Environmental impact of cadmium: a review by the Panel on Hazardous Trace Substances.

This report is the result of a review by a 
Panel on Hazardous Trace Substances, as 
part of a report to an ad hoc Committee on 
Environmental Health Research whose chairman 
was Dr. David Rall, Director of the 
National Institute of Environmental Health 
Sciences, NIH.

The Panel undertook as one of its charges an in-depth examination of several groups of chemicals. This examination was aimed at defining sources of environmental contamination by these chemicals, there distribution in the environment, their transport and alternation, and their biological effects on humans and on other components of the biosphere.
The Panel has also taken the view that it will be important to develop quantitative means for understanding the patterns of the movement of these materials into, and their alteration and persistence within the biosphere; tentative models aimed at these objectives have been developed.
The first report by the Subpanel on polychlorinated biphenyls has been published (1). The present report on Cadmium is an extensive review of this element, prepared by the Subpanel on Cadmium, identified above. It has been reviewed in detail both by the Subpanel and by the entire Panel on May 1974 2J53 Hazardous Trace Substances. A number of suggestions that emerged from this review have been incorporated into the report. To this extent, the report is a collective effort. However, the chapters or parts of chapters were each the direct responsibility of one or more authors.
The sections on General Properties and Uses and Environmental Forms and Sources were contributed primarily by M. Fleischer; the section on Cadmium Material Flows was contributed primarily by A. F. Sarofim; the section on Effects on Man and Animals was prepared mainly by D. W. Fassett  General Summary Cadmium was recognized many years ago to be a highly toxic element, and the need for precautions in industrial operations in which workers were exposed to dusts and vapors of the element or its compounds had long been known. It was not, however, until comparatively recently that concern began to be expressed over the possible effects on human health of exposure over long periods to low concentrations of cadmium, in part because of its steadily increasing consumption and consequent increase in the general environment, and in part because of the outbreak of the itai-itai disease in Japan in the late 1940's and early 1950's. Exposure to cadmium, originating from smelter wastes and concentrated by the rice plant to levels far above those of the normal environment, has been shown to have been one of tne causative factors in this disease.
The present report reviews and assesses the information available to November 1972 on the environmental occurrence, transport, and biological effects of cadmium with special attention to significant gaps in our knowledge, and recommends areas of research for future studies.
Occurrence, Transport, and Biological Ef- fects Cadmium is a relatively rare element that is rather uniformly distributed in the most abundant rocks of the Earth's crust, which has an average content of 0.15-0.2 ppm Cd. It is slightly concentrated in shales, especially in those rich in organic matter, in lacustrine and oceanic sediments, in manganese nodules, and in marine phosphorites; the latter average about 25 ppm Cd. The only natural concentrations of commercial interest are those in sulfide deposits, especially those containing zinc, lead, and copper, from which it is recovered as a by-product. Total production in the United States and in the world reached all-time highs in 1969 of 5,736 and 17,576 metric tons, respectively. Production decreased in 1970 and 1971, but it has been estimated that consumption in the United States will nearly double by the year 2000 to a level of about 13,600 metric tons.
Cadmium metal has an appreciable vapor pressure, higher than that of zinc, at the temperatures used in smelting ores, in the manufacture of metallic alloys, and in the reprocessing of cadmium-containing alloys and of cadmium-plated materials. These processes account for about 90%o of the cadmium of atmospheric emissions, estimated by Davis et al. (2) to be about 2300 metric tons annually in the United States. The only other major sources of atmospheric emissions of cadmium are the burning of coal and oil, the burning of cadmiumweighted plastics, and the burning of sewage sludge. This estimate of total emission of cadmium to the atmosphere is in qualitative accord with published measurements of the cadmium content of the air in various parts of the United States, although data on Environmental Health Perspectives 254A residence time in the atmosphere are very scanty. The data on the contents of cadmium in soils near point sources of cadmium emissions, such as smelters and metallurgical plants, show that the fall-out has resulted in high concentrations of cadmium close to those sources.
Most fresh waters contain less than 1 ,Ag/ 1. Cd; sea water averages about 0.15. Data on the transport of cadmium in aqueous systems are too fragmentary to permit reliable balances of the flow of cadmium to be constructed, but they suggest that erosion and weathering of rocks and soils contribute far less cadmium to streams in the environment than that contributed by man's activities. Sewage sludge contains notable concentrations of cadmium, and the leaching of such sludge that has been used for land fill could contribute appreciable amounts of cadmium to the drainage system. Information is lacking on the behavior in soil and in drainage systems of the appreciable amounts of cadmium added to the soil in superphosphate fertilizers.
Plants exposed to concentrations of cadmium above those of normal background contain higher than normal concentrations of cadmium. Mosses appear to be especially good indicators of exposure to high concentrations. Damage to plants from excess cadmium has been reported, but the concentrations of cadmium required were higher than even those in soils of contaminated areas.
Only scattered data are available on levels of cadmium content in wild or domestic animals. The levels found in animals are generally much lower than those found in adult humans. No clear geographical correlation has been demonstrated between levels in herbivorous animals and levels in vegetation. In marine animals, the highest concentrations recorded have been in pelagic zooplankton (13 ppm, dry weight), molluscs (locally up to 73 ppm, wet weight), and plankton-eating birds (20-53 ppm, wet weight, in livers). There is no evidence that cadmium concentrates in marine food chains. Adverse effects on reproduction of fish have been reported at concentrations of cadmium similar to those of moderately polluted waters.
The average intake of cadmium by humans is generally estimated to be about  Ag/ day, mostly from food. Intestinal absorption is low, probably about 3-87o; the cadmium is notably concentrated in the kidneys and liver, which contain 50-75%o of the total body burden. A higher proportion of the cadmium reaching the respiratory tract is absorbed, but the total amount so absorbed is less than that from foods, except perhaps for smokers. The high content of cadmium reported for tobacco may cause smokers to have considerably higher body burdens of cadmium than nonsmokers. The total burden of cadmium in humans increases with age from very little at birth to an average of about 30 mg in the age range 40-50; it may decrease slightly after that.
Exposure to fumes or dusts of cadmium metal or cadmium oxide is known to cause acute pulmonary edema. Chronic exposure to cadmium through the respiratory tract produces a number of toxic effects, the most important of which is chronic emphysema, accompanied by renal disturbance. The effects noted in the itai-itai disease in Japan differ in many respects from the foregoing. The victims were almost all women over 50 who had borne several children; the disease was characterized by osteomalacia and osteoporosis, as well as renal damage. Although cadmium has been implicated as a causative agent, it seems probable that there was profound disturbance of calcium metabolism and that deficiencies of calcium, vitamin A, and vitamin D played important roles in the disease.
It has repeatedly been suggested that build-up of cadmium in the body (or perhaps an increase of the ratio cadmium/zinc) is related to the occurrence of hypertension in man. Experimental studies of rats and rabbits indicate that these animals develop hypertension after intake of cadmium orally or by injection. However, epidemiological studies of persons occupationally exposed to May 1974 cadmium and post-mortem studies of cadmium levels in the kidney have not yielded unequivocal statistical evidence of a relation between cadmium content or cadmium/ zinc ratio and hypertension.
Carcinogenic effects of cadmium have not been recognized in humans.
In summary, cadmium has not yet been proved to be a hazard to the average individual exposed to the average levels now present in the environment. The data from Japan, however, strongly suggest that part of the population is more sensitive to these hazards because of dietary deficiencies. Further research is needed on the hazards of average levels of cadmium to such individuals and on the hazards due to higher than average intakes of cadmium by persons who live in areas of high emission, who are heavy smokers, or who eat much shellfish. In the meantime, precautions are necessary to decrease emissions from the major sources and to insure that present levels of cadmium in the enviroment are not increased.

General Recommendations
One of the major purposes of this review was to identify gaps in our knowledge of the impact of cadmium in the environment and to suggest researches aimed at filling such gaps. Specific recommendations are made at the end of each chapter of this report. More generally, however, the Subpanel has found, as have other groups that have made similar reviews, that assessments of this kind are greatly hampered by a notable lack of interdisciplinary coordination of research on environmental pollutants. This is, of course, a natural consequence of the fact that so many disciplines and special fields of research, commonly with very limited contact with one another, are involved in such research, thus greatly increasing the difficulty of assembling the results into a coherent picture.
The Subpanel concurs with the Subpanel on Polychlorinated Biphenyls that there is a need for the development of general systems models to describe the transport of environmental pollutants. It proved to be difficult to prepare a model for the transport of cadmium which satisfactorily explained ambient concentrations of cadmium, particularly in surface run-off. However, new technological developments-such as the large increase in recent years of the production of cadmium-weighted plasticsemphasize the need for generalized models applicable to a variety of substances, yet flexible enough to meet changing conditions. Our review has also emphasized the need to study environmental problems as a whole, rather than in piecemeal fashion. Thus, the effects of emissions from metal smelters, a major source of emissions of cadmium, are the sum of the effects due not only to cadmium, but also to zinc, lead, arsenic, and sulfur dioxide, and possibly also to other elements such as copper and thallium, so that conclusions based on the study of a single element must be extremely uncertain. It is therefore important that such problem be considered as a whole, not only in systems models, but also in generalizing monitoring systems and in designing experiments.
As with the PCB problem, a major gap in our understanding of the cadmium problem is that we do not know how to apply data from laboratory studies on the toxicity of cadmium to its effects on animal populations in the natural environment. The difficulty of showing that adverse effects are occurring in the field indicates the need for fuller studies of the effects of environmental contaminants on animal communities, both in model ecosystems (microcosms) in the laboratory and in real ecosystems in the field. Similar difficulties in evaluating the consequences of low-level exposure of humans over long periods of time emphasize the urgent need for improved epidemiological studies of these effects. of eight stable isotopes of abundance: 106Cd, 1.22%b; 108Cd, 0.88%o; "0Cd, 12.39%o; 111Cd, 12.75 %; 112Cd, 24.07%o; 118Cd, 12.26%o; 1"4Cd, 28.86%o; 116Cd, 7.58.o-Like zinc and mercury, cadmium is a transition metal in Group IIb of the periodic table of elements. Cadmium and zinc, however, differ from mercury in that the latter has 14 additional electrons in the fourth orbital, which probably accounts for the high stability of compounds with mercury-carbon bonds, whereas the similar alkyl-cadmium compounds are extremely unstable, reacting rapidly with water and moist air under environmental conditions. Unlike mercury, cadmium and zinc show only valence + 2 in their compounds; they are also generally similar in reactivity, zinc being the more reactive, and cadmium showing a slightly greater tendency to form covalent bonds, especially with sulfur.
The ionic radius of cadmium is 0.88 A for fourfold coordination, 1.03A for sixfold coordination. The sulfide, CdS (dimorphous, like the corresponding zinc and mercury sulfides), and the carbonate, CdCO,, are less soluble than the corresponding zinc compounds, but the hydroxide, Cd(OH)2, is more soluble than Zn(OH)2. Cadmium forms a wide variety of soluble complexes, notably with cyanides and ammines.
Cadmium metal is a bluish-white to silverwhite metal, density 8.645, Brinell hardness 21, that melts at 321°C and boils at 765°C. Its vapor pressure, greater than that of zinc, is 1.4 mm at 400°C and 16 mm at 500°C, so that losses by vaporization are to be expected during metallugrical processing. The vapor is very reactive, quickly forming finely divided CdO in the air.
The geochemistry of cadmium has been reviewed by Ivanov (3) and by Wakita and Schmitt (4). It is a strongly chalcophilic element, i.e., it is concentrated in sulfide deposits, in which it follows zinc and mercury, and to a much lesser extent lead and copper. The abundance of cadmium in the Earth's crust is generally estimated to be 0.15-0.2 ppm. Its concentration is low in all igneous rocks and its content in them shows no clear relations to those of any May 1974 major element, nor to that of zinc, the ratio Zn/Cd varying widely in all types of igneous rocks. Higher concentrations of cadmium occur in shales, oceanic and lacustrine sediments, phosphorites, and oceanic manganese nodules; in these the ratio Zn/Cd is generally lower than in igneous rocks.

Use Patterns of Cadmium
The principal uses of cadmium are as alloys, in plating metals, pigments, as a stabilizing material for polyvinyl plastics, and in batteries. Two different estimates of the material flow of cadmium in the USA in 1968 are shown in Figures 1 and 2, as given by Davis et al. (2) and by Heindl (5). The values for miscellaneous or other uses ( Fig. 1) include uses in fungicides, nuclear control rods, phosphors, ceramics, and others.
Data from the U. S. Bureau of Mines (6) on U. S. and world production of cadmium are summarized in Table 1. Consumption in the United States was in metric tons 6046 in 1968, 6845 in 1969, with a sharp decrease to 4218 in 1970. It has been estimated that U. S. annual consumption will rise to about 13,600 metric tons by the year 2000.
Environmental Forms and Sources
Nearly all the cadmium in primary sulfide ores occurs as a minor element in sulfides of other elements, especially those that have fourfold covalent bonding. Cadmium occurs mainly in the zinc sulfides sphalerite and wurtzite, the maximum cadmium content of sphalerite being about 5%, and the entire series wurtzite-greenockite being known.  The average cadmium content of zinc sulfides of individual deposits ranges from 0.02 to 1.4%o, with a median content about 0.3% Cd. Fourteen other sulfides have been reported to contain more than 500 ppm Cd; the most important of these, with maximum and median reported contents of cadmium respectively, are: galena, PbS, 5000, 10 ppm; tetrahedrite-tennartite, (Cu, Zn)2 (Sb, As) 4S13, 2400, 600 ppm; metacinnabar, HgS, 11.7%, <10 ppm; chalcopyrite, CuFeS2, 600, <10 ppm.
Cadmium is readily dissolved from sulfides by acid waters formed during the oxidation of sulfide ores. Such mine waters have been reported to contain up to 42 ppm Cd, although values of 0.1-2 ppm are much more common. The subsequent fate of the dissolved cadmium depends largely on Eh-pH conditions and on the type of country rock encountered by the solutions, but much if not most of the cadmium appears to be precipitated before travelling very far. The precipitates generally have lower Zn/Cd ratios than the primary sulfides of the same deposits, that is, they are enriched in cadmium. This is true of secondary sulfides (sphalerite-hawleyite series, wurtzite-greenockite series), the zinc carbonate smithsonite (maximum Cd, 0.87%o), the zinc silicate hemimorphite (maximum Cd, 1.2%o), the basic iron sulfate jarosite (maximum Cd, 1.0%), and iron oxides (maximum Cd, >0.1%). The possibility of pollution of natural waters by cadmium dissolved from sulfide ores is discussed in a later section.

Cadmium in Igneous and Metamorphic Rocks
Relatively few analyses of igneous rocks for cadmium are available; these are summarized in Table 2. The range reported is <0.001 to 1.6 ppm, with very few analyses above 0.5 ppm; alkalic rocks, alkalic granites, and rhyolites have cadmium contents slightly above average. Vincent and Bilefield (8) found a distinct, but slight enrichment in cadmium with increasing magmatic differentiation. The ratio Zn/Cd varies widely in all types of rocks, with recorded extremes of 27 to 7000; no clear variation of cadmium content with any major constituent is apparent.
Few determinations have been made of the distribution of cadmium among coexisting minerals of igneous rocks. In three gabbroic rocks analyzed by Nesterenko et al. (10), cadmium was present in all the rock-forming minerals, in amounts of 0.015-0. 10   Marowsky and Wedepohl (9) found that This association of cadmium with organic regional metamorphism has reduced the matter plus the known existence of organic cadmium content of metamorphosed shales compounds of cadmium, such as dimethylof the Swiss Alps by at least one order of cadmium and diethylcadmium, raises the magnitude.
specter that cadmium, like mercury, might be recycled geochemically in such a form.

Cadmium in Sedimentary Rocks
However, Fassett (13) has pointed out that The available data are summarized in such compounds are much less stable than Table 3. The content of cadmium is apthe analogous mercury compounds. It would parently low in limestones and sandstones, be highly desirable to have further study but cadmium is appreciably concentrated in of this possibility.
shales, oceanic, and lacustrine sediments, oceanic manganese oxide nodules, and in Cadmium in Coal and Oil phosphorites. Of special interest are the data Data on the cadmium content of coal of Tourtelot et al. (11), which indicate that and oil are scanty. Lagerwerff and Specht cadmium, like mercury, is especially con-(15) found 1-2 ppm Cd in eight bituminous centrated in shales with high content of coals from Kentucky and Illinois. Aberorganic matter. Gulbrandsen (12) states nathy et al. (16) found cadmium in only that cadmium and zinc in the phosphorites 70 of 827 coal samples from the U.S. in of the Phosphoria formation are associated amounts of 2-100 ppm Cd in ash (approxwith organic matter. imately 0.2-10 ppm Cd in the coal). These  20 for Fairfax, Ohio (suburban). Harrison (29) found an average of 19 for the Chicago area. Friberg et al. (30) quote weekly means of 5 in the center of Stock-holm and a monthly mean of 0.9 in a rural area in Sweden, also 1.5 in Erlangen, Germany. Just and Kelus (31) reported 2-51 in the air of 10 Polish towns in 1967. Dudley et al. (32) found 3-8.7 in five marine aerosols.
All these data emphasize the probability that nearly all airborne cadmium is due to man's activities. The highest concentrations are reported from cities with considerable industry, especially metallurgical and smelting operations. Towns in the United States reporting maximum Cd values >100 ng/m3 are given in Table 4.
Friberg et al. (30) quote weekly means of 500 ng Cd/m8 at a distance of 100 m and 200 ng/ms at a distance of 400 m from a Japanese smelter, and 160-320 ng/ms at a distance of 500 m from another Japanese smelter. Weekly means of 600 (maximum 54000) mg/m8 at a distance of 100 m and of 300 at a distance of 500 m were recorded in Sweden near a factory using coppercadmium alloys. These are several hundreds times as great as concentrations in uncontaminated areas.
Data are not yet available on the fate of air-borne cadmium and its residence time in the atnosphere. It is presumed to be carried down by rain and snow, but the few determinations made so far are inadequate. The data of Lagerwerff (35) indicate that near highways nearly half the cadmium taken up by plants is from airborne sources.

Cadmium in Atmospheric Emissions
The recent estimates by Davis et al. (2) are summarized in Table 5. It should be noted that these are not based on actual measurements, but are calculated from production data plus assumptions as to the losses of cadmium in each type of operation.
The emissions in smelting operations are undoubtedly considerable (see above), yet the figure in Table 5 for emissions from this source is nearly 20%o of the total production of the United States in 1968 (Table  1), which seems high. The emissions from reclamation of steel and radiator scrap assume that all of the cadmium in these materials, none of which is recovered, reaches the atmosphere. The emissions from burning cadmium-stabilized plastics were cal-Culated on the assumptions that 40%o of the plastic went to waste, that 15%o of the waste was burned, and that all of the cadmium in burned material reached the atmosphere. It seems likely that these estimates are somewhat high.
On the other hand, Table 5 does not include estimates of possible cadmium emissions from the burning of coal and heating oil; these might amount, respectively, to 100 tons and 30 tons/yr (2); Bertine and Goldberg (21) estimate about 2.2 tons/yr from burning oil.
If cadmium were present in appreciable amounts in gasoline, the emission from this source could be serious. Reports of its presence in gasoline have been published (36,37), but these have not been verified. Lagerwerff and Specht (15,22) reported finding 6 parts per billion (ppb) Cd in one of 12 gasolines tested, the rest containing less than 0.01 ppb.

Cadmium in Soils
A selection of representative data on the cadmium content of soils is given in Table 6. Recent analyses of uncontaminated soils indicate that normal contents of cadmium are less than 1 ppm, perhaps about 0.4 ppm on the average. A few analyses that show considerably higher values are unexplained; Malyuga (38) reported 16 and 45 ppm in chernozems developed on serpentinite, USSR; Environmental Health Perspectives Year 1960Year 1962Year 1960Year 1963Year 1960Year 1960Year 1970Year 1964Year 1969Year 1962Year 1964Year 1961Year 1964Year 1961Year 1960Year 1964Year 1961Year 1969Year 1960Year 1961Year 1961Year 1961Year 1962Year 1954Year -1959Year 1961 Table 6 gives means of concentrations of cadmium in soils of varying environments in the area of Grand Rapids, Michigan (43). The contents are significantly higher in the industrial and airport zones than in the residential.
Considerable amounts of cadmium may be added to soils by the addition of phosphate fertilizers or by the application of sewage sludge.

Cadmium in Waters
Determinations of cadmium in waters are summarized in Table 7 Many more analyses of contaminated soils are given in the references cited in that section of Table 6 and by Burkitt et al. (40). The data of Lagerwerff and Specht (15) and of Lagerwerff (35) show clearly the extent of soil contamination near highways. Contamination from emissions from smelters and metallurgical plants is even more striking. The data of Miesch and Huffman (41) show that the effects of contamination from a smelter that had been operating for 80 years were discernible at distances of at least 11 miles away, and probably further, as shown in Figures 3 and 4.
Cadmium is evidently mobile in these smelter-contaminated soils and works its way downward to a depth of at least 30 cm, as illustrated by the data of Kobayashi (42), which show at 5 cm, 44   averaging about 0.15 Jug/l. or about 0.15 ppb. Many of the samples analyzed are near-shore samples and may have been contaminated (56). Thus the recent analyses by Jaakkola et al. (59) are higher than the average; the samples from the Gulf of Bothnia had been contaminated by discharges from a leadzinc smelter. Krauskopf (74) has calculated that sea water now contains less than 0.1%' of the cadmium that has been supplied by erosion of the land surface and is far below the concentration calculated from solubility data, so that sea water is distinctly undersaturated with respect to cadmium. The data in Tables 3 and 7 indicate that nearly all the cadmium has been removed by coprecipitation with or adsorption on clays, hydrous manganese oxides, and phosphorites. Posselt (75) has shown that the removal of cadmium from waters by precipitation of hydrous oxides of Fes+ and Mn"4 is rapid and effective.
Most fresh waters contain less than 1 ppb Cd. The chemistry of cadmium in surface and ground waters has recently been reviewed by Hem (75) (67), who made ultracentrifuge separations of coarse and colloidal particulate matter from four samples in a mineralized area, showed that the particulate matter had concentrations of cadmium some thousands of times that dissolved in the waters, yet that 85-96% of the total cadmium was present as dissolved material.
These data indicate that cadmium is precipitated on stream sediments under some conditions, the extent of precipitation depending on pH, degree of complexing, and many other factors. Abdullah and Royle (72) report that "clean" streams in Wales contain 0.41 jug Cd/l.; streams affected by old mining activities contained 1.1-3.4 ,ig/A.
It has been mentioned that mine waters have been reported to contain up to 42 ppm Cd; such waters might reach the drainage system. Mink et al. (77,78) found traces to 0.45 mg/l. (ca. 450 ppb) Cd in waters of the South Fork of the Coeur d'Alene River, Idaho, corresponding to the transport of up to 394 lb Cd/day. This stream drains an area in which thousands of tons of ground tailings of Pb-Zn-Ag ores were dumped decades ago.
The surface waters that contain more than a few ppb Cd near urban areas have almost certainly been contaminated by industrial wastes from metallurgical plants, plating works, or plants manufacturing cadmium pigments, cadmium-stabilized plastics, or nickel-cadmium batteries, or by effluent from sewage treatment. Few data are available. Tenny and Stanley (79) examined 1123 samples of industrial wastes of the Chicago area and reported that 1.43%o contained more than 10 mg/l. Cd, 0.27%o contained more than 50 mg/I. Cd; all these were from metal treatment and plating. Among the few well-documented studies of industrial contamination of water supplies by cadmium are those of Lieber and Welsch (80) and Perlmutter and Lieber (81), who traced the spread of cadmium from a plating plant over an area up to 0.8 mile long and 0.2 mile wide, finding ground waters containing up to 3200 ppb Cd.  Another potential source of cadmium in surface waters is the leaching of land fill or soils to which sewage sludge has been added. Tenny and Stanley (79) reported an average content of 0.003 mg/l. in the final effluent of sewage treatment plants in the Chicago area, nearly all the cadmium having been precipitated in the sludge. Digested sewage from two plants contained 70  Cadmium is a relatively rare element; its abundance in the Earth's crust is generally estimated to be 0.15-0.2 ppm. It is not markedly concentrated in any igneous rocks, and its concentration in them shows no clear correlation with that of any major constituent, nor with that of zinc. Higher concentrations of cadmium occur in shales (especially in those rich in organic matter), oceanic and lacustrine sediments, oceanic manganese nodules, and in marine phosphorites; in these ratio Zn/Cd is generally lower than in igneous rocks.
All the cadmium production of the world comes from sulfide ores of zinc, lead, and copper, from which it is recovered as a byproduct. Total production in the United States and in the world reached all-time high in 1969 of 5736 and 17,576 metric tons, respectively.
The principal uses of cadmium are in electroplating of metals, in alloys, in pigments, as a stabilizing material for polyvinyl plastics, and in batteries. It has been estimated that consumption in the United States will nearly double by the year 2000 to a level of about 13,600 metric tons.
Data on the cadmium contents of coal and oil are scanty; the range in coal is 0.2-10 ppm; in motor oils, mostly 0.1-0.5 ppm, due at least in part to the addition of zinc dithiophosphite.
Measurable amounts of cadmium are present in the atmosphere. In rural areas the concentration is generally from tenths of a nanogram to 10 ng/m3; in urban areas it is 2-420 ng/ms. Higher values have been reported near smelters and metallurgical plants. Soils near such sources also show much higher concentrations of cadmium than the normal content of about 0.4 ppm. Recommendations 1. Improvement of our knowledge of the natural transport of cadmium requires better estimates of the distribution of cadmium in igneous and sedimentary rocks and their constituent minerals, and information, now entirely lacking, on the nature of the binding of cadmium in organic-rich shales and in phosphorites.
2. Data are needed on the cadmium content of coal and fuel oil and on the proportion of the cadmium present that reaches the atmosphere during burning, especially from large installations such as power plants. 3. Continued monitoring of the air of our cities is essential, and special studies of the emissions of cadmium by metallurgical plants and zinc smelters should be undertaken. 4. In conjunction with these, study should be made of the fate of atmospheric emissions of cadmium with respect to the time of residence and fall-out with rain or snow, especially in areas of high emission.
5. More data are needed on the emission of cadmium from processes of scrap reclamation and from the disposal of cadmiumweighted plastics.
Environmental Health Perspectives 6. The fate of cadmium in the hydrologic cycle is little known. Studies of the cadmium content of surface waters should be accompanied by study of its content in the associated particulate matter and in stream sediments. Data are needed on the rate of release of cadmium from soils treated with phosphate fertilizers and on the possible leaching of cadmium from galvanized plumbing.
7. Discharges of water from plating works and from plants manufacturing Ni-Cd batteries, Cd pigments, etc., should be monitored carefully.

Cadmium Material Flows
In this section an attempt is made to trace the flow of cadmium from its sources to the areas in which it is accumulated in the environment. The discussion is restricted to fluxes in the U.S. and adjacent seas, but reference is made to European or Japanese studies where these help fill gaps in the picture. Most of the data cited are for the fiveyear period 1967-1972. Although there is not an exact correspondence in dates of the information drawn from different sources, it is felt that the time span covered is small enough to permit the neglect, for the purposes of formulating gross models of flow, of variations in flow rates and concentration over the period considered.
In view of the near constancy of the ratio of cadmium to zinc in many ores and the similarity in the major uses of the two elements, extensive reference is made to zinc flows and zinc/cadmium ratios. Zinc/cadmium ratios in environmental samples that differ significantly from those in geological reserves are indicative either of errors in sampling and analysis or of selective mechanisms for the flow and accumulation of one of the two metals. Whenever anomalous ratios are encountered, a careful examination of the cause for departure from the norm is warranted.
The section has drawn heavily on the estimates by Davis et al. (2) of atmospheric emissions and on the summary by Chizhikov (86) of cadmium fluxes in smelters and refiners. Subsequent to preparation of a draft of this section, a very detailed description of the societal flow of Cadmium was published by Fulkerson and Goeller (87).
The processing of cadmium will be considered in some detail in order to obtain information on such factors as physical and chemical state that may influence the flow and accumulation of cadmium in the environment. Wherever possible, estimates of emission art tested against measured ambient concentrations.

Mining and Concentration of Ores
Cadmium is recovered with zinc usually from polymetallic ores containing lead and copper. The classification of the ore as zinc, lead-zinc, lead, copper-lead-zinc, or other permutations is determined by the relative concentrations of the metals of primary interest. In 1969, of the 501,760 tons (metric) of zinc mined domestically, 63% was in zinc ores, 19%o in lead-zinc ores, 9%o in lead ores, 6% in copper-lead-zinc ores, and 4%l in other ores. The average zinc content of the domestic ores was 3.93%o. Imported ores are generally richer in the prime metals. In 1969, 512,560 tons of zinc were imported in foreign ores, and 297,560 tons were imported as slab metal (88). Zinc/cadmium ratios in selected ores are reported in Table 8. Chizhikov (86) quotes a value of 200 for the mean of zinc/cadmium ratios in geological reserves, a value which is approximately twice the ratio of the rates of production of zinc and cadmium in the U.S. (Table 1). This difference may be explained in part by the significant U.S. imports of zinc and cadmium, as metallic slabs and concentrated ores and flue dust, in ratios different from that encountered in the primary ores. Imports of the metal and ores have changed markedly in the past years, as the U.S. producers, faced with a cost-price squeeze and high costs for pollution-control equipment, have closed several domestic smelters (five in the period 1969-1971).
Little information is available on the contamination of the environment in the mining of the polymetallic ores containing cadmium. The major potential sources of contamination are fr-om the particulate emission during mining and the leaching of cadmium from the overburden. The zinc content is, however, low, ranging from less than 2%o in the overburden to 3.9% in the ore, so that the amount of cadmium lost via these mechanisms is expected to be low. Acid waters will, however, dissolve the sulfide ores and have been reported to accumulate as much as 42 ppm Cd, compared to more common values for mine waters of 0.1-2 ppm.
The ores are enriched primarily by flotation to yield concentrates containing 40-60% prime metal. Since the initial concentration of prime metal is of the order of several percent, most of the original mass of the ore will be discarded in the tailings. The cadmium in the ores will generally concentrate with the zinc but there may be a slight fractionation between the cadmium and zinc between concentrate and tailings. Illustrations of the recovery of cadmium in the concentration of lead-zinc ores are presented in Table 9. These values and similar studies by Vinogradov (44) suggest that 18-36%o of the cadmium in polymetallic ores is retained in the tailings but that the fraction of zinc retained in the tailings is lower. By comparison, it is estimated in the ORNL study on cadmium (87) that 8%o of the zinc and llo of the cadmium in the original ore are retained by the tailings (Table 10). The cadmium concentration in tailings is low, and losses from the tailings to the atmosphere should not be significant. Davis and his associates (2) estimate that wind loss from tailings amount to 0.2 lb/ton of cadmium mined, a figure that is unsubstantiated. However, the conclusion reached by Davis et al. (2) that wind losses contribute a negligible fraction of the cadmium emissions is undoubtedly valid. Leaching by acid waters is of greater concern than wind losses, but no statistics are available on the rate of dissolution of cadmium from tailings.  Metallurgical Processing Major losses of cadmium occur during the processing or concentrations of zinc ores, lead ores, and, to a minor extent, copper ores. The extent of emission will depend on the metallurgical process employed and the age and condition of the processing plant, and whether cadmium is recovered at a particular installation. Of the 14 plants that produced primary zinc in the U.S. in 1969, only eight produced cadmium metal. There is no information on the fate of the cadmium content of the ores processed in the remaining six plants.
The principal source of cadmium is the zinc ores. These are first roasted to oxidize the sulfide and then treated by either pyrometallurgical or electrolytic processes.

Pyrometallurgical Recovery of Zinc
Roasting is carried out at temperatures of 700-1200°C. The amount of cadmium vaporized during roasting increases with temperature and ranges from 10 to 25%. The dust from the roasting operations is collected in electrostatic precipitators or bag filters, and the SO2 either dispersed from a high stack or piped to a sulfuric acid plant. Comparison of the representative analysis of dust from an electrostatic precipitator with that in the ore concentrate to a roasting plant (Table 11) shows a considerable enrichment of both cadmium and lead in the dusts. The sulfur in the dusts is mainly in the form of sulfates (Table 12).
The roasted ore is sintered to drive off residual sulfur and to agglomerate the ore. The dust from the sintering operation is led to collectors and then vented to the atmosphere. Typically 15-20% of the initial cadmium accumulates in the dust obtained during sintering (86). Examination of Table 11 shows that the cadmium/zinc ratio in the dust from the collector is somewhat higher than in the ore concentrate and that the fine dust passing through the cyclone is considerably enriched in cadmium and lead. The plant for which data are presented in Table 11 employs a relatively inefficient cyclone collector to remove the dust from the flue gases leaving the sinter plant; more efficient electrostatic precipitators and bag filters are more representative of industry practice. It should be noted that the sulfur content of the dust escaping the cyclone collector is less than half the value of that in the dust (mainly sulfates) from the roasting plant. The water-soluble fraction of the cadmium in the dust is expected to be cor- May 1974 respondingly lower. The amount of cadmium recovered in the flue dust from the sintering machine may be increased by adding chlorides to the charge fed the sintering machines. In this case, the degree of cadmium removal reaches 70-80%b. The agglomerate from the sinter plants will typically retain more than 50% of the cadmium content of the ore. The agglommerate containing the zinc and cadmium oxides is reduced with carbon and carbon monoxide. The resulting metals are distilled either batchwise (in plants utilizing horizontal or Belgian retorts) or continuously (in plants utilizing vertical retorts or blast furnaces). In the batch process most of the cadmium is carried over during the initial stages of the distillation (Table 13). In the continuous process the cadmium is distributed between the blue powder (approximately 5% of the distillate) and the condenser dust in the concentrations shown in Table 14. The condenser dust is redistilled. According to Chizhikov (86), approximately half of the cadmium present in the sinter passes into the zinc metal to yield an average cadmium concentration of 0.12%o, and the concentration of cadmium in the residual slag is approximately 0.05%o %. The residual slag is retreated when high recoveries of cadmium are sought.

Hydrometallurgical Recovery of Zinc
In the hydrometallurgical process, the ore concentrates are roasted in a single multiple-hearth unit, in which approximately 20% of the cadmium is volatilized (86). The zinc concentrate is leached with dilute sulfuric acid. Most of the cadmium passes into solution as the sulfate; 9-22% remains in the residue. The cadmium in the sulfate solution is deposited together with copper and lead by addition of zinc dust 62-86o of the cadmium being precipitated (86). The zinc is electrodeposited on aluminum cathodes from the purified sulfate solution. In 1969, electrolytic recovery accounted for approximately 41%o of the primary zinc production (89).

Cadmium Distribution in Lead and Copper Production
Lead and copper concentrates recovered from polymetallic ores or from residues from zinc production contain varying quantities of zinc and cadmium. The extent of recovery of cadmium from the processing of these ores and residues depends on the initial cadmium concentration. This will be illustrated for a few case studies selected by Chizhikov (86).
In a lead production plant, 70-80%o recovery is claimed for the cadmium in the  Environmental Health Perspectives dust from the sintering and smelting operations. The concentration of the cadmium in the dust from the sintering operation was 0.35.-0.367%, and the water-soluble fraction of the dust ranged from 6 to 25%o. The average particle size of the sinter dust collected in an electrostatic precipitator varied from 0.4 to 1.2 u. The dust from the smelters had cadmium concentrations that varied from 1.6 to 6.5%o and water-soluble fractions that ranged from 0 to 1%o; the dust was submicron in size. Information, admittedly of a qualitative nature, supplied by smelter operators to Davis et al. (2) suggested that emissions were in the 4-8 MA size range. It is not possible to judge whether these data are spurious or indicative of growth of the emitted particles by condensation and agglomeration in the stack. Cadmium vaporizes to a much smaller extent in copper plants, and a significant fraction of the cadmium in the ore concentrate passes into the slag from the smelter. Balances on smelters of different design are presented in Table 15 and on a Bessemer converter in Table 16. From Tables 15 and 16 it can be estimated that roughly one third of the cadmium losses passes into the slag, and the remaining two thirds are lost to the atmosphere.

Production of Cadmium Metal
Cadmium is recovered from the blue powder collected during the distillation of zinc, the cadmium fraction from zinc distillation units, the copper-cadmium filter cakes deposited from sulfate solutions, and the dusts from roasters, sintering machines, and fuming furnaces used in zinc, lead and copper production.
Prior to 1910, cadmium was produced by distillation in retorts similar to those used for zinc recovery. Cadmium losses from the distillation process were high, often 40-50% of the initial quantity (86).
At present, most of the cadmium is recovered by hydrometallurgical processes. The cadmium is dissolved in sulfuric acid, and cadmium sponge is deposited by addition of zinc. The process may be repeated several times to increase the purity of the cadmium. Finally the sponge is redissolved, and the cadmium is deposited electrolytically. Recoveries of cadmium in the hydrometallurgical process are reported to exceed 90%o (86,89).

Summary on Emissions from Smelters
The amount and physical and chemical state of the cadmium emitted by smelters will depend on plant design and on the efficiency of the particulate collection devices on the individual units within a plant. The emission from sinter units and retort stills will consist mainly of the oxide (or the chloride when chlorides are added to the charge to the sinter unit), whereas sulfates may constitute a significant portion of the emissions from roasters. The particles will vary in size from the submicron range to about 8 u and have a Zn/Cd ratio varying from the extremes encountered in the cadmium and zinc purification stills. On the average, because of the higher volatility of cadmium, the Zn/Cd ratio in the atmospheric losses from smelters will be lower than those in the ores processed. Values for smelter recovery efficiencies of 75% for cadmium and 89-97.5% for zinc are quoted by Heindl (90), suggesting that, for an ore with a Zn/Cd ratio of 200, the Zn/Cd ratio in the portions not recovered (losses to the atmosphere and in the slag) will range from 20 to 88.

Transport of Emissions from Smelters
The fraction of the stack emissions in the vicinity of the smelter will depend upon particle size, varying from a negligibly small value for submicron particles to practically complete deposition for particles greater than 10 u. Uncertainty regarding the particle size distribution at the source and the dynamics of particle growth in the atmosphere preclude any attempt at quantitative modelling of the particle deposition rates. Of the cadmium deposited, the sulfates and chlorides may be leached from the soil by rainfall, but the oxides, which have a negligible vapor pressure, will accumulate in the soil, other than the small amounts taken up by vegetation or windblown from the surface. It is, therefore, expected that the soils surrounding smelters may have high concentrations of cadmium and zinc corresponding roughly to the cumulative deposits of emissions from the local smelter, less the small amounts taken up by the vegetation and leached by precipitation. Some selected studies of soil concentrations and dustfall surrounding smelters support this conclusion, at least to an order of magnitude approximation.
East Helena, Montana: An extensive study (41) of the pollution surrounding lead and zinc smelters in East Helena, Montana, revealed concentrations of zinc up to 5200 ppm and cadmium up to 160 ppm in the soils in the vicinity of the smelter. From measurements at different radii from the smelter and at depths up to 25 cm it was estimated (41) that 5100 tons of zinc and 260 tons of cadmium had been added to the soils within a radius of 1.0-16 km from the smelter stack. The Zn/Cd ratio of about 20 is, as anticipated, considerably smaller than the ratio in geological reserves. The zinc production at the East Helena plant was about 80 tons/day at the time of the survey, and the plant was reported to have a zinc recovery of 91%b. If the zinc in the slag is assumed to be a third of the total losses, a rate of zinc atmospheric emission of about 5 tons/day is estimated.
The solubility of the dust emitted by the different units in the smelter ranges from near zero to 70%o from the roasters, with a mean of probably less than 15%o. The diffusion of the insoluble fraction into the soil is low, and most of the dustfall is retained in the surface layers. The amount of zinc accumulated in the soil, therefore, corresponds to the cumulative emission of about 1000 days at full production.
These calculations support the conclusion that a major fraction of the emission from smelters accumulates in the surrounding soil. The estimates of emission and acculation given above are too approximate to permit a closure of a material balance and to determine by difference the amount of cadmium transported to distant locations either directly by long-range aerial transport or indirectly by plant uptake, leaching, and transport in surface waters.
The distribution of cadmium and the Zn/ Cd ratio in the top 25 cm of the soil is shown in Table 17, taken from the smoothed data of the East Helena study. It can be seen that the Zn/Cd ratio increases with depth but is lower than that in geological reserves, even at the 15-25 cm depth. Soils collected outside the Helena valley as part of the study had a mean cadmium concentration of 0.8 ppm and zinc concentration of 58 ppm (Zn/Cd = 74). Low Zn/Cd ratios were also observed in the settleable dust and suspended particulate providing support for the postulate that the relatively low Zn/Cd ratio in the soil reflected the low Zn/Cd ratio at the source and not a selective leaching of zinc from the soil.  (41).

Environmental Health Perspectives
The uptake by vegetation may be estimated from the concentration of cadmium in the crops. The concentration of cadmium in plants sampled from ranches within 6 km of East Helena reported in the EPA study (41) include 0.3-3.2 ppm (wet-weight basis) for alfalfa, 6.3 ppm for barley straw, 0.1-1.2 ppm barley kernel, and 1.2-9.8 ppm pasture grass. As a rough gauge of cadmium removal from the soil by crops, the above value will be assumed to be an average value for the region, and a crop yield of 2 tons/ acre for alfalfa selected. Based on an assumed average concentration of 2 ppm in alfalfa, the cadmium removal per crop in an area with a 6-km radius is 1.2 tons/yr. The plant uptake per crop is less than a percent of the cadmium accumulation in the soil and does not provide a significant mechanism for cadmium removal from the soil, although locally it may be a major route for human and animal cadmium intake.
Annaka City Smelter, Japan: Distributions of cadmium and zinc in the soil 900 m from the smelter chimney (42) are shown in Table 18. The ratio Zn/Cd in the surface layers is again lower than in geological reserves, suggesting a preferential emission of cadmium from the smelter stack. As was the case for East Helena, the depth of penetration of zinc and cadmium in the soil is seen to be small.

Metal Processing Furnaces
The major source of atmospheric emissions of cadmium other than smelters was identified by Davis et al. (2) as metal treatment furnaces. Fulkerson and Goeller (87)  have more recently reevaluated the cadmium emission from the treatment of scrap steel. Based on figures of 2830 tons of cadmium for electroplating, 416,000 tons of zinc with an average cadmium content of 0.023% for zinc galvanizing, 76 million tons of finished steel, and 26 million tons of steel scrap, they estimate that 1000 tons of cadmium were volatilized during the remelting of steel scrap in 1968. The amount emitted will depend on the pollution abatement devices. For an assumed 90%b efficiency in the dust collectors they obtain an emission estimate of 100 tons, approximately one tenth of the value estimated by Davis et al. (2). The amount of cadmium emitted may, however, be much higher if the collected dust is recycled to the furnaces. The dusts from steel remelting furnaces have iron oxide contents which range from 25%o to over 90%o. Ores with a high concentration of iron oxide may be recycled, thus providing an opportunity for enhancing the emission of the volatile cadmium compounds. The fraction of the collected dust that is recycled is not known, so that the emission of cadmium from furnaces treating plated steels cannot be estimated with any accuracy. It will here be assumed, somewhat arbitrarily, that as a consequence of inadequate air pollution control from older units and some recycle that 500 tons of cadmium were emitted and 500 tons deposited in dumps in 1968.
It should be noted that, in addition, zinc products contain small amounts of cadmium and that the disposal of zinc products also contributes to cadmium emission. Table 19 gives the maximum permissible cadmium concentrations in the different grades of zinc marketed. If the different grades of zinc contained the maximum permissible cadmium concentration, they would contain an amount of cadmium equivalent to 36%o of the total U.S. consumption. The cadmium content is, however, lower than the maximum limit for the different grades and corresponds to approximately 7%o of production (87).
The zinc content of particles discharged from different metal treatment furnaces is relatively high, even for ferrous metal furnaces (Table 20). The particle sizes of the zinc fumes from these furnace types is reported (91) to be in the range 0.03-0.3 u. Although cadmium emissions were not measured, they are expected to show a similar pattern.

Coal Combustion
Estimates of the emission of cadmium from coal fired units must depend on the efficiency of collection of the cadmium in the particulate collection equipment. A study by Bolton et al. (92) of the efficiency of collection of trace elements in the fly ash of a cyclone-fired boiler equipped with electrostatic precipitators indicated that the collection efficiency was high for cadmium and most other trace elements other than mercury. The collection efficiency for cadmium can therefore be approximated by that for total particulate collection. If it is assumed that for the 455 million tons of coal consumed in 1968, the average particulate collection efficiency was 80%o and the average cadmium concentration was 1.0 ppm, an annual emission of 91 tons is obtained, close to the value of 100 tons estimated by Davis et al. (2) but lower than the upper bound of 1000 tons given by Fulkerson and Goeller (87) based on the maximum concentration of cadmium in coal. Obviously, the emission of cadmium from coal-fired units is strongly dependent on the cadmium concentration in coal and on the collection efficiency for particulate. More data are needed on cadmium concentrations in coal and in the different size fractions of the ash emitted from furnace combustion chambers of representative design.
Estimations by Klein and Russell (93) of fallout surrounding a 650 MW power plant in service since 1962 showed a cumulative cadmium emission of 18 tons for a cumulative coal consumption of 107 tons. Prorating these values to a coal consumption in the United States of 455 million tons/yr would yield a rate of cadmium emission of 820 tons/yr. It should be noted, however, that the Zn/Cd ratio in the estimated fallout was only 9, suggesting that the estimated cadmium emission may be mode than an order of magnitude too large.

Cement Production
Goldberg (E. D. Goldberg, Univ. Calif., San Diego, personal communication) estimated the world wide emission of cadmium from cement manufacture was 100 tons in 1962 assuming that the concentration of cadmium was 0.3 ppm in shale and 0.035 ppm in carbonates, and that all the cadmium was volatized. It is here estimated that the U.S. emission in 1968 was 3 tons, assuming that 80% of the emissions were captured in particulate collection devices. Incineration Davis (2) estimated that 86 tons of cadmium were incinerated in 1968, based on a disposal of 40%o of the annual average consumption of cadmium for paints, pigments, and miscellaneous uses and incineration of 15%o of the total wastes. A rough check of this estimate may be made using measurements by Levins (Arthur D. Little, unpublished data, 1972) of trace metals in the particulate from a municipal incinerator. Four samples were found to contain 250, 400, 200, and 1700 ppm (average 640 ppm) cad-mium and to have Zn/Cd ratios of 100, 90, 410, and 40 (average 160). For a national municipal waste generation in 1968 of 160 million tons, 15%o burned in incinerators and dumps, a particulate emission of 0.007 tons/ton, corresponding to municipal unit with wet baffles ($4); the total cadmium emission is 107 tons, in fair agreement with the Davis estimate. With the closing of openburning dumps and stricter air-pollution regulations on incinerators this source of emission should be reduced significantly.

Ambient Concentration in Urban Areas
The total emission of cadmium to the urban atmosphere has been estimated as 706 metric tons/yr, with major contributions from metal processing (500 tons/yr), coal and oil combustion (120 tons/yr), and incineration of plastics and pigments (86 tons/ yr). Since roughly three quarters of total suspended particulate in urban areas are of nonagricultural origin and the emission of particulate within cities is of the order of 2 x 107 tons/yr (95) the emission should yield a mean cadmium level of the order of 25 ppm. In urban air with particulate concentrations typically in the range 80-120 jug/ mi, the concentration of cadmium should have an average for all urban areas of approximately 0.002 jug/m8, with departures from the mean determined primarily by the concentration of metal processing industries in a given city. Table 21 provides values of cadmium and Zn/Cd ratios in selected cities. The concentrations are in fair agreement with the rough number gauged from emission estimates. The Zn/Cd ratio is frequently below that in the geological reserves, possibly due to the selective vaporization of cadmium in the processing furnaces.
Measurements of dustfall in Cincinnati by Creason et al. (23) show a Zn/Cd ratio of 150 and an average cadmium concentration in the dustfall of 13 ppm, about twice the above estimate for a national average. The cadmium concentration in air in Cincinnati varied from 0.008 ,Ag/m3 in an urban area to 0.0035 jtg/m3 in an industrial area. The values are higher than the average value derived earlier, possibly due to a concentration of metallurgical industry in Cincinnati.
The rate of deposition of cadmium from the atmosphere will depend upon the particle size, which is a function of size distribution at the source and agglomeration and sedimentation in the atmosphere. Measurements by Lee et al. (97,98) of the cadmium and zinc concentrations in sized fractions of the suspended particulate in St. Louis are summarized in Table 22. The mass mean diameter of the particles containing zinc and cadmium are slightly larger than the mean diameter of the total suspended particulate, but the difference is not sufficiently large to effect any appreciable fractionation of cadmium between dust fall and suspended particulate. Cadmium and zinc will have half-lives comparable to that of total suspended particulate of the order of a few days-and they will, therefore, deposit mainly in the vicinity of the source. Small particles may be transported considerable distances before deposition, and longrange transport will also occur at a slow rate by the periodic suspension of deposits by wind action and redeposition. [The ambient concentrations of particulate in rural areas has been increasing slowly in a period when the concentrations in urban areas have been decreasing (99).]

Water Transport
Quantitative data on cadmium inputs to surface and coastal waters are fragmentary. It is expected that the major inputs will be from the leaching of mine tailings, particularly by acid waters, from the emission into the atmosphere of soluble sulfates and chlorides, from the disposal of waste streams by hydrometallurgical and cadmium-plating installations, from surface runoffs in urban areas, and from the disposal of sewage sludge and effluent into water streams.
Mine Drainage: Streams and sediment in the Conway Catchment in Wales show anomalous concentrations of zinc and cadmium in both water and sediment as a consequence of drainage from old mines and periodic sediment transport by heavy rains (100). The maximum zinc concentration in the sediment exceeded one percent in the area of the mines and had a mean value of approximately 900 ppm in the estuary. Concentrations of soluble zinc compounds, probably in an ionic or weakly complexed form, were as high as 3000 ppb in the tributary streams draining the mines and 100 ppb in the estuary, compared with values of 25 ppb in streams draining unmineralized areas (Table 23). The decrease in concentration from the tributary to estuary is prob-ably a consequence of dilution rather than self-cleansing of the stream by ion exchange with the bottom sediment (this postulate is speculative and should be tested out by field studies). Although cadmium flows were not measured, they are expected to be 0.5-1% those of zinc. No quantitative estimates are available of the amounts of cadmium and zinc leached from mines, but these are expected to be a major source of water pollution and may be inferred from concentrations in surface flows in mineralized areas.
As an example of the enhancement of pollution from mined areas, the drainage from strip mines has been found to yield an average of 30,000 tons of sediment per square mile annually, which is 10 to 60 times the yield from agricultural lands (101). Such a large enhancement of sediment drainage is not expected for zinc mines, since these are mostly subsurface, but the potential for significant drainage from mines, particularly those that are abandoned, should be noted.
Smelter Losses: Soluble salts emitted by smelters include the sulfates from the roasters, and the chlorides from sinter units.
From consideration of the process details, it appears unlikely that the sum of the emissions of the soluble salts could exceed 15% of the estimated 955 tons of cadmium emitted to the atmosphere by smelters. Additional losses are expected in waste waters from the hydrometallurgical plants, but these will not exceed 10% of production, i.e., 240 tons/yr, based on consideration of the reported cadmium recovery efficiencies.  (79).
Sewage: From consideration of the flow of cadmium in sewage treatment plants, it appears that, practically all the cadmium in the entering stream will be discharged in the effluent for installations employing only primary treatment, whereas most of the cadmium will be retained in the sewage sludge for those plants utilizing secondary treatment. Over the time interval considered, approximately half of the sewage underwent secondary treatment, so that 50%o of the cadmium content of sewage would have been discharged into surface waters. The cadmium-containing sludge is disposed of by a combination of burial, addition to soil as fertilizer, incineration, or wet oxidation. The major route of cadmium into sewage is from industries processing cadmium products and from surface runoff in cities where the storm drains and sewers are combined. The estimates of industrial losses are of the order of 150 tons/ yr, and the maximum value for surface runoff is 1000 tons/yr. These estimates may be compared with selected measurements of cadmium in sