Mercury in the Environment: Sources and Health Risks

Mercury in the Environment:
Sources and Health Risks
Updated January 19, 2006
Linda-Jo Schierow
Specialist in Environmental Policy
Resources, Science, and Industry Division

Mercury in the Environment: Sources and Health Risks
Concern about mercury in the environment has increased in recent years due to
emerging evidence that exposure to low levels of mercury may harm the developing
nervous systems of unborn children. At least five bills in the 109th Congress aim to
reduce mercury emissions from coal-fired electric utilities. The various proposals
and a final regulation promulgated by the U.S. Environmental Protection Agency
(EPA) on March 15, 2005, differ in how much and how soon emission reduction
would be required, and in whether reductions would be achieved through controls at
each plant or through a nationwide cap and trade system. The latter approach could
allow individual plants to continue emitting current levels of mercury, potentially
worsening conditions at nearby “hot spots.” Analysis of competing proposals raises
questions about the sources, fate, and toxicity of mercury in the environment. This
CRS report provides background information about mercury and summarizes recent
scientific findings. For information about regulatory proposals to reduce
environmental emissions of mercury, see CRS Report RL32868, Mercury Emissions
from Electric Power Plants: An Analysis of EPA’s Cap-and-Trade Regulations.
Mercury is a natural element found in rocks, soil, water, air, plants, and animals,
in a variety of chemical forms. Natural forces move mercury through the
environment, from air to soil to water, and back again. Industrial activities have
increased the portion of mercury in the atmosphere and oceans, and have
contaminated some local environments. Coal-fired electric utilities are the largest
single source of U.S. mercury emissions, according to EPA, but mobile sources also
are important. The chemical form of mercury generally determines how it moves
through the environment, but mercury can and does change form relatively rapidly
where bromine and other oxidizing substances (e.g., ozone) are abundant. In soil or
sediments of lakes, streams, and probably oceans (especially where water is oxygen-
poor and acidic, and sulfate is present), bacteria convert inorganic mercury to more
toxic methylmercury, which can accumulate in fish. Newly deposited mercury
seems to be more readily converted than older deposits.
People and wildlife who eat contaminated fish can be exposed to toxic levels
of methylmercury. In people, methylmercury enters the brain, where it may cause
structural damage. Methylmercury also crosses the placenta. The National Research
Council has reported that the human fetus is sensitive to methylmercury exposure,
and the current risk to U.S. women who eat large amounts of fish and seafood during
pregnancy is “likely to be sufficient to result in an increase in the number of children
who have to struggle to keep up in school.” Some studies indicate that the
cardiovascular system may be even more sensitive. Mercury concentrations generally
are low, but the estimated safe blood-mercury level is exceeded in about 6% of U.S.
women between the ages of 16 and 49 years. EPA and the Food and Drug
Administration advise women of child-bearing age to avoid certain large fish, and to
limit the amount eaten of other fish. In making choices about fish consumption, the
health benefits of eating fish also should be considered. Fish-eating wildlife also are
exposed to methylmercury, but it is not clear whether typical current levels of
environmental contamination are harmful. This report will be updated as warranted
by significant scientific discoveries.

In troduction ..................................................1
Sources of Mercury in the Environment............................2
Fate of Mercury Released to the Environment.......................6
Transport, Deposition, Re-emission, and Transformation...........6
Methylmercury Formation and Accumulation....................9
Risks of Methylmercury Poisoning...............................12
Toxicity of Methylmercury.................................12
Environmental Methylmercury Exposure......................15
Recommended Exposure Limits.............................18
U.S. Fish Consumption, Methylmercury Exposure,
and Health Risk......................................21
Wildlife Exposure and Health Effects.............................24
Conclusion ..................................................26
List of Figures
Figure 1. Emission and Deposition of Pollutants.........................2
List of Tables
Table 1. Estimates of U.S. Mercury Emissions from Major Sources..........5
Table 2. Geometric Mean and Selected Percentiles of Total Blood Mercury
Concentrations (ppb) for U.S. Children Aged 1-5 Years and Women
Aged 16-49 Years............................................16
Table 3. Mercury Concentrations in Some Popular Fish (ppm).............17
Table 4. Federal Upper Limits for Methylmercury Exposure...............21
Table 5. Recommended Number of Meals per Month of Fish Containing
Various Methylmercury Concentrations, Based on EPA RfD...........22
Table 6. Various Estimates of Fish Consumption and Mercury Exposure,
Assuming a Concentration in Fish of 0.3 ppm......................24
Table 7. Relative Fatty Acid Content and Mercury Concentration
in Some Popular Fish..........................................25

Mercury in the Environment:
Sources and Health Risks
Congressional concern about mercury in the environment has greatly increased
in recent years due to emerging scientific evidence that exposure to low levels of
mercury may harm the developing nervous systems of young children. At higher
levels of exposure, mercury is known to be a potent neurotoxin. People in the United
States are exposed to mercury primarily by eating large, predatory fish. Risks of
health problems for people who consume mercury in fish have caused wide public
concern and prompted the U.S. Environmental Protection Agency (EPA) and the
Food and Drug Administration (FDA) to issue consumer alerts, warning women of
child-bearing age and young children to avoid certain fish altogether and to limit the
number of meals for other fish.
Numerous legislative proposals in the 109th Congress aim to reduce levels of
mercury in the environment — in consumer products, in solid waste, in utility and
other emission sources, and in surface water. Most of these proposals focus on
sources of mercury emissions to air, because atmospheric mercury deposition
accounts for most of the mercury in U.S. freshwater lakes and streams. At least five
proposals target emissions from coal-fired electric utilities, because they are thought
to be the last remaining major uncontrolled source of mercury emissions. These
various proposals and a final regulation promulgated by the U.S. Environmental
Protection Agency (EPA) on March 15, 2005, differ in how much and how soon
emission reduction would be required, as well as in the extent to which reductions
would be distributed geographically across the United States.
Analysis of the competing policy proposals for reducing mercury emissions
raises questions about the urgency of a need for emission controls, the likelihood that
they will reduce mercury contamination of fish, and the possibility that overall
reductions might be achieved at the expense of local “hot spots” of mercury
contamination. To answer such questions requires an understanding of the sources,
fate, and toxicity of mercury in the environment — an understanding that is growing
quickly as the results of numerous scientific studies are being reported. This CRS
report provides background information about mercury, and summarizes recent
scientific findings. It discusses the sources (i.e., natural versus industrial, historic
versus modern) and chemical forms of mercury in the environment; how mercury
moves through the environment and concentrates in fish (i.e., the fate of mercury);
and the risks to human health and wildlife of mercury exposure through fish
consumption. Each of these major sections of the report aims to summarize scientific
evidence relevant to specific arguments and questions that have emerged in the policy
context. For example, the section on mercury in the environment addresses the
question “Are utility emissions deposited locally or regionally, or do they rise to
merge with the global atmospheric mercury pool?” For information about specific

regulatory proposals to reduce environmental mercury, see CRS Report RL32868,
Mercury Emissions from Electric Power Plants: An Analysis of EPA’s Cap-and-
Trade Regulations; CRS Issue Brief IB10137, Clean Air Act Issues in the 109th
Congress, both by James E. McCarthy; or CRS Report RL31908, Mercury in
Products and Waste: Legislative and Regulatory Activities to Control Mercury, by
Linda G. Luther.
Sources of Mercury in the Environment
Mercury is a natural element, a silver-colored, shiny, liquid metal that is found
in a variety of chemical forms in rocks, soil, water, air, plants, and animals.
Sometimes mercury occurs in its elemental, relatively pure form, as a liquid or vapor,
but more commonly mercury is found combined with other elements in various
compounds, which may be inorganic (e.g., the mineral cinnabar, a combination of
mercury and sulfur) or organic (e.g., methylmercury).1
Figure 1. Emission and Deposition of Pollutants
Source: EPA, Frequently asked questions about atmospheric deposition: A handbook for watershed
managers, EPA-453/R-01-009, Sept. 2001, p. 4, at [
Natural forces move mercury through the environment, from air to soil to water,
and back again. Volcanoes and deep sea vents release tons of mercury to the
atmosphere and oceans. Mercury in the air falls to earth with dust, rain, and snow.
Mercury evaporates from the oceans, leaves of plants, and other surfaces back into2
the air. Depending on geologic and meteorologic conditions, the relative amounts

