mega? LEVELS. 23% SGME
S-ELECT'EE 900% AND EVALUE‘i'l-SN
9F ASSAY TECBREQEES

Them gov Hm Dogma a? M. 5.
Mi‘CHiGAR’ SMTE ”NWERSET?
Manefi L Gomez
WU.

\\\\\\E\\\\\\\\\S\W

3 10682 3796

WWW

\\\\\\\\\\\\\\\\

J‘W‘ZQ

\

WY

 

ABSTRACT

MERCURY LEVELS IN SOME SELECTED FOODS
AND EVALUATION OF ASSAY TECHNIQUES

BY

Manel I. Gomez

The primary objective of the study was to assess total
mercury levels in selected foodstuffs representative of the
average diet. Foods with no direct exposure to mercury
contamination were selected to represent background levels
of mercury. The study was confined to foods of Michigan
origin, in order to relate background mercury levels to geo-
graphic location.

Total mercury measurements were made by flameless atomic
absorption spectrophotometry after wet acid digestion of
samples. Concentrated H2804 digestion was applicable to
most animal products with the exception of beef and pork
liver and high-fat foods, such as cheese. Nitric acid:
sulfuric acid digestion mixtures were employed on all plant
products. Refined products such as salt and sugar were digested
with HNO3 alone. The major modifications in digestion procedures

were the use of 35 percent w/v HNO3 in place of concentrated HNO3

and the use of steam bath temperatures throughout the

Manel I. Gomez

digestion and subsequent KMnO oxidation. The moderation and

4
control of digestion conditions thus accomplished, permitted
the use of simple digestion equipment with no attached con-
denser systems and the handling of a large number of samples
at one time.

Digestion procedures were evaluated by recovery studies
on mercury added as mercuric chloride and methylmercuric
chloride to food samples of known mercury content. A recovery
of 83-87 percent of mercury added as methylmercuric chloride
in the 0.01—0.2 fig. range and 97 percent of mercury as mer-
curic chloride in the same range indicated satisfactory
efficiencies in digestion. In addition, the analytical pro-
cedures were evaluated by inter-laboratory comparative studies
on reference samples of fish. The results were in good agree-
ment.

Losses of mercury of 71 and 85 percent were observed
in preliminary lyophilization and vacuum drying of egg
samples respectively, precluding the use of these methods
in the preliminary concentration of samples prior to
wet-digestion.

The concentrations of mercury found in major foods
indicate that the average diet contains a very low level of
mercury in the range 0.01-0.03 ppm. Amounts of mercury found
in vegetables and fruits Show that residues are barely
detectable and are at or below the 0.01 ppm level. The

average concentration of mercury in dried plant products

such as grains and cereals was 0.02 ppm while those in animal

Manel I. Gomez

products was 0.03 ppm. Fish was the only food showing sig-
nificant levels of mercury with a mean concentration of 0.17
ppm. This level was however still below the F.D.A. guide-
line of 0.5 ppm. No substantial differences were observed
between different strains of plant and animal products
except in the case of fish.

Results of this study suggest that mercury levels
present in the major foods do not represent overt contamina-
tion above a mean background level of 0.01—0.03 ppm. The
digestion techniques employed were shown to have merit for

the routine analysis of a large number of food samples.

MERCURY LEVELS IN SOME SELECTED FOODS
AND EVALUATION OF ASSAY TECHNIQUES

BY ‘,
.v“
Manel If Gomez

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

. MASTER OF SCIENCE
Department of Food Science and Human Nutrition

1972

ACKNOWLEDGEMENTS

The author wishes to extend her sincere thanks and
appreciation to Dr. P. Markakis, her major professor, for
his continuous encouragement, advice, able guidance and
personal interest throughout her graduate program.

Thanks and appreciation are also expressed to
Dr. F. M. D'itri and Dr. L. E. Dawson for having served on
the Guidance Committee and for their counsel, guidance and
valued suggestions at all times. The author is especially
indebted to Dr. F. M. D'itri for generous access to ref-
erence material throughout the research project and for
the collaboration provided by the Laboratory of the Institute
of Water Research.

Thanks are also due to fellow graduate students for
their cooperation and assistance.

The cooperation of all staff members of the Departments
of Crop Science, Dairy Science, Food Science, Horticulture
and Poultry Science who were responsible for the provision
of and assistance with collection of samples is greatly
appreciated.

The author also expresses her appreciation to her

husband Modestus and children, Vaughan and Sharon, for their

ii

patience and understanding and the encouragement and inspira-

tion which made this undertaking possible.

iii

TABLE OF CONTENTS

Page
INTRODUCTION 0 O O I O O O O l 0 O O O O O O O O O O O 1

REVIEW OF LITERATURE O I C O O O O O O O O O O O O O O 3

Industrial Uses of Mercury . . . . . . . . . . . 3
Agricultural Uses of Mercury . . . . . . . . . . 5

Translocation . . . . . . . . . . . . . . . 7
Biomethylation . . . . . . . . . . . . . . . . . 9

Magnification in the Food Chain . . . . . . 11
Mercury in Foods . . . . . . . . . . . . . . . . 12
Standards . . . . . . . . . . . . . . . . . 13
MATERIALS AND METHODS . . . . . . . . . . . . . . . . 15
Food Samples . . . . . . . . . . . . . . . . . . 15
Sampling Procedure . . . . . . . . . . . . . 15
Storage of Samples . . . . . . . . . . . . . 16
Physical Pre-Treatment of Samples . . . . . 20
Mercury Determination . . . . . . . . . . . . . . 20
Preparation of Standards . . . . . . . . . . 22
Mercury Standards . . . . . . . . . . . . . 22
Intermediate Standards . . . . . . . . . . . 22
Calibration of Instrument . . . . . . . . . 25
Analytical Method . . . . . . . . . . . . . 25

Reagents . . . . . . . . . . . . . . . . . . 26

iv

TABLE OF CONTENTS (Cont'd.)

Page

Glassware . . . . . . . . . . . . . . . . . 26

Digestion Procedures . . . . . . . . . . . . 27

Digestion I . . . . . . . . . . . . . . 27

Digestion II . . . . . . . . . . . . . 28

Digestion III . . . . . . . . . . . . . 30
Reduction-Aeration . . . . . . . . . . . . . 31

Recovery Studies . . . . . . . . . . . . . . . . 31
Preparation of Standards . . . . . . . . . . 31
Intermediate Standard . . . . . . . . . . . 32

Working Standard . . . . . . . . . . . . . . 32

Recovery of Methylmercuric Chloride . . . . 32

Recovery of Mercuric Chloride . . . . . . . 33

RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . 34

Levels of Mercury in Foods . . . . . . . . . . . 34

Animal Foods . . . . . . . . . . . . . . . . 34

Fish . .'. . . . . . . . . . . . . . . 35
Dairy Products . . . . . . . . . . . . 35
Poultry Products . . . . . . . . . . . 40
Meats . . . . . . . . . . . . . . . . . 42

Plant Foods . . . . . . . . . . . . . . . . 43

Cereals and Legumes . . . . . . . . . . 43
Fruits and Vegetables . . . . . . . . . 45

Bread, Salt and Sugar . . . . . . . . . . . 47
MethOdOlogY O O O ‘0 O O O O O O O I O O O O O 0 O 5 0

Comparison of Mercury Recoveries from
Digestion Procedures I, II and III . . . . . 50

Recovery Results . . . . . . . . . . . . . . 52

General Discussion . . . . . . . . . . . . . . . 58

TABLE OF CONTENTS (Cont'd.)

Methodology
Digestion
Oxidation
Equipment
Recovery
Precision and Sensitivity

SUMMARY AND CONCLUSIONS

LITERATURE CITED

vi

Page
61
62
63
64
65
65
67

70

Table

10

11

12

13

14

15

LIST OF TABLES

U.S. Mercury Consumption for 1969 . . . . . . .

Food Items of Animal Orgin Assayed for
Total Mercury . . . . . . . . . . . . . . . . .

Food Items of Plant Origin Assayed for
Mercury 0 O O O O O I O O O O O O O O O O O O 0

Three Subsidiary Food Items Assayed for
Total Mercury . . . . . . . . . . . . . . . . .

Physical Pre-Treatment of Samples for
AnalYSiS O O O O O O O O O O O O O O O O O O 0

Total Mercury Content of Animal Foods--Fish . .
Total Mercury Content of Animal Foods--Dairy .

Total Mercury Content of Animal Foods--
Poultry 0 O O O O O O O O O O O O O O C O O O 0

Total Mercury Content of Animal Foods-~Meats .

Comparison of Total Mercury Contents of Fresh,
Vacuum Dried and Lyophilized Eggs . . . . . . .

Total Mercury Content of Plant Foods-—Cereals
and Legumes . . . . . . . . . . . . . . . . . .

Total Mercury Content of Plant Foods--Fruits
and Vegetables . . . . . . . . . . . . . . . .

Total Mercury Content of Foods--Sugar, Salt
and Bread 0 O O O O O O O O O O O O O O O O I 0

Summary of Base-Line Levels of Mercury in Major
Food Groups—-Comparison with Reported Values .

Levels of Mercury in Fish and Eggs Determined
by Three Digestion Procedures . . . . . . . . .

Vii

17

18

19

21
36

37

39

41

42

44

46

48

51

LIST OF TABLES (cont'd.)

Table
16 Recovery of Mercury Added As.Methylmercuric
Chloride to 1 gm. Samples of Chicken Mean
(Leg Meat) by H SO :KMnO Digestion--
. . 2 4 4
Digestion I . . . . . . . . . . . . . . . . .
17 Recovery of Mercury Added As Methylmercuric
Chloride to 1 gm. Samples of Barley by
HNO3:HZSO4 digestion--Digestion II . . . . .
18 Recovery of Mercury Added As Mercuric
Chloride to 1 gm. Samples of Food . . . . . .
19 Interlaboratory Comparison of Mercury Concen-

tration of Fish (ppm) . . . . . . . . . . . .

viii

Figure
1

LIST OF FIGURES

Instrument Calibration . . . .

ix

INTRODUCTION

In the last two decades mercury has come to be recognized
as a major and persistent environmental contaminant, as a
result of its widespread use in industry and agriculture and
its emission into the atmosphere from smelting operations
and burning of fossil fuels. The toxicity of mercury had
been known and recognized in early times. However, the
evaluation of its toxicity in the light of its behavior and
interactions in natural systems came to be considered only
recently.

Present day large scale uses of mercury have resulted
in its widespread dissemination throughout the ecosystem,
contaminating the most basic ingredients of the environment,
water, soil, air and food. This element and its compounds
have been demonstrated to have the unique property of bio-
transformation into one of its most toxic forms, methyl-
mercury, which is readily magnified in the food chain. Food
is a major source of intake and therefore the importance
of close and continuous vigilance on the levels of mercury
present in foods cannot be overemphasized. Plant and animal
products used as foodstuffs by man may contain mercury in

quantities detrimental to health.

At present there are few, if any, studies which are
monitoring the relative importance of the environmental
mercury sources to its total concentration in foodstuffs.
Data presently available on the levels of mercury in food
pertain mostly to individual components of the diet or
single categories of food, such as grains, dairy products,
meat or fish. Only limited data are available on mercury
contents of total diets or food items representative of total
diets.

Natural or background levels of mercury in sea water
have been associated with levels of mercury found in marine
fish. Background levels of mercury in soils and ground
water have been determined to evaluate the extent of environ-
mental contamination. However, no systematic monitoring of
the background level of mercury in foods has been reported
and the limited data available present wide variations.

Such studies have been largely handicapped by lack of
methodology involving a minimum of sample preparation and
accompanying losses of mercury.

The present study was undertaken to determine total
mercury levels in some selected foods of Michigan origin,
'with no direct exposure to mercury contamination. The pri-
.mary objective of the study was to determine background levels
<If mercury in foods. Concurrently, evaluations were made of
thee assay techniques for determining submicrogram levels of
.mezncury, with a View to developing a simple, rapid and

rel¢iable method of analysis for mercury in a variety of plant

and animal materials.

REVIEW OF LITERATURE

The current concern over the environmental contamina-
tion of mercury has directed attention to its sources,
transformations and mechanisms of concentration in the food
chain. The two major areas of present day use are industry

and agriculture.

Industrial Uses of Mercury

Chlor-alkali production by the mercury cell method has
been estimated to contribute the largest measure to the en-
vironmental burden of mercury. About 25 percent of the annual
U.S. consumption of 6 million pounds of mercury in 1969
(Table l) was used in mercury cell chlor-alkali production
(U.S. Bureau of Mines, 1970). It is estimated that in the
process, close to one-third pound of mercury can be lost for
every ton of chlorine produced and plants with a daily out-
put of 100 tons of chlorine could have an annual discharge
rate of 10,000 pounds of mercury (Bligh, 1972).

Large amounts of mercury are also used in the manufac-
ture of a variety of products such as paints (anti-fouling
and marine), electrical apparatus and batteries, cosmetics,
fluorescent and neon lights, acetaldehyde and plastics

(U.S. Bureau of Mines, 1970).

 

 

 

 

Table l. U.S. Mercury Consumption for 19691
Consumption
Industry Thousands of Pounds
Chlor-Alkali Industry 1572
Electrical Apparatus 1382
Paint 739
Instruments 391
Catalysts 221
Dental Preparations 209
Agriculture 204
General Laboratory Use 126
Pharmaceuticals 52
Pulp and Paper Making 42
Amalgamation 15
Other 1082
Total 6035

 

lSource: U.S. Dept. of Interior, Bureau of Mines.

Though organomercurial slimicides were used in the past
in the paper and pulp industries, in many countries regula-
tions controlling the use of mercurials in the manufacture
of paper, intended for use in food packaging, have led to
limited use of organomercurials (Bligh, 1972).

In addition to being intentionally used in industry,
mercury is also an unintentional by-product in a number
of processes involving natural products containing traces
of mercury. Bailey £3.31. (1961) reported that native
mercury and possibly other forms of mercury occur in petro-
leum, natural gas and crude oils. Many petroleum deposits
were shown by Bertine and Goldberg (1971) to contain mercury
in the ppm range and thus petroleum is another significant
source due to its large-scale use. Weiss gt_al, (1971) also
estimated that significant quantities of mercury could con-
Ceivably be discharged into the environment from the heating
of shale and limestone components to temperatures of 1,500°C
in the manufacture of cement.

The incineration or indiscriminate disposal of many
industrial and consumer products containing mercury constitutes
still another source of mercury in the environment (D Itri,

1972).

Agricultural Uses of Mercury
Organomercurial fungicides have been used in agriculture
since 1914. The three basic types of organomercurials are

the alkyl, aryl and alkoxyalkyl derivatives of mercury. Of

these the alkyl derivatives are the most toxic and the
alkoxyalkyl derivatives the least. Since the beginning of
the '503 mercurial seed dressings were suspected in Sweden

as the cause of diminishing bird and wildlife populations
(Borg 22 31,, 1966). The most toxic alkylmercurials were
taken off the market in Sweden in 1966 (Lofroth, 1969) and
recently in the U.S. However though the use of organomercu-
rials has been diminishing and though they constitute only

a fraction of the industrial uses of mercury they represent

a more direct and significant source of contamination in
respect to foods. Smart (1968) observed that in Britain,

the officially specified zero levels for residues were ex-
ceeded even when crops were treated in accordance with strict
agricultural practice. The most widely treated plant pro-
ducts are grains, which constitute an important component

of animal feeds and through which route mercury may be
carried in the food chain to meats and animal products.

In experimental studies, hens fed seeds treated with methyl-
Imercury had high mercury content in their eggs (Smart and
Lloyd, 1963). A family in New Mexico suffered severe mercury
intoxication from the ingestion of pork from hogs fed treated
seed (Curley gt 31., 1971) , while a number of epidemics of
fulisoning have been reported from Guatemala, Pakistan, Iraq
ant! Iran from the direct ingestion of treated seeds by

humans (Eyl, 1971).

Translocation

 

Organomercurial fungicides applied as seed and soil
dressings and foliar sprays have been reported to enter the
food chain directly as residues on plant material and by
systemic translocation within plant tissues. Translocation
of mercury from treated rice was shown to produce significant
levels of residues in harvested grain by Tomizawa 32 El.
(1966) in Japan. Epps (1966) reported higher levels of
mercury in rice from treated seed than from untreated con-
trols. However, Westermark (1967) found comparable levels
of 0.008-0.012 ppm in wheat and barley whether or not they
were grown from treated seeds. Saha gt El: (1970) found
significant amounts of mercury in grain from wheat grown in
soil treated with methylmercury dicyandiamide.

Translocation of mercury to edible parts of plants from
the foliar application of mercurials has been demonstrated in
fruits and vegetables. Martin and Pickard (1957) reported
mercury residues in the range of 0.02-0.12 ppm resulting from
the experimental spraying of apples. Stone gt_al. (1957)
reported measurable traces in the skin and pulp of treated
apples. Ross and Stewart (1960) suggested translocation
from other parts of the plant (other than from surface resi-
dues) to account for the residues of 0.05 ppm at harvest time.
They further established a translocation mechanism (1962)
by demonstrating the presence of residues in fruit enclosed
in plastic bags during spraying. Szkolnik et;§l, (1965)

added credence to the persistence of mercury in biological

systems with evidence of low levels of residues in trees
that had received no spray treatment in 10 years or more.

Similar mercury residue accumulations have been observed
in potatoes and tomatoes as with apples. Smart (1964)
demonstrated translocation of mercury by the root, skin and
leaves of potatoes, though soil treatment with inorganic
mercury gave rise to only negligible levels in the tubers at
harvest.

Furtani and Osajima (1965-1966), investigating the con-
tent of mercury in rice, fruits and vegetables, inferred that
mercury in food products is partly the residue of fungicides
sprayed on crops and partly due to absorption from the soil
through the roots. Mercury levels in soils may represent
cumulations of direct mercurial applications or fallout from
atmospheric sources superimposed on naturally occurring
levels.

Aomine §t_§l, (1967) observed that mercury permeates
the soil profile in a form more soluble than the sulfide and
is retained by the soil for considerable periods of time.
Ross and Stewart (1962) found no mercury in apples in an
orchard where surface soil contained as much as 1.8 ppm and
concurred with Booer (1944) that the insoluble mercuric
sulfide was the end product of transformation of mercury
compounds in the soil. In disagreement with these observa-
tions Furtani and Osajima (1966) found abnormally high
mercury contents in brown rice cultivated with no applica-

tions of fungicides.

The overall evaluation of the translocation of mercury
as a factor contributing to the contamination of foods is
difficult on the basis of data involving different mercury
compounds, rates of application, dosages and harvesting
schedules. However it appears that some of the mercury
applied to seeds and foliage is translocated to grain and
edible parts of plants though not in significant amounts.
The more persistent and important sources of mercury are the
soil accumulations resulting from the direct application of
fungicides, indirect fallout of atmospheric mercury and

run-off from terrestrial sources.

Biomethylation
While the toxicity of organomercurials has been recog-
nized for a long time, the inorganic mercury wastes of
industry were until recently believed to be innocuous and
safely disposed of in water where they were supposed to re—
main relatively inert. The epidemics of poisoning in the
Minamata and Niigata incidents in the years 1953-1965 were
the first indication that industrial wastes were not as
innocuous as hitherto assumed.
Japanese workers (Irukayama gt 31., 1966) were the first
'bo identify the causative factor in Minamata disease as
HEEthylmercury present in the shellfish of.Minamata bay. It
W343 suspected that the source of mercury was the conversion
0f inorganic mercury wastes to methylmercury by the action

0f 1microbes in the mud or by plankton. Jensen and Jernelov

10

(1967) demonstrated this conversion using micro-organisms
of bottom sediments.

The underlying mechanism of the methylation was worked
on by Wood gt 21° (1968) who showed that methylcobalamine
was involved in the methyl transfer, in the synthesis of both
monomethyl and dimethyl mercury in enzymatic and nonenzymatic
systems. Immura gt_al. (1971) found supportive evidence for
the mechanism and emphasized the essentiality of methyl-
cobalamine as methyldonor in the reaction. Landner (1971),

in experiments with NeurospOra crassa was able to demonstrate

 

methylation by an organism with no known requirement for
vitamin 312’ While inorganic mercury is essential for the
biomethylation reaction, Jernelov (1969) has shown that aryl
and alkoxyalkyl mercurials undergo breakdown to intermediate
inorganic mercury before conversion to methylmercury.

From the foregoing it is evident that whatever the form
of mercury discharged into aquatic and other systems sup-
porting methylating organisms, the ultimate end-product
is the highly toxic methylmercury. Studies of aquatic and
other food products bear out the widespread presence of
methylmercury. In Sweden, regardless of the nature of the
mercurial pollutant, only methylmercury has been identified
in fish (Westoo, 1967). In addition, Westoo ( 1967 and 1969)
reported that 80 to 100 percent of the mercury was present
as methylmercury in a number of animal products as well.

In the context of biomethylation it is of interest that the

same workers demonstrated in vitro methylation of mercury

11

by liver homogenates (Westoo, 1968). Kiwimae'gtyal. (1969)
showed that hens fed seeds treated with inorganic, alkoxyalkyl
or aryl compounds laid eggs containing only methylmercury,
although this finding was not confirmed by Stoewsand gt_al,
(1971). What is of greatest concern is the estimate that
already existing deposits of mercury in bottom sediments in
lakes and rivers can continue to generate methylmercury for
many years to come and be a continuing source of contamination
of the aquatic food chain (Johnels, 1967). Beasley (1971)
identified significant levels of mercury in fish protein
concentrates. Since fish is used as a source of protein in
animal feeds, it can be a continuing source of contamination

of many animal products.

Magnification in the Food Chain

 

Besides the unique property of biomethylation, mercury
has the tendency to undergo concentration in the aquatic
food chain. With the aid of labelled compounds Hannerz
(1968), studying uptake and accumulation of mercury in fish,
concluded that in comparison with other mercury compounds
methylmercury had the greatest concentration factor. The
ability to concentrate mercury though common to all fish
varies in relation to age, Species and trophic levels.
Larger predatory fish such as pike and swordfish have higher
concentration factors (Johnels and Westermark, 1969).
Tejning and Vesterberg (1964) reported a similar concentra—

tion effect in eggs when hens fed on treated seed laid

12

eggs containing twice the concentration of mercury as was

present in the fodder.

Mercury in Foods

Through all of the mechanisms discussed in the fore-
going, aquatic food would be expected to have the highest
concentrations of mercury. Fish have been shown in Japan,
Canada and Sweden to have comparatively high background
levels of mercury. In Minamata and Niigata, mercury levels
in fish, of the order of 5-20 ppm were identified (Bligh,
1972). Most of the fish investigated from fresh waters and
coastal areas in Sweden were reported to have concentrations
in the range of 0.2-1.0 ppm (Lofroth, 1969). These mercury
concentrations were mostly related to industrial sources of
pollution. For example, Johnels gt El. (1967) found that
organisms downstream from paper mills had higher levels of
mercury than those upstream. Bligh (1972) similarly ob-
served that Canadian fish in the vicinity of every chlor-
alkali plant had levels of mercury over 0.5 ppm.

Swedish workers also related the concentration of mer-
cury in other foods and animal tissues with industrial and
agricultural uses of mercury. Westob (1966, 1967) correlated
a decline in the mercury content of Swedish eggs with the ban
on alkylmercury seed dressings. Johnels gt_al, (1968)
noticed increased concentrations of mercury in feathers of
recent Ospreys as compared to those of pre-industrial

times.

13

Many experimental studies have been undertaken to
evaluate mercury levels in relation to known and identi-
fiable sources of contamination. Smart (1968) examined a
number of grain, fruit and vegetable crops in order to
evaluate pesticide residues. Observations derived from such
studies are restricted to specific foods or categories of
food and "normal" levels of mercury are based on a limited
number of samples used as controls. Stock (1934) initiated
analysis of mercury concentrations in foods in the 19305
and since then other workers have reported results of mer-
cury concentrations in foods (Goldwater, 1964). There is
a dearth of data on more comprehensive surveys of mercury
in foods carried out in recent times and unrelated to spe-
cific sources of contamination. Jervis (1970) examined mer-
cury levels in over 300 different Canadian foods, many of
which were reported to contain mercury levels approaching
or above tolerance levels. Some of Jervis' data were
questioned by Somers (1971) whose re-examination of some
samples yielded results lower by a ten fold factor. The
studies of Corneleiussen (1969) on a number of common foods
in various U.S. cities serves as the only basis for com-

parison of results from the present study.

