higher. Quantitative studies relating
    frequency of signs and symptoms to various indices of exposure have
    been reported for all three categories so that it is now possible to
    compare the effects of differences in intensity and duration.

        Before discussing these studies, it would be well to point out
    some of the attendant difficulties. Methylmercury and the other
    short-chain alkylmercurials produce effects unique among the mercury
    compounds. However, some, if not all, of these effects can be caused
    by agents other than mercury or by certain disease states. Thus, in
    examining the population for neurological changes, it must be borne in
    mind, that there could be, at least in theory, many causes of the
    observed neurological effects other than methylmercury itself.
    Dose-response relationships derived from these epidemiological studies
    imply a cause-effect relationship. In fact, the only proof we have
    that methylmercury caused certain effects in these populations is that
    (1) these effects coincide in time with exposure to methylmercury, (2)
    the frequency of these effects in a given population increases with
    increasing exposure to methylmercury, and (3) the major signs have
    been reproduced in some animal models. One of the key problems in
    these studies is to distinguish between the background frequency of a
    sign or symptom and the increase in that frequency due to increased
    exposure to methylmercury.

        The studies in the Iraqi population (typical of category one) were
    made 2-3 months after cessation of exposure and in most cases,
    sometime after the onset of signs and symptoms in the patients. Thus
    the investigators were faced with the problem of recapitulation of
    exposure and of determining the dose received by the individual.

    FIGURE 5

    FIGURE 6

        Bakir et al. (1973) classified the population into cohorts
    according to blood levels of mercury. The first blood samples were
    collected at various times after cessation of exposure. They were
    corrected to an average point in time corresponding to 65 days after
    cessation of exposure using a clearance half-time of 65 days. Back
    extrapolation to times earlier than 65 days was not attempted because
    the pattern of mercury clearance from blood was not established for
    this period. In fact it was noted that in 11 cases, blood mercury
    concentrations did not exhibit statistically significant decline
    during the first 20 days of March 1972. These individuals had stopped
    consumption of bread 45 days before collection of the first blood

        Thus, instead of attempting to back extrapolate the blood samples
    to the time of onset of symptoms (for example in the Niigata studies
    to be discussed below), a different approach was made to recapitulate
    exposure. It was noted that 58 individuals (approximately half the
    population studied by Bakir et al., 1973) gave sufficient information
    on their consumption of contaminated bread to allow estimation of the
    ingested dose of mercury. When the blood levels in these individuals,
    corrected to the time point of 65 days after exposure, were plotted
    against the ingested dose as reported by the patients, the
    relationship between blood levels and estimated dose was linear for
    both adults and children, the results for the children giving a
    steeper slope consistent with a smaller volume of distribution of
    methylmercury. The correlation coefficient for people over 18 years of
    age was 0.85 and for people of 10-15 years it was 0.89 (Fig. 2). This
    empirical correlation between blood mercury and ingested dose was used
    to estimate ingested dose from observed blood levels (corrected to 65
    days after exposure).

        Bakir et al. (1973) proceeded to estimate the average amount of
    methylmercury for each group of the people by use of the exponential
    equation (equation 1) discussed in section 6. They assumed complete
    absorption of methylmercury from the diet and that the average
    elimination half-time was 70 days. This estimated body burden, plotted
    on a logarithmic scale, was related to the frequency of signs and
    symptoms in each group of the population (see Fig. 3). The signs and
    symptoms were paraesthesia, ataxia, visual changes, dysarthria,
    hearing defects, and death. It was noted that there was a background
    frequency of signs or symptoms that were not related to the mercury

    level, and that the frequency increased in relation to the mercury
    levels, at high doses of mercury. The thresholda body burden for this
    mercury-related increase in frequency was estimated for each sign of
    symptom. Paraesthesia had the lowest threshold body burden. From the
    dose-response curve, the onset of symptoms was estimated to occur at
    approximately 25 mg of mercury or 0.5 mg/kg body weight.

        Bakir et al. (1973) noted that the empirical relationship between
    blood level and ingested dose did not conform with the relationship
    expected from Miettinen's tracer studies on volunteers. Bakir et al.
    (1973) pointed out that this difference may be due "either to
    differences in conditions of exposure between the Iraqi patients and
    the volunteers given labelled methylmercury, or to underestimations of
    dose in Iraq, or to both causes". They noted that, in 14 patients, the
    average exposure period reported by the patients was 48 days as
    compared with 66 days calculated from hair analysis of the same
    patients. Thus, an alternative dose-response relationship was plotted
    by increasing the estimates of ingested dose by a factor of about 1.6,
    which made their empirically observed blood-ingested dose relationship
    identical to that calculated from Miettinen. This yielded a threshold
    body burden for paraesthesia of 40 mg of mercury or 0.8 mg/kg body

        An alternative calculation may be made directly from the blood
    level, to try to estimate the minimum concentration of mercury in
    blood at which paraesthesia became detectable. If the mean values
    (estimated as geometric means) of the blood mercury for each cohort,
    as reported by Bakir et al., are plotted on a logarithmic scale
    against the frequency of paraesthesia, the relationship has basically
    the same pattern as observed when body burden was used (Fig. 4). The
    horizontal portion of the line relates to the background frequency at
    a mean blood mercury level of between 1.0 and 22 g/100 ml. The next
    points lie significantly above the background frequency level. A
    least-squares linear regression line through these points intercepts
    the background level at a mercury level of approximately 24 g/100 ml.
    However, the estimated threshold value is for mercury concentrations
    in the blood 65 days after cessation of exposure. If correction is
    made using a 65 day clearance half-time, the actual threshold level
    could be twice as high, i.e. 48 g/100 ml. Evidence noted above
    suggested that blood mercury may have been cleared at longer
    half-times than 65 days. Thus the actual threshold mercury value for
    blood probably lies between 24 and 48 g/100 ml.


    a The phrase "threshold body burden" is meant to indicate the value
      of the body burden at which effects due to methylmercury become
      detectable above the background frequency. It is not intended to
      mean that methylmercury does not produce effects in some
      individuals below this level.

    FIGURE 7

        Mufti et al. (1976) and Kazantzis et al. (1976) have reported the
    results of a survey of 956 persons in a heavily affected village in
    Iraq, on an additional 207 persons living nearby, and on 1014 persons
    in a control village that did not receive the treated grain. Mufti et
    al. (1976) reported on 427 persons who had eaten contaminated bread.
    They were divided into groups according to the total consumption of
    contaminated bread (loaves per day x period of consumption) and the
    frequency of parasthesia was reported for each group. The total
    quantity of mercury consumed can be estimated from the mercury content
    of each load.

        Figure 5 is a plot of the frequency of paraesthesia in this
    population against the log of the mean total ingested dose for each
    group plotted on a logarithmic scale. A horizontal line is drawn
    through the point corresponding to 4% paraesthesia assuming this to be
    the background frequency in this population. The points at higher
    ingested doses lie significantly above this line. A linear regression
    line drawn through these points intercepts the background frequency at
    a total infested dose of 37 mg of mercury. Assuming 50 kg to be the
    average body weight for this population (Mufti et al., 1976), this
    threshold dose would be 0.7 mg/kg. However, during the period of
    consumption (average 32 days), some excretion of methylmercury took
    place, so that the maximum body burden must have been less than the
    total ingested. The difference would be small in view of the short
    period of consumption.

        Thus the study of this large population, carried out approximately
    6 months after the study by Bakir et al. (1973), would also be
    compatible with a body burden of mercury in the range of 0.5-0.8 mg/kg
    body weight.

        Shaharistani et al. (1976) have reported on 184 persons in rural
    Iraq, 143 of whom consumed the contaminated bread.a The signs and
    symptoms were classified as mild, moderate, and severe. People
    classified as mild cases complained of numbness of the extremities and
    had slight tremors and mild ataxia. Moderate cases had difficulty of
    hearing, tunnel vision, and partial paralysis. The severe cases
    generally suffered from a combination of the following; complete
    paralysis, loss of vision, loss of hearing, loss of speech, and coma.
    The dose was expressed as the peak hair concentration of mercury
    determined by neutron activation of consecutive 1-cm segments of the
    hair sample, and graphical determination of the peak concentration as
    described by Giovanoli & Berg (1974).


    a The clinical observations were made by a local physician in the
      rural district and by a resident physician of the hospital where
      the hair samples were collected (Shahristani, personal
        Shahristani et al. (1976) did not formulate the usual
    dose-response relationships. Instead the population was classified
    according to the signs and symptoms of poisoning into four groups (no
    symptoms, mild, moderate, and severe symptoms). Figure 6 is redrawn
    from a similar figure presented by Shahristani et al. to indicate the
    range of peak mercury concentrations in hair for each group. The group
    having no signs and symptoms attributable to mercury poisoning had
    hair values in the range 1-300 mg/kg, the mild group in the range of
    120-600 mg/kg, the moderate in the range 200-600 mg/kg and the severe
    in the range 400-1600 mg/kg. Unfortunately, insufficient information
    is given in the paper to allow formulation of the usual dose-response
    relationship in which the population is classified according to dose,
    and so these results cannot be compared quantitatively with the
    results of Bakir et al. (1973) and Mufti et al. (1976). However, they
    do indicate that mild cases have been reported with peak hair levels
    as low as 120 mg/kg.

        Shahristani et al. gave sufficient information on the 30 cases in
    the group they studied to allow estimation of their daily intake of
    methylmercury from contaminated bread. It was possible to compare the
    mercury concentrations in hair, as they increased during the ingestion
    period, with the daily intake of mercury. It was found that the
    concentration of mercury in hair was related to daily intake by an
    exponential equation similar to that described in section 6 relating
    body burden to daily intake. Thus they were able to calculate the
    ratio of mercury concentration in hair (mg/kg) to average
    concentration of mercury in the body (mg/kg). The average ratio was
    found to be 137 with a range of 82-268 in the 30 individuals. Using
    this average ratio, the hair level of 120 mg/kg at which mild symptoms
    were first observed would be equivalent to a body burden of mercury of
    0.8 mg/kg body weight.

        Observations on the Niigata outbreak of methylmercury poisoning
    included figures on concentrations of mercury in samples of blood and
    hair as well as detailed clinical reports. A Swedish Expert Group
    (1971) estimated the blood levels in patients at the time of onset of
    symptoms. This involved a graphical procedure in which the
    concentration of mercury in samples of blood, collected from the
    patient at various times after admission to hospital, were plotted
    against the time of collection on semilogarithmic paper (Fig. 7). The
    decline in blood levels of mercury corresponded to a clearance
    half-time of 70 days, although there was considerable scatter of the
    data. Back extrapolation of time of onset of symptoms revealed that
    the lowest group (about 3 patients) had blood mercury levels in the
    range of 20-40 g/100 ml.

        Data on hair concentrations in the Niigata outbreak were analysed
    and discussed by the Swedish Expert Group (1971). One hair sample,
    collected close to the time of onset of symptoms, contained mercury at
    52 mg/kg. Unfortunately, no corresponding blood sample was available.


    Accurate comparison of hair to blood concentrations in the Niigata
    samples was not possible because the hair specimens were not
    representative of the current blood levels. Tsubaki (1971) has
    presented data relating fish consumption to hair values in the Niigata
    outbreak which also identified cases of poisoning (also published by
    the Swedish Expert Group, 1971). The hair concentrations cannot be
    related to the signs and symptoms because they were not extrapolated
    back to the time of onset of symptoms.

        Several fish-eating populations (category 3) have been examined
    for signs and symptoms of poisoning, and determinations made of
    concentration of mercury in samples of blood and hair. A Swedish
    population of fish eaters had blood levels of mercury up to
    56 g/100 ml. Skerfving (1972) reported a dose-response curve
    calculated from data for this population, which had no cases of
    mercury poisoning, and for the Niigata cases reviewed above. The
    frequency of signs and symptoms was related to concentrations of
    mercury in the hair. The apparent threshold effect corresponded to
    50-90 mg/kg in the hair and thus to a blood mercury level of
    20-36 g/100 ml.