1 Organic compounds consist of carbon combined with other substances. Organic
compounds, such as methylmercury, are created by, and generally are more readily absorbed
by, living things.
2 Depending on the mercury compound, either a liquid or solid may turn into gas. In the

of mercury in the atmosphere, surface water, or soil may vary from one year, decade,
century, or millennium to another.
During the past 500 years or so, human activities have released mercury from
its relatively stable and water-insoluble form (cinnabar) in rocks and soil through
mining, fossil fuel combustion, and other activities, and so have increased the portion
of mercury that is actively cycling through the atmosphere, surface waters, plants,
and animals as it changes chemical and physical form. Released mercury may enter
the air, persist in the atmosphere and travel great distances or be deposited locally,
dissolve in water droplets, settle back onto the land or water, re-enter the air (i.e., be
re-emitted), be buried in lake or ocean sediments, or be taken into plants and animals.
The generally accepted estimate is that roughly three to five times as much mercury
is mobilized today as was mobile before industrialization.3 However, the author of
one recent study argues that the mercury deposited from the atmosphere today is at
least 10 times the amount of mercury that was being deposited 500 years ago.4
In 1995, about 1,913 metric tons (roughly 2,104 U.S. tons)5 of mercury were
newly emitted globally as a result of stationary combustion, metal production, cement
production, and waste disposal.6 Roughly another 514 metric tons (565 U.S. tons)
were emitted from other human sources, including chlor-alkali plants, gold
production, and mercury uses.7 Thus, 2,427 metric tons (2,670 U.S. tons) of mercury
were released due to human activities in 1995, according to recent estimates.8 These
and other mercury emissions from human activities (past and present) account for at
least 50% and perhaps as much as 75% of current, annual, global mercury emissions
from all sources (including natural sources), but a large, unknown portion of those

2 (...continued)
latter case, the correct term is sublime, rather than evaporate.
3 Tom Atkeson and Don Axelrad, 2004 Everglades Consolidated Report (2003), Chapter 2B,
“Mercury Monitoring, Research and Environmental Assessment,” p. 2B-7; C. H. Lamborg,
H. Balcom, D. R. Engstrom, et al., “Modern and historic atmospheric mercury fluxes in both
hemispheres: Global and regional mercury cycling implications,” Global Biogeochemical
Cycles, v. 16, n. 4 (2002), pp. 51-1 to 51-11.
4 R. Bindler, “Estimating the natural background atmospheric deposition rate of mercury
utilizing ombrotrophic bogs in southern Sweden,” Environmental Science & Technology,
v. 37, no. 1 (2003), pp. 40-46.
5 A metric ton is 1,000 kilograms (one million grams), or about 2,200 pounds.
6 E. G. Pacyna and J. M. Pacyna, “Global emission of mercury from anthropogenic sources
in 1995,” Water, Air, and Soil Pollution, v. 137 (2000), pp. 149-165.
7 J. M. Pacyna, E. G. Pacyna, F. Steenhuisen, et al., “Mapping 1995 global anthropogenic
emissions of mercury,” Atmospheric Environment, v. 37, supp. no. 1 (2003), pp. S109-S117.
8 Corrected emission data for 1995 and new data for 2000 are available on the internet, but
have not yet been published. These indicate that global mercury emissions for 1995 totaled
2,317 metric tons (2,549 US tons). For 2000, total global emissions were estimated to be

2,188 metric tons (2,407 US tons). See Pacyna, E., J. Pacyna, F. Steenhuisen, and S.

Wilson, “Global Anthropogenic Mercury Emission Inventory for 2000,” Atmospheric
Environment (in prep. 2005), at [
html], visited Jan. 19, 2006.

mercury emissions is due to past rather than current human activities, according to
EPA estimates.9 The most recent estimates of global, natural mercury emissions
range between roughly 1,600 and 3,200 metric tons (1,960 and 3,520 U.S. tons) per
year. 10
People have released mercury to the environment primarily through mining and
smelting of minerals, burning of fossil fuels (e.g., coal, oil, and diesel fuel), use and
disposal of mercury, certain industrial processes (e.g., chlorine production and
cement production), and burning of municipal and medical wastes. In some parts of
the world such activities are increasing, but in the United States, annual mercury
emissions are decreasing. Most of the largest and most direct sources of U.S.
mercury releases to water and air have been eliminated. Among the remaining U.S.
industrial sources, coal-fired electric utilities are the most important, accounting for
about 40% of current U.S. mercury releases.11
Three estimates of U.S. national emissions are presented in Table 1. The first12
two estimates were made by EPA for the National Emissions Inventory. CRS
added 12 tons of emissions from gold mines to the EPA emission inventory that was13
conducted for 1995, at the suggestion of EPA. EPA was unaware of the emissions
from that source at the time the inventory was conducted. The “other” category
encompasses emissions from various unidentified industries, including most iron and
steel mills. EPA advised CRS to note that there are some sources not accounted for
in the 1999 EPA inventory, such as iron and steel production using mercury-
contaminated scrap, which probably accounts for 7-10 tons of emissions per year.14
These emissions are not included in the “other” category. EPA also does not include
mobile source emissions in its inventory, although these might be significant, because
the agency is still developing an estimate.

9 EPA, Mercury Study Report to Congress, vol. 1, Executive Summary, EPA-452/R-97-003,
(Washington: GPO, 1997), pp. 3-4.
10 C. Seigneur, K. Vijayaraghavan, K. Lohman, et al., “Global source attribution for
mercury deposition in the United States,” Environmental Science & Technology, v. 38, n.

2 (2004), pp. 555-569.

11 However, emissions from two additional sources are not well quantified and may be larger
contributors: mobile sources and chlor-alkali plants. In a recent regulatory action, EPA
stated that 65 tons of mercury were consumed by nine chlor-alkali plants but were not
reported to have been released from chlorine production plants in the year 2000. That
amount of mercury is greater than the amount released by all coal-fired utilities annually,
and is equivalent to 124 gallons of mercury per plant. Although industry personnel claim
that a large proportion of the “consumed” mercury condenses and accumulates in pipes,
tanks, and other equipment, EPA considers the discrepancy between mercury purchased,
consumed, and released to be unexplained (68 Federal Register 70920, Dec. 19, 2003).
12 EPA, National Emissions Inventory, at [
html#final3haps], visited Jan. 19, 2006.
13 Alexis Cain, personal communication, March 12, 2004.
14 Ibid.

Since the time that EPA completed its 1999 inventory, the medical waste
incinerator rules promulgated under the Clean Air Act have been fully implemented,
which may have further reduced emissions from that source, and gold mining
emissions have decreased due to a voluntary project. Chlorine production emissions
also may have declined since the 1999 inventory, because some facilities closed, but
one additional facility was identified and included in emission estimates by Seigneur
et al.,15 which appear in the third column. These latter estimates were calculated by
researchers with Atmospheric & Environmental Research, Inc., and published in
2004, but represent emissions in the year 1998. It is not clear why the Seigneur
estimates for 1998 emissions from waste incineration are so much larger than EPA
estimates for emissions from that category in 1999. Seigneur included emissions
from landfills and electric arc furnaces in the “other” category. The Electric Power
Research Institute (EPRI) provided the estimates used in that article for utility
emissions. Both the EPRI calculations and the EPA estimate for 1999 utility
emissions were based on measurements of mercury content in coal and stack
emissions that were collected for the year 1999, in response to an information
collection request issued by EPA.16
Table 1. Estimates of U.S. Mercury Emissions
from Major Sources
(U.S. tons per year)
SourceEPA 1995EPA 1999Seigneur et al. 1998
Electric Power514846
Industrial Boilers121214
Gold Mining12127
Waste Incineration/771532
Chl o r i ne 17 87 7
Mobile Sources18 — — 27
Source: EPA, National Emissions Inventory, at [
html#final3haps], visited April 9, 2004.

15 Seigneur et al., 2004.
16 Recommendations for the Utility Air Toxics MACT Final Working Group Report, Oct.
2002, at [], visited
Jan. 19, 2006.
17 See footnote 9, above.
18 Ibid.

Fate of Mercury Released to the Environment
Transport, Deposition, Re-emission, and Transformation. Chemical
form generally determines the ease with which mercury moves through the air, water,
and soil and over distances. For example, elemental mercury emissions may remain
airborne for more than a year, traveling around the world as part of the so-called
“global pool” of atmospheric mercury. About 95% of atmospheric mercury is
elemental. Particulate and reactive gaseous mercury (both organic and inorganic) are
found in the atmosphere in smaller amounts, because they travel shorter distances
from the point of emission and are more quickly deposited. Reactive gaseous
mercury typically is deposited within about 100 kilometers of the point of emission.19
Coal-fired electric utility emissions vary depending on the technology and coal used
at each plant, but are roughly 50% elemental mercury, according to EPA.20
However, the chemical form of mercury emissions can and does change in the
atmosphere, making it difficult to predict the fate of particular emissions, including
utility emissions. Elemental mercury emitted to the atmosphere can attach to
particles or change to a water-soluble form (i.e., a reactive gas) that more easily
combines with other chemicals and deposits relatively quickly. Reactive, gaseous
mercury is more likely to form (and to be deposited) in the presence of sunlight. This
explains why measured concentrations of atmospheric mercury generally are lower
during the day than they are at night.
Mercury deposition in North America increases in spring and peaks in summer,
according to data from the mercury deposition network.21 Higher summer deposition
probably results, at least in part, from the increase in solar energy that is available to
spark key chemical reactions (i.e., oxidation). For example, scientists have shown
that in the lower layers of the atmosphere (i.e., roughly 400 meters of land or 1,000
meters of the ocean surface), elemental mercury gas may be quickly oxidized by
bromine, chlorine, ozone, or hydroxide in the presence of sunlight, leading to local
“mercury depletion events.”22 In such cases, concentrations of elemental gaseous

19 Gwendolyn Judson (undated), “Analysis of mercury speciation profiles currently used for
atmospheric chemistry modeling,” Wisconsin Department of Natural Resources, Madison,
WI, at [],
visited Jan. 19, 2006.
20 69 Federal Register 4674, Jan. 30, 2004.
21 Environment Canada, Ecological Monitoring and Assessment Network, Meeting the
Challenges of Continental Pollutant Pathways, Mercury Case Study, at
[], visited
Jan. 19, 2006.
22 P. A. Ariya, A. Ghalizov, and A. Gidas, “Reactions of gaseous mercury with atomic and
molecular halogens: Kinetics, products studies and atmospheric implications,” Journal of
Physical Chemistry, v. 106 (2002), pp. 7310-7320; B. Pal and A. P. Ariya, “Studies of ozone
initiated reactions of gaseous mercury: Kinetics, product studies, and atmospheric
implications,” Physical Chemistry and Chemical Physics, v. 6, n. 3 (2004), pp. 572-579; I.
M. Hedgecock and N. Pirrone, “Chasing quicksilver: Modeling the atmospheric lifetime of
HgO(g) in the marine boundary layer at various latitudes,” Environmental Science &

mercury in the atmosphere decrease rapidly as the oxidized forms of mercury are
deposited to the surface in dry deposits (i.e., without the help of rain or snow).23 This
has been observed during the summer in the Arctic and Antarctic regions, and over
the oceans.
Summer mercury deposition also might be a result of increased oxidation by
ozone. Higher ozone concentrations occur in summer, also due to the action of
Mercury that is deposited onto plants or soil can be re-emitted to air, attached
to soil, dissolved, washed away, buried, or ingested. It may again change chemical
form. Mercury often attaches to soil particles, especially humus. Recent research
indicates that soil may be a repository for the largest portion of mercury emitted in
the past.25
Mercury may be delivered to surface water bodies by air, in soil, or in streams
and rivers. For many isolated lakes, very large lakes, and the oceans, atmospheric
deposition (wet and dry) accounts for the largest portion of mercury contamination.26
Mercury deposited or delivered to surface water may be re-emitted to air, remain
suspended or dissolved in the water column, be deposited in sediments, or absorbed
or ingested by living things. Re-emission rates from the ocean surface to air may be
very large.27 For example, some experts believe that as much as 90% of the mercury
deposited to the ocean surface might be re-emitted.28 Nevertheless, the concentration
of mercury in the mixing layer of the deep oceans probably is increasing by a few
percent per year.29

22 (...continued)
Technology, v. 38, n. 1 (2004), pp. 69-76.
23 United Nations Environment Programme, Inter-Organization Programme for the Sound
Management of Chemicals, Global Mercury Assessment (2003), p. 28, at [http://www.chem.], visited Jan. 19, 2006.
24 P. Weiss-Penzias, D. A. Jaffe, A. McClintick, et al., “Gaseous elemental mercury in the
marine boundary layer: Evidence for rapid removal in anthropogenic pollution,”
Environmental Science & Technology, v. 37, n. 17 (2003), pp. 3755-3763.
25 R. P. Mason and G. R. Sheu, “Role of the ocean in the global mercury cycle,” Global
Biogeochemical Cycles, v. 16, n. 4 (2002), pp. 40-1 to 40-14; James G. Wiener, et al.,
“Ecotoxicology of Mercury,” in David J. Hoffman, et al, Handbook of Ecotoxicology, 2nd
ed. (Boca Raton, FL: Lewis Publishers, 2003), p. 418.
26 Mason and Sheu.
27 M. S. Landis and G. J. Keeler, “Atmospheric mercury deposition to Lake Michigan during
the Lake Michigan mass balance study,” Environmental Science & Technology, v. 36, n. 21
(2002), pp. 4518-4524.
28 Robert Mason, personal communication, April 1, 2004.
29 Mason and Sheu.

Mercury in the air eventually will fall back to land or surface water. A recent
analysis of deposition data collected for both hemispheres indicates that total gaseous
mercury increased in the late 1970s, peaked in the late 1980s, decreased somewhat
until the mid-1990s, and has remained constant since then.30 At present,
approximately 5,000 metric tons (5,500 U.S. tons) of mercury are deposited globally
each year.31
Layered samples (known as cores) of glaciers and peat provide historical records
of mercury deposits that clearly show the contemporary impact on land of major
trends in mercury emissions. That is, cores record the historical rise in mercury
emissions and deposition due to mining and industrialization. However, while such
records inform us about relative changes in global, regional, and local emissions over
a scale of years, even centuries, they provide little information about the precise
relationship between particular emissions and particular deposits. This is because the
path and time taken by emitted mercury to cycle through environmental media
depends on its chemical form, as well as on physical conditions like height of
emission, temperature, sunlight, wind speed and direction, humidity, and the presence
of certain other substances, such as ozone.
Atmospheric deposition tends to be greater in areas closer to emission sources
and in locations with more rainfall. Thus, EPA has estimated that about 60% of
mercury deposited in the United States is from local or regional U.S. sources, and
deposition increases from west to east.32 Local or even regional deposition can result
in areas of relatively high deposition, or “hot spots.” Deposition of mercury in
particular cases varies, however, depending on many factors, including regional and
local climate and weather patterns, soil types, topography, vegetation, and local or
regional sources of mercury emissions.33 Thus, mercury may be deposited near to or
far from an emission source.34
The relative contribution of various sources to mercury deposition also can
change over time. For example, the record of mercury deposition in ice cores from
Fremont Glacier, Wyoming, shows peaks of high mercury deposition following
volcanic eruptions in the northern and southern hemispheres, as well as during the
California Gold Rush.35 Such cores are difficult to interpret, however, because they
reflect local as well as global influences.

30 R. Slemr, E. G. Brunke, R. Ebinghaus, et al., “Worldwide trend of atmospheric mercury
since 1977,” Geophysical Research Letters, v. 30, n. 10 (2003), p. 1516.
31 Ibid.
32 EPA (undated) Draft Report, Mercury Sources and Regulations, 1999 Update, “The
Binational Toxics Strategy — Canada and United States,” p. 4, at [
glnpo/bns/mercury/stephg.html], visited Jan. 19, 2006.
33 Ibid.
34 Everglades Consolidated Report, p. 2B-7; Seigneur et al.
35 U.S. House of Representatives, Committee on Science, Subcommittee on Environment,
Technology, and Standards, Mercury Emissions: State of the Science and Technology,
hearing, Nov. 5, 2003, statement of David P. Krabbenhoft, at [
science/hearings/ets03/nov05/krabben.htm], visited Jan. 19, 2006.

Only a few ecosystems have been studied in sufficient detail to determine the
sources of mercury contamination. However, additional information about emission
sources and deposition is being gathered through monitoring and modeling across the
continental United States, particularly as states undertake detailed analyses of steps
needed to restore the quality of waters that are impaired by mercury. According to
EPA, more than 700 bodies of water throughout the United States are listed as
impaired by mercury; in most cases, the source of the mercury contamination is air
deposition. To address these impairments, states are developing Total Maximum
Daily Loads (TMDLs), which are plans to bring those waters into attainment with
water quality standards. The Florida Everglades and Devil’s Lake in Wisconsin were
selected as pilot TMDL projects for mercury.
Scientists studying the Florida Everglades have estimated that at least half of the
mercury deposited in the Everglades is emitted locally, while between 5% and 29%
is emitted regionally (from within the southeastern United States). The remainder
derives from sources outside the United States.36 EPA has estimated that 80% of
deposition to Pines Lakes, New Jersey, comes from U.S. sources.37 In contrast,
almost all of the mercury found in remote regions of the Arctic is believed to have
traveled from distant sources.38
Methylmercury Formation and Accumulation. The most biologically
significant transformation of mercury occurs in soil or sediments of lakes or streams,
where bacteria (primarily sulfate-reducing bacteria) are capable of converting
inorganic mercury to methylmercury.39 The significance of methylation is that
relative to inorganic mercury, methylmercury is more easily absorbed by living
tissues, more likely to be ingested in food, and much more toxic to animals.
Methylmercury is easily absorbed by the digestive tract and accumulates in the bodies
of fish and other animals, when it is ingested faster than it can be excreted. Because
methylmercury tends to be stored in muscle tissue (i.e., the edible meat of fish and
other animals), animals higher on the food chain tend to have higher levels of
exposure. Predatory fish (e.g., walleye, large-mouthed bass, or tuna), fish-eating
birds (e.g., loons, ospreys, or eagles), and fish-eating mammals (e.g., raccoons, otters,
or mink) which top the longest food chains accumulate the greatest concentrations
of methylmercury. In the Florida Everglades, methylmercury concentrations in fish40
are up to ten million times greater than concentrations of mercury in water.
Inorganic mercury is not easily transferred through the food chain and does not
concentrate to higher levels with each nutritional link.

36 Everglades Consolidated Report, p. 2B-11.
37 Seigneur et al.
38 Krabbenhoft statement.
39 There are several processes by which mercury can become methylated and demethylated.
The role of sulfate-reducing bacteria generally is thought to be the most important, however.
J. G. Wiener, D. P. Krabbenhoft, G. H. Heinz, et al., “Ecotoxicology of mercury,” in D. J.
Hoffman, B. A. Rattner, G. A. Burton, Jr., et al. (eds.), Handbook of Ecotoxicology, 2nd ed.
(Boca Raton, FL: Lewis Publishers, 2003), pp. 420-421.
40 Everglades Consolidated Report, p. 2B-16.

Generally, the more mercury that is added to an ecosystem, through direct
discharge to water, runoff from the surrounding watershed, or deposition from air,
the more mercury that will be found in fish.41 However, the rate of methylmercury
formation and accumulation is highly variable, even within relatively small
geographic areas, because it depends on many factors, in addition to the abundance
of inorganic mercury. Recent research indicates that some ecosystems are
particularly sensitive to relatively small mercury inputs, and are more likely to
experience high rates of methylmercury production and accumulation. Sensitive
ecosystems include low-alkalinity (i.e., low capacity for neutralizing acid) and humic
lakes and streams (which are characterized by an abundance of dissolved,
decomposed, plant or bacterial matter), wetlands, surface waters connected to
wetlands, and waters linked to areas subjected to flooding.42 Methylmercury
formation by sulfate-reducing bacteria and bioaccumulation is favored in ecosystems:
!that are oxygen-poor and acidic;
!that contain sulfate (the most common form of sulfur in surface
waters), but not too much sulfide (the form of sulfur rendered by
sulfate-reducing bacteria;43 and
!in which mercury is recently deposited, rather than older mercury.44
In a Wisconsin lake, researchers found that levels of both sulfate and mercury
determined levels of production and bioaccumulation of methylmercury, and that
“modest changes in acid rain or mercury deposition can significantly affect mercury
bioaccumulation over short-time scales.”45 In response to a significant decrease in
mercury deposition between 1994 and 2000, methylmercury in yellow perch
decreased by roughly 30% (5% per year).
The link between industrial emissions and mercury levels in the oceans is less
clear, because the role of the oceans in mercury cycling is poorly understood. On the
one hand, significant quantities of reactive inorganic mercury are deposited in the
oceans, and methylmercury is found in marine fish and their predators, sometimes at
very high concentrations. And, although methylmercury levels are very low in the
surface layer of the open oceans, concentrations are greater, perhaps as much as

41 Ibid., p. 2B-2.
42 Wiener et al., p. 440.
43 Everglades Consolidated Report, pp. 2B-11-12, 16-18; J. M. Benoit, C. C. Gilmour, and
R. P. Mason, “Sulfide controls on mercury speciation and bioavailability to methylating
bacteria in sediment pore waters,” Environmental Science & Technology, v. 33, n. 6 (1999),
pp. 951-957.
44 Everglades Consolidated Report, appendix 2B-3, p. 2; H. Hintelmann, R. Harris, A.
Heyes, et al., “Reactivity and mobility of new and old mercury deposition in a boreal forest
ecosystem during the first year of the METAALICUS study, Environmental Science &
Technology, v. 36, n. 23 (2002), pp. 5034-5040.
45 T. R. Hrabik and C. J. Watras, “Recent declines in mercury concentration in a freshwater
fishery: Isolating the effects of de-acidification and decreased atmospheric mercury
deposition in Little Rock Lake,” The Science of the Total Environment, v. 297 (2002), pp.


three-fold higher than they were prior to industrialization (assuming that insignificant
amounts descended to the ocean depths).46 So we know that organic (methyl)
mercury is formed in the oceans. What we do not know is where the mercury in
ocean fish originated — in industrial emissions deposited to the oceans or in the
natural reservoir of the ocean depths — nor where it was transformed into
Some scientists believe that methylmercury probably is formed in the deep
sediments of oceans or in the areas surrounding deep thermal vents in the ocean
floor.47 In that case, they argue, deposition of atmospheric mercury cannot account
for current methylmercury levels in ocean fish, given the relatively large size of the
deep sea reservoir of mercury and the time it takes for the ocean depths to mix with
the surface layers where fish feed, an estimated 400 years. If all the mercury
deposited into the oceans due to human activities over the past hundred years were
mixed into the ocean even to its greatest depths, the mercury concentration of ocean
water would have increased only an estimated 1% to 10% over pre-industrial
Other scientists believe that sulfate-reducing bacteria form methylmercury in
coastal sediments where it is taken up by tiny plants and animals at the bottom of
aquatic food webs. Small fish and other animals feeding in near-shore waters
concentrate the mercury, then venture far enough from shore to be prey for larger
fish, seabirds, and mammals.48
At this time, not enough information is available to determine whether mercury
levels in ocean fish and fish-eating marine mammals have increased or decreased
over the past hundred years or so, much less whether levels rose and declined as a
result of changes in atmospheric emissions.49 Although most scientists who study
mercury agree that deposition of atmospheric mercury has increased, and therefore
the total amount of mercury in the oceans probably has increased, particularly in the
surface layer, and one study (described below) has found increased mercury levels

46 Robert Mason, personal communication, April 1, 2004.
47 A. M. L. Kraepiel, K. Keller, H. B. Chin, et al., “Sources and variations of mercury in
tuna,” Environmental Science & Technology, v. 37, n. 24 (2003), p. 5551-5558.
48 J. K. King, J. E. Kostka, M. E. Frischer, et al., “A quantitative relationship that
demonstrates mercury methylation rates in marine sediments are based on the community
composition and activity of sulfate-reducing bacteria,” Environmental Science &
Technology, v. 35, n. 12 (2001), pp. 2491-2496.
49 A famous 1972 study by G. E. Miller et al. (Science, v. 175 (1972), pp. 1121-1122)
purported to lend support to the contention that mercury in ocean fish derived from natural
sources. However, because Miller reported the study in a letter to the editor of Science, it
was not peer-reviewed, and is an insufficient basis for drawing any conclusions: A single
swordfish head preserved in 1946 was the source of the “historical” data for swordfish;
while five skipjack tuna, one bluefin tuna, and one albacore tuna, all less than two-thirds
meter long, originating from two different oceans, served as the historical reference
specimens for tuna. The modern specimens were one fresh albacore tuna, one fresh skipjack
tuna, three cans of albacore tuna, and six fresh swordfish. No information was provided
about the age or length of these specimens.

in feathers of fish-eating seabirds,50 measurements of mercury in ocean water and fish
are lacking or inconclusive. In part, this lack of data is due to the difficulty of
measuring mercury: measurement of methylmercury has only been possible since
about 1985, and past measurements of total mercury often were inaccurate because
samples were so easily contaminated.
A recent study that compared total mercury concentrations in yellowfin tuna
captured in 1971 with methylmercury in yellowfin tuna caught in 1998, both in the
vicinity of Hawaii, found no significant differences in mercury concentrations.51
However, the significance of these measurements is unclear, given the historical
trend in atmospheric deposition, which peaked in the mid 1980s.
Another study compared feathers over time from two kinds of fish-eating birds
that live in the northern Atlantic Ocean.52 Feathers were obtained from museum
specimens taken as long ago as 1885. The study found a significant increase in
concentrations of methylmercury over time. Among birds that eat fish living near the
ocean surface, concentrations of methylmercury in feathers increased at an estimated
rate of 1.1% annually between 1885 and 1994. According to study authors, this
increase is consistent with the estimated three-fold increases in concentrations of
mercury in the atmosphere and surface oceans due to human industry over the same
period of time. Among birds that eat fish living in a deeper, darker ocean layer,
methylmercury concentrations increased at an estimated rate of 3.5 to 4.8% per
year. 53
Risks of Methylmercury Poisoning
Toxicity of Methylmercury. Methylmercury is highly toxic to the central
nervous system of humans and many animals. The observed effects of toxic levels
of exposure generally have been similar in laboratory animals, domestic pets,
wildlife, and people. Typically, there is a lag time of weeks or even months between
exposure to mercury and the onset of health effects.54
In human adults, absorbed methylmercury is dispersed throughout the body in
blood and enters the brain, where it may cause structural damage. The physical
lesions may lead to tingling and numbness in fingers and toes, loss of coordination,
difficulty in walking, generalized weakness, impairment of hearing and vision,
tremor, and finally loss of consciousness and death. At high levels of exposure,
effects on the brain are easily observed and irreversible. Damage to the brain may

50 L. R. Monteiro and R. W. Furness, “Accelerated increase in mercury contamination in
North Atlantic mesopelagic food chains as indicated by time series of seabird feathers,”
Environmental Toxicology and Chemistry, v. 16, n. 12 (1997), pp. 2489-2493.
51 A. M. L. Kraepiel, K. Keller, H. B. Chin, et al., “Sources and variations of mercury in
tuna,” Environmental Science & Technology, v. 37, n. 24 (2003), pp. 5551-5558.
52 Monteiro and Furness.
53 Ibid.
54 T. W. Clarkson, “The three modern faces of mercury,” Environmental Health
Perspectives, v. 110 (2002), supp. 1, p. 11-23.

exist, however, in the absence of these observable symptoms of toxicity.55 Nervous
system damage (indicated by tingling and/or numbness in the fingers and toes) has
been estimated to occur in about 5 % of adults whose hair is found to contain 50 parts
of methylmercury per million parts of hair (ppm).56 This condition is predictive of
more severe toxicity. Lower levels of exposure may have more subtle adverse
impacts on coordination, ability to concentrate, and thought processes.57
Methylmercury readily crosses the placenta of pregnant women.58 Levels of
methylmercury in the fetal brain are roughly five to seven times the levels in maternal
blood.59 Compared to the adult brain, the fetal brain is more sensitive to
methylmercury. In the fetus, methylmercury exposure can affect brain development,
as evidenced during childhood by a child’s ability to learn and function normally
after birth.60 Human poisoning incidents in Iraq and Japan caused severely exposed
children to be born with cerebral palsy and mental retardation, and in a few cases
infants died. In Japan, poisoning occurred because local fish were poisoned by
industrial mercury releases to Minamata Bay. The average mercury content of fish
samples there ranged from 9 to 24 ppm. Recent research indicates that exposure to
much lower levels of methylmercury also leads to developmental effects on cognitive
There is general agreement that as little as 10 ppm methylmercury in maternal
hair indicates a level of exposure that may produce prenatal effects.62 Some believe
effects occur at even lower exposure levels. For example, a study of women and
their infants in eastern Massachusetts indicated that there might be adverse effects
when mothers have less than 3 ppm methylmercury in hair.63 At very low levels of
exposure, effects may be very subtle, and detectable only on a population basis — for

55 National Research Council, Toxicological Effects of Methylmercury (Wahington: National
Academy Press, 2000), 344 pp.
56 World Health Organization, Environmental Health Criteria 101: Methylmercury (1990).
57 E. M. Yokoo, J. G. Valente, L Grattan, et al., “Low level methylmercury exposure affects
neuropsychological function in adults,” Environmental Health: A Global Access Science
Source, v. 2, n. 1 (2003), pp. 8-19, at [], visited Jan.

19, 2006.

58 Wolfe, Schwarzbach, and Sulaiman, p. 149.
59 E. Cernichiari, R. Brewer, G. J. Myers, et al., “Monitoring methylmercury during
pregnancy: maternal hair predicts fetal brain exposure,” Neurotoxicology, v. 16, n. 4 (1995),
pp. 705-710.
60 EPA, Mercury Study Report to Congress, vol. 1, pp. 3-23.
61 P. Grandjean, P. Weihe, R. F. White, et al., “Cognitive deficit in seven-year-old children
with prenatal exposure to methylmercury,” Neurotoxicology and Teratology, v. 19, n. 4
(1997), p. 417.
62 Clarkson.
63 Oken, E., R.O. Wright, K.P. Kleinman, et al. Maternal fish consumption, hair mercury,
and infant cognition in a U.S. cohort, Environmental Health Perspectives, v. 113, n. 10,
(2005) p. 1376-1380.

example, by an increase in the proportion of an exposed population that falls below
a level of function defined as impaired.
In response to a mandate from the U.S. Congress, EPA contracted with the National
Research Council (NRC) to review available research on methylmercury toxicity.
The NRC Committee issued a report in 2000.64 It concluded that scientific studies
have demonstrated the sensitivity of the human fetus to pre-natal methylmercury exposure,
and that the risk to women who eat large amounts of fish and seafood during pregnancy
is “likely to be sufficient to result in an increase in the number of children who have
to struggle to keep up in school.”65
A study published in 2003 strengthened and extended the findings of the single
major study of children which failed to find any adverse effects in children exposed
to mercury before they were born.66 However, one NRC Committee member testified
before a House subcommittee in November 2003 that although those findings had not
been published at the time, they only confirmed results already considered and would
not have led to a different Committee conclusion.67 This conclusion has since been
confirmed in a peer-reviewed publication by four members of the original NRC
Human sensitivity to cardiovascular toxicity might be even greater than to
developmental neurotoxicity, given recent research results. For example, a study of
1,833 Finnish men found that those who had at least 2 ppm of mercury in hair had
twice the risk of acute myocardial infarction compared to men with less mercury in
hair.69 ( Two ppm of methylmercury roughly corresponds to the upper 10th percentile
of current methylmercury exposure among adult men in the United States.)
A follow-up study of the Finnish men also looked at levels of fish-derived fatty
acids. Results suggested that the adverse effect of mercury exposure resulted from
its interference with the protective effect of fatty acids in the fish. Men who ate fish
appeared to benefit from a protective effect of the acids against heart disease, but among

64 National Research Council, Toxicological Effects of Methylmercury.
65 Ibid. p. 325.
66 G. J. Myers, P. W. Davidson, C. Cox, et al, “Prenatal methylmercury exposure from ocean
fish consumption in the Seychelles child development study,” Lancet, v. 361 (2003), pp.


67 U.S. House of Representatives, Committee on Science, Subcommittee on Environment,
Technology, and Standards, Mercury Emissions: State of the Science and Technology,
hearing, Nov. 5, 2003, statement of Thomas A. Burke, at [
hearings/ets03/nov05/burke.htm], visited Jan. 19, 2006.
68 A. H. Stern, J. L. Jacobson, L. Ryan, et al., “Do recent data from the Seychelles Islands
alter the conclusions of the NRC report on the toxicological effects of methylmercury?”
Commentary, Environmental Health: A Global Access Science Source, v. 3, n. 1 (2004), pp.

2-5, at [], visited Jan. 19, 2006.

69 J. T. Salonen, K. Saponin, K. Nyyssonen, et al., “Intake of mercury from fish, lipid
peroxidation, and the risk of myocardial infarction and coronary, cardiovascular, and any
death in Eastern Finnish men,” Circulation, v. 91, n. 3 (1995), pp. 645-655.

those with more than 2 ppm mercury in their hair the protective effect was reduced
by half.70 Other studies generally are consistent with these results, but one major study
failed to find an association between total mercury exposure (measured in toenail
clippings) and cardiovascular disease.71 More research is needed to explore interactions
among the various risk factors, fish-derived fatty acids, and mercury exposure with
respect to heart disease.
Environmental Methylmercury Exposure. People may be exposed to mercury
by eating or drinking, inhaling, or simply absorbing it through their skin. The level
of recent (within a month or two) individual exposure to mercury may be determined
based on measured concentrations of mercury in blood. For a slightly longer exposure
history (e.g., over several months), mercury concentrations in human hair several inches
from the scalp may be useful. However, there is no way to measure exposure that
occurred more than a few years ago, because methylmercury breaks down in the bodies
of animals, and both organic and inorganic mercury are excreted over time.
Although rates of physiological processes vary widely among individuals, in general,
people eliminate about half the mercury taken in within a period of roughly 44-80 days.72
In this way, mercury differs from many other pollutants such as lead, which may be
measured in the bone or teeth years after exposure has ceased. If mercury exposure
ends (because mercury is excreted) before a toxic amount of mercury has accumulated
in the body, adverse health effects would not be expected to occur. However, effects
would not necessarily subside after excretion, if a toxic level of exposure had occurred.
The 1999-2002 National Health and Nutrition Examination Survey (NHANES)
collected data on blood mercury levels for a representative sample of U.S. women of
child-bearing age. The results for the first two years (1999-2000) are summarized in
Table 2. Because mercury is present in much lower levels in blood than in tissues
such as hair, concentrations are expressed as parts of mercury per billion parts blood
(ppb), by weight. Based on these data, the Centers for Disease Control and Prevention
(CDC) concluded that mercury concentrations generally were low among women of
child-bearing age and children in the U.S. population. These results were confirmed
by data collected in 2001-2002.73 However, study authors noted that the survey was
designed to gather baseline data, and that there were too few people interviewed to
provide reliable estimates of blood mercury levels for individuals at the highest levels
of exposure.74

70 T. Rissanen, S. Voutilainen, K. Nyyssonen, et al., “Fish oil-derived fatty acids,
docosahexaenoic acid and docosapentaenoic acid, and the risk of acute coronary events: the
Kuopio ischaemic heart disease risk factor study,” Circulation, v. 102, n. 22 (2000), pp.


71 K. Yoshizawa, E. B. Rimm, J. S. Morris, et al., “Mercury and the risk of coronary heart
disease in men,” New England Journal of Medicine, v. 347, n. 22 (2002), pp. 1755-1760.
72 EPA, Mercury Study Report to Congress, vol. 1, pp. 3-23.
73 Centers for Disease Prevention and Control. MMWR Weekly. Nov. 5, 2004, v. 53, n. 43,
p. 1018-1020.
74 S. E. Schober, T. H. Sinks, R. L. Jones, et al., “Blood mercury levels in U.S. children and

In the United States, most people are exposed to mercury primarily through eating
the flesh (muscle) of fish.75 People who eat a lot of predatory fish, such as bass, pike,
tuna, or swordfish, which may be highly contaminated, may increase the risk of adverse
health effects for themselves or, in the case of women who become pregnant, for any
unborn children.76 Thus, NHANES 1999-2000 found that women who ate three or
more servings of fish within a month had almost four times the level of mercury in
their blood as women who ate no fish that month.77 Nevertheless, 95% of the 448 women
who ate fish relatively frequently (at least three times during the previous 30 days)
had blood mercury levels less than about 11 ppb. About 25% of the study population
ate no fish or shellfish at all.78 Generally, their blood contained levels of mercury that
were below 2 ppb.
Table 2. Geometric Mean and Selected Percentiles
of Total Blood Mercury Concentrations (ppb) for
U.S. Children Aged 1-5 Years and Women Aged 16-49 Years
Selected percentiles
Gr oup GeometricMean th th th th th th
10 25 50 75 90 95
Children0.3< 0.14< 2.3
Women1.2 0.2
Source: 1999-2000 National Health and Nutrition Examination Survey.79
The amount of mercury in fish varies with the species, age, and size of the fish.
Uncontaminated fish contain less than 0.01 ppm methylmercury in muscle, while very80
contaminated swordfish in U.S. waters have more than 3 ppm mercury. (Grossly
contaminated fish in Minamata Bay, Japan, contained between 9 and 24 ppm mercury.)81
Even higher levels have been found where there is a local source of water pollution.

74 (...continued)
women of childbearing age, 1999-2000,” JAMA, v. 289, n. 13 (2003), pp. 1667-1674.
75 Note that mercury is stored in muscle rather than skin, fat, or bone, and so it cannot be
avoided by removing those parts before eating.
76 J. M. Hightower and D. Moore, “Mercury levels in high-end consumers of fish,”
Environmental Health Perspectives, v. 111, n. 4 (2003), pp. 604-608.
77 S. E. Schober, T. H. Sinks, R. L. Jones, et al.
78 Ibid.
79 Ibid; K. R. Mahaffey, R. P. Clickner, and C. C. Bodurow, “Blood organic mercury and
dietary mercury intake: National Health and Nutrition Examination Survey, 1999 and 2000,”
Environmental Health Perspectives, v. 112, n. 5 (2004), p. 568.
80 FDA, Mercury Levels in Commercial Fish and Shellfish, at [
~frf/sea-mehg.html], visited Jan. 19, 2006.
81 FDA, “Mercury In Fish: Cause for Concern?” Consumer Magazine, Sept. 1994, as
updated May 1995.

Diverse species of fish differ in sensitivity to mercury. Significant toxic effects and
death are associated in adult fish of various species with between 6 ppm (e.g., for
walleyes) and 20 ppm (e.g., for salmon) in muscle tissue. However, individual fish
within species also differ in sensitivity, and fish seem able to tolerate higher
concentrations of mercury if it is accumulated slowly.82 In general, older, larger fish
of the same species will have more mercury. Table 3 provides the average concentration
found in recent years in selected species popular with American consumers.
Concentrations are given in parts of mercury per million parts of fish (ppm). Freshwater
fish are in italic type. Methylmercury levels in particular species of fish are highly
variable, however, reflecting the chemistry and methylation potential of the bodies
of water in which they live.
Table 3. Mercury Concentrations in Some Popular Fish (ppm)
SpeciesAverage Mercury Level (ppm)
Tuna, white, canned (solid and chunk 0.36
Tuna, light, canned (chunk)0.12
Crab (blue, king, snow)0.06
Flatfish (flounder, sole, plaice)0.05
Sources: FDA websites Mercury Levels in Commercial Fish and Shellfish, at [http://www.] and Mercury in Fish: FDA Monitoring Program (1990-2003), at
[ h t t p : / / www. c f s a n . f d a . g o v / ~ f r f / s e a m e h g 2 . h t m l ] .
Note: Italics indicate freshwater fish. All other fish are marine.

82 Wiener et al., p. 427.

Recommended Exposure Limits. A key question for Congress is whether
there is currently a potential for adverse health effects among individuals who regularly
consume fish. Federal agencies have estimated the risk associated with methylmercury
exposure at current levels of environmental (i.e., fish) contamination. Of particular
relevance is the reference dose (RfD) set by EPA, which is discussed in some detail
below. Because there has been some controversy surrounding the EPA RfD, it is
compared to two other maximum allowable concentration levels established by federal
agencies, the minimum risk level (MRL) set by the Agency for Toxic Substances and
Disease Registry, and the Acceptable Daily Intake (ADI) level established by the Food
and Drug Administration. As explained below, the apparent inconsistency among the
FDA, ATSDR, and EPA estimates of a “safe” exposure level for methylmercury is
primarily due to the agencies’ diverse responsibilities and actions that are triggered
when contamination is found to occur.
EPA Reference Dose for Methylmercury. The EPA Reference Dose (RfD)
is a risk assessment tool, used to estimate daily intake levels of chemicals that are
expected to be “without an appreciable risk of deleterious health effects,”83 even if
exposure persists over a lifetime. The risk associated with exposure to methylmercury
above the RfD is uncertain, but likely to increase with increasing exposure levels.
The RfD is intended to account for sensitive members of the human population, such
as pregnant women and infants, but not individuals with unusual sensitivity due to
conditions such as genetic disorders or severe illness. To calculate the RfD, EPA
generally uses a “no observed adverse effect level” (NOAEL), which may be observed
or estimated using a model. A NOAEL estimates the threshold level of exposure below
which adverse effects do not occur. Then the RfD is established by dividing the NOAEL
by uncertainty factors which account for the need to extrapolate from limited data sets
to the general U.S. population.
In 1985, EPA established its first RfD for people who eat methylmercury-
contaminated fish at 0.3 micrograms of methylmercury (µg) per kilogram of body
weight (kgbw) per day. This is equivalent to about 126 µg of methylmercury per week
(roughly the amount in two 7-ounce servings of fish containing 0.3 ppm mercury) for
a person weighing 132 pounds. This dose is based on the lowest level of exposure that
produced adverse effects on the nervous systems (i.e., numbness and tingling in the
extremities) of adult Iraqis after they were poisoned by eating contaminated grain during
1971-1972 and adult Japanese who ate contaminated fish from Minamata Bay during
the mid-1950s.
Two years after EPA set its RfD, data were published showing adverse effects
of maternal mercury exposure on the development of Iraqi children who were exposed
in the womb. In 1995, EPA revised its RfD, basing it on these developmental effects.
This second RfD of 0.1 µg/kgbw/day (42 µg per week for a person weighing 132 pounds)
remains in effect. This level would be exceeded if a 132-pound person ate more than
one fish meal per week, and the fish contained more than 0.21 ppm of mercury.

83 EPA Fact Sheet, Mercury Update: Impact on Fish Advisories, June 2001, at [http://www.], visited Jan. 19, 2006.

To calculate the current RfD, EPA used a benchmark dose approach. The benchmark
dose for methylmercury estimates the level of exposure that has a 5% chance of doubling
the number of children (from 5% to 10% of the exposed population) who function
at an abnormally low level on a standardized measure. In 1997, the benchmark dose
calculated was 11 parts methylmercury per million parts maternal hair (ppm), by weight,
based on all the adverse health effects observed in Iraqi children who were exposed
to methylmercury before birth. The findings of other human studies as well as toxicity
data collected from animals in scientific laboratories, supported the validity of the EPA
calculated benchmark dose.84 Benchmark doses calculated based on data from studies
of island populations with heavy seafood consumption produced similar values (11
to 17 ppm).85
EPA used the benchmark dose to conclude that consumption of 1.1 µg/kgbw/day
of methylmercury probably was safe for the unborn children of women who ate
contaminated grain in Iraq. At this level of mercury intake, Iraqi women who weighed
an average of 60 kg (about 132 pounds) had about 11 ppm mercury in maternal hair
and 44 µg methylmercury per liter of blood. (However, individual ratios of hair to
blood concentrations varied widely.) EPA divided that daily dose (1.1 µg/kgbw/day)
by an uncertainty factor of 10, accounting for the lack of data on reproductive effects
and differences among individuals, to establish the RfD at 0.1 µg/kgbw/day. At this
level of exposure, a mercury concentration of approximately 4 to 5 parts mercury per
billion parts blood (ppb), by weight, and 1 part mercury per million parts of hair (ppm),
by weight, would be expected to accumulate in an adult.
According to EPA’s independent advisory group, the Science Advisory Board
(SAB), the1997 EPA RfD was strongly supported by multiple studies based on different
ethnic populations and species, exposures, and developmental endpoints, all suggesting
similar RfDs.86 However, the SAB advised EPA to consider an additional uncertainty
factor to account for the need to extrapolate from the observed effects of an acute, short-
term exposure to effects that might result from low-level, life-long exposure; the difficulty
of detecting subtle population effects; and evidence from animal and human studies
suggesting possible neurological degeneration in the elderly and high mercury exposure
of the fetus compared to the mother’s exposure.
Soon after the results of long-term studies were published, the NRC recommended
that EPA base its RfD on a evidence of chronic toxicity among island dwellers who
were exposed to methylmercury through fish and other seafood.87 The NRC panel
concluded in its 2000 report that there is a 5% chance that maternal exposure to1.0
µg/kgbw/day of methylmercury would double the proportion of children functioning
at an abnormally low level. Mothers eating that amount of mercury (in contaminated
fish), on average, would have about 12 ppm methylmercury in their hair (and 58 ppb
in their blood); fetuses would be exposed to about 58 ppb in cord blood. Recent analyses

84 EPA, Mercury Study Report to Congress, vol. 1, pp. 3-26.
85 Lynn R. Goldman and William H. Farland, letter, Science, v. 279, n. 5351, pp. 640-641.
86 Science Advisory Board, An SAB Report: Review of the EPA Draft Mercury Study Report
to Congress, EPA-SAB-EC-98-001, Oct. 1997, p. 91.
87 National Research Council, Toxicological Effects of Methylmercury, p. 325.

indicate that these numbers may need to be revised to incorporate research results
indicating that the relationship between cord blood and maternal mercury intake is
highly variable.
Based on the NRC report, EPA revised the RfD for methylmercury. The value
of the RfD did not change from 0.1 µg/kgbw/day, but the basis for the RfD was updated
using the most current data and analyses. This RfD is considered to be protective of
all populations in the United States, including sensitive subpopulations. Based on
that RfD, pursuant to section 304(a)(1) of the Clean Water Act, EPA established in
2001 a water quality criterion for methylmercury of 0.3 parts of methylmercury per
one million parts of fish tissue (ppm). (This is the first time that EPA based a water
quality criterion on a concentration of a pollutant in fish rather than in the water column.)
EPA indicated that to protect consumers of fish and shellfish among the general
population, this concentration of methylmercury in fish and shellfish tissue should
not be exceeded.
Agency for Toxic Substances and Disease Registry Minimum Risk
Level. The Agency for Toxic Substances and Disease Registry (ATSDR), a branch
of the Public Health Service, has health-related authority under the Comprehensive
Emergency Response, Compensation, and Liability Act (CERCLA, better known as
Superfund). One of the agency’s responsibilities is to study hazardous substances found
at sites on the national priority list (NPL) and to publish and periodically update
toxicological profiles of those most frequently found. In revising the toxicological
profile for mercury, ATSDR evaluated available data and concluded in 1999 that they
supported a Minimum Risk Level (MRL) for chronic exposure to methylmercury of
0.3 µg/kgbw/day. (This is the same as EPA’s 1985 RfD.) ATSDR uses the MRL as
a screening tool to determine when the risks posed by a hazardous waste site require
additional study.
Food and Drug Administration Action Level. The Food and Drug
Administration (FDA) established an action level in 1984 at a concentration of 1 ppm
methylmercury in fish or seafood products sold through interstate commerce. At this
level, the Acceptable Daily Intake for an adult in the general population is 0.42
µg/kgbw/day, slightly higher than 0.3 µg/kgbw/day, the RfD established by EPA in 1985.
The FDA action level is based on the mid-point of the estimated range of the “lowest
observed adverse effects level” (LOAEL), or 300 µg of methylmercury/day, at which
level of exposure Japanese adults who ate contaminated fish experienced paresthesia
(numbness and tingling in extremities). FDA divided this value by 10 to account for
scientific uncertainties and to provide a margin of safety. FDA chose not to use the
Iraqi data on the effects of fetal exposure as a basis for revising its action level, due
to concerns about uncertainties (in contrast to the relative certainty of the health benefits
of consuming fish.) The FDA action level is enforceable; the Administration may seize
interstate shipments of fish and shellfish containing more than 1 ppm of methylmercury,
and may seize treated seed grain containing more than 1 ppm of mercury.88 For the
purpose of advising the general public about fish consumption, FDA has used EPA’s

88 Mercury formerly was used as a pesticide to treat seed grain. However, the last mercury-
based pesticides registered for use in the United States (to control mold) were voluntarily
canceled by the manufacturer in November 1993.

RfD, recommending that women of child-bearing age avoid certain fish and limit
consumption of other fish.89
The inconsistency among the FDA, ATSDR, and EPA estimates of a “safe” exposure
level for methylmercury is more apparent than real: the differences are less than the
uncertainty factor, and the reference levels serve different purposes. In addition, the
EPA number assumes a lifetime of exposure, while the ATSDR level is for chronic
exposure of 365 days or longer, and the FDA level is for consumption of particular
Table 4 consolidates the quantitative information provided above to facilitate
comparisons among agencies.
U.S. Fish Consumption, Methylmercury Exposure, and Health Risk.
By comparing methylmercury concentrations for popular fish (Table 3) with federal
guidelines (Table 4), it is possible to assess the relative safety of eating different fish
and shellfish.90 Table 5 provides estimates of the numbers of meals of fish with different
average levels of contamination that one could eat without increasing methylmercury
exposure beyond the EPA RfD. It is important to note, however, that these
recommendations assume that the size of meals, the age and size of particular fish,
and the age and size of the consumer are “average.” Generally, if other factors are
held constant, risks of poisoning increase to the extent that consumers are younger
or smaller than average, eat larger amounts, or eat older and larger fish (and risks decrease
if the reverse is true). For example, a fish lover who consumed one 7-ounce meal
of freshwater fish (roughly 200 grams) containing 0.3 ppm of methylmercury (the level
permitted by the EPA water quality criterion) seven days in a row could be exposed
to ten times the level of EPA’s RfD, a level equal to the benchmark dose level. But,
because different fish contain different levels of methylmercury, daily consumption
of 7 ounces of fish could result in much lower or much higher levels of methylmercury
exposure, depending on the types of fish consumed.
Table 4. Federal Upper Limits for
Methylmercury Exposure
AgencyLevel of DailyExposureLevel in Fish
EPA0.1 µg/kgbw/day 0.3 ppm
ATSDR0.3 µg/kgbw/day
FDA0.42 µg/kgbw/day 1 ppm

89 EPA/FDA Joint Federal Advisory for Mercury in Fish, at [
fishadvice/advice.html], visited Jan. 19, 2006.
90 Guides to fish consumption may be found on the Internet. For example, see the one
produced by the state of Maine at [
presentations/sunday/frohmberg.pdf], visited Jan. 19, 2006.

Table 5. Recommended Number of Meals per Month of Fish
Containing Various Methylmercury Concentrations,
Based on EPA RfD91
Methylmercury Level Allowable Meals per Month
in Fish (ppm)(8 ounce or 232 grams)
0.1 9
Average U.S. fish consumption is 7-14 ounces (200-400 grams) per month, according
to EPA, when those who do not eat fish are included.92 On average, that level of fish
consumption would expose fish eaters to 4 µg of mercury per day, a level below the
RfD for anyone weighing more than 88 pounds (40 kilograms). Fish consumption
rates in the United States are estimated annually by the National Marine Fisheries Service
(NMFS). Rates are estimated based on total fish and shellfish in commerce (edible
weight) divided by the total population in the middle of the census period. No adjustments
are made for waste or spoilage of the fish or for people who do not eat fish. Sport-caught
fish are not included. For 2002, NMFS estimated per person consumption at 15.6 pounds
of fish.93 Of this quantity, 11 pounds were fresh or frozen, including 6 pounds of finfish
and 5 of shellfish. Cured fish accounted for 0.3 pounds and canned fish for 4.3 pounds
per capita. Seventy-seven percent of the fish consumed was imported.
Consumption rate estimates are higher when only those who eat fish are considered.
Unfortunately, data are limited. In the Mercury Study Report to Congress, EPA estimated
85% of adults in the United States consume fish and shellfish at least once a month
with about 40% of adults selecting fish and shellfish as part of their diets at least
once a week (based on food frequency data collected among more than 19,000
adult respondents in the NHANES III conducted between 1988 and 1994). This
same survey identified 1-2% of adults who indicated they consume fish and shellfish94

almost daily.
91 Methylmercury: EPA Update 06/02, presentation by Rita Schoeny, Mercury Conference.
92 EPA, Mercury Study Report to Congress, vol. 1, pp. 3-23.
93 National Marine Fisheries Service, Fisheries of the United States — 2002, at [http://www.], visited Jan. 19, 2006.
94 EPA, Mercury Study Report to Congress, vol. 4, p. ES-2.

Data from NHANES 1999-2002 indicates that exposure to methylmercury is greater
than the RfD for approximately 6% of women of child-bearing age.95 This percentage
is based on four years of data; it is lower than was found by NHANES in the first two-year
reporting period,1999-2000. However, a declining trend should not be inferred, because
the difference is not statistically significant. At least two more years of data are needed
to determine whether the apparent decline in blood mercury levels is a real trend. For
study subjects who identified themselves as Asian, Pacific Islander, Native American,
or multiracial, approximately 16% had levels greater than the reference level.96
Data for certain areas of the California coast indicate that although half of all
consumers surveyed ate 21 grams per day or less, 5% of consumers ate more than 161
grams per day (more than 10 pounds per month) of fish that consumers caught
themselves.97 At 0.3 ppm methylmercury, such consumers would be taking in about

48 µg per day of methylmercury, an amount close to the benchmark dose. Similarly,

a 1988 study of Michigan anglers who eat the fish they catch found that they ate on
average 45 grams of freshwater fish per day, but 5% of those surveyed ate 98 grams
per day.98 That amounts to 1.5 pounds of fish per week per person, much more than
is recommended for contaminated species of fish, but not an implausibly large amount.
Table 6 illustrates the general relationship between plausible levels of fish consumption
and methylmercury exposure for various segments of the U.S. population, assuming
that fish contain methylmercury at the level of the water quality criterion established
by EPA.
In making choices about fish consumption, factors other than, or in addition to,
methylmercury concentration should be considered. In particular, the health benefits
of eating fish high in omega fatty acids are important, especially for cardiovascular
health and fetal development. The benefits of fish consumption for the development
of intellectual abilities in infants was supported recently by a study of 130 mother-child99
pairs. The study measured maternal fish consumption, hair mercury levels, and infant
scores on tests of visual recognition memory (VRM) and found that VRM scores rose
significantly with fish consumption, falling only when mercury levels in maternal hair
rose above 1.2 ppm. The study authors concluded that pregnant women should eat
at least two fish meals each week, but that they should choose fish species that tend
to be high in fatty acids but low in mercury content. As shown below, lake trout

95 Centers for Disease Prevention and Control, MMWR Weekly, Nov. 5, 2004, v. 53, n. 43,
pp. 1018-1020.
96 Hightower, J.M., A. O’Hare, and G.T. Hernandez. Blood mercury reporting in NHANES:
Identifying asian, Pacific islander, native American, and multiracial groups. Environmental
Health Perspectives, v. 114, n. 1 (2006) p. 173-175.
97 Office of Environmental Health Hazard Assessment, Chemicals in Fish: Consumption of
Fish and Shellfish in California and the United States, Final Report, Pesticide and
Environmental Toxicology Section, Office of Environmental Health Hazard Assessment,
California Environmental Protection Agency, Oakland, CA, 2001, p. 92.
98 Ibid.
99 Oken, E., R.O. Wright, K.P. Kleinman, et al. 2005. Maternal fish consumption, hair
mercury, and infant cognition in a U.S. cohort, Environmental Health Perspectives, v. 113,
n. 10, p. 1376-1380.

Table 6. Various Estimates of Fish Consumption and Mercury
Exposure, Assuming a Concentration in Fish of 0.3 ppm
PopulationMean Monthly Consumption Rate (ounces)Methylmercury Exposureper Day (µg/day)100
Per Capita U.S. (All)21 6
Average U.S. Fish
Consumer7 - 14 2 - 4
Average MI Angler/
Upper 5% CA Sport
Fish Angler/Consumer170 49
and salmon would fit those requirements. Table 7 provides average mercury levels
and relative fatty acid content for some popular fish.
Wildlife Exposure and Health Effects
Fish consumption also is the dominant pathway for wildlife exposure to
methylmercury. Fish-eating predators in North America generally have relatively high
concentrations of mercury. Toxic mercury levels have been found in individual mink,
otters, loons, the Florida panther, and other U.S. birds and wildlife. However, it is
not clear whether typical levels of environmental contamination are stressful for wildlife.
Fish-eating birds annually eliminate much of their accumulated methylmercury
when they form new feathers. Moreover, seabirds seem to be able to demethylate
methylmercury, rendering it less toxic. Nevertheless, methylmercury exposure may
harm sensitive species at levels found in certain local environments. Many scientists
suspect that the immune system is weakened as a result of methylmercury exposure.
The most likely adverse impact on birds of methylmercury exposure is impaired ability
to reproduce.101
In common loons, which have been studied extensively, concentrations of mercury
in blood correlate with mercury levels in the fish they eat. Mercury levels in loon blood
increase from west to east in Canada, with the highest levels being found in southeast
Canada.102 A recent study of mercury in 577 loon eggs collected across eight U.S.
states from Alaska to Maine found a similar trend of increasing mercury concentrations

100 Assumes chronic exposure at a constant level, which may be invalid.
101 Wiener et al., p. 430.
102 D. C. Evers, J. D. Kaplan, M. W. Meyer, et al., “A geographic trend in mercury measured
in common loon feathers and blood,” Environmental Toxicology and Chemistry, v. 17, n.

2 (1998), pp. 173-184.

Table 7. Relative Fatty Acid Content and Mercury Concentration
in Some Popular Fish
SpeciesAverage MercuryLevel (ppm)Relative Fatty AcidContent
Shrimp<0.01 Moderate
Tuna, white, canned (solid0.36High
and chunk albacore)
Tuna, light, canned0.12Moderate
Salmon0.01 High
Pollock 0.06 Moderate
Catfish0.05Low (Channel)
Moderate (Brown
Crab (blue, king, snow)0.06Moderate
Clams<0.01 Low
Flatfish (flounder, sole,0.05Low
Halibut 0.26 Moderate
Trout0.03Moderate (Rainbow)
High (Lake)
Swordfish 0.97 Lo w
Sources: FDA websites Mercury Levels in Commercial Fish and Shellfish, at [
~frf/sea-mehg.html], and Mercury in Fish: FDA Monitoring Program (1990-2003), at [http://www.cfsan.]; Purdue University, Food Safety and Quality, Angling Indiana, 2004 Fish
Consumption Advisory, Nutritional Content of Fish, at [
NutritionalContentofFish/omega3.pdf], visited Jan. 19, 2006.
from west to east.103 These blood and egg concentrations are consistent with the pattern104
of mercury deposition for North America (i.e., increasing from west to east). A study
reported in 2003 declining egg volume, but no effect on fertility, with increasing mercury
concentrations in New England. However, eggs were collected only if abandoned,

103 D. C. Evers, K. M. Taylor, A. Major, et al., “Common loon eggs as indicators of
methylmercury availability in North America,” Ecotoxicology, v. 12 (2003), pp. 69-81.
104 EPA, Mercury Study Report to Congress, vol. 7, “Characterization of human health and
wildlife risks from mercury exposure in the United States.”

which might have biased the results.105 Reduced egg laying has been associated with
concentrations greater than 0.4 ppm methylmercury in prey fish.106
Mink and otter exposed over a long period of time to more than 1 ppm
methylmercury in their diets exhibit classic signs of poisoning and may die. Higher
concentrations cause earlier but similar health effects.107 Less than half that concentration
is not lethal; data are lacking for more subtle effects on mink of mercury exposure.
There are no field data indicating that the wildlife species most at risk (because they
eat fish) currently are experiencing adverse health effects from mercury exposure.
Current scientific knowledge can inform the debate about competing legislative
and administrative proposals to reduce mercury emissions from utilities, but it cannot
provide firm answers to all of the specific questions that have been raised. Neither
can science resolve policy controversies that revolve around value judgments, for
example, questions about how urgent the need is for utility emission controls. However,
recent scientific studies have provided potentially useful information for policy makers,
about chemical changes to mercury emissions that may take place in the atmosphere;
rates of mercury deposition to, and re-emission from, the earth’s surface; the relationship
between mercury emissions and mercury levels in freshwater fish in various specific
ecosystems; and the potential effects of low level, chronic exposure to methyl mercury
through fish consumption.
Scientific studies have clearly demonstrated that levels of mercury in the atmosphere
and in deposits to earth have at least doubled and probably tripled due to human activities,
even in places that are remote from human influence. Although most of the largest
and most direct U.S. sources of mercury releases to water and air have been controlled,
and levels of U.S. mercury deposition are declining, levels of mercury in fish continue
to be a concern. Electric utilities are the only uncontrolled major stationary source
of U.S. mercury emissions. As a result, control of utility emissions might be the most
direct step that could be taken to reduce mercury deposition in the United States.
However, there are uncertainties in chemistry and transport, leading to current debates
among policy makers.
Local and regional emissions from various sources have caused mercury deposition
to increase as much as tenfold in some locations, indicating that there is a possibility
that local “hot spots” of mercury contamination might persist, despite overall reductions
in mercury emissions. In sensitive experimental lakes and wetlands, when local and
regional mercury emissions decreased, deposition decreased proportionately, and levels
of methylmercury in freshwater fish dropped quickly. This indicates that controls on
mercury emissions from electric power plants (particularly those plants with emissions
that tend to be deposited locally) could lead to substantial reductions in deposition
at some nearby hot spots. It remains to be determined whether there is a link between

105 Evers et al., 2003.
106 Wiener et al., p. 430.
107 Wiener et al., p. 435.

mercury emissions and mercury in ocean fish. However, scientists have shown that
significant quantities of emitted mercury are deposited in the oceans; methylmercury
is found in marine fish and predatory seabirds, sometimes at very high concentrations;
and sulfate-reducing bacteria are active in coastal sediments.
As yet unquantifiable but potentially significant risks from emissions exist, to
people and wildlife locally, but also in areas distant from emission sources. Research
continues to find evidence of subtle impacts on human health of low levels of
methylmercury exposure, levels close to current levels of exposure for people who
eat large amounts of certain large, predatory fish. In considering the potential adverse
effects of mercury, however, the potential nutritional benefits of eating fish that are
not heavily contaminated by mercury should not be overlooked.