Standards

 

From Swedish toxicological studies and data from the
Niigata incident an upper maximum acceptability limit for
mercury in fish was set at 1.0 ppm in Sweden, with a non-

enforced recommendation to limit consumption of fish to one

14

meal a week (Lofroth, 1969). In both the U.S. and Canada,
an interim safety limit of 0.5 ppm was adopted as a matter
of expediency on the basis of the Swedish standard.

According to the Miller Amendment to the Federal In-
secticide and Rodenticide act of 1955, a zero tolerance is
specified for mercury residues in foods, resulting from the
application of any mercury containing materials (Zweig, 1963).
The tolerance limits adopted in many countries range from the
1 ppm limit adopted by Sweden and the U.K. to 0.03 ppm in
the Benelux countries (Smart and Hill, 1968). However, no
firm basis has yet been established for determining a safe
standard for mercury in foods. The only available guideline
at present is that of 0.05 ppm as permissible upper limit
for mercury in foods, proposed by WHO in 1967 (FAQ/WHO,
1967). The latter guideline seems to find justification in
the light of recent evidence of c-mitotic effects observed
in human leukocytes exposed to methylmercury chloride
concentrations as low as 0.25 ppm (Fiskesjo, 1970). No
information is presently available on the sub-chromic
effects of the ingestion of low levels of mercury in food
over long periods of time. Such effects need to be considered

in the stipulation of a valid standard for mercury in foods.

MATERIALS AND METHODS

Food Samples

Foods were selected to represent a cross-section of the
average diet and included three categories of food, animal
products, cereals and legumes and fruits and vegetables.
Three basic subsidiary items of diet, sugar, salt and bread
were included in the study. Foods of animal origin included
dairy, fish, meat and poultry products. Fruit and vegetable
samples were representative of market samples as to ripeness
and maturity. Vegetables examined included stem and root
vegetables.

The survey was intended to cover foods of Michigan
origin and to relate mercury levels to geographic location.
For purposes of documentation and access to history of sam-
ples, nearly all samples were obtained from Michigan State
University, Experimental Research Station sources. Some
food items not available from the Experimental Station were
acquired from known private farms or retail stores in

Michigan.

Sampling Procedures
Equal portions of three to four random samples of ma-

terial were drawn from bulk samples (bins of cereal, carcasses

15

16

of animals etc.) and pooled to give representative sub-samples
(Tables 2,3 and 4). The quantity and size of samples were

as recommended for routine pesticide analysis by Zweig (1963).

Storage of Samples
Fresh produce was packed in polyethylene bags and held
in cold storage at 320 F. Pillay gt_al, (1971) found no
detectable traces of mercury in polyethylene stored blanks
and recommended the latter as a suitable medium for storage
of samples for trace analysis. Fresh produce samples were
analyzed within a week to 14 days of collection. However,
some samples such as potatoes and apples had already been
held for varying periods under refrigeration from time of
harvest to time of collection. Refrigeration of these sam-
ples was continued up to time of analysis. Dried vegetable
products, cereals and legumes were packed in polyethylene
bags and stored at 320 F. A representative sample of each
(sub-sample of 500 gm.) was ground in a Wiley mill to a
consistency which would pass through a 40 mesh seive, and
stored in stoppered containers for analysis. Milk was re-
frigerated at 320 F and analyzed the day following collec-
tiona Processed products such as salt and sugar were shelf
stcxred in original containers at room temperature, while bread
and cheese were held under refrigeration at 320 F. All
aniLDnal tissues were packed in cryovac bags and frozen. Eggs
were; refrigerated at 320 F and fish samples were obtained from
the IMSU Institute of Water Research Laboratory as frozen

fillets or homogenized samples.

17

 

 

 

 

 

 

 

Table 2. Food Items of Animal Origin, Assayed for Total Mercury
Sample Collected
Quantity
Food Item Pooled Drawn Condition Source
Dairy Products
Milk (i) 4 milkings 1 qt. Fresh, Whole
Unpasteurized MSU Dairy Farms
Milk (ii) 2x1 gallon 1 qt. Homogenized
Pasteurized,
2% Fat Retail Store
Milk Dried 1 lb. Spray Dried Commercial Source
Cheese 1 lb. Processed, MSU Dairy Dept.
Cheddar
PoultryiProducts
Chicken Meat 3 carcasses 3 1b.
(Breast)
Chicken Livers 3 livers 10 oz. Fresh MSU Poultry Farms
Eggs 24 24
Meats
Pork
( i) Duroc 2x1 lb. portions 2 1b.
(2 carcasses)
(ii) York 2x1 lb. portions 2 1b.
. (2 carcasses) Fresh MSU Swine Farms
Pork Livers
( i) Duroc 2x1 lb. portions 2 lb.
(2 carcasses)
(ii) York 2x1 lb. portions 2 lb.
(2 carcasses)
Beef 3x1 lb. portions 3 lb. Fresh MSU Beef Barns
(3 carcasses)
Beef Liver 3x1 1b. portions 3 1b. Fresh MSU Beef Barns
(3 carcasses)
Fish
Large Mouth Portion of Long. 10 gm. MSU Institute of
Bass Back Muscle Homo enized or Water Resources,
Yellow Perch Portion of Long. 10 gm. . 9 Laboratory, Winter
Filleted and
Back Muscle Frozen Green Lake,
Sunfish Pumpkin Portion of Long. 10 gm. Kellogg Bird

Seed

Back Muscle

Sanctuary

18

 

 

 

 

 

 

 

 

Table 3. Food Items of Plant Origin, Assayed for Total Mercury
Sample Collected
Food Item Quantity Condition Source
Cereals
Wheat
( i) Dickson 2 lb.
(ii) Beizes
Barley Unhulled MSU Crop SCience
( 1) Avon Research
(ii) Talbot 2 lb.
Oats
( i) Garry 2 lb.
Rice, long grain 2 1b. Polished
white. (2x1 lb. pkg.) Packaged Retail Store
Legumes
( i) Navy Beans 2 1b. MSU Crop Science
. Research,
Dried Saginaw Valley
(ii) Red Kidney 2 lb. Bean Farms
Vegetables
Asparagus 2 lb. Fresh, washed Private Farm
trimmed Hart, Mich.
Potatoes
( i) Russet Burbank 3 1b. MSU Horticul-
Stored under
refri eration ture Research
(ii) Merrimak 528 3 lb. 9 Center.
Fruits
Apples
( i) Golden Delicious 3 lb Stored under
refrigeration
(ii) Johnathan 3 lb. Stored under MSU Horticul-
. . ture Research
refrigeration
Center.
Tomatoes 3 lb. Freshly
harvested
Strawberries 2 qt. Freshly Retail Produce
(2x1 qt. pkg.) harvested Market.

(grown in Grand
Rapids, Mich.)

'f"‘

'u'-

19

 

 

 

 

 

Table 4. Three Subsidiary Food Items, Assayed for Total Mercury
Sample Collected
Food Item Quantity Condition Source
Bread, White 2 1b. Day-old Retail Store
(2x20 oz. pkg.)
Sugar, White 2 lb. Refined, Retail Store
Granulated (2x5 1b. pkg.) packaged *(Saginaw)
(i) and (ii)
Salt, uniodized 1 lb. Refined Retail Store

(2x1 lb. pkg.)

*(Detroit)

 

*Source of raw material.

20

Physical Pre-Treatment of Samples for Analysis
Samples were prepared for analysis as detailed in
Table 5. Frozen animal tissues were thawed sufficiently to
permit slicing and macerated at low speed in a waring blender.
Frozen tissues were not thawed completely in order to avoid
loss of "drip." Hand-chopping and slicing of samples were
performed on a glass chopping plate with stainless steel
knife. Meat, pork and poultry samples were trimmed of fat
as completely as possible to represent residue levels in
"lean" meat.
All equipment used in sample pre-treatment was thoroughly
washed using
1. Alconox detergent
2. 8-10 rinses in tap water
3. 2 rinses in distilled, de-ionized water
Stringent precautions were observed in handling and trans-

fer of samples to avoid mercury contamination.

Mercury Determination

Mercury in all samples was determined as total inorganic
:mercury using the flameless atomic absorption spectro-
photometry technique. The instrument used in this study was
a direct reading Coleman Mercury Analyzer - MAS 50, equipped
'with.closed system aeration described by Hatch and Ott (1968).
The range of detection was between 0.0 and 9.0 micrograms
t0t£il.mercury, on two scales of measurement (0.00-0.25 pg.

and 0.25-9.0 pg.) and the specified sensitivity of the

21

Table 5. Physical Pre-Treatment of Samples for Analysis

 

 

Food Item Treatment Conditions
Milk (i) — _
Milk (ii) lyophilized(a) 10 hours

Milk Spray Dried - -

 

Cheese - -
Chicken Meat Chopped and Waring blender
Macerated (low speed)
Chicken Liver Chopped and Waring blender
Macerated (low speed)
Eggs ( i) Homogenized Waring blender
(low speed)
( ii) Vacuum Dried Room temperature
(12 hours)
(iii) Lyophilized 10-12 hours
Pork, Pork Liver, Chopped and waring blender
Beef and Beef Liver Macerated (low speed)
Fish Homogenized Ploytron
homogenizer
Cereals Ground Wiley Mill
Legumes Ground Wiley Mill
vegetables Chopped, sliced Manually
or cubed
Fruits
( i) Flesh Macerated Waring blender
(low speed)
(ii) Peel Chopped Manually
(a)

\firtis Freeze-mobile -.Model 10-145-MR-BA.
(b)

PVaring blender with glass container.

22

instrument is given to be 0.01 pg. mercury. Readings were
estimated to the closest 0.005 pg. of total mercury.

Instrument calibration was checked at the beginning of the
study and thereafter periodically using mercury standards de-
scribed below. Typical calibration results from four sets of
independent readings are shown in Figure 1. All analyses were
performed against 3 to 4 standards in each run of samples. Day
to day reproducibility of instrument calibration was excellent.

Preparation of Standards. Mercury Standard: Mercuric chloride

 

 

was dried at 1000 C for 3 hours and stored in a desiccator.
1.3538 gm. were weighed out and dissolved in approximately
100 ml. distilled de-ionized water, 55 m1. of 18 N HZSO4 were
added and the solution diluted to 1 liter (mercury stock solu-

tion, 1 mg./ml.--stable up to 1 month).

Intermediate Standard: One ml. of stock solution was pipetted

 

into a 100 ml. volumetric flask, 10 m1. of 18 N H2804 were added
and the solution diluted to volume with distilled de-ionized
water. Each m1. of standard contained 10 pg. mercury.

Working Standard: One ml. of intermediate solution was

 

pipetted into a 100 ml. volumetric flask, 10 ml. of
18N H2804 were added and the solution diluted to volume
with distilled de-ionized water to make a solution con-
taining 0.1 pg. mercury/ml. Fresh intermediate and

working standards were prepared from stock solution at

the time of use.

23

Figure l. Instrument Calibration

mcg Hg readings

.ll

.07

. 03

.01

24

Sets of
Calibration Data

. x1

" .2

A 03
A4

.01 .02 .03 .04 .05 .06 .07 .08 .09 .1

mcg Hg - Standards
Fig. I. Instrument Calibration

25

Calibration of Instrument

 

Volumes of the working standard ranging from 0.1 to
5.0 ml. were pipetted into aeration flasks to which were
previously added 100 ml. distilled de-ionized water and
2 drops of 5 percent KMnO4. Then, 5 ml. of 10 percent
hydroxylamine hydrochloride were added and the flasks swirled
to effect complete decolorization of KMnO4. The aeration
head was inserted into the flask and readings noted at the
maximum recording of the meter. Readings were corrected
for reagent blanks which were in the range 0.00 to 0.005 pg.

mercury.

Analytical Method

 

Analysis was based on the Hatch and Ott (1968) pro-
cedure for determining mercury at the nanogram level.

The sample was acid digested (specific choice of acid
or combination of acids were used depending on the nature
of the matrix) and the mercury present was oxidized to the
mercuric form with a suitable oxidizing agent. Potassium
permanganate was the oxidizing agent of choice. The ex-
cess KMnO4 was reduced with hydroxylamine hydrochloride and
the mercury was reduced to the metallic state with stannous
chloride. The pump and aerator system circulated mercury
vapor into the absorption cell where absorption at 253.7nm
took place and the resulting change in energy was transmitted

to the photo detector.

26

Reagents

The application of the method of analysis at very low
mercury levels greatly depends on the precision with which
reagents blanks can be reproduced. Reagents were screened
in preliminary experiments to obtain reagents which would
yield low and constant blank readings. All reagents were
of analytical grade (ACS) and were from sources detailed
below.