        The studies by Turner et al. (1974) and Marsh et al. (1974) on
    populations of fish eaters in Peru and Samoa are consistent with the
    relationship published by Skerfving (1972). Thus Turner et al. noted
    that 8 people had mercury blood levels between 20 and 30 g/100 ml and
    Marsh et al. also found 2 people having blood levels in the same

        Paccagnella et al. (1974) could not find any connexion between the
    prevalence of neurological defects and concentrations of mercury in
    samples of blood and hair in the high fish consumers on S. Pietro
    Island. No neurological deficits were reported in the three
    individuals having blood mercury levels between 20 and 24 g/100 ml.
    However the dose-response data reviewed above indicate that the risk
    of poisoning at these levels is small.

        Table 4 records the results obtained in studies on the populations
    discussed above. The studies on the Niigata population and the report
    of Shahristani et al. on the Iraqi population identify cases of
    methylmercury poisoning. The quoted levels in hair and blood are those
    seen in the most sensitive individuals in that population. The results
    (Table 4) indicate that such "sensitive" individuals may exhibit
    symptoms of methylmercury poisoning at blood mercury levels in the
    range of 20-40 g/100 ml and at hair levels of 50-60 mg/kg.
    Unfortunately these studies do not indicate the percentage of the
    general population that is sensitive. Studies on high fish consumers
    in Sweden, Peru, Samoa, and Italy, suggest a low probability of
    symptoms at mercury levels of 20-40 g/100 ml blood. At least 15
    persons having blood levels in this range did not exhibit symptoms of

        Table 4.  Summary of concentrations of mercury in samples of blood and hair and the body
              burden of mercury associated with effects (usually paraesthesia) in the most
              sensitive group in the populationa
                             Mercury concentration
                                                     Mercury in
                 Total No.   Blood         Hair      the body
    Population   studied     (g/100 ml)   (mg/kg)   (mg/kg)           References

    Niigata      17          20-40         52        --            Swedish Expert Group
    Iraq         184         --            120       0.8           Shahristani et al.

    Iraq         125         24-48         --        0.5-0.8        Bakir et al. (1973)

    Iraq         427         --            --        0.7            Mufti et al. (1976)

    a The numbers quoted in this table should not be considered independently of the
      accompanying test.
        The studies by Bakir et al. (1973) and Mufti et al. (1976) on the
    Iraqi population followed a different approach. Individual cases of
    methylmercury poisoning were not reported. Instead, the population was
    divided into groups according to observed levels of mercury in the
    blood and according to estimated dose, and the frequency of specific
    signs and symptoms was reported for each group. Using this approach,
    it is necessary to distinguish between the background frequency of
    signs and symptoms and the increase of frequency due to methylmercury.
    The figures quoted in Table 4 represent a graphical estimate (see
    Fig. 3, 4, and 5) of the blood level, a body burden or dose of
    methylmercury where the frequency of paraesthesia emerges above the
    background level. Variation in observed background frequencies ranged
    from 4% (Fig. 5) to 9.5% (Fig. 3), thus setting a practical limit to
    the accuracy of such graphical estimates. Thus the numbers quoted in
    Table 4 are compatible with frequencies of paraesthesia due to
    methylmercury of 5% or less.

        The data in Table 4 apply only to neurological signs and symptoms
    in adults. They do not apply to infants exposed either prenatally or
    in the early postnatal period.

        Skerfving et al. (1974) have reported on the cytogenetic effects
    of methylmercury in 23 people exposed through intakes of various
    amounts of fish containing methylmercury (0.5-7 mg/kg) and in 16
    people with a low or moderate intake of mainly oceanic fish. The
    mercury levels in blood in the "exposed" subjects ranged from 1.4 to
    11.6 g/100 ml and in the "non-exposed" from 0.3 to 1.8 g/100 ml. A
    statistical relationship was found between frequency of chromosome
    breaks and blood mercury levels. In a study carried out in Iraq
    (Firman, unpublished report), no statistically significant difference
    was noted in an exposed group compared with a control population with
    regard to chromosome damage.

    8.1.3  Children and infants with  in utero exposure

        Of the cases of poisoning reported from Minamata (Harada, 1968),
    23 were due to prenatal exposure to methylmercury. These infants had
    severe cerebral involvement (palsy and retardation) whereas their
    mothers had mild or no manifestations of poisoning. However, there is
    a possibility that slight symptoms in the mothers might have been
    overlooked (Harada, 1971). It should also be noted that the infants
    were examined some years after birth and that no mercury levels are
    available either for the mothers or for the affected infants. A case
    of prenatal methylmercury poisoning has been reported for a family
    that consumed meat from pigs that ate grain treated with methylmercury
    fungicide (Pierce et al., 1972). The mother was exposed in early
    pregnancy and had a hair mercury level of 186 mg/kg. It was stated
    that the mother had no symptoms other than a slight slur of speech
    which occurred during two weeks in early pregnancy. The child, which
    was delivered normally after an uneventful pregnancy, exhibited
    tremulous movement of the extremities in the first few days of life
    and subsequently developed myclonic convulsions (Snyder, 1971). At one
    year of age the infant exhibited normal physical growth but could not
    sit up and was blind.

        Cases of prenatal poisoning have been referred to in a preliminary
    report on the Iraqi epidemic where it was noted that blood levels in
    the infants at birth and in the first few months after birth could be
    considerably higher than those of the mother (Bakir et al., 1973).
    Quantitative data on the prenatal exposure of the infants is not yet
    available. These observations, however, have led to the belief that
    prenatal life in man is more sensitive than adult life but the
    difference in the degree of sensitivity has not yet been
    quantitatively established.

        Amin-Zaki et al. (1974a) have reported clinical examinations of 15
    infant-mother pairs in which the infant was exposed prenatally to
    methyl-mercury. Clinical manifestations were evident in 6 of the
    mothers and in at least 6 of the infants. Five of the infants were
    severely affected having gross impairment of motor and mental
    development. However, in only 1 infant-mother pair was the infant
    affected and the mother free of signs and symptoms.

        These observations from the 1971-72 Iraq outbreak must be regarded
    as preliminary. It may take time for some of the consequences of
    prenatal exposure to manifest themselves. Harada (1971) scrutinized
    the chromosomes of 7 victims with congenital Minamata disease and 1
    infant victim with severe noncongenital Minamata disease, and reported
    that the chromosome patterns were within normal range.

    8.2  Clinical Studies of Effects of Mercury-binding Compounds

        Attempts to treat mercury poisoning have generally involved the
    use of antidotes that reduce the amount of mercury in the target
    tissue, either by forming an inactive complex with mercury, or by
    enhancing its removal from the tissue. Such antidotes are of course
    used in conjunction with general supportive therapy. Ideally the
    antidote should have a sufficiently high affinity for mercury so that
    nontoxic doses are able to remove mercury from tissue binding sites.
    The mercury chelate so formed should be less toxic than mercury and
    preferably should be rapidly excreted. The agent should be
    metabolically stable so that dosing should not be too frequent and
    preferably the agent should be given by mouth. These antidotes are
    most effective when given early after exposure to mercury. Clearly the
    removal of mercury is without much advantage if irreversible damage
    has already occurred.

        The first effective antidote, 2,3-dimercaptopropanol (British
    Antilewisite-BAL), is a sulfur-containing compound (a dithiol
    molecule, possessing a remarkably high affinity for divalent ionic
    mercury) and was developed on the basis that mercury and other heavy
    metals combined with sulfur groups in the body (for detailed review,
    see Levine, 1970). This compound is life saving in cases of acute
    mercury(II) chloride poisoning, alleviates symptoms from overdoses of
    mercury diuretics, and dramatically relieves certain symptoms of
    acrodynia. For alkylmercury poisoning BAL is contraindicated since it
    increases brain mercury levels (Berlin & Rylander, 1964; Magos, 1968).
    It also does not alleviate neurological disorders caused by mercury
    vapour exposure (Hay et al., 1963; Glomme & Gustavson, 1959). Unithiol
    (2,3, dimercaptopropansulfonate) is a water soluble derivative of BAL
    that is apparently more effective in mobilizing mercury (Trojanowska &
    Azendzikowski, in press; Dutkiewicz & Oginski, 1967). Furthermore
    unithiol does not produce redistribution to the brain, as has been
    observed after BAL treatment. Unithiol is effective in the treatment
    of occupational mercurialism (Fesenko, 1969), but there are no reports
    on its effects on alkylmercury poisoning.

        The penicillamines (D-penicillamine and  N-acetyl-DL-
    penicillamine) are effective in increasing the excretion of mercury
    after exposure to mercury vapour and in relieving the symptoms of
    chronic mercury vapour poisoning (Smith & Miller, 1961; Parameshvera,
    1967). Fatal brain damage can be prevented in the offspring of rats
    treated with methylmercury, by D-penicillamine according to the
    experiments of Matsumoto et al. (1967). Recent literature reports
    (Bakir et al., 1973; Suzuki & Yoshino, 1969) indicate that the
    penicillamines are capable of mobilizing mercury from tissues and
    increasing the excretion of mercury in cases of methylmercury
    poisoning in man. Thus it appears that the penicillamines offer
    advantages over BAL in that they are orally effective, less toxic, and
    effective in treating mercury vapour poisoning and probably
    alkylmercury poisoning, when administered immediately following

        A slight increase in the urinary excretion of mercury has been
    noted in methylmercury poisoned patients with "Minamata disease"
    (Katsuna, 1968), after administration of EDTA.

        Thioacetamide increases urinary excretion of mercury in animals
    dosed with mercury(II) chloride but probably one of the major causes
    of this effect is kidney damage caused by the combination of the toxic
    effects of thioacetamide and mercury leading to an increased
    exfoliation of renal tubular cells (Trojanowska et al., 1971).

        Some interesting new ideas in the realm of antidotes to mercury
    are worth noting. Aaseth (1973) described the application of large
    molecular weight mercaptodextran in the successful treatment of
    mercury(II) chloride poisoning in animals. This agent does not enter
    the intracellular spaces and achieves removal of mercury from the body
    without redistribution. However, its effectiveness is limited by the
    time of administration. For example, if given more than 2 hours after
    exposure to mercury, this compound is totally ineffective whereas BAL
    is still useful.

        A second approach is to give a nonabsorbable mercury-binding
    compound in the diet, in order to trap the mercury that is secreted in
    the bile, to prevent its reabsorption, and to greatly increase the
    faecal elimination (Takahashi & Hirayama, 1971; Clarkson et al.,
    1973a). A polystyrene resin containing fixed sulfhydryl groups has
    been shown to increase mercury excretion in experimental animals given
    methylmercury, and to reduce blood mercury levels in the victims of
    methylmercury poisoning in Iraq (Clarkson et al., 1973a; Bakir et al.,
    1973). There is, however, variation in response among patients. More
    recently it has been demonstrated that phenobarbital can increase the
    biliary excretion of methylmercury compounds (Magos et al., 1974).

        A new technique for the removal of methylmercury directly from
    blood has been proposed for use in methylmercury poisoning (Kostyniak
    et al, 1975). The simultaneous combination of extracorporeal regional
    complexation of methylmercury with haemodialysis has been effective in
    producing a rapid removal of mercury in both experimental animals and
    man (Kostyniak et al., 1974; Al-Abbasi et al., 1974, unpublished

    8.3  Pathological Findings and Progression of Disease

        The signs and symptoms of acute toxicity, severe gastrointestinal
    damage, shock, cardiovascular collapse, and acute renal failure, after
    large doses of divalent mercury and pulmonary irritation after
    inhalation of massive doses of the vapour, reflect the fact that all
    mercury compounds are chemically reactive, can denature proteins,
    inactivate enzymes, and disrupt cell membranes, leading to cellular
    death and the destruction of any tissue with which they come into
    contact in sufficient concentration. The pathological effects
    following long-term exposure to lower doses of mercury compounds are
    more subtle and depend upon the type of mercury compound to which the
    subjects are exposed. The remainder of this section will be concerned
    with long-term exposure, as this type of exposure is most important in
    helping to assess the hazards to man of the presence of mercury in
    food and air.

        Pathological findings on human subjects exposed to mercury and its
    compounds have been reviewed by a Swedish Expert Group (1971) and by
    Friberg & Vostal (1972).

        Pathological findings demonstrate that methyl- and ethylmercury
    compounds are primarily neurotoxic and produce similar types of lesion
    in man. The main pathological features consist of the destruction of
    neurological cells in the cortex particularly in the visual areas of
    the occipital cortex and various degrees of damage to the granular
    layer in the cerebellum. Damage to the peripheral nerves may occur as
    indicated by clinical signs but no definitive pathological
    observations are available for man. Takeuchi (1970) has reported on
    changes in the diameter of the peripheral nerves in patients suffering
    from heavy exposure to methylmercury in the Minamata Bay epidemic.
    However, Von Burg & Rustam (quoted by Bakir et al., 1973) could not
    find any changes in conduction velocities in the patients in Iraq who
    received very high exposure to methylmercury.

        Brain concentrations of mercury associated with the onset of
    pathological changes following exposure to the vapour or to doses of
    the alkylmercury compounds are not fully known. Takahata et al. (1970)
    demonstrated mean mercury levels of 11 mg/kg wet weight in the brains
    of two workers, poisoned by exposure in a mercury mine, who had had no

    known exposure to mercury for 10 years prior to death. Studies on
    occupationally poisoned individuals (Swedish Expert Group, 1971), and
    the patients who died in the Minamata epidemic indicate that the onset
    point of signs and symptoms corresponds to an average brain level of
    approximately 5 mg/kg wet weight. These findings are in general
    agreement with threshold methylmercury concentrations in studies on
    experimentally poisoned animals.

    8.3.1  Psychiatric and neurological disturbances

        Trahtenberg (as reviewed by Friberg & Nordberg, 1972) reported
    minor psychiatric disturbances such as insomnia, shyness, nervousness,
    and dizziness in workers exposed to elemental mercury vapour
    concentrations of the order of 0.1 mg/m3. The classical literature
    (for review, see Friberg & Nordberg, 1973) contains detailed accounts
    of the consequences of long exposure to higher concentrations of
    elemental mercury vapour where the full syndrome of erethism is seen.
    Individual variation in exposed people is the rule but the most
    commonly reported syndrome includes loss of memory, insomnia, lack of
    self-control, irritability and excitability, anxiety, loss of
    self-confidence, drowsiness, and depression. In the most severe cases
    delirium with hallucinations, suicidal melancholia, or even
    manic-depressive psychoses have been described.

        The presence of tremor is one of the most characteristic features
    of mercurialism and usually follows the minor psychological
    disturbances referred to above. With continuing exposure to elemental
    mercury vapour, the tremor develops gradually in the form of fine
    trembling of the muscles interrupted by coarse shaking movements every
    few minutes. It may be seen in the fingers, but also on the closed
    eyelids, lips, and on the protruding tongue. The frequency is of the
    order of 5-8 cycles per second. It is intentional and stops during
    sleep. On cessation of exposure, the tremor gradually disappears.
    Dramatic alterations in the steadiness of the handwriting may be seen
    in persons suffering from mercurial tremor.

        The most common signs and symptoms in cases of poisoning due to
    methyl- or ethylmercury compounds are paraesthesia, loss of sensation
    in the extremities and around the mouth, ataxia, constriction of the
    visual fields, and impairment of hearing. In the Japanese experience
    the effects of alkylmercury poisoning are usually irreversible but the
    coordination may improve after rehabilitation (Kitagawa, 1968). On the
    other hand, in Iraq, improvement in motor disturbances was often
    spontaneous. In addition, paraesthesia was often reversible in Iraq
    but a persistent symptom in Japan (Damluji, 1974; Tsubaki, 1971).

    8.3.2  Eye and visual effects

        Occupational exposure to elemental mercury vapour causes the
    appearance of a greyish-brown or yellow haze on the anterior surface
    of the lens of the eye (Atkinson, 1943). It appears usually after
    long-term exposure and the depth of colour depends upon the length of
    time and the air concentration of mercury to which the worker has been
    exposed. The presence of this coloured reflex may or may not be
    associated with signs and symptoms of poisoning.

        The narrowing of the visual fields is a classic sign of poisoning
    due to short-chain alkylmercurial compounds (Hunter et al., 1940). In
    cases of severe exposure, the constriction may proceed to complete

    8.3.3  Kidney damage

        Kazantzis et al. (1962) reported four cases of proteinuria in two
    groups of workmen exposed to elemental mercury vapour. Exposure
    conditions were not specified, but all four cases were excreting
    mercury in urine in excess of 1000 g/litre at the time of the first
    examination. The proteinuria disappeared after the workers were
    removed from exposure. Joselow & Goldwater (1967) noted that the mean
    urinary protein concentration (90 mg/litre) in a group of workers
    exposed to elemental mercury vapour was significantly higher than the
    mean protein concentrations (53 mg/litre) in a nonexposed group. The
    urinary protein correlated with urinary mercury levels.

        Renal involvement after exposure to methylmercury compounds is
    very rare (Bakir et al., 1973). Cases of renal damage have been
    reported in an outbreak of ethylmercury poisoning (Jalili & Al-Abbasi,

    8.3.4  Skin and mucous membrane changes

        Dermatitis has been reported after occupational exposure to
    phenylmercurials (for review, see Goldwater, 1973). However dermatitis
    may result from exposure to inorganic mercury (Hunter, 1969).
    Sensitivity to metallic mercury in tooth fillings has resulted in
    facial and intra-oral rashes. Skin reactions due to sensitivity to
    phenylmercury have also been reported (for review, see Clarkson,

        Dermatitis has been reported after skin contact with the
    alkylmercurials (for review, see Swedish Expert Group, 1971). Oral
    ingestion of methyl- and ethylmercury compounds may also result in
    this condition, as observed in the Iraqi epidemics (Jalili &
    Al-Abbasi, 1961; Damluji et al., 1976).


    9.1  General Considerations

        In order to control risks to human health from the presence of
    mercury in the environment, it is necessary to attempt to define the
    degree of risk in any given environmental situation. The health risk
    evaluation depends in part on a knowledge of dose-effect,
    dose-response relationships. It also depends on a knowledge of the
    variation in exposure (or intake) in any given situation. A general
    review of sampling and analytical techniques (section 2), and of
    environmental sources and exposure levels (sections 3 and 5) has been
    presented in this document. However, specific assessment of exposure
    or intake must be made by the public health authorities responsible
    for any given population or group. This section of the report is
    concerned, therefore, primarily with a summary of those dose-effect,
    dose-response relationships that are relevant to human populations in
    both occupational and general environmental exposure to mercury and
    its compounds.

        Species differences in the metabolism and toxicity of mercury and
    its compounds are so great that this health risk evaluation is based
    primarily on data for man. Animal data have been included or
    considered only when data for man are lacking, and have generally been
    used in a qualitative way. Thus animal data may in certain cases
    indicate the potential for genetic effects, or that a certain stage of
    the life cycle, for example, the fetus appears to be the most
    sensitive stage. General patterns of deposition of mercury in the
    body, as seen from animal experiments, have been considered to be
    broadly applicable to man in a qualitative sense.

        Quantitative estimates of the dose-response relationship in man
    are fraught with many difficulties. Observations of
    occupationally-exposed persons should offer the best possibilities for
    well controlled studies. However, the exposure range and population
    sizes are limited and difficulties are usually encountered in
    obtaining accurate estimates of time-weighted average exposure to
    airborne mercury. The most difficult situation for controlled study
    arises in the case of methylmercury where our knowledge derives
    primarily from observations on the unexpected outbreaks of poisoning
    from contaminated food in Japan and Iraq. Generally the studies
    commenced some time after the start of the outbreak or after the end
    of the exposure. Attempts had to be made either to recapitulate
    exposure by the back-extrapolation of blood levels, or to estimate the
    ingested dose from the patient's ability to remember. Doubts have also
    been expressed on the reliability of analytical methods in the earlier
    Japanese outbreaks.

        Estimates will be made of minimum effect exposures or
    concentrations in indicator media since these numbers may be of value
    to authorities in setting safety standards. These minimum effect
    figures (e.g. Tables 5 and 6) should be viewed in terms of the overall
    dose-response relationship, namely, that as the dose is decreased, so
    also is the probability of poisoning, and that the minimum effect
    level is that dose (expressed as exposure level, daily intake,
    concentration in indicator media) that is associated with the first
    detectable effect in the population under study. The effect will be
    present at some specified frequency in the population. The value of
    the minimum effect dose is the dose derived from the observed
    dose-response relationship. (It will be subject to more statistical
    uncertainty than, for example, the dose giving 50% frequency of the
    effect.) Its value will depend on the size of the population under
    study; the larger the population, the more likely it will be that
    effects will become evident in the more sensitive individuals.

    9.1.1  Elemental mercury vapour

        The central nervous system is the critical organ for the toxic
    effects of inhaled elemental mercury. Effects on the kidney such as
    proteinuria have been reported but only at doses higher than those
    associated with the onset of signs and symptoms from the central
    nervous system. No information is available with regard to mercury
    levels in the central nervous system at the time of onset of signs and
    symptoms or at death in man. Furthermore, we do not have indicator
    media such as urine or blood the mercury levels of which would reflect
    those in the brain. Observations on animals exposed to elemental
    mercury vapour reveal large regional differences in distribution
    within the brain so that average brain concentrations, even if they
    were available for man, might not be of much value.

        Our evaluation of the health risks from exposure to elemental
    mercury vapour in man has therefore to be based on an empirical
    relationship between air levels (exposure) and the frequency of signs
    and symptoms in exposed populations. Concentrations of mercury in
    urine and blood are related, on a group basis, to average air

        A complete metabolic model relating air levels to absorption,
    accumulation, and excretion of mercury in man following exposure to
    elemental mercury vapour is not available. However, some useful
    generalizations are beginning to emerge from recent observations. Most
    of the inhaled vapour (approximately 80%) is retained in the lung.
    Once absorbed into the blood stream, it is rapidly oxidized to ionic
    mercury. The limited information on biological half-times in man
    suggests that a workman exposed to a constant average concentration of

    mercury vapour in his working environment would not reach a state of
    balance (steady state) until after one year of exposure. Consequently
    one would expect the concentrations of mercury in blood or urine to
    exhibit a consistent relationship to air levels after the worker had
    been exposed for at least one year. Unfortunately, most of the
    publications in the literature do not indicate the period of
    employment of the worker. However, general experience in occupational
    health studies (for review, see MAC Committee, 1969) reveals that
    exposure of workers to an average air concentration of mercury of
    0.05 mg/m3 are associated, on a group basis, with blood levels of
    approximately 3.5 g/100 ml, and with urinary concentrations of
    150 g/ litre. Linear relationship has been observed between urinary
    and blood concentrations of mercury in people occupationally exposed
    to elemental mercury vapour. Thus, measurements of mercury
    concentrations in the working atmosphere and in samples of blood and
    urine from the workmen may be used as an index of exposure. These
    values in turn may be compared with the frequency of clinical signs
    and symptoms in workers experiencing different degrees of exposure.
    Such studies, reported in detail in section 8 and briefly reviewed
    below, form the basis for establishing "threshold limit values" or
    maximal allowable air concentrations of mercury in occupational

        Early studies dating back to the 1930s indicate that cases of
    poisoning occurred at atmospheric mercury levels above 0.1 mg/m3 (for
    details, see section 8.1.1). Recent data also demonstrate that there
    was an increase in complaints of appetite-loss and insomnia in a group
    of workers exposed to time-weighted average air concentrations of
    mercury between 0.06 and 0.1 mg/m3 as compared with two lower
    exposure groups (0.01 and 0.05 mg/m3). These findings are in
    agreement with data published in the 1940s and 1950s indicating that
    mercury intoxication occurred in workers exposed to mercury in air
    concentration less than 0.2 mg/m3, but no data were given on the
    lower exposure limits.

        At this time it is difficult, if not impossible, to establish a
    lower limit at which no effects occur. Studies reviewed in section
    8.1.1 indicate that mental disturbances may be seen at extremely low
    mercury concentrations in air. However, the problem in the
    interpretation of these reports is that, as the air concentration of
    mercury decreases, it becomes more difficult to correlate effects with
    exposure with any degree of confidence. For example, it would appear
    that effects of mercury levels below 0.05 mg/m3 have not been
    unequivocally established.

        Concentrations of mercury in blood and urine equivalent to average
    mercury concentrations in air of 0.05 and 0.1 mg/m3 are given in
    Table 5. In general these relationships are not seen in indivduals but
    only when averaged over a substantial number of workers.

        An occupational limit for mercury in air concentration of
    0.05 mg/m3 would be equivalent to an ambient air level for the
    general population of approximately 0.015 mg/m3. This calculation is
    based on the assumptions of a daily ventilation of 10 m3 during
    working hours and 20 m3 for a 24-hour day and that there are 225
    working days in the year. The data of Smith et al. (1970) would
    indicate that the probability of seeing adverse effects at air levels
    of, 0.05 mg/m3 for occupational exposure, and 0.015 mg/m3 for
    continuous environmental exposure, is low for such symptoms as loss of
    appetite, weight loss, and shyness. However, this calculation does not
    take into account the sensitive groups in the general population. It
    should be noted that concentrations found in ambient air are far below
    these levels (see section 5).

    Table 5.  The time-weighted average air concentrations associated
              with the earliest effects in the most sensitive adults
              following long-term exposure to elemental mercury vapour.
              The table also lists the equivalent blood and urine


    Air         Blood            Urine        Earliest effects
    (mg/m3)     (g/100 ml)      (g/litre)   

    0.05        3.5              150          non-specific symptoms
    0.1-0.2     7-14             300-600      tremor

    a Blood and urine values may be used on a group basis owing to
      gross individual variations. Furthermore, these average values
      reflect exposure only after exposure for a year or more. After
      shorter periods of exposure, air concentrations would be
      associated with lower concentrations in blood and urine.

    9.1.2  Methylmercury compounds

        The estimate of risks to human health from methylmercury compounds
    is important for several reasons. First, many thousands of people have
    been poisoned following accidental consumption of food contaminated
    with methylmercury fungicides or the consumption of fish contaminated
    by industrial release of methylmercury. Second, methylmercury probably
    accounts for a significant part of mercury in the human diet and is
    especially important in fish and fish products. Third, the
    risk-benefit calculations with regard to methylmercury in fish are of
    critical importance in those countries and areas of the world where
    fish is an important dietary source of protein, or where the fish
    industry is of economic importance.

        This evaluation of risk from exposure to methylmercury compounds
    is based primarily upon data in man. The data consist of observations
    on the frequency of signs and symptoms in populations exposed to a
    wide range of mercury intake, observations on the concentrations of
    mercury in hair and blood samples, and on estimates of dietary intake
    (section 8).

        In estimates of risks to human health, the custom has been
    followed of attempting to determine the lowest concentration in
    indicator media or the lowest daily intake associated with the onset
    of toxic signs and symptoms in man, along with information on maximal
    intakes which produce no effects. This procedure has been followed in
    estimating the data presented in Table 6. The levels in blood and
    hair, and the amount of the body burden of mercury associated with the
    onset of signs and symptoms are taken from Table 4.

        The reports on the Minamata outbreak to the effect that infants
    were born having cerebral palsy due to methylmercury whereas their
    mothers lacked or had only slight symptoms led to the belief that the
    fetus was the stage of the life cycle that was most sensitive to
    methylmercury. Studies on animals, exposed during the gestation period
    to methylmercury led to qualitative confirmation of this conclusion.

        Table 6 lists our conclusions on concentrations of total mercury
    in indicator media associated with the earliest effects of
    methylmercury in the most sensitive group in the adult population. As
    discussed below, the prevalence of the earliest effects would be
    expected to be approximately 5%. The equivalent long-term daily intake
    quoted in Table 6 was calculated on the most conservative relationship
    (quoted in Table 3) i.e. that the steady state blood concentration
    (ng/ml) is numerically equal to the average daily intake (g/day/70 kg
    body weight).

        However, Nordberg & Strangert (1976) have proposed an alternative
    approach to estimating risks of poisoning on long-term exposure. The
    relationship between long-term daily intake and steady state blood
    levels (and hair levels and body burden) depends,  inter alia, on the
    biological half-time of methylmercury in man. Biological half-time
    times are subject to individual variation as discussed in section 6.
    Thus an individual having a long biological half-time (a slow excretor
    of methylmercury) would accumulate higher steady state levels than one
    having a short biological half-time. Thus the statistical distribution
    of biological half-times should be taken into account in estimates of
    risk of poisoning.

        Variations also occur in threshold concentrations for the
    appearance of signs and symptoms in individuals in the population.
    These may be estimated from empirical relationships relating frequency
    of signs and symptoms to concentrations of methylmercury in indicator
    media (blood and hair) or body burdens (e.g. Fig. 3, 4, and 5). On the
    assumptions that the distribution of biological half-times was normal,

    that the distribution of individual threshold values for paraesthesia
    was log-normal, and that these distributions were independent of each
    other, Nordberg & Strangert (1974) estimated the overall probability
    of an individual developing symptoms of paraesthesia. Their
    calculations were based on data of Shahristani & Shihab (1974) and
    gave the distribution of biological half-times estimated from analysis
    of hair samples after the Iraqi outbreak, and the statistical
    distribution of individual threshold values estimated from the data of
    Bakir et al. (1973) on the relationship of frequency of paraesthesia
    to body burdens. Their results indicated that with a long-term daily
    intake of 4 g/kg would yield the risk of paraesthesia of about 8%.
    Subsequent calculations based on the data by Mufti et al. (1976) would
    indicate a risk of between 3% and 4% for the same daily intake.a

        The figures estimated by Nordberg & Strangert may somewhat
    overestimate the risk of poisoning. The errors to be expected in both
    field and laboratory observations would tend to decrease the slope in
    dose-response relationships. This would lead to an overestimate of the
    variance of threshold values in the general population. Nevertheless,
    the estimate of risk by Nordberg & Strangert is in reasonable
    agreement with the more empirical estimates discussed in section 8 and
    indicates that, with a long-term daily intake as listed in Table 6,
    the prevalence of the earliest effects could be expected to be 5% or

        Occupational hazards have arisen mainly from airborne
    concentrations of alkylmercurials. Skin contact has also been noted,
    but the quantitative importance of skin contact and percutaneous
    absorption cannot be estimated. Hazards from occupational exposures
    can be estimated only by reference to data already discussed above
    with respect to dietary intake of methylmercury compounds. The data
    summarized in Table 6 indicate that the first effect (paraesthesia)
    associated with long-term intake of methylmercury arises at an intake
    level of approximately 5 g/kg body weight per day. Assuming a daily
    ventilation at work of 10 m3 of air, 80% retention of the inhaled
    mercurial, 225 working days to the year, the average time-weighted air
    concentrations that would give rise to this intake would be
    0.07 mg/m3. Consideration of occupational risks should also take into
    account the possibility of a high but brief exposure to methylmercury


    a Nordberg & Strangert, personal communication.

    Table 6.  The concentrations of total mercury in indicator media
              and the equivalent long-term daily intake of mercury as
              methylmercury associated with the earliest effects in
              the most sensitive group in the adult populationa,b


    Concentrations in indicator media

    Blood          Hair      Equivalent long-term daily intakec
    (g/100 ml)    (g/g)    (g/kg body weight)

    20-50          50-125    3-7

    a The prevalence of the earliest effects could be expected to be
      approximately 5%.

    b The WHO Task Group specifically urged that this table should
      not be considered independently of the text in section 8.

    c A Japanese group has recently concluded that a daily intake of
      mercury of 5 g/kg is the "minimal toxic dose", following a
      ten-year follow-up study of the Minamata outbreak (Research
      Committee on Minamata Disease, 1975).

        The estimates in Table 6 apply only to adults. As discussed
    previously, prenatal life may be the stage of the life-cycle most
    sensitive to methyl-mercury. It would be prudent therefore to follow
    the advice of the MAC Committee (MAC, 1969), not to expose females of
    child bearing age occupationally to methylmercury compounds.

        Studies by Skerfving et al. (section 8) have indicated that
    chromosome breaks may be associated with exposure to methylmercury.
    There are, however, other studies in which no such relationship was
    found. Furthermore the health significance of chromosome breakage is
    not known. However, as reviewed elsewhere in this criteria document,
    experiments on animals and other forms of life do indicate the
    potential for genetic damage by methylmercury.

    9.1.3  Ethylmercury compounds and other short-chain

        Insufficient information is available to allow risk calculations
    based on data from human exposure to ethyl- or higher short-chain
    alkylmercurial compounds. Suzuki et al. (1973) have reported
    observations on five patients poisoned with ethylmercury compounds.
    The picture of distribution between plasma and red cells and
    observations on autopsy tissue indicate that the metabolism and
    patterns of distribution of ethylmercury are generally similar to
    those of methylmercury compounds. However, in one individual the
    clearance-time from blood was only 10 days. Evidence reviewed in
    section 6.6 indicates that ethylmercury compounds may be more rapidly
    converted to inorganic mercury in the body than methylmercury
    compounds. Thus, these limited observations suggest that ethylmercury
    compounds are probably less hazardous than the methylmercury compounds
    in so much as they remain in the body for a shorter time because of
    transformation to inorganic mercury and more rapid excretion. Thus,
    standards set for methylmercury compounds will probably be sufficient
    to control the hazards from other short-chain alkylmercurial

    9.1.4  Inorganic mercury, aryl- and alkoxyalkylmercurials

        The risk to human health from long-term ingestion of inorganic,
    aryl-, and alkoxyalkyl- compounds of mercury in the diet is difficult
    to estimate because there are not any recorded cases of human
    poisoning under these circumstances. Data from animals cannot be used
    for exact quantitative extrapolation because of species differences in
    the metabolism and toxicity of these compounds. However, certain
    qualitative conclusions based on animals can probably be extrapolated
    to man. For example, the arylmercurials are rapidly converted to
    inorganic mercury in mammals. The penetration of mercury across the
    blood-brain and placental barriers is less after doses of inorganic
    and aryl compounds than after equivalent doses of elemental mercury
    vapour and short-chain alkylmercurials. Animal studies indicate that
    the kidney is the critical target organ for exposure to inorganic,
    alkyl-, and alkoxy-alkylmercurials. Kidney involvement appears to be
    minimal in workers exposed to concentrations of mercury vapour
    (0.05-0.1 mg/m3) that elicit the first signs and symptoms of damage
    to the central nervous system. Thus guidelines for health protection,
    set for long-term exposure to elemental mercury vapour, should offer
    an even greater safety margin for equivalent exposures to inorganic,
    alkyl, and alkoxyalkyl compounds. In fact, recognizing the lower toxic
    potential of these forms of mercury, the MAC Committee (1969) advised
    a maximum allowable concentration for occupational exposure of
    0.1 mg/m3 -- twice as high as that for elemental mercury vapour.

        In the discussion of guidelines for exposure to elemental mercury
    vapour, it was concluded that long-term exposure of the general
    population to 0.015 mg/m3 was equivalent, in terms of average daily
    mercury intake, to the occupational limit of 0.05 mg/m3. Assuming an
    average pulmonary retention of 80%, the average daily amount entering
    the blood stream is 280 g based on a daily ventilation of 20 m3.

        The equivalent daily intake in diet of phenylmercury compounds
    would also be about 240 g, as animal studies indicate virtually
    complete absorption in the gastrointestinal tract. The alkoxyalkyl and
    aryl compounds are probably absorbed equally well from food. Tracer
    studies on volunteers indicate that approximately 10% of inorganic
    mercury compounds are absorbed from the diet, so that dietary intakes
    approximately ten times greater than those of phenylmercury compounds
    would offer no greater risk of poisoning.

        A daily intake of mercury of 240 g is in the same range as the
    daily intake listed for methylmercury compounds in Table 6. The
    biological half-time for inorganic mercury, based on tracer studies in
    man, appears to be less than that for tracer doses ofmethylmercury.
    Animal data suggest that aryl and alkoxyaryl compounds have biological
    half-times lower than that for methylmercury and similar to that for
    inorganic mercury. Thus the long-term ingestion of inorganic, aryl,
    and alkoxyaryl compounds should offer no greater hazards and probably
    substantially less than the hazards from ingestion of methylmercury

    9.2  Summary and Guidelines

        In the case of elemental mercury and alkylmercury the Task Group
    was able to construct tables (Tables 5 and 6) that related exposure to
    symptoms as well as to concentrations in indicator media in the human
    body. In the case of inorganic mercury, arylmercurials, and
    alkoxyalkyl mercurials, this could not be done because of the
    inconsistency in the animal data available as well as a paucity of
    data in man.

        The tables for elemental mercury and alkylmercury are given above.
    In constructing these tables, the Task Group evaluated the results of
    studies summarized in this document, and drawing on their experience
    and judgement, identified the concentration and amounts of mercury
    associated with certain observed effects. There is insufficient
    information available to permit precise quantification of this risk.
    Usually, the proportion that may be expected to be affected is small.

        The expected health effects for elemental mercury at a level in
    air of 0.05 mg/m3 have been quoted only for occupational exposure
    assuming 8 hours per day and 225 working days a year. The equivalent
    environmental mercury levels in air for continuous exposure would be

    approximately 0.015 mg/m3 to give the same degree of risk. The urine
    and blood values, of course, would be the same as those quoted in
    Table 5. Even though the figures do not only take into account
    specific sensitive groups, it is highly unlikely that the
    concentration in the general environment approaches levels of
    toxicological significance.

        The ranges of minimum effect values quoted in Table 6 for
    methylmercury reflect the uncertainty in estimations.

        Although it was not possible to identify even approximate minimum
    effect values for inorganic, aryl-, and alkoxyalkylmercurials, the
    Task Group concluded that the limited experience for occupational
    exposure suggested that these forms of mercury were probably less
    hazardous than either elemental mercury vapour or methylmercury
    compounds. Thus the figures for occupational exposure to elemental
    mercury vapour given in Table 5 would serve as conservative figures
    for occupational exposure to these forms of mercury and those in Table
    6 would offer conservative figures for dietary intake.

        A Joint FAO/WHO Expert Committee on Food Additives (1972)
    established a provisional tolerable weekly intake of 0.3 mg of total
    mercury per person of which no more than 0.2 mg should be present as
    methylmercury (expressed as mercury); these amounts are equivalent to
    5 g and 3.3 g, respectively, per kg of body weight. Where the total
    mercury intake in the diet is found to exceed 0.3 mg per week, the
    level of methylmercury compounds should also be investigated. If the
    excessive intake is attributable entirely to inorganic mercury, the
    above provisional limit for total mercury no longer applies and will
    need to be reassessed in the light of all prevailing circumstances.


    AASETH, J. (1973)  Acta Pharm. Toxicol., 32: 430.

    ABERG, B. ET AL. (1969)  Arch. environ. Health, 19: 478.

    ACGIH (1976)  Threshold limit values for chemical substances and
         physical agents in the workroom environment with intended changes
         for 1976. Cincinnati, American Conference of governmental
        industrial hygienists, pp. 94.

    ALBANUS, L. ET AL. (1972)  Environ. Res., 5: 425.

    AMIN-ZAKI, L. ET AL. (1974a)  Pediatrics, 54: 587.

    AMIN-ZAKI, L. ET AL. (1974b)  J. Pediatr., 85: 81.

    AMIN-ZAKI, L. ET AL. (in press)  Am. J. Dis. Children.

    ANDERSSON, A. (1967)  Grundforbattring, 20: 95.

    ARMSTRONG, R. D. ET AL. (1963)  J. Am. ind. Hyg. Assoc., 24: 366.

    ASHE, W. F. ET AL. (1953)  Arch. ind. Hyg. occup. Med., 7: 19.

    ATKINSON, W. (1943)  Am. J. Ophthalmol., 26: 685.

    BACHE, C. A. ET AL. (1971)  Science, 172: 951.

    BAKIR, F. ET AL. (1973)  Science, 181: 230.

    BAKULINA, A. V. (1968)  Sov. Med., 31: 60.

    BALDI, G. ET AL. (1953)  Med. Lav., 44: 161.

    BARBER, R. T. ET AL. (1972)  Science, 178, 636.

    BERLIN, M. & RYLANDER, R. (1964)  J. Pharm. exp. Ther., 146: 236.

    BERLIN, M. ET AL. (1966)  Arch. environ. Health, 12: 33.

    BERLIN, M. ET AL. (1969)  Arch. environ. Health, 18: 42.

    BERLIN, M. ET AL. (1973)  The uptake and distribution of methyl
         mercury in the brain of Saimiri sciureus  in relation to
         behavioral and morphological changes. In: Miller, M. W. &
        Clarkson, T. W. ed.  Mercury, mercurials and mercaptans.
        Springfield, C. C. Thomas, p. 187.

    BERTINI, K. K. & GOLDBERG, F. D. (1971)  Norw. Hyg. Tidschr., Suppl.
        4, p. 65.

    BIDSTRUP, P. L. (1964)  Toxicity of mercury and its compounds,
        Amsterdam, Elsevier Publishing Co., p. 34.

    BIDSTRUP, P. L. ET AL. (1951)  Lancet, 2: 856.

    BIRKE, G. ET AL. (1972)  Arch. environ. Health, 25: 77.

    BORNMANN, G. ET AL. (1970)  Arch. Toxicol., 26: 203.

    BOUQUIAUX, J. (1974) In:  Proceedings of the International Symposium
         on the problems of contamination of man and his environment by
         mercury and cadmium, Luxembourg, 3-5 July, 1973, CEC Luxembourg,
        p. 23.

    BRAMEN, R. S. & JOHNSON, D. L. (1974) In:  XIth Annual NSF-RANN Trace
         Contaminants Conference, California, 29-31 August, 1974,
        National Technical Information Service, US Department of Commerce,
        Springfield, Virginia, p. 75.

    BRAR, S.S. ET AL. (1969)  Thermal neutron activation analysis of
         airborne particulate matter in Chicago metropolitan area,
        Washington DC, National Bureau of Standards, p. 43.

    BRUNE, D. (1969)  An. Chem. Acta., 44: 15.

    BURROWS, W. D. (1975) In: Krenkel, D. A., ed.  Heavy metals in the
         aquatic environment, Oxford, Pergamon.

    BURTON, J. D. & LEATHERLAND, T. M. (1971)  Nature (Lond.), 231: 440.

    BYRNE, A. R. & KOSTA, L. (1974)  Talanta, 21: 1083.

    CARDOZO, R. L. (1972)  Chem. Weekbl., 68: 9.

    CAVANAGH, J. B. & CHEN, F. C. K. (1971)  Acta Neuropathol., 19: 216.

    CEC, Working Group of Experts (1974)  Non-organic Micropollutants of
         the Environment. Methods of Analysis. Report. Luxembourg,
        Commission of the European Communities, VF/1966/74.

    CEMBER, H. (1962)  Am. ind. Hyg. Assoc. J., 23: 304.

    CHANG, L. W. & HARTMANN, H. A. (1972a)  Acta Neuropathol., 20: 122.

    CHANG, L. W. & HARTMANN, H. A. (1972b)  Acta Neuropathol., 20: 316.

    CHANG, L. W. & HARTMANN, H. A. (1972c)  Acta Neuropathol., 21: 179.

    CHANG, L. W. & YAMAGUCHI, S. (1974)  Environ. Res., 7: 133.

    CHANG, L. W. ET AL. (1974)  Acta Neuropathol., 27: 171.

    CHARBONNEAU, S. M. ET AL. (1974)  Toxicol. appl. Pharmacol., 27: 569.

    CHAU, Y. K. & SAITOH, H. (1970)  Environ. Sci. Technol., 4: 839.

    CHILDS, E. A. (1973)  Arch. Environ. Health, 27: 77.

    CLARKSON, T. W. (1972a)  Recent advances in toxicology of mercury with
         emphasis on the alkyl mercurials. In: Goldberg, L. ed.  Critical
         reviews in toxicology, Vol. 1, Cleveland, The Chemical Rubber
        Co., issue 2, p. 203.

    CLARKSON, T. W. (1972b)  Ann. Rev. Pharmacol., 12: 375.

    CLARKSON, T. W. ET AL. (1961)  Studies on the equilibration of mercury
         vapor with blood, Rochester, University of Rochester Atomic
        Energy Project, U.R. 582, p. 64.

    CLARKSON, T. W. & ROTHSTEIN, A. (1964)  Health Phys., 10: 1115.

    CLARKSON, T. W. & VOSTAL, J. (1973)  Mercurials, mercuric ion and
         sodium transport, In: Lant, A. F. & Wilson, G. M., ed.  Modern
         diuretic therapy in the treatment of cardiovascular disease,
        Amsterdam, Excerpta Medica, p. 229.

    CLARKSON, T. W. ET AL. (1965)  Br. J. Pharmacol., 24: 1.

    CLARKSON, T. W. ET AL. (1972)  Biol. Neonat., (Basel), 21: 239.

    CLARKSON, T. W. ET AL. (1973a)  Arch. environ. Health, 26: 173.

    CLARKSON, T. W. ET AL. (1975) In: Krenkel, D. A., ed.  Heavy metals in
         the environment, Oxford, Pergamon.

    CLEGG, D. J. (1971) In:  Proceedings of the Symposium on Mercury in
         Man's Environment, Ottawa, 15-16 February 1971, Ottawa, Royal
        Society of Canada, p. 141.

    COPPLESTONE, J. F. & MCARTHUR, D. A. (1967)  Br. J. ind. Med., 
        24: 77.

    COTTON, F. A. & WILKINSON, L. (1972)  Advanced inorganic chemistry --
         a comprehensive text, New York, Wiley Inter-Science
        Publications. Third edition. DALGAARD-MIKKELSEN, S. (1969)  Norw.
         Hyg. Tidschr., 50: 34.

    DALL'AGLIO, M. (1968)  The abundance of mercury in 300 natural water
         samples from Tuscany and Altium. In: Ahrens, L. H., ed.  Origin
         and distribution of the elements, New York. Pergamon Press,
        p. 1065.

    DAMLUJI, S. (1962)  J. Fac. Med. Baghdad, 4: 83.

    DAMLUJI, S. ET AL. (1976) In:  World Health Organization Conference on
         Intoxication due to Alkyl mercury Treated Seed, Baghdad, 9-13
        November 1974, Geneva, World Health Organization, p. 11 (Suppl.
        to  Bull. Worm Health Organ., Vol. 53).

    DANZINGER, S. J. & POSSICK, P. A. (1973)  J. occup. Med., 15: 15.

    DDR Standards (1968) TGL 22310, Berlin, German Democratic Republic.

    DIEHL, J. F. & SCHELLENZ, R. (1974) In:  Proceedings of the
         International Symposium on the problems of contamination of man
         and his environment by mercury and cadmium, Luxembourg, 3-5 July
         1973, CEC, Luxembourg, p. 401.

    D'ITRI, F. M. (1972)  The environmental mercury problem. Cleveland,
        Chemical Rubber Co., p. 69.

    DOHERTY, P. E. & DORSETT, R. S. (1971)  Anal. Chem., 43: 1887.

    DONOVAN, P. P. (1974) In:  Proceedings of the International Symposium
         on the problems of contamination of man and his environment by
         mercury and cadmium, Luxembourg, 3-5 July 1973, CEC, Luxembourg,
        p. 573.

    DURUM, W. H. ET AL. (1971)  Reconnaissance of minor elements in
         surface waters of the United States, Washington, DC, p. 49
        (US Geological Survey Circular 643).

    DUTKIEWICZ, T. & OGINSKI, M. (1967)  Int. Arch. Gewebepath.
         Gewebehygiene, 23: 197.

    EINARSSON, K. ET AL. (1974)  Acta med. scand., 195, 527.

    EL-BERGEAMI, M. M. ET AL. (1973)  Fed. Proc., 32: 886.

    ENGLESON, G. & HERNER, T. (1952)  Acta pediatr. scand., 41: 289.

    ERIKSSON, E. (1967)  Oikos, Suppl. 9, p. 13.

    FAGERSTROM, R. & JERNELOV, A. (1971)  Water Res., 5: 121.

    FERNANDEZ, M. DE N. ET AL. (1966)  Rev. de Sanid Hyg. Publica, Madrid,
        40: 1.

    FESENKO, S. (1969)  Vrac. Delo, 19: 85.

    FILBY, R. H. ET AL. (1970)  Mikrochim. Acta., 6: 1130.

    FISHMAN, M. J. (1970)  Anal. Chem., 42: 1462.

    FITZHUGH, O. G. ET AL. (1950)  Arch. ind. Hyg. occup. Med., 2: 433.

    FLEWELLING, F. J. (1975) In: Krenkel, D. A., ed.  Heavy metals in the
         aquatic environment, Oxford, Pergamon.

    FONDS, A. W. (1971)  TNO-Nieuws., 26: 375.

    FORRESTER, C. R. ET AL. (1972)  J. Fish. Res. Board Can., 29: 1487.

    FOWLER, B. A. (1972a)  Science, 175: 780.

    FOWLER, B. A. (1972b)  Am. J. Pathol., 69: 163.

    FRIBERG, L. (1951)  Norw. Hyg. Tidschr., 32: 240.

    FRIBERG, L. (1956)  Acta pharmacol. toxicol., 12: 411.

    FRIBERG, L. & VOSTAL, J. ed. (1972)  Mercury in the environment -- a
         toxicological and epidemiological appraisal, Cleveland, The
        Chemical Rubber Co.

    FRIBERG, L. & NORDBERG, G. F. (1972)  Inorganic mercury -- relation
         between exposure and effects. In: Friberg, L. & Vostal, J. ed.
         Mercury in the environment -- a toxicological and epidemiological
         appraisal, Cleveland, Chemical Rubber Co., p. 7.

    FRIBERG, L. & NORDBERG, G. F. (1973)  Inorganic mercury --
         a toxicological and epidemiological appraisal. In: Miller, M. W.
        & Clarkson, T. W., ed.  Mercury, mercurials and mercaptans,
        Springfield, C. C. Thomas, p. 5.

    FRIBERG, L. ET AL. (1961)  Acta derm. venereol., 41: 4952.

    FUJIMURA, Y. (1964)  Jpn. J. Hyg., 18: 10.

    FUJITA, M. ET AL. (1968)  Anal. Chem., 40: 2042.

    FURUKAWA, K. ET AL. (1969)  Agric. biol. Chem., 33: 128.

    GAFAFER, W. M. ed. (1966)  Occupational diseases -- a guide to their
         recognition, US Department of Health, Education and Welfare,
        Public Health Service, pp. 175-176 (Public Health Service
        Publication 1097).

    GAGE, J. C. (1961)  Br. J. ind. Med., 18: 287.

    GAGE, J. C. (1964)  Br. J. ind. Med., 21: 197.

    GAGE, J. C. (1974)  The metabolism of methoxyethyl mercury and phenyl
         mercury in the rat. In: Miller, M. W. & Clarkson, T. W., ed.
         Mercury, mercurials and mercaptans, Springfield, C. C. Thomas,
        p. 346.

    GANTHER, H. E. & SUNDE, M. L. (1974)  J. Food Sci., 39: 1.

    GANTHER, H. E. ET AL. (1972)  Science, 175: 1122.

    GAVIS, J. & FERGUSON, J. F. (1973)  Water Res., 6: 989.

    GIOVANOLI, T. & BERG, G. G. (1974)  Arch environ Health, 28: 139.

    GIOVANOLI, T. ET AL. (1974)  Clin. Chem., 30 (2): 222.

    GLOMME, J. & GUSTAVSON, K. H. (1959)  Acta med. scand., 164: 175.

    GOLDWATER, L. J. (1964)  J. R. Inst. publ. health Hyg., 27: 270.

    GOLDWATER, L. J. (1973)  Aryl and alkoxyalloyl mercurials. In:
        Miller, M. W. & Clarkson, T. W., ed.  Mercury, mercurials and
         mercaptans, Springfield, C. C. Thomas, p. 56.

    GOLDWATER, L. J. ET AL. (1963)  Arch. environ. Health, 7: 586.

    GOULD, E. S. (1962)  Inorganic reactions and structures, New York,
        Holt, Rinehart and Winston.

    GRANT, C. A. (1973)  Pathology of experimental methyl mercury
         intoxication. In: Miller, M. W. & Clarkson, T. W., ed.  Mercury,
         mercurials and mercaptans, Springfield, C. C. Thomas, p. 294.

    GREENWOOD, M. R. & CLARKSON, T. W. (1970)  Am. ind. Hyg. Assoc.,
        31: 250.

    GUTENMANN, W. H. & LISK, D. J. (1960)  J. Agric. Food Chem., 8: 306.

    HAQ, I. U. (1963)  Br. med. J., 5335: 1579.

    HARADA, Y. (1968)  Congenital (or fetal) Minamata Disease. In:
        Katsuna, M., ed.  Minamata Disease, Kumamoto University, p. 93.

    HARADA, Y. ET AL. (1971)  Clinical Study (Rinsho to Kenkyu),
        48: 1431.

    HASANEN, E. (1974) In:  Proceedings of the International Symposium on
         the problems of contamination of man and his environment by
         mercury and cadmium, Luxembourg, 3-5 July 1973, CEC, Luxembourg,
        p. 109.

    HATCH, W. R. & OTT, W. L. (1968)  Anal. Chem., 40: 2085.

    HAY, W. J. ET AL. (1963)  J. Neurol. Neurosurg. Psychiatr., 26: 199.

    HAYES, A. L. & ROTHSTEIN, A. (1962)  J. Pharmacol. exp. Ther.,
        138: 1.

    HEINDRYCKX, R. ET AL. (1974) In:  Proceedings of the International
         Symposium on the problems of contamination of man and his
         environment by mercury and cadmium, Luxembourg, 3-5 July 1973,
        CEC, Luxembourg, p. 135.

    HINKLE, M. E. & LEARNED, R. E. (1969)  U.S. Geol. Survey Prof. Paper,
        No. 650, p. 1251.

    HOLDEN, A. B. (1972)  Present levels of mercury in man and his
         environment. In:  Mercury contamination in man and his
         environment, Vienna, IAEA, p. 143.

    HOSOHARA, K. (1961)  J. chem. Soc. Jpn., 82: 1107.

    HUGHES, W. L. (1957)  Ann. N.Y. Acad. Sci., 65: 454.

    HUNTER, D. (1969)  Diseases of occupations, London, Little, Brown and
        Co., p. 306.

    HUNTER, D. ET AL. (1940)  Quart. J. Med., 9: 193.

    HUNTER, D. & RUSSELL, D. S. (1954)  J. Neurol. Neurosurg. Psychiat.,
        17: 235.

    HURSH, J. ET AL. (1976)  Arch. environ. Health (in press).

    INTERNATIONAL ATOMIC ENERGY AGENCY (1972)  Mercury contamination in
         man and his environment, Vienna, International Atomic Energy
        Agency, Technical Report Series No. 137.

    INTERNATIONAL LABOUR ORGANISATION (1970)  Permissible levels of toxic
         substances in the working environment, Geneva, International
        Labour Organisation, Occupational Safety and Health Series 20.

    JACOBS, M. B. ET AL. (1964)  Arch. environ. Health, 9: 454.

    JAKUBOWSKI, M. ET AL. (1970)  Toxicol. appl. Pharmacol., 16: 743.

    JALILI, M. A. & ABBASI, A. H. (1961)  Br. J. ind. Med., 18: 303.

    JENSEN, S. & JERNELOV, A. (1967)  Biocidinformazion, 10: 4.

    JENSEN, S. & JERNELOV, A. (1968)  Biocidinformazion, 16: 3.

    JENSEN, S. & JERNELOV, A. (1969)  Nature (Lond.), 223: 753.

    JENSEN, S. & JERNELOV, A. (1972)  Behavior of mercury in the
         environment. In:  Mercury contamination in man and his
         environment, Vienna, International Atomic Energy Agency,
        Technical Report Series No. 137, p. 43.

    JERNELV, A. (1968)  Fatten, 24: 456.

    JERNELV, A. (1973)  A new biochemical pathway for the methylation of
         mercury and some ecological considerations. In: Miller, M. W. &
        Clarkson, T. W., ed.  Mercury, mercurials and mercaptans,
        Springfield, C. C. Thomas, p. 315.

    JOENSUU, O. L. (1971)  Science, 172: 1027.

    JOHNELS, A. G. (1971)  Observed levels and their dynamics in the
         environment, results from Sweden. In:  Proceedings of the
         Symposium on Mercury in Man's Environment, Ottawa, 15-16
        February 1971, Ottawa, Royal Society of Canada, p. 66.

    JOHNELS, A. G. ET AL. (1967)  Oikos, 18: 323.

    JOHNSON, W. C. & VICKARY, C. (1970)  Analyst., 95: 357.

         mercury, lead, cadmium and the food additives amaranth,
         diethylpyrocarbamate and octyl gallate, WHO Food Additives
        Series, No. 4, FAO Nutrition Meeting Report Series No. 51a.

    JOINT FAO/WHO MEETING (1974)  The use of mercury and alternative
         compounds as seed dressings, WHO Technical Report Series,
        No. 555, 29 pp.

    JONASSON, I. R. & BOYLE, R. W. (1971) In:  Proceedings of the
         Symposium on Mercury in Man's Environment, Ottawa 15-16 February
         1971, Ottawa, Royal Society of Canada, p. 5.

    JONSSON, E. ET AL. (1972)  Var Foda, 24: 59.

    JOSELOW, M. M. & GOLDWATER, L. J. (1967)  Arch. environ. Health, 15:

    JULIUSBERG, F. (1901)  Arch. Derm. Syph., 56: 5.

    KANAZAWA, J. & SATO, R. (1959)  Jpn. Analyst. 8: 322.

    KATSUNA, M. ed. (1968)  Minamata disease, Kumamoto University, Japan.

    KAZANTZIS, G. ET AL. (1962)  Quart. J. Med., 31: 403.

    KAZANTZIS, G. ET AL. (1976) In:  World Health Organization Conference
         on Intoxication due to Alkyl mercury Treated Seed, Baghdad,
         9-13 November 1974, Geneva, World Health Organization, pp. 37,
        49 (Suppl. to  Bull. World Health Organ., Vol. 53).

    KESIC, B. & HAEUSLER, V. (1951)  Ind. Med. Surg., 20: 485.

    KHERA, K. S. (1973)  Toxicol. appl. Pharmacol., 24: 167.

    KITAGAWA, T. (1968)  Clinical investigations in Minamata Disease.
         Rehabilitation in Minamata Disease. In: Katsuna, M., ed.
         Minamata Disease, Kumamoto University, p. 127.

    KLEINERT, S. J. (1972)  Mercury levels in Wisconsin fish, In:
        Hartung, R. & Dinman, E. D, ed.  Environmental mercury
         contamination, Ann Arbor, Ann Arbor Science Publishers, p. 58.

    KLEIN, R. H. & HERMAN, S. P. (1971)  Science, 172: 872.

    KLEIN, R. ET AL. (1972)  Arch. Pathol., 93: 408.

    KLEIN, R. ET AL. (1973)  Arch. Pathol., 96: 83.

    KLOKE, A. (1974) In:  Proceedings of the International Symposium on
         the problems of contamination of man and his environment by
         mercury and cadmium, Luxembourg, 3-5 July 1973, CEC Luxembourg,
        p. 83.

    KOEMAN, J. H. ET AL. (1972)  TNO Nieuws, 27: 570.

    KOEMAN, J. H. ET AL. (1973)  Nature Lond, 245: 385.

    KOPP, J. F. ET AL. (1972)  J. Am. Water Works Assoc., p. 40.

    KORONOWSKI, P. (1973)  Nebenwirkungen yon Quecksilberverbindungen auf
         Mensch und Tier. Mitt. d. Biolog. Bundesanstalf f. Land-und
        Forstwirtschaft, No. 153, Berlin.

    KORRINGA, P. & HAGEL, P. (1974) In:  Proceedings of the International
         Symposium on the problems of contamination of man and his
         environment by mercury and cadmium, Luxembourg 3-5 July 1973,
        CEC, Luxembourg, p. 279.

    KOSTA, L. ET AL. (1975) In:  Recent advances in the assessment of the
         health effects of pollution, WHO/EPA/CEC Symposium, Paris 8-11
         June 1974, Paris, p. 245.

    KOSTYNIAK, P. J. ET AL. (1975)  J. Pharmacol. exp. Ther., 192: 260.

    KRENKEL, D. A., ed. (1975)  Heavy metals in the aquatic environment,
        Oxford, Pergamon, pp. 352.

    KUDSK, F. N. (1964)  Scand. J. clin. Lav. Invest., 16: 575.

    KUDSK, F. N. (1965a)  Acta Pharmacol., 23: 250.

    KUDSK, F. N. (1965b)  Acta Pharmacol., 23: 263.

    KUDSK, F. N. (1973)  Biological oxidation of elemental mercury, In:
        Miller, M. W. and Clarkson, T. W., ed.  Mercury, mercurials and
         mercaptans, Springfield, C. C. Thomas, p. 355.

    KURKLAND, L. T. (1973)  An appraisal of the epidemiology and
         toxicology of alkyl mercury compounds. In: Miller, M. W. &
        Clarkson, T. W., ed.  Mercury, mercurials and
         mercaptans, Springfield, C. C. Thomas, p. 23.

    LADD, A. C. ET AL. (1964)  Arch. environ. Health, 9: 43.

    LADD, A. C. ET AL. (1966)  J. occup. Med, 8: 127.

    LAMM, C. G. & RUZICKA, J. (1972)  The determination of traces of
         mercury by spectrophotometry, atomic absorption, radioisotope
         dilution and other methods. In:  Mercury contamination in man
         and his environment. Vienna, International Atomic Energy Agency,
        Technical Report Series No. 137, p. 111.

    LANDNER, L. (1971)  Nature (Lond.), 230: 452.

    LEATHERLAND, T. M. ET AL. (1971)  Nature (Lond.), 232: 112.

    LEHOTZKY, K. (1972)  Int. Arch. Arbeitsmed., 30: 193.

    LEVINE, W. G. (1970)  Heavy metal antagonists. In: Goodman, L. S. &
         Gilman, A., ed.  The pharmacological basis of therapeutics, 4th
        ed., London, Macmillan, p. 944.

    LIDMUS, V. & ULFVARSON, U. (1968)  Acta chem. scand., 22: 2150.

    LOFROTH, G. (1969)  Methyl mercury. A review of health hazards and
         side effects associated with the emission of mercury compounds
         into natural systems, Stockholm, Swedish Natural Research
        Council, Ecol. Research Bull., No. 4.

    LOVEJOY, H. B. ET AL. (1974)  J. occup. Med., 15: 590.

    LUCIER, G. ET AL. (1971)  Chem. biol. Interactions, 4: 265.

    LUCIER, G. ET AL. (1972)  Pestic. Biochem. Physiol., 2: 244.

    LUCIER, G. ET AL. (1973)  Molecular Pharmacol., 9, 237.

    LUNDGREN, K. ET AL. (1967)  Scand. J. clin. Lab. Invest., 20: 164.

    MCCARTHY, J. H. (1968)  Mining Eng., 20: 46.

    MCCARTHY, J. H. (1970)  Mercury in the atmosphere. In:  Mercury in
         the environment, Washington, DC, US Government Printing Office,
        US Geological Survey, Professional Paper No. 713, p. 37.

    MCDUFFIE, B. R. (1973)  Discussion. In: Miller, M. W. & Clarkson, T.
        W., ed.  Mercury, mercurials and mercaptans, Springfield, C. C.
        Thomas, p. 50.

    MCGILL, C. M. ET AL. (1964)  J. occup. Med., 6: 335.

    MAC COMMITTEE (1969)  Arch. environ. Health, 19: 891.

    MAGOS, L. (1967)  Environ. Res., 1: 323.

    MAGOS, L. (1968)  Br. J. ind. Med., 25: 152.

    MAGOS, L. (1971)  Analyst., 96: 847.

    MAGOS, L. & BUTLER, W. H. (1972)  Food, Cosmet. Toxicol, 10: 513.

    MAGOS, L. & CLARKSON, T. W. (1972)  J. Assoc. Offic. Anal. Chem,
        56: 966.

    MAGOS, L. ET AL. (1964)  Br. J. ind. Med., 21: 294.

    MAGOS, L. ET AL. (1973)  Toxicol. appl. Pharmacol., 26: 180.

    MAGOS, L. ET AL. (1974)  Toxicol. appl. Pharmacol., 28: 367.

    MARSH, D. O. ET AL. (1974) In:  Proceedings of the 1st International
         Congress on Mercury, Barcelona 6-10 May 1974, Barcelona, p. 235.

    MATSUMOTO, H. ET AL. (1967)  Life Sci, 6: 2321.

    MIETTINEN, J. K. (1973)  Absorption and elimination of dietary mercury
         (H-g++ ) and methyl mercury in man. In: Miller, M. W. &
        Clarkson, T. W., ed.  Mercury, mercurials and mercaptans,
        Springfield, C. C. Thomas, p. 233.

    MILLER, G. E. ET AL. (1972)  Science, 175: 1121.

    MILLER, M. W. & Clarkson, T. W. (1973)  Mercury, mercurials and
         mercaptans, Springfield, C. C. Thomas.

    MORROW, P. ET AL. (1964)  Health Phys., 10: 543.

    MUFTI, A. W. ET AL. (1976) In:  World Health Organization Conference
         on Intoxication due to Alkylmercury Treated Seed, Baghdad 9-13
         November 1974, Geneva, World Health Organization p. 23 (Suppl.
        to  Bull. World Health Organ., Vol. 53).

    NEAL, P. A. ET AL. (1937)  Study of chronic mercurialism in the
         hatters, furcutting industry. Washington, DC, USPHS, Public
        Health Bulletin 234.

    NIIGATA REPORT (1967)  Report on the cases of mercury poisoning in
         Niigata, Tokyo, Ministry of Health and Welfare.

    NIOSH (1973)  Criteria for a recommended standard. Occupational
         exposure to inorganic mercury. Washington, Public Health
        Service, US Department of Health, Education and Welfare.

    NORDBERG, G. F., ed. (1976) Effects and dose-response relationships of
        toxic metals. Proceedings from an international meeting organized
        by the Sub-committee on the Toxicology of Metals of the Permanent
        Commission and International Associations on Occupational Health,
        Tokyo, 18-23 November 1974. Amsterdam, Elsevier.

    NORDBERG, G. & SERENIUS, F. (1969)  Acta Pharmacol. Toxicol.,
        27: 269.

    NORDBERG, G. & SKERFVING, S. (1972)  Metabolism. In: Friberg, L. &
        Vostal, J., ed.  Mercury in the environment, Cleveland, Chemical
        Rubber Co., p. 29.

    NORDBERG, G. & STRANGERT, P. (1976) In:  Effects and dose-response
        relationships of toxic metals. A report from an International
        Meeting arranged by the Sub-committee on the Toxicity of Metals,
        Tokyo, 18-23 November 1974, Tokyo, p. 273.

    NORDBERG, G. ET AL. (1971)  Proceedings of the International Congress
         of Occupational Health, 16: 234.

    NORSETH, T. & CLARKSON, T. W. (1971)  Arch. environ. Health, 22: 568.

    OKINAKA, S. ET AL. (1964)  Neurology, 14: 69.

    OKUNO, I. ET AL. (1972)  J. Assoc. Offic. Anal. Chem., 55: 96.

    ORDONEZ, J. V. ET AL. (1966)  Bull. Offic. Sanit. Panamer., 60: 510.

    OSTLUND, K. (1969a)  Norw. Hyg. Tidschr., 50: 82.

    OSTLUND, K. (1969b)  Acta Pharmacol., Suppl. 1: 1.

    PACCAGNELLA, R. ET AL. (1974)In:  Proceedings of the International
         Symposium on the problems of contamination of man and his
         environment by mercury and cadmium, Luxembourg 3-5 July 1973,
        CEC, Luxembourg, p. 463.

    PARAMESHVERA, V. (1967)  Br. J. ind. Med., 24: 73.

    PARIZEK, J. & OSTADALOVA, J. (1967)  Experientia, 23: 142.

    PARIZEK, J. ET AL. (1969) In: Barltrop, D. & Burland, W. L., ed.
         Mineral metabolism in paediatrics, Oxford and Edinburgh,
        Blackwell Scientific Publications, p. 117.

    PARIZEK, J. ET AL. (1971) In: Mertz, W. & Cornatzer, W. E., ed.  Newer
         trace elements in nutrition, NY, M. Dekker Inc., pp. 85-122.

    PARIZEK, J. ET AL. (1974) In: Hoekstra, W. G., Ganther, H. E. & Mertz,
        W., ed.  Trace element metabolism in animals, University Park
        Press, p. 119.

    PASSOW, H. ET AL. (1961)  Pharmacol. Rev. 13: 185.

    PETERS, R. A. (1963)  Biochemical lesions and lethal synthesis, New
        York, Macmillan.

    PETERSON, C. L. ET AL. (1973)  Fisheries Bull., 71: 603.

    PHILLIPS, R. & CEMBER, H. (1969)  J. occup. Med., 11: 170.

    PIERCE, P. E. ET AL. (1972)  J. Am. med. Assoc., 220: 1439.

    PILLAY, K. K. ET AL. (1971)  Anal. Chem., 43: 1419.

    PIOTROWSKI, J. & BOLANOWSKA, W. (1970)  Med. Pracy., 21: 338.

    PIOTROWSKI, J. ET AL. (1969)  Med. Pracy., 20: 589.

    PIOTROWSKI, J. ET AL. (1974a)  Arch. Toxicol., 32: 351.

    PIOTROWSKI, J. ET AL. (1974b)  Toxicol. appl. Pharmacol., 27: 11.

    PIOTROWSKI, J. ET AL. (1975)  Int. Arch. Arbeits. Med. (in press).

    PISCATOR, M. & LIND, B. (1972)  Arch. environ. Health, 24: 426.

    POLESAJEV, N. (1936)  N. Gigena truda i technicka bezopastnos i,
        No. 6, p. 86.

    POTTER, S. D. & MATRONE, G. (1973)  Fed. Proc., 32: 929.

    PRICKETT, C. S. ET AL. (1950)  Proc. Soc. exp. biol. Med., 73: 585.

    RAEDER, M. G. & SNEKVIK, E. (1940)  Kgl. Norske Videnskab. Selskabs
         Forhandl., 13: 169.

    RAEDER, M. G. & SNEKVIK, E. (1948)  Kgl. Norske Videnskab. Selskabs
         Forhandl., 21: 102.

    RAHOLA, T. ET AL. (1973)  Ann. clin. Res., 5: 214.

    RAMEL, C. (1972)  Genetic effects. In: Friberg, L. & Vostal, J., ed.
         Mercury in the environment -- a toxicological and epidemiological
         appraisal, Cleveland, Chemical Rubber Co., p. 9.

    RAUHUT, A. & WILD, L. (1973)  Metallurgy, 27: 933.

    REICHERT, J. K. (1973)  Z. Gewasserschultz, Wasser, Abwasser,
        10: 277.

    RESEARCH COMMITTEE ON MINAMATA DISEASE (1975)  Pathological, clinical
         and epidemiological research about Minamata Disease, 10 years
         after, Kumamoto, Kumamoto University.

    RIVERS, J. B. ET AL. (1972)  Bull. environ. Contamin. Toxicol.,
        8: 257.

    ROCHOW, E. L. ET AL. (1957)  The chemistry of organometallic
         compounds, New York, John Wiley & Sons.

    ROTHSTEIN, A. (1973)  Mercaptans, the biological targets for
         mercurials. In: Miller, M. W. & Clarkson, T. W., ed.  Mercury,
         mercurials and mercaptans, Springfield, C. C. Thomas, p. 68.

    ROTHSTEIN, A. & HAYES, A. (1960)  J. Pharmacol. exp. Ther., 130: 166.

    SAHAGAIN, B. M. ET AL. (1966)  J. Nutr., 90, 259.

    SAHAGAIN, B. M. ET AL. (1967)  J. Nutr., 93: 291.

    SAPOTA, A. ET AL. (1974)  Med. Pracy, 25: 129.

    SCHAMBERG, J. ET AL. (1918)  J. Am. med. Assoc., 70: 142.

    SCHRAMEL, P. ET AL. (1973)  Intern. J. environ. Stud., 5: 37.

    SEIFERT, P. & NEUDERT, H. (1954)  Zbl. Arbeitsmed., 4: 129.

    SELYE, H. (1970)  Science, 168: 775.

    SERA, K. ET AL. (1962)  Kumamoto med. J., 15: 38.

    SERGEEV, YE. A. (1967)  Methodology of mercuriometric investigations.
        In: Kransnikov, V. I., ed.  Proceedings of the first all-union
         conference on geochemical methods of prospecting for ore
         deposits. Moscow.

    SHAHRISTANI, H. & SHIHAB, K. (1974)  Arch. environ. Health, 28: 342.

    SHAHRISTANI, H. ET AL. (1976) In:  World Health Organization
         Conference on Intoxication due to Alkyl mercury Treated Seed,
         Baghdad 9-13 November 1974, Geneva, World Health Organization,
        p. 105 (Suppl. to  Bull. World Health Organ., Vol. 53).

    SHAIKH, Z. A. ET AL. (1973)  Sequestration of mercury by
         cadmium-induced metallothioneine. In: Hemphill, D. D., ed.
         Trace Substances in Environmental Health, Columbia, University
        of Missouri, Vol. 7, p. 12.

    SILLEN, L. G. (1963)  Svensk. Kern. Tidskr., 75: 161.

    SJOSTRAND, B. (1964)  Anal. Chem., 36: 814.

    SKERFVING, S. (1974a)  Conference on Environmental Effects of Mercury,
         Brussels, November 1973. In: Publicaties om het Institut voor
        Weltenschappen om het haefmilieu, p. 103, Brussels, Vrije

    SKERFVING, S. (1974b)  Toxicol, 2: 3.

    SKERFVING, S. ET AL. (1974)  Environ. Res., 7: 83.

    SKOG, E. & WAHLBERG, J. E. (1964)  J. Invest. Derm., 43: 187.

    SMART, N. A. (1968)  Residue Rev., 23: 1.

    SMART, N. A. ET AL. (1969)  Analyst., 94: 143.

    SMITH, A. D. M. & MILLER, J. W: (1961)  Lancet, 1: 640.

    SMITH, A. R. & MOSKOWITCH, S. (1948)  Monthly Rev. N.Y. State Dept.
         Labor, Div. Ind. Hyg. Safety Stand., 27: 45.

    SMITH, A. R. ET AL. (1949)  Monthly Rev. N.Y. State Dept. Labor, Div.
         Ind. Hyg. Safety Stand., 28: 17.

    SMITH, R. G. ET AL. (1970)  Am. ind. Hyg. Assoc. J., 31: 687.

    SMITH, J. C. ET AL. (1971a)  Nature (Lond.), 232: 393.

    SMITH, J. C. ET AL. (1971b)  Fed. Proc., 30: 221.

    SNYDER, R. D. (1971)  New Eng. J. Med., 284: 1014.

    SOLOMON, J. & UTHE, J. F. (1971) In: Buhler, D. R., ed.  Mercury in
         the western environment, Corvallis, Continuing Education
        Publications, p. 265.

    SOMERS, E. (1971) In:  Proceedings of the Symposium on Mercury in
         Man's Environment, Ottawa 15-16 February 1971, Ottawa, Royal
        Society of Canada, p. 99.

    SOMJEN, G. G. ET AL. (1973a)  J. Pharmacol. exp. Ther., 186: 579.

    SOMJEN, G. G. ET AL. (1973b)  J. Pharmacol. exp. Ther., 187: 602.

    SPRAGUE, J. B. & CARSON, W. G. (1970)  Spot-checks of mercury residues
         in some fishes from the Canadian Atlantic coast, Ottawa,
        Fisheries Research Board, Canada, Manuscript Rep. No. 1085, p. 16.

    SPYKER, J. M. ET AL. (1972)  Science, 177: 621.

    STOCK, A. & CUCUEL, F. (1934)  Naturwissenschaften, 22: 319.

    SUMINO, K. (1968)  Kobe J. med. Sci., 14: 115.

    SUMINO, K. (1975) In: Krenkel, D. A., ed.  Heavy metals in the aquatic
         environment, Oxford, Pergamon.

    SUZUKI, T. & YOSHINO, K. (1969)  Ind. Med., 11: 21.

    SUZUKI, T. & TANAKA, A. (1971)  Ind. Med., 13: 52.

    SUZUKI, T. ET AL. (1979)  Ind. Health, 8: 39.

    SUZUKI, T. ET AL. (1971)  Bull. environ. Contam. Toxicol., 5: 502.

    SUZUKI, T. ET AL. (1973) In: Miller, M. W. & Clarkson, T. W., ed.
         Mercury, mercurials and mercaptans, Springfield, C. C. Thomas,
        p. 209.

    SWEDISH EXPERT GROUP (1971)  Norw. Hyg. Tidschr., Suppl. 4, p. 65.

    SWENSSON, A. & ULFVARSON, U. (1968)  Acta Pharmacol. Toxicol.,
        26: 273.

    SWENSSON, A. & ULFVARSON, U. (1969)  Poult. Sci., 48: 1567.

    TAKAHATA, N. ET AL. (1970)  Polio Psychiat. Neurol. Jpn, 24: 59.

    TAKEUCHI, T. (1970) In:  International Conference on Environmental
         Mercury Contamination, Ann Arbor 3-7 September, 1970, Ann Arbor,
        p. 50.

    TAKAHASHI, H. & HIRAYAMA, K. (1971)  Nature (Lond.), 232: 201.

    TASK GROUP ON LUNG DYNAMICS (1966)  Health Phys., 12: 173.

    TASK GROUP ON METAL ACCUMULATION (1973)  Environ. Physiol. Biochem.,
        3: 65.

    TATTON, J. O. (1972)  Identification of mercurial compounds. In:
         Mercury contamination in man and his environment, Vienna,
        International Atomic Energy Agency, Technical Report Series
        137, p. 131.

    TATTON, J. O. & WAGSTAFFE, P. J. (1969)  J. Chromatog., 44: 284.

    TEISINGER, J. & FIEROVA-BERGEROVA, V. (1965)  Ind. Med. Sura.,
        34: 580.

    TEINING, S. (1970)  Mercury contents in blood corpuscles and in blood
         plasma in non-fisheaters.

    Department of Occupational Medicine, University Hospital, Lund, Report
        No. 700406. TEJNING, S. & OHMAN, H. (1966) In:  Proceedings of the
         XVth International Con9ress of Occupational Health, Vienna 1966,
        Vienna, p. 239.

    THOMPSON, J. A. J. & MCCOMAS, F. T. (1973)  Environ. Lett., 5: 189.

    TONOMURA, K. & KANZAKI, E. (1969)  Blochem. Biophys. Acta, 184: 227.

    TONOMURA, K. ET AL. (1968a) J.  Ferment. Technol., 46: 685.

    TONOMURA, K. ET AL. (1968b) J.  Ferment. Technol., 46; 506.

    TORIBARA, T. Y. ET AL. (1970)  Talanta, 17: 1025.

    TRAHXENBERG, I. M. (1969)  Zdorov'e Kiev, p. 292 (in Russian, German

    TROJANOWSKA, B. (1966)  Med. Pracy., 17:  535.

    TROJANOWSKA, B. & AZENDZIKOWSKI, S. (in press)  Toxicol. appl.

    TROJANOWSKA, B. ET AL. (1971)  Toxicol. appl. Pharmacol., 18: 374.

    TSUBAKI, T. (1971) In:  Special Symposium on Mercury in Man's
         Environment, Ottawa 15-16 February 1971, Ottawa, p. 131.

    TSUDA, M. ET AL. (1963)  Yokohama med. Bull., 14: 287.

    TURNER, M. D. ET AL. (1974) In:  Proceedings of the 1st International
         Congress on Mercury, Barcelona 5-10 May 1974, Barcelona (in

    UI, J. (1967)  Norw. Hyg. Tidschr., 50, 139.

    UKITA, T. ET AL. (1963)  J. Hyg. Chem., 9: 138.

    ULFVARSON, U. (1962)  Int. Arch. Gewerbepath., 19: 412.

    UTHE, J.P. ET AL. (1970)  J. Fish. Res. Board, Canada, 27: 805.

    VOEGE, F. A. (1971) In:  Proceedings of the Symposium on Mercury in
         Man's Environment, Ottawa 15-16 February 1971, Ottawa, Royal
        Society of Canada, p. 107.

    VON BURG, R. ET AL. (1974)  J. Chromatogr., 97: 65.

    VOSTAL, J. (1972)  Transport and transformation of mercury in nature
         and possible routes of exposure. In: Friberg, L. & Vostal, J.
        ed.  Mercury in the environment, Cleveland, Chemical Rubber Co.,
        p. 3.

    VOSTAL, J. (1972) In:  Proceedings of the XVIIth International
         Congress on Occupational Health, Buenos Aires, October 1972,
        Buenos Aires.

    VOUK, V. B. ET AL. (1950)  Hr. J. Med., 7: 168.

    WALLACE, R. A. ET AL. (1971)  Mercury in the environment, the human
         element, Oak Ridge, Oak Ridge National Laboratory, ORNL

    WATANABE, S. (1971) In:  Proceedings of the YVIth International
        Congress on Occupational Health, Tokyo, 22-27 September 1969,
        Tokyo, p. 553.

    WAHLBERG, J. E. (1965)Arch.  environ. Health, 11:201.

    WEAVER, J. S. (1973)  Anal. Chem., 45: 1950.

    WEBB, J. L. (1966)  Enzyme and metabolic inhibitors, New York,
        Academic Press, Vol. 2. WEISS, H. V. ET AL. (1971)  Science,
        174: 692.

    WERSHAW, R. L. (1970)  Sources and behavior of mercury in surface
         waters. In: US Geological Survey,  Mercury in the environment,
        Washington, DC, US Government Printing Office, Professional paper
        No. 713.

    WEST, I. & LIM, J. (1968)  J. occup. Med., 10: 697.

    WESTERMARK, T. & LJUNGGREN, K. (1972)  The determination of mercury
         and its compounds by destructive neutron activation analysis.
        In:  Mercury contamination in man and his environment, Vienna,
        IAEA, p. 99.

    WESTOO, G. (1966)  Acta chem. scand., 20: 2131.

    WESTOO, G. (1967)  Acta chem. scand., 21: 1790.

    WESTOO, G. (1968)  Acta chem. scand., 22: 2277.

    WESTOO, G. (1973)  Heavy metals in the aquatic environment, Nashville
         4-7 December 1973, Nashville (in press).

    WHITE, M. N. & LISK, D. J. (1970)  J. Assoc. Offic. Anal. Chem.,
        53: 530.

    WHO REGIONAL OFFICE FOR EUROPE (1973)  The hazards to health of
         persistent substances in water. Copenhagen, p. 147.

    WILLFORD, W. A. ET AL. (1973)  J. Assoc. Offic. Anal. Chem.,
        56: 1006.

    WILLISTON, S. H. (1968)  J. geophys. Res., 73: 7051.

    WILLISTON, S. H. (1971)  J. geophys. Res., 28: 1285.

    WINTER, D. ET AL. (1968)  Arch. environ. Health, 16: 626.

    WINTER, J. A. & CLEMENTS, H. A. (1972)  Analyses of mercury in water
         -- a preliminary study of methods. Cincinnati, US Environmental
        Protection Agency, EPA-R4-72-303, p. 56.

    WISNIEWSKA, J. M. ET AL. (1970)  Toxicol. appl. Pharmacol., 16: 754.

    WISNIEWSKA, J. M. ET AL. (1972)  Acta Biochim. Pol., 19:11.

    WOLF, D. ET AL. (1974)  Quecksilbermessunoen am Arbeitsplatz bei der
         Herstellung von Leuchtstoffrohren. STF report 1-74
        Staubforschungsinstitut der gewerblichen Berufsgenossenschaften,
        Bonn 1974.

    WOOD, K. ET AL. (1968)  Nature (Lond.), 220: 173.

    WORLD HEALTH ORGANIZATION (1971)  International standards for drinking
         water, 3rd ed., Geneva

    YOSHINO, Y. ET AL. (1966)  J. Neurochem., 13: 1223.

    See Also:
       Toxicological Abbreviations
       Mercury (ICSC)
       Mercury (WHO Food Additives Series 4)
       Mercury (WHO Food Additives Series 13)
       MERCURY (JECFA Evaluation)
       Mercury (UKPID)