Mercuric chloride (Speciality Chemicals Division,
Baker Chemical Company)

Sulfuric acid--Sp. Gr. 1.84 (Speciality Chemicals
Division, Baker Chemical Company)

Nitric acid--Sp. Gr. 1.47 (Mallinckrodt) ‘
Hydrochloric acid--Sp. Gr. 1.18 (Mallinckrodt)
Acetone (Baker Chemical Company)

Hydroxylamine hydrochloride (Sigma Chemical Com-
Pany)

Stannous chloride (Mallinckrodt)
Potassium permanganate (Mallinckrodt)

Methylmercuric chloride--97-99% (K & L Laboratories
Inc.)

All aqueous solutions of reagents were made using distilled

de—ionized water.

Glassware

 

Reproducibility of results at low levels of determina—
tion is dependent on strict attention to cleanliness of
glassware. The following procedure was used in routine

cleaning of glassware.

27

1. Wash in "Alconox" detergent.
2. 8 rinses in tap water.
3. Soak in warm 3 N HNO3.

4. 4-5 rinses in tap water.

5. 2 rinses in distilled de-ionized water.

Digestion Procedures

 

In biological samples containing mercury, destruction
of organic matter without loss of mercury is a major concern
because of the volatility of mercury and its covalent com-
pounds. The moderation of digestion conditions on the other
hand could result in incomplete release of mercury. Diges-
tion procedures were modified to obtain a satisfactory
compromise between these two factors. Due to the variety
of materials analyzed and differences in response of the
materials to the oxidizing mixture, the following three

different digestion procedures were adopted.

Digestion I.

 

Sulfuric acid (Sp. Gr. 1.84) digestion, with KMnO4
(5% w/v) oxidation was found applicable to most animal

foods with the exception of beef liver, pork liver and
cheese. A limitation of this digestion procedure is the
incomplete digestion of fats which are reported to contain
negligible amounts of mercury (Barett, 1956) and are normally

removed by filtration prior to the oxidation procedure.

The digestion procedure was based<x1Uthe and Armstrong's

28

(1970) modification of the Hatch and Ott (1968) method.
A double oxidation was introduced as recommended by Barett
(1956) and in the Dow Chemical Co. Method (1970).

One to three grams of sample were placed in the bottom
of a tared 100 ml. volumetric flask using a transfer tool
of the type recommended by Uthe and Armstrong (1970). Five
ml. of concentrated (36N) H2804 were added, a 10 ml. beaker
inverted over the mouth of the flask and the flask heated
for lk-Z hours on the steam bath. Although the dissolution
of organic matter appeared to take place in a shorter time,
heating periods of less than an hour gave rise to excessive

foaming in the KMnO oxidation step as well as in the final

4
reduction-aeration step indicating incomplete oxidation of
all organic matter. At the end of this period digestion
flasks were removed from the steam bath, allowed to stand
one hour at room temperature before they were cooled in an
ice bath and undigested fats removed by filtration through
a pledget of glass wool. The filter was rinsed repeatedly
with 1:1 H2504 and the digest further cooled in an ice bath

before gradual addition of 15 ml. of KMnO The flasks

4.
were allowed to stand at room temperature until the initial
reaction subsided, after which heating was continued as

before for 20-30 minutes. The flasks were cooled, loosely

stoppered and stood overnight for final analysis.

Digestion II.

 

A modification of the method reported by Jeffus and

Elkins (1970) employing a H2804:HNO3 digestion mixture,

29

was found applicable to all the plant foods examined. Some
plant materials, notably legumes, rice and apples gave ex-
cessive foaming with consequent loss of some of the sample,

when concentrated HNO (Sp. Gr. 1.47) was used in the di-

3
gestion mixture. Use of concentrated HNO3 also necessitated
higher temperatures and prolonged heating to effectively
remove all nitrous fumes. The reaction was more easily
controlled with 5.6 N HNO3 (35 w/v).

In addition samples high in carbohydrate tended to char
and carbonize if excess HNO3 was not used. Therefore, in—

stead of the 1:1 digestion mixture recommended by Jeffus and

Elkins, a 3:1 (HNO :H 504) mixture was used. Fifteen ml.

3 2

of HNO were added (in one step or in increments depending

3
on sample reaction) to 1 to 3 grams of sample in a 100 ml.
volumetric flask and the flask held at room temperature for
30-45 minutes. To samples of low moisture content up to

5 ml. of water were added prior to the addition of HNO3

to prevent charring. The flasks were set on a steam bath and
heated until dissolution of solids was nearly complete

after which the flask was removed from the bath and cooled

in ice. Five ml. H2804 were added slowly and the flask held
at room temperature till the initial reaction subsided.

Then heating was resumed on the steam bath for an additional
1% hours until evolution of brown nitrous fumes ceased.

The flask was cooled in ice and subjected to the KMnO4

oxidation as described in Digestion I above.

30

With some plant materials such as unhulled cereals,
there was incomplete breakdown of cellulose and evidence
of waxy undigested residues. The residues were not fil-
tered since some loss of mercury could occur through entrap-
ment on the cellulose fibers and the whole sample was sub-

mitted to aeration and reduction.

Digestion III.

 

The method of Hoover, Melton and Howard (Hoover gt 31.,
1970) was tried on eggs, fish and some plant materials in
preliminary experimentation. Thirty-five percent (w/v)

HNO was substituted for concentrated HNO3 as in Digestion

3
II. To avoid a vigorous initial reaction samples were added
in small increments of 0.5 gm. at a time to 10 m1. of

HNO3 in a 125 ml. Erlenmeyer flask, with heating between
additions to complete the initial reaction. One to two grams
of sample were added in this manner and digestion continued
for one to two hours till evolution of nitrous fumes ceased.
The digest was cooled, diluted with 20 ml. distilled water
being careful to rinse down the sides of the flask, and
oxidation carried out as in Digestion I.

With most plant materials, the dissolution of organic
matter was much less complete than with Digestion II. With
animal tissues more complete dissolution of solid matter was
observed, though in comparative digestions higher recoveries

of mercury were obtained from eggs and fish, with Digestion

I. Digestion III was thus found inadequate for most of the

31

materials studied, though it was applied to the relatively
refined food products, salt and sugar with which Digestion
I resulted in precipitation of sodium sulfate and charring

respectively.

Reduction-Aeration

 

The method of reduction-aeration was uniform for all
samples. The digested materials were transferred quanti-
tatively to the aeration flask making sure that any adhering
manganous oxides in the digestion flask was solubilized
with l to 2 drops of 10 percent hydroxylamine hydrochloride.
Then the final volume was made up to 100 ml. with de-ionized
distilled water. Five m1. hydroxylamine hydrochloride was
added and the flask swirled briefly to effect dissolution of
all the manganous oxides. This was followed by the addition
of 2 ml. of 10 percent stannous chloride and immediate in-
sertion of the aeration head. Readings were noted with the

"memory" switch on to retain the highest reading recorded.

Recovery Studies
Samples of representative items of the food categories
examined were spiked with known additions of mercury as
mercuric chloride and methylmercuric chloride and subjected

to digestion and subsequent analysis.

Preparation of Standards

 

l. Mercuric chloride: The same standards were used

as for instrument calibration.

32

2. Methylmercuric chloride: 12.78 mg. of methylmer-
curic chloride was weighed direct into a stoppered 10 ml.
volumetric flask. The reagent was held in an ice bath during
weighing to minimize volatilization. Transfer of reagent
to the weighing flask was done as rapidly as possible and the
flask stOppered immediately. All subsequent operations were
conducted under the hood.

Methylmercuric chloride was dissolved in acetone and
made up to 10 ml. with acetone to give a primary standard
containing 1 mg. mercury/ml. (Magos, 1971). Linstedt (1971)
has confirmed that after digestion all interferences (by

absorption at 253.7nm) from acetone were eliminated.

Intermediate Standard (10 pg. mercury/ml.)

 

One ml. of stock solution was pipetted into a 100 ml.
volumetric flask and made to volume with distilled de-ionized

water.

Working Standard (0.1 pg. mercury/ml.)

 

A working standard was prepared by appropriate dilution
of the intermediate standard.

All standards were stored in the refrigerator and
intermediate and working standards prepared at time of use.
Volumes of working standard ranging from 0.1 to 2.0 ml.
were added to 1 gm. samples of chicken meat and barley and
were subjected to digestions I and II respectively. Two
to four replicate samples were run at each level and three
unspiked blanks of each food material were run with each

digestion procedure.

33

Recoveries of Mercuric Chloride

In routine analysis, 2 to 3 samples of foods from each
of the categories examined were spiked with known additions
of mercuric chloride in each run of analyses to determine
recoveries. Even when mercury levels determined by the
different digestion techniques on the same sample differed,
good and comparable recoveries of added mercury (as mer-
curic chloride) were obtained under different digestion con-

ditions.

RESULTS AND DISCUSSION

Levels of Mercury in Foods

Twenty-eight varieties of foods were examined for total
mercury content. Mercury content of all samples were ex-
pressed on a fresh weight or "as received" basis and results
were uncorrected for recovery.

In Tables 6-13, the results of assays for mercury in food
are grouped according to the class of foods analyzed. The
mercury levels are expressed in ppm and are the mean of four
to six replicate analyses.

Table 14 summarizes the base-line levels of mercury in
the major food groups examined and compares these results

with some recently reported values.

Animal Foods

 

Results of analysis of the four categories of animal
foods studied, Fish, Dairy, Poultry, and Meat are shown in
Tables 6, 7, 8 and 9. Of the animal foods studied only
fish had levels approaching the action level of 0.5 ppm
adopted by F.D.A. If the W.H.O. guideline or tolerance
level for mercury in foods of 0.05 ppm is considered, all

fish samples examined exceeded this level.

34

35

Fish used in this study were drawn from an isolated
inland lake with no known source of direct industrial con-
tamination. However, a "natural" level of mercury may have
accumulated through fallout from coal burning power plants
and through the ecological chain involving migrant bird
populations on the lake. Birds feeding on treated seed in
the fields or on fish in contaminated lakes may carry a
body burden of mercury which may be transposed to other
habitats through droppings, and feathers and other excretory
material.

Pappas and Rosenberg (1966) reported background levels
of mercury in haddock, in the range of 0.017-0.023 ppm.
However, the background level of 0.13 ppm reported by Bligh
(1972) for perch from Lake Winnipegosis in Canada, is more

consistent with the findings of this study.

Dairerroducts

 

Results of analysis of dairy products are summarized in
Table 7. Samples of whole milk from two different sources
were analyzed for mercury content and mean levels in both
sets of samples were found to be 0.01 ppm. Goldwater
(1964) reported values in the range 0.003-0.001 ppm for whole
milk in the U.S. while Corneleiussen (1969) reported values
ranging from 0.002-0.02 ppm in various cities in the U.S.
Similar concentrations were found by Reigo (1970) in Sweden.

The levels of mercury obtained on freeze-dried samples of

.mCOADmcHEHouoU mo HTQEDG ou momma momospcoumm ca monsmflm

 

Ema ha.o I swam How osam> com:

36

 

Aommm cflxmfism x emamcsmv

moo.o v + NH.o Amy ma.OIHH.o msmonnam mfifiommq x msaamcmmo mwfiomoq

Anonom zoaawwv

 

moo.o v + mH.o Amy ma.onma.o mcooco>mam mouom

I Ammmn zusoE mmumqv

mmo.o + mm.o Amv mm.o:-.o mmofioEHmm mououmouoflz
.

cows mmcmm EmuH boom

 

Aemmv coflumuucmocou musonmz

 

emamuumooom finesse

mo ucoucoo wnsouwz Hmuoe .m magma

 

 

c d I I: l D .OA. II at. u l I, b I.
I .
I .I. .. In I .l l I .l I. I. . . I
III! I..- ' i I I I
.4. . n.
I x
1.1 t I. I. I, b ' .
u I . I 4| . l I no. I I.
I. I | I It . . u all I I I
a I I 1| I _q
.. I. I. l
. h I .I o u l
. . l . I
I I I I 4 I I
.. I I I II
.. .\
I
. . I _.n I. . a n I . \ ”.1. U
.. . .. I . s ‘ . I I. “I I . I u
l l ‘ o D O I. u I
- ... ~ _ “riflfi . 4 .\
. . . n I . I». r r n
.. . )I 'I' .1 ' II. I! I '

 

37

Table 7. Total Mercury Content of Animal Foods--Dairy

 

Mercury Concentration (PPm)

 

 

Food Item Range Mean

0
Milk Whole (i) 0.01 ~0.02 (5) 0.01 i 0.005
Unpasteurized
Milk Whole (ii) 0.005-0.01 (6) 0.01 i} 0.005
Homogenized, pasteurized
Milk, (i) Freeze Dried 0.005-0.01 (6) 0.01 i < 0.005
Milk Spray Dried 0.01 -0.03 (6) 0.02 i 0.01
Cheese Cheddar 0.015-0.030 (6) 0.02 i. 0.01

 

Mean for Dairy Products - 0.01 ppm

 

I
Figures in parentheses refer to number of determinations.

YIEL—‘u

"_ (D ’1‘ 0(DV I

--~-.----_ -._ ._ .—

 

1". I} .3

noijtjflnocro?

( I

if.” i :1’.

Viuezo”

- ——

emnifi

38

whole milk when expressed on a wet-weight basis represented
losses of mercury in freeze-drying. These losses were not
quantitated since the levels found were generally below the
lower limit of detection. Samples of spray-dried milk from

a commercial source contained levels close to the upper limit
reported by Corneleiussen (1969). Similar values were ob-

served for the cheese sampled in the present study.

Poultry Products

 

Chicken meat was found to contain higher levels of mer—
cury than pork or beef. Smart and Lloyd (1963) reported
0.01 ppm or less in tissues from hens fed untreated seed,
while Westoé (1966) found mercury residues in normal Swedish
commercial samples of poultry to be in the range of 0.028 to
0.031 ppm for liver and 0.009-0.022 ppm for breast meat.
Results of the present study are consistent with those of
Westoa. Liver levels of mercury obtained in this study
were not substantially higher than those of breast meat,
in agreement with the observation of Jervis (1970) that in-
creased liver levels were not a consistent feature with
young birds.

Results of Table 10 indicate that in both vacuum drying
and lyophilization of eggs in the pre-treatment for analysis,
considerable losses of mercury occur, with greater losses
in vacuum drying. Similar results were obtained by Pappas
and Rosenberg (1966) in vacuum drying of fish and egg

samples prior to their combustion analysis. However, with

39

Table 8. Total Mercury Content of Animal Foods--Poultry

 

Mercury Concentration (ppm)

 

 

Food Item Range Mean

Chicken Meat 0.015-0.030 (6). 0.025 i 0.005
(Breast Muscle)

Chicken Liver 0.025-0.030 (6) 0.030 :'< 0.005
Eggs, Fresh, Whole 0.030-0.040 (6) 0.035 1 0.005
Eggs Lyophilized 0.03 -0.07 (6) 0.05 i 0.015
Eggs Vacuum Dried 0.005-0.01 (6) 0.01 i < 0.005

 

Figures in parentheses refer to number of determinations.

40

lyophilized eggs these workers obtained values in the range
0.0-0.003 ppm which are lower by at least one order of
magnitude than those of the present study. While the results
of Table 10 indicate a 71 percent loss of mercury in
lyophilization, Pillay gt_gl. (1971) reported even higher
losses ranging from 81-98 percent from fish homogenates
(using mercury-free vacuum gauges), which in part is thought
to explain the low background levels reported by Pappas and
Rosenberg on lyophilized eggs. Hens fed grain treated with
methylmercury were shown to produce eggs having twice the
mercury concentration as was present in the fodder (Westoo,
1966). Tejning (1967) also pointed out that in the female
chicken a considerable amount of excretory mercury normally
accumulating in feathers is lost to the egg. Due to this
capacity for biological concentration, eggs like fish may
probably have higher levels than other foods as is suggested

by the results of this study.

b12222
Mercury levels in meats are represented in Table 9.
Levels were below 0.01 ppm in both pork and beef although
corresponding liver samples showed higher values in the range
0.01-0.03 ppm representing some degree of liver accumula-
tion of mercury relative to flesh. The values obtained
for pork and beef in the present study were below those
reported by Westoo (1966) for Swedish pork (0.06 ppm) and

by Somers for Canadian meat (0.04 ppm) respectively.

41

Table 9. Total Mercury Content of Animal Foods--Meats

 

Mercury Concentration (ppm)

 

 

Food Item Range Mean

Pork ( i) i 0.005~ (6) 0.005 : < 0.005
Pork (ii) i 0.005-0.01 (6) 0.01 i < 0.005
Pork Liver ( i) 0.02 —0.04 (6) 0.03 i 0.01
Pork Liver (ii) 0.02 -0.03 (5) 0.03 i 0.005
Beef 1 0.005-0.02 (6) 0.01 i 0.005
Beef Liver 0.01 -0.015 (6) 0.01 i < 0.005

 

I
Figures in parentheses refer to number of

determinations.

_-—-.

o

 

. o '
_’ _ ,.,> . u-p
I ,. o . -
'
_..'. ' c.-
, e—. 4 v
-_.. _ - - - - .
r .' a.
I 4

 

_..* ”~-* .-0

— p
. _ . c- -
— "
h 9 ‘
a - '
i
.-
u
. . .
, -

.—_.

 

42

amazon NHIoH How Umufiawnmoma Ame wumcmmoeoc mo coauuom m>Humucmmohmom

“muson ma How ousumuomfion Eoou um oowuo Edsoo> .m>onm Eoum onEMm m>aumucmmohmom

.usmfio3 who .mE oom n panTB uo3 .Em H
Ave

.usmflm3 omwho Essom> .mE omm u named? #03 .Em H

any
.mmmo NH mo mumcmmoeog Eonm onEMm m>wumucmmmummmAmV
wahIIcowumNHHHemoma cw mmoa ommucmouom

wmmII anawuo Edsow> CH mmoa ommucmouom

 

 

 

 

Ho.o H mo.o moo.o H. Ho.o Ame no.olmo.o No.0IHo.o “ovooNflHHSQOSA

moo.o H Ho.o moo.o H moo.o Ame Ho.oIoo.o Ho.OIoo.o Anvooauo Essom>

I moo.o_H mmo.o Ame I ao.oumo.o imvmmmm ammum

“nowoz who unmfloz #03 I usmfloz mun unmwmz umz ucofiuomua
com: mmcwm

 

Aammv cowumuucmocoo mwsonoz

 

mama canaaaamomq

one omwna Edsom> .nmmum mo mucoucou hunches Hmuoa mo GOmfiHmmfiou .oa GHQMB

43

Plant Food

Cereals and Legumes

 

Amounts of mercury found in cereals and legumes are
shown in Table 11. The overall mean of all values for cereals
was 0.02 with a range of 0.01 to 0.03 ppm. There was no
substantial difference in values between different strains
of the same grain. It has been reported that the residues
in ears of grain grown from dressed seed are very small.
Westermark (1967) found mercury levels in grains of wheat
or barley ranging from 0.008—0.012 ppm whether or not they
were grown from dressed seed, while Somers (1971) reported
lower ranges of 0.005-0.009 in Canadian wheat. Pappas and
Rosenberg (1966) analyzed wheat samples from various parts
of the U.S. and reported values ranging from 0.013-0.127
ppm. However, it was not ascertained whether the wheat was
grown from treated wheat or not.

In Japan, where organomercurials have been extensively
used in rice cultivation, Tomizawa (1966) found comparatively
high background levels of mercury of 0.227-0.238 ppm in rice
from unsprayed fields, while Smart and Hill (1968) reported
negligible levels (:_0.005) of mercury in rice imported to
U.K., occasionally rising to 0.01-0.015 ppm. It was not
ascertained whether the rice used in this study was of do-
mestic origin (U.S.) or was imported. However, levels of
mercury obtained were in agreement with those reported by
Smart. The mean value of 0.03 ppm obtained for legumes

was the highest level observed in plant products, but was of

44

Table 11. Total Mercury Content of Plant Foods--Cereals
and Legumes

 

Mercury Concentration )ppm)

 

 

 

 

 

Food Item Range Mean
Cereals
I
Wheat (Dickson) 0.01 —0.025 (6) 0.02 i. 0.01
Wheat (Beizes) 0.02 -0.03 (6) 0.02 i 0.005
Barley (Avon) 0.02 —0.025 (4) 0.02 :_< 0.005
Barley (Talbot) 0.02 -0.04 (6) 0.03 i. 0.01
Oats (Garry) 0.01 (5) 0.01 i < 0.005
Rice, polished
Long grain i 0.005-0.01 (7) 0.01 i 0.005
Mean value for Cereals 0.02 ppm
Legumes
Beans-—Navy 0.025-0.03 (6) 0.03 i, 0.005
Beans--Red Kidney 0.02 -0.035 (6) 0.03 i 0.01
Mean Value for Legumes 0.03 ppm

 

I
Figures in parentheses refer to the number of determinations.

45

the same order of magnitude as those reported by Corneleiussen

(1969) and Somers (1971) of 0.01 and :_0.02 respectively.

Fruits and Vegetables

 

Table 12 summarizes the results of analysis of fruits
and vegetables. Of 39 samples of fruit and vegetables ex-
amined in 4 to 6 replicate analyses, 24 gave mercury con—
centrations less than 0.01 ppm and they did not show sig-
nificant traces of mercury when considered individually.
However, these samples treated as a group showed a mean level
approaching 0.01 ppm revealing a detectable trace of mercury
in the majority of the samples.

In work on experimentally sprayed apples, residues ranging
from 0.02 to 0.12 ppm were reported by Martin and Pickard
(1957), while values for unsprayed controls were usually
i 0.01 ppm. Jacobs and Goldwater (1961) found residues of
0.07-0.06 ppm in sprayed Red Delicious apples while unsprayed
controls had only 0.01 ppm at harvest and values as high as
0.16 ppm earlier in the season. The latter values were re-
garded as being unusually high for background levels of mer-
cury.

Table 12 also shows a comparison of results obtained
from unwashed peel, washed peel and pulp samples of potatoes,
apples and tomatoes. The consistent difference between un-
washed and washed peel of potatoes and apples probably
represents surface contamination. The comparable values of

mercury in washed peel and pulp of these plant products

46

 

 

 

 

Table 12. Total Mercury Content of Plant Foods——Fruits and Vegetables
Mercury Concentration (ppm)

Food Item Range Mean

Fruits

Apples (Golden Delicious)

Flesh §_0.005-0.01 (6) 0.01 :_< 0.005

Peel washed §_0.005-0.01 (6) 0.01 :_ 0.005

Peel Unwashed 0.015-0.03 (4) 0.02 :_ 0.01
Apples (Johnathan)

Flesh 0.01 (6) 0.01 :_< 0.005
Strawberries 0.01 (6) 0.01 __< 0.005
Tomatoes

Flesh :_0.005—0.01 (5) 0.01 :_< 0.005

Peel Washed 0.01 (5) 0.01 i_< 0.005

Peel Unwashed 0.01 -0.02 (5) 0.015 :.< 0.005
Vegetables
Potatoes (Russet)

Flesh 0.01 -0.015 (4) 0.01 :_< 0.005

Peel washed 0.01 -0.02 (6) 0.015 :_ 0.005

Peel Unwashed 0.02 -0.03 (6) 0.03 :.< 0.005
Potatoes (Merrimack 528)

Flesh §.0.005-0.010 (6) 0.01 :.< 0.005

Peel Unwashed 0.01 - .020 (6) 0.015 :_< 0.005
Asparagus 0.01 -0.015 (6) 0.01 :_< 0.005

 

Mean Value for Fruits and Vegetables-0.01

 

fl

Figures in parentheses refer to number of determinations.

47

indicates the lack of a distribution differential at back-
ground levels. Smart (1964) extended this observation to
treated crops as well and noticed no difference in residues
in peel and flesh of potatoes receiving foliar sprays in
the growing season. He further adduced that the residue
present in peel was evidence that soil mercury could be
taken up by the skin of the tuber as well as by the root

system.

Bread, Salt and Sugar

 

Results of analysis of bread, salt and sugar are shown
in Table 13. Of the three subsidiary foods analyzed, salt
was the only dietary item found to have levels of mercury
higher than most of the other foods examined. This is of
further interest since salt was the only mineral food ex-
amined and its higher mercury content may be related to geo-
logical levels of mercury. It is also probable that the
higher values recorded could have resulted from interfering
absorption (at 253.7nm) by some inorganic or organic come
ponent or additive of the salt. The values obtained, how—
ever, also necessarily incorporate a possible negative error
arising from incomplete reduction of the mercury in the final
reduction step by presence of traces of chlorine liberated

from the previous KMnO oxidation. The high degrees of vari-

4
ation between replicate readings was thought to result from
the varying degrees of incomplete reduction due to the

presence of residual chlorine. Further experimentation with

48

Table 13. Total Mercury Content of Foods: Sugar, Salt and

 

 

 

Bread
Mercury Concentration (PPm)
Food Item Range Mean
I

Sugar (White) ( i) 0.01 (6) 0.01 t:_ 0.005
Sugar (ii) 0.01 (6) 0.01 :_< 0.005
Salt (non-iodized) 0.03 -0.09 (6) 0.06 i, 0.02
Bread (White) : 0.005-0.01 (6) 0.005 :_< 0.005

 

I
Figures in parentheses refer to number of determinations.

49

Table 14. Summary of Base-Line Levels of Mercury in Major Food Groups--
Comparison with Reported Values

Food Class

Mercury Concentration (ppm)

(a)

Reported Mercury Conc.

Quan)::.30%
Minimum Maximum

 

Red Meat
(Pork and Beef)

White Meat
(Chicken)

Eggs
Fish
Milk

Liver
(Pork and Beef)

Cereals
Legumes

Fruits and
Vegetables

I
(18)

( 6)
( 6)
(17)

(21)

(18)
(31)

(12)

(39)

0.01

0.024

0.010

0.17

0.01 0.002

0.02

0.02 0.02

0.03
0.002

0.01

0.02

0.05

0.010

 

(a)

major cities in the U.S.

Mercury levels (by food class) reported by Corneleiussen (1969) for

Figures in parentheses refer to the total number of samples analyzed.

50

technique is required to completely eliminate all traces of
chlorine before the final reduction step, to verify the
results of this study.

Levels of mercury obtained for sugar and bread are con-
sistent with those reported by other workers (Lee gt_§l,,

1972).

Methodology

Comparison of Mercury Recoveries from Digestion PrOcedures
I, II and III

 

 

The results of comparative recoveries of mercury from
fish and egg with digestion procedures I, II, and III are
shown in Table 15. Already monitored reference samples of
fish (of known mercury content) were subjected to the three
digestion procedures. In digestion I and II samples were
analyzed at intervals of 1 hour and 12 hours following
digestion.

The results of Table 15 indicate comparable recoveries
of mercury from fish, with digestions I and II. Concentrations
of mercury in both cases showed excellent agreement with the
monitored value. Mercury concentrations in fish and egg
samples, obtained with digestion III, were lower by over
50 percent than those from digestions I and II and for this
reason digestion III was considered inadequate. The lower
results are probably attributable not to losses of mercury
in digestion but to incomplete release of organically bound

mercury in the foods examined when HNO (35%) was used alone.

3

51

.hnsoume Ema ¢N.o Gamucoo
UHEoud we ..bwq noumommm

on .cowvwmmfio v

oczxueommm cam muumeouonmouuommm
umpmz mo manuaumcH .s.m.2 may we omuouacos swam

GOHHQHOQO
mo moHQEmm

 

 

 

 

 

Abe
.moumoaammu m on v mo coo: Ame
Awmme
. m
we o I ozm
I mm~.o .mfla scum mason NH Hum >\>
I A~.o .mao soup “so: H sommmumozm HH .mao
Ho.o mH.o .mao some muse: NH i>\3 ammo
I mH.o .meo Eoum H50: a mozm HHH .mHo
mmo.o mm.o .mflo Scum meson ma hem.a .Hw.mmv
I mo.o .mfln Eoum Moon a wOmmm H .mHQ
mama names .wflxo v05,2 weave ooeumz
AMVAEmmv coaumuucoocou hunches coflumomao
mousooooum coflpmmmwo momma an conflEHmumo mmmm one Swan Cw musouoz mo mam>oq .mH magma

52

A difference in mercury concentrations between samples
of fish, at 1 hour and 12 hour intervals from digestion,
was observed with digestion I. Final oxidation with KMnO4
was carried out on the steam bath instead of by direct heat-
ing on a hot—plate as in the Barret (1956) and Dow (1970)
procedures and these oxidation conditions were probably not
vigorous enough to complete the oxidation. Oxidation was
apparently completed in the 12 hour standing period as sug-
gested by the higher mercury concentration at the end of
this interval. Overnight standing of the digests from di-
gestion I before final analysis was therefore adopted as a

routine practice.

Recovery Results

 

The recovery data for digestion procedures I and II
using methylmercuric chloride are shown in Tables 16 and 17.
Table 16 shows the recovery data obtained with digestion I
using 1 gm. samples of chicken meat spiked with methyl-
mercury standards in the range 0.01 to 0.2 pg. mercury. An
average recovery of 83 percent was obtained in this range.
Recoveries of mercury from methylmercuric chloride spiked
samples of barley, using digestion II are Shown in Table 16.
Average recoveries by this method were 87 percent. Recovery
of mercury from mercuric chloride spiked samples of food
are shown in Table 18. An average recovery of 97.6 percent
of mercury added to milk, barley, potatoes, beans and chicken

meat was obtained.

53

Table 16. Recovery of Mercury Added As Methylmercuric Chloride to 1 gram
Samples of Chicken Meat (Leg Meat) by H2804:KMnO4Digestion--

Digestion I

 

#1

pg. Hg. Net pg. Hg pg. Hg Recovered Percent Recovered

 

 

 

Added Estimated MaximumTMinimum2 Maximum Minimum Mean % Recovered

0.01 0.035 0.01 0.005 100 50
0.030 0.005 0.00 50 - 63

0.035 0.01 0.005 100 50

0.035 0.01 0.005 100 50

0.03 0.055 0.03 0.025 100 83
0.060 0.035 0.030 116 100 91

0.055 0.030 0.025 100 83

0.050 0.025 0.020 83 66

0.05 0.07 0.045 0.040 95 80
0.08 0.055 0.050 110 100 98

0.08 0.055 0.050 110 100

0.075 0.050 0.045 100 90

0.10 0.110 0.085 0.070 85 70
0.100 0.075 0.070 75 70 79

0.110 0.085 0.075 85 75

0.115 0.090 0.085 90 85

0.20 0.20 0.175 0.170 88 85
0.220 0.195 0.190 98 95 84

0.190 0.165 0.160 83 80

0.180 0.155 0.150 78 75
Average Recovery 83%

 

lNet-minimum blank reading--.025 pg.
2Net-maximum blank reading--.030 pg.

Blank includes a reagent blank and mercury content of the chicken meat.

Table 17.

54

Recovery of Mercury Added As Methylmercuric Chloride to

1 Gram Samples of Barley-—(HNO

3

:H SO

2

4

) Digestion II

 

pg. Hg Net pg. Hg pg. Hg Recovered Percent Recovered

 

 

 

 

 

Added Estimated Maximum1 Minimum2 Maximum Minimum Mean % Recovered
0.01 0.045 0.01 0.005 100 50
75
0.045 0.01 0.005 100 50
0.03 0.06 0.025 0.020 83 66
83
0.065 0.030 0.025 100 83
0.05 0.08 0.045 0.040 90 80
90
0.085 0.050 0.045 100 90
0.10 0.135 0.100 0.095 100 95
92
0.125 0.090 0.085 90 85
0.20 0.230 0.195 0.190 97 95
95
0.225 0.190 0.185 95 93
Average Recovery 87%
1Net pg. Hg--minimum blank reading-0.035 pg.

2Net pg. Hg--maximum blank reading-0.040 pg.

Blank includes a reagent blank and mercury content of the barley.

55

 

 

 

m.m.H mm.hm wum>oomm ommum>m
mm a m
m.em mm o~.o AH .mHoe em m 6mm: cmoneo
mm a m m
om mm Ho.o AHH .mHov om m" ozm m>mz .mcmmm
em a m m
m.NOH HHH mo.o LHH .mHav om m" ozm mmoumpom
mOH a m m
m.mm om eo.o AHH .mHov 0m m“ ozm mmHumm
mOH v m
m.m0H AOH mo.o 1H .mHoe om m emHua madam xHHz
mno>ooom. ooum>ooom wooed mm .m: GOHpmomHa EmuH poom
w com: mm ucoouom

 

ooom mo monEmm Emnu H ou moHuoHsu UHHDUHTS m4 coupe mnsouoz mo

mum>oomm .mH mHQme

56

The variation in replicate recovery values at the 0.01 pg.
level need to be interpreted in terms of limitations of in-
strument readings, on account of which a difference of scale
reading of 0.005 pg. results in a‘: 50 percent difference
in recovery.

Results of interlaboratory comparison of the mercury
content of fish samples is represented in Table 19. Excellent
agreement between the values obtained (using atomic absorp-
tion spectophotometry and H2804:KMnO4 digestion) reflects
the accuracy of the analytical techniques and the degree of

instrument reliability.

57

.coHpmmmHo 0ocsmueommm
mcho SHHoEowozmouuoomm GOHHQHOmnm UHEou< we .QoH souwommm umuma um omuouHcoE monEmm

 

 

 

3v

.mmum0HHmmH mo Hones: 0p Home“ mommeucoumm CH mousmHmAHv

.D.m.z .muoumnonmq noummmom “mum: mo ousuHumcHAmv

00.0 000.0.“ 0~.0100 0~.0I~m.0 H .0Ha mmmm eusosmmumq

0H.o oo.o H_mH.o Ame mH.oINH.o H .mHQ noumm BoHHow

I- H0.0 H 0H.0 100 0H.0I0H.0 HH .maa 0660

cmeESm .anmcsm

HH.0 00.0 H ~H.0 100 0H.0IHH.0 H .mHa 066m

cmeEsm .anmcom

coo: .cwoz mmcmm coHummmHQ anm
Haioouoymcoz HHV omcHEHoqu

 

Hemmv COHpmuucoocou musouoz

 

Hammv anm mo :oHponucmocoo wnsouoz mo GOmHHmmEoo SHOHMHOQMHHTDGH .mH oHnoB

A3

GENERAL DISCUSSION

The primary objective of the study was the determina-
tion of trace mercury background levels in foods of Michigan
origin, in which there was no direct exposure to mercury
contamination. The results indicate that practically all
the foods examined contained traces of mercury and for the
most part, these levels were barely above detection limits.
The results of this study are consistent with those of sur-
veys carried out in the U.S. (Corneliussen, 1969) and more
recently in the U.K. (Ministry of Agriculture, Fisheries
and Food, U.K., 1971), indicating no overt contamination of
foods above the base levels ranging from 0.01 to 0.03 ppm.
The levels detected probably represent a background level of
naturally occurring mercury derived from soil, ground water
or atmospheric sources.

However, it is difficult to differentiate between mercury
of social and industrial origin, from the naturally occurring
geologically related concentrations of mercury in nature.
The wide variations in values of background mercury levels
in soils, reported by different workers illustrates this.
Martin (1963) reported natural background levels of some
English soils to be in the range of 10-60 ppb while Anderson

(1967) reported values of 20-920 ppb in Swedish soils.

58

59

Kimura and Miller (1962) found levels of 116 ppm for soils
from Washington, in the U.S. Relatively high mercury con-
centrations were reported in soil samples taken from Central
Michigan (U.S. Dept. of Interior, 1971), in the range 200-
1,500 ppb as compared to a 71 ppb national average.

The major atmospheric sources of mercury contamination
arise from the burning of fossil fuels such as shale oil and
coal. Mercury from such sources could be transported con-
siderable distances by the wind and be deposited on vegetation
situated windward from such sources. Atmospheric concen-
trations of mercury further add to soil and ground water
levels by being washed down in rain water. Erikson (1967)
reported concentrations of mercury in rain water ranging from
0.000-0.20 ppb. Background levels in ground water have
been reported by Dall'aglio (1968) in Italy ranging from
0.01-0.05 ppb, while Hinkle and Learned (1969) reported higher
ranges of 0.2-0.7 ppb in the U.S. While soil and ground water
levels of mercury contribute in greatest measure to the
background levels of systemic mercury in foods, mercury
from atmospheric sources cannot be entirely discounted par-
ticularly in relation to physical contamination of exterior
surfaces and entrappment. Additionally, many parameters
have been found to influence the movement of mercury and metal
ions in the environment, including pH, temperature, presence
of biological agents, etc. Variations among these and other

unknown parameters can increase the mercury available to a

60

system many times beyond the normal background level
(D'itri, 1972).

The foregoing evidence offers a basis for the presence
of a background level of mercury in most common foods. It
is important therefore to set and evaluate tolerance levels
of mercury in relation to background levels inasmuch as it
is important to continuously monitor foods and food sources
to ensure that the background levels are well within safety
standards.

The overall mean of mercury concentrations in animal
products was higher than that in plant products. The gen-
erally higher background level of mercury in animal products
is probably a result of bioaccumulations within the food
chain. A recent survey of mercury levels in foods (Ministry
of Agr. Fisheries & Food, 1971) carried out in Britain in—
dicates that mercury in animal feeds was higher in compari-
son to most human foods specially in fish meals and some
grains. It is conceivable that animals could concentrate
the mercury derived from such feed sources.

Dried plant products, yiz cereals and legumes had
higher levels in comparison to fruits and vegetables. This
latter difference could be a reflection of the differing
water contents of these two categories of food or could be
related to the higher protein content of the cereals and
legumes as compared to fruits and vegetables.

Though different species of animal and plant materials

were analyzed there was no substantial difference in levels

61

between the different species except in the case of
fish.

The results of the present study also confirm the
findings of Pillay 25 El. (1971) in relation to mercury losses
in vacuum drying and freeze drying of samples as a preliminary
step to digestion. Neither of these treatments is therefore
desirable for the concentration of samples prior to digestion.
However, the losses of mercury in freeze-drying may find
application in the preparation of mercury-free fish-meals or

fish protein concentrates from mercury contaminated fish.

Methodology

 

While the main objective of this study was the evaluation
of mercury contents of foods, it was necessary to develop
a suitable analytical method for determining sub-microgram
quantities of mercury in a wide variety of organic matrices.
A rapid, simple and adequately sensitive analytical procedure
capable of being applied to the routine monitoring of a wide
variety of matrices was sought.
The basic demands on the analytical method were
1. Use of a small sample size
2. Limited digestion time
3. Moderate digestion conditions
4. Use of simple equipment, permitting several
samples to be processed concurrently with
little attention.
Choice of sample size was dictated by the need to

minimize reagent volumes and digestion time, while permitting

62

the use of small 100 ml. capacity digestion flasks, without
loss of material in the initial reaction with the digesting
acids. After preliminary experimentation a sample size

of l to 3 gm. was found optimum both in terms of digestion

time and of giving results within the limits of sensitivity

of the method.

Digestion

 

The H SO digestion was found applicable to most animal

2 4
products except beef liver, pork liver and cheese. .While
liver samples tended to foam in the final reduction-aeration
step either from incomplete breakdown and/or from the forma-
tion of surface active compounds, cheese samples gave too
high a residue of undigested fat in acid digestion. Though
the latter limitation was observed also with eggs, both
digestions II and III yielded much lower recoveries of mercury
from eggs. The mixed acid digestion (Digestion II) was how-
ever used for the analysis of both liver and cheese.

Due to the high content of carbohydrate in plant materials,
Digestion I generally resulted in charring and carbonization.
Mixed acid digestion was therefore used on all plant materials.
Dried plant materials, notably cereals and legumes, reacted
violently in the early stages of digestion when concentrated
HNO3 (Sp. Gr. 1.47) was used. Lang and Nelson (1942) pointed
out that a violent reaction and excessive foaming in the

early stages of the reaction may lead to losses of mercury.

Abbott and Johnson (1957) showed that mercury may be carried

63

away though the digestion mixture by the carbon dioxide which
was evolved, but Gorsuch (1959) working with different ex-
perimental conditions found no such losses. Nevertheless

in all cases, 35 percent HNO in the mixed acid digestion

3
was found to give a smoother reaction with minimal foaming
and was used in place of concentrated HNO3. Digestion was

started with HNO to a point of partial dissolution before

3
addition of H SO4 to moderate the initial reaction that

2

otherwise occurred on addition of the mixed acids together.
Pickard and Martin (1960) used the same expedient in the
digestion of tomatoes and coffee beans. Cooling of the
digest in ice during the addition of the H2504 followed by
gradual heating eliminated the sudden and violentevolution
of nitrous fumes which otherwise occurs.

The use of selenium as a fixative to prevent losses
of volatile mercury has been reported by a number of workers
(Kunze, 1948; Abbott and Johnson, 1957), with analyses involv-
ing dithizone complexing and colorimetric estimation of
mercury. Limited experimentation with selenium suggest its
use is not applicable when analysis is accomplished by the
flameless atomic absorption technique, because of spurious

high signals either from absorption by selenium or some

volatile compound of selenium at 253.7nm.

Oxidation
The oxidising agents normally used in the wet digestion

procedure are HclO4, H O and KMnO Gorsuch (1959) first

2 2 4'

64

found that mixed HNO3:HZSO4:HclO4 acid could be used in de-
termining traces of mercury in foods. The method was also

proposed by the Analytical Methods Committee of the Society
for Analytical Chemists (1960) but was not recommended for

trace residue levels due to losses arising from the forma-

tion of volatile chloro compounds.

Polley and Miller (1955) used H202 in microdetermination
of mercury in soils and biological materials. The method
was investigated by Kudsk (1964) who reported low recoveries
as did Campbell and Head (1955). In this study, poor re-
coveries of mercury added as mercuric chloride in the 0.05-

0-15 pg. range, were obtained when H202 oxidation was tried

in preliminary experimentation with wheat.

Equipment

 

A number of authors have stressed the necessity of
using special digestion vessels with complicated condenser
and recovery systems in order to avoid losses of mercury.
Kudsk (1964) recommended that a Leibig condenser was com-
pletely adequate when used in conjunction with long-necked
digestion flasks. Lee and Laufman (1971) used capped centri-
fuge tubes for digestion of 1 gm. samples of paper pulp with
aqua regia, while Hoover 32 EL- (1970) used 125 ml. Erlenmeyer
flasks and heated directly on a hot-plate with no attached
recovery systems. They were able to obtain recoveries of
95-100 percent of added mercury. The recoveries obtained

in this study under different digestion conditions indicated

65

that the use of long-necked volumetric flasks with moderate
digestion temperatures removes the need for elaborate re-
covery equipment, thereby permitting the analysis of a large

number of samples at one time.

Recovery

There is only limited knowledge of the possible variation
in chemical forms of mercury present in various foods. Re— -
covery data using methylmercury spiked samples of food served
as an indication of the extent of methylmercury breakdown
as well as losses of mercury in digestion. The recovery data
from the mercuric chloride spiked samples served as an indi-
cation of the extent of mercury losses in digestion. Organic
and inorganic compounds of mercury were used in recovery
studies to obtain an estimate of the extent of mercury release
from organically bound forms and the extent of losses from
volatilization during the digestion process. However, the
use of one or the other chemical form of mercury in re-
covery studies does not necessarily indicate the extent
of breakdown or release of mercury from the forms actually
present in the sample. The differences in recoveries of
mercury from eggs and fish under identical digestion conditions
serves to illustrate this. Recovery data have to be applied

and interpreted with a recognition of this limitation.

Precision and Sensitivity

 

Good agreement with results of total mercury evaluation

on reference samples of fish, in interlaboratory comparative

66

studies served to confirm and validate the accuracy of the
analytical and instrumental procedures. The variation in
the standard deviation of the replicate analyses reflect
only partly the sampling and experimental errors. Since all
readings and calculations based thereon were approximated
to the closest 0.005 pg., the standard deviation also re-
flects the limitations of instrument calibration. For this
reason it was not attempted to differentiate results below
0.005 pg. with any degree of certainty. Maximum deviation
of :_50 percent was observed in replicate analyses at the
0.01 pg. level. However, this was thought to be acceptable
in view of the low levels of determination and for the limits

of precision sought in this study.

SUMMARY AND CONCLUSIONS

The results of this limited survey of Michigan foods
indicate that practically all the foods sampled contained
detectable traces of mercury and that a background level
of a very low order ranging from 0.01 to 0.03 ppm is commonly
present. The only significant levels of mercury observed
occurred in fish, though these levels were within the
F.D.A. tolerance limit of 0.5 ppm. From the data collected,
levels of mercury in fresh fruits and vegetables were con-
sistently i 0.01 ppm. In comparison levels in cereals and
legumes were in the range 0.01 to 0.03 (mean value 0.02)
ppm and in the animal products levels were in the range
0.01 to 0.17 ppm with a mean of 0.03 ppm. From the samples
analyzed it appears that levels in fruits and vegetables
tend to be very low and at or below the lower limits of
detection, while those in cereals and legumes are higher
and were comparable to those found in animal products.
Whether these differences are statistically significant has
to be established by more large-scale and statistically valid
sampling. While the levels of mercury in foods observed
were within the 0.05 ppm guideline (WHO, 1963) for mercury
in foods, they suggest that a zero tolerance would be unreal—

istic. The results of this study should permit some estimates

67

68

of the average daily intake of mercury in combination with
data on the approximate composition of diets based on the food
items examined.

Losses of mercury were noted in the vacuum drying and
lyophilization of egg samples preparatory to analysis, pre-
cluding the use of these procedures for preliminary concen—
tration or pre-processing of samples to limit bulk or reduce
moisture content.

Quantitative monitoring of mercury in foods is handi-
capped largely by lack of suitable methods of analysis in-
volving a minimum of sample preparation without destruction
of sample and that would be adaptable to speedy handling of
a large number of samples.

Modifications in methodology were adopted towards these
ends. The main features of the methodology used were

1. The use of moderate and controlled digestion
temperatures throughout the digestion.

2. The use of conditions minimizing the rate of
initial reaction of sample and digesting acids.

3. Use of simple digestion equipment enabling a
large number of samples to be handled concurrently.

4. Use of 35 percent HNO in place of concentrated

3
HNO3 in the mixed acid digestion.

5. Use of one reaction vessel throughout the
digestion and subsequent analysis up to the.
point of reduction-aeration in the aeration

flasks, thus minimizing manipulative losses and

contamination possibilities.

69

6. Maintaining the digest under oxidized condi-
tions in excess of KMnO4 up to the final re-
duction step thus minimizing losses of mercury
(at very low dilutions) in the standing period
between digestion and reduction-aeration
(Uthe €3.2l3' 1970)

Results of recoveries of added mercury with the diges-
tion techniques used, on a range of food materials indicates

that the methods have merit for determining trace mercury

levels in foods.

 

LITERATURE CITED

LITERATURE CITED

Abbott, D. C., and E. I. Johnson, 1957. The determination
of traces of mercury in apples. Analyst 82: 206.

Analytical Methods Committee, 1960. Report prepared by
the metallic impurities in organic matter. Sub-
Committee. 85: 643.

Andersson, A., 1967. Mercury in Swedish soils--The Mercury
Problem Chem. Abstracts 69: 51225 j. Swedish.

Aomine, S., H. Kawasaki and K. Inque, 1967. Mercury resi-
dues of Paddy and orchard soils. Soil Science and
Plant Nutr. 13: No. 6, 186.

Bailey, E. H., P. D. Suavely and D. E. White, 1961. Chemical
Analysis of Brine, Crude Oil--Cymric, Ken County,
Calif. U.S. Geol. Surv. Paper 424 D, D 306-D 309.

Barret, F. R., 1956. Microdetermination of mercury in
biological materials. Analyst. 81: 249.

Beasley,T. M., 1971. Mercury in selected fish protein
concentrates. Envir. Sci. and Tech. 5: 634-635.

Bertine, K. K. and E. D. Goldberg, 1971. Fossil fuel combus-
tion and major sedimentary cycle. Science 173: 233.

Bligh, E. G., 1972. Mercury in Canadian fish. Journ. Inst.
Can. Sci. & Tech. 5: A6.

Booer, J. R., 1944. Behavior of Mercury Compounds in Soil.
Ann. Appl. Biol. 31: 340.

Borg, K., H. Wanntorp, K. Erne and E. Hanko, 1966.
Alkylmercury poisoning in terrestrial Swedish wildlife.
J. Applied. Ecol. 3, Suppl., 171.

Campbell, E. E. and B. M. Head, 1955. The determination of
mercury in urine, single extraction method. Amer.
Indus. Hyg. Ass. Quart. 16: 275.

Corneleiussen, P. E., 1969. Pesticide Residues in Total
Diet Samples (iv) Pest. Monit. Jour. 2: 140.

70

71

Curley, A. C., V. A. Sedlah and E. F. Girling, 1971. Organic
mercury identified as the cause of poisoning. Science
72: 65.

Dall'-aglio, M., 1968. The abundance of mercury in 300
natural water samples from Tuscany and Latium, in
'Origin and Distribution of the Elements.‘ Pergammon
Press, Oxford. Ed. L. H. Ahrens, 1065-1081.

D'Itri, F. M., 1972. The Environmental Mercury Problem.
CRC Press, Cleveland, Ohio, 124 pp.

Dow Chemical Co., 1970. The determination of Mercury (Atomic
absorption spectrophotometric method) CAS-AM-70. 13,
P 13.

Epps, E. A., 1966. Colorimetric determination of mercury
residues on rice. Journ. Ass. Off. Agr. Chem. 49::
793.

Erickson, E., 1967. 'Mercury in Nature' in The Mercury
Problem. Oikos (Suppl.) 9: 13.

Eyl, T. B., 1971. Organic mercury food poisoning. New
Eng. Journ. of Med. 284: 706.

Fiskesjo, G., 1970. The effect of two organic mercury
compounds on human leukocytes in vitro. Hereditas 64:
142-146.

Furtani, S. and Y. Osajima, 1966. Residual components from
agricultural chemicals II Mercury in Rice. Sci. Bull.
Fac. Agr. Kyushu Univ. 21: 363.

Goldwater, L. J., 1964. Occupational exposure to mercury-—
Lecture 'Historical Background' in the Harben Lectures,
Roy. Inst. of Pub. Health and Hyg. 27: 279.

Goldwater, L. J., 1971. Mercury in the Environment.
Scientific American 224: 15-21.

Gorsuch, T. T., 1959. Radiochemical investigations on the
recovery for analysis of trace elements in organic and
biological materials. Analyst. 84: 135.

Hannerz, L., 1968. Experimental investigations on the
accumulation of mercury in water organisms. Inst. of
Water Res. 48: 120-176.

Hatch, W. R. and W. L. Ott, 1968. Determination of submicrogram
quantities of mercury by atomic absorption spectro-
photometry. Anal. Chem. 40: 2085—2087.

72

Hinkle, M. E. and R. E. Learned, 1969. Determination of
mercury in natural waters by collection on silver
screens. U. S. Geol. Surv. Prof. Paper 650-D,
D251-D254.

Hoover, N. L., J. R. Melton and P. Howard, 1970. 'Determina-
tion of trace amounts of mercury in foods by flame-
less atomic absorption. Jour. of the A.O.A.C. 54:
No. 4, 1971.

Imura, N., E. S. Shoe-Kung, Pan., K. N. Jong Yoon Kim and
T. K. T. Ukita, 1971. Chemical methylation of inorganic
mercury with methylcobalamine or Vit. B12 analog.
Science 172: 1248.

Irukayama, K., 1966. The pollution of Minamata Bay and
Minamata Disease. Adv. Water Pollut. Res. 3: 153-165.

Jacobs, M. B. and L. J. Goldwater, 1961. Ultramicrodeter-
mination of mercury in apples. Food. Technol. 15:
357.

Jeffus, M. T., J. S. Elkins and C. T. Kenner, 1970. Deter-
mination of mercury in biological materials. Jour.
of the A.O.A.C. 53: No. 6, 1172-1175.

Jensen, 8., and A. Jernelov, 1969. Biological methylation
of mercury in aquatic organisms. Nature 223: 753.

Jernelov, A., 1968. Conversion of mercury compounds in
'Chemical Fallout.‘ Edtd. by M. W. Miller and G. G.
Berg. Charles Thomas Publishers, Springfield, Illinois.

Jervis, R. E., D. Debrun, W. LePage and B. Tiefenbach, 1970.
Mercury residues in Canadian foods, fish and wildlife.
Reprint of a paper presented at the 'Trace minerals in
Environmental Materials' Conference.

Johnels, A. G. and T. Westermark, 1969. Mercury contamina-
tion of the environment in Sweden' Chemical Fallout
edited by M. W. Miller and G. G. Berg. Charles C. Thomas,
Publisher, Springfield, Illinois.

Johnels, A. G., 1967. Cited in "A review of health hazards
and side effects associated with the emission of mercury
compounds into natural systems." Rept. Natural Sci.
Res. Council, Stockholm, Sweden, by G. Lofroth, 1969.

Johnels, A. G., 1968. Cited in "A review of health hazards
and side effects associated with the emission of mer-
cury compounds into natural systems." Rept. Natural
Sci. Res. Council. Stockholm, Sweden, by G. Lofroth,
1969.

 

73

Kimura, Y. and V. L. M. Miller, 1962. Mercury determina-
tion at the microgram level by a reduction-aeration
method of concentration. Anal. Chem. Acta. 27: 325.

Kiwimae, A., A. Swensson, U. Ulfvarsson and G. westoa, 1969.
Methylmercury compounds in eggs from hens after oral
administration of mercury compounds. Agr. Food Chem.
17: 1014.

Kudsk, F. N., 1964. Determination of mercury in biological
materials. Scand. J. of Clin. & Lab. Invest. 16: 575-
583.

Kunze, F. M., 1948. Addition of Selenium to wet-ash procedures
for the determination of mercury in apple peel. J.
Ass. Off. Agr. Chem.

Landner, L., 1971. Biochemical model for the biological
methylation of mercury suggested from methylation studies
in vivo with Neurospora -rassa. Nature 230: 452.

Laug, E. P. and K. W. Nelson, 1942. Report on Mercury. J.
Ass. Off. Agr. Chem. 25: 399.

Lee, D. C. and C. W. Laufman, 1971. Determination of
sub-microgram quantities of mercury in pulp and paper
board by flameless atomic absorption spectrometry.
Anal. Chem. 43: No. 8, 1127.

Lee, D. F., B. Thomas, J. A. Roughan and E. D. waters, 1972.
Mercury content of some foodstuffs of vegetable origin.
Pest. Sci. 3: 13-17.

Lindstedt, G., 1970. A rapid method for the determination
of mercury in urine. Analyst. 95: 2641.

Lofroth, G., 1969. A review of health hazards and side ef-
fects associated with the emission of mercury compounds
into natural systems. Rept. Natural Sci. Res. Council,
Stockholm, Sweden.

Magos, L., 1971. Selective atomic absorption determination
of inorganic mercury and methylmercury in undigested
biological samples. Analyst 96: 847-853.

Martin, J. T. and J. A. Pickard, 1957. Spray application prob-
lems XLIII. Mercury deposits on apple fruits and
foliage. Ann. Rept. Agr. Hort. Sta. Long. Ashton,
Briston, p. 76.

Martin, J. T., 1963. Mercury residues in plants. Analyst.
88: 413.

74

Ministry of Agriculture, Fisheries and Food, 1971. Survey
of mercury in foods. Her Majesty's Stationary Office,
Lond.

Pappas, E. G. and L. A. Rosenberg, 1966. Determination of
sub-microgram quantities of mercury in fish and eggs
by cold vapour atomic absorption photometry. Ass.
Off. Anal. Chem. 49: 782-792.

Pickard, J. A. and J. T. Martin, 1960. Determination of
mercury in plant material. J. Sci. Food and Agric.
11: 374.

Pillay, K. K. S., C. C. Thomas, J. A. Sandell and C. M. Hyche,
1971. Determination of mercury in biological and en-
vironmental samples by neutron activation analysis.

Anal. Chem. 43: No. 13, 1417.

Polley, D. and V. L. Miller, 1955. Rapid microprocedure
for determination of mercury in biological and mineral
materials. Anal. Chem. 27: 1162.

Reigo, J., 1970. Mercury and fluorine content in Swedish
milk. Svenska Mejeritidu. 62: 312.

Ross, R. G. and D. K. R. Stewart, 1962. Movement and accuum-
lations of mercury in apples, trees and soil. Can. Jour.
Plant. Sci. 42: 280.

Ross, R. G. and D. K. R. Stewart, 1960. Mercury residues on
apple fruit and folliage. Can. Jour. Plant Sci. 40:
117.

Saha, J. G., Y. W. Lee, R. D. Tinline, S. H. F. Chinn andv
H. M. Austensen. 1970. Mercury residues in cereal
trains from seeds or soil treated with organomercury
compounds. Can. Journ. Plant. Sci. 58: 597-599.

Smart, N. A. and M. K. Lloyd, 1963. Mercury residues in eggs,
flesh and livers of hens fed on wheat treated with
methylmercury dicyandiamide. J. Sci. Food and Agr.

14: 734.

Smart, N. A., 1968. Residues of mercury compounds. Residue
Rev. 23: 1.

Smart, N. A. and A. R. C. Hill, 1968. Pesticide residues in
foodstuffs in Great Britain, vi. Mercury residues in
rice. J. Sci. Food. Agr. 19: 315-316.

75

Somers, E., 1971. Mercury contamination of foods. Proc.
of Roy. Soc. of Canada, International Symposium on
'Mercury in mans environment,‘ p. 99.

Stock, A. and F. Cucel, 1934. Die Verbreitung des Quecksilbers.
Naturwissenschaften. 22: 390.

Stoewsand, G. S., J. L. Anderson, W. H. Guteman, C. A. Bache
and D. J. Lisk, 1971. Metabolic studies of mercuric
chloride in developing Japanese quail. Sci. (In Press).
Cited in "Mercury its occurrence and effects in the eco-
system" by Peakall, D. B. and R. Lovett, 1972. BioSci.
22: No. 1, 20-25.

Stone, H. M. P., P. J. Clark and H. Jack, 1957. Mercury
content of apples. New. Zea. J. Sci. & Techn. 38:
843.

Szkolnik, M., K. D. Hichey, E. J. Broderich and D. J.
Lisk, 1965. Mercury residues of apple fruit as related
to application schedule and to colorimetric and neutron
activation analysis. Plant Dis. Rept. 49: 568.

Tejning, S. and R. Vesterberg, 1964. Alkyl mercury treated
seed in food grain. Poultry Sci. 43: 6.

Tejning, S., 1967. (a) Biological effects of methylmercury
dicyandiamide treated grain in the domestic fowl. Oikos
(Suppl.) 8: 7-16. (b) Mercury in pheasants derived
from seed grain dressed with methyl and ethyl mercury
compounds. Oikos 18: 334-344.

Tomizawa, C. A. Kobayashi, M. Shibuya, Y. Koshimizu and Y.
Oota, 1966. Studies on residue analysis of pesticides
in plant materials II. Neutron activation analysis of
mercury in rice grain. Shokuhiu Eiesigaku Zasshi 7:
33. Chem. Abstr. 65: l4330f.

U.S. Bureau of Mines, 1970. Mineral Industry Surveys, U.S.
Government Printing Office, Washington, D. C.

U.S. Dept. of Interior, 1971. Cited by D'itri, F., 1971.
The Environmental Mercury Problem--Legislative Report,
p. 242.

Uthe, J. F., F. A. J. Armstrong and M. P. Stainton, 1970.
Mercury determination in fish samples by wet digestion
and flameless atomic absorption spectrophotometry. J.
Fish. Res. Bd. of Canada 27: 805-811.

76

Westermark, T., 1967. Private Communication. Cited in
Residues of mercury compounds by Smart, N. A. Residue
Reviews 23: l.

Westoo, G., 1966. Kvicksilver i agg fran svenska hons i koft
fran svenska hons broiler och kyckligar samti svensk
kycklinglever. Var. Foda 18: 85.

Westoo, G., 1966. Kvicksilver i koft och lever fran svin
kal och oxe samt ren. Var. Foda 18: 88.

WestoB, G., 1967. Determination of methylmercury compounds
in foodstuffs II. Determination of methylmercury in
fish, eggs, meat and liver. Acta. Chem. Scand. 21. 1790.

Westoo, G., 1968. Determination of methylmercury salts in
various kinds of biological material. Act. Chem. Scan.
22: 2277.

Westoa, G., 1969. Mercury and methylmercury levels in some
animal products. Var. Foda 7: 138.

Weiss, H. V., M. Loide and E. D. Goldberg, 1971. Mercury
in a Greenland Ice-Sheet. Evidence of recent input
by man. Science 1971, 174: 692.

WHO/FAQ, 1967. Pesticide residues in food. WHO Tech.
Rept. Ser. 370.

Wood, J. M., F. S. Kennedy and C. G. Rosen, 1968. Synthesis
of methylmercury compounds by extracts of methanogenic
bacterium. Nature 220: 173.

Zweig, G., 1963. Analytical methods for pesticides, plant
growth regulators and food additives. Vol. 1. Acad.
Press., Lond.

MICHIGAN STATE UNIV. LIBRQRIES

31293106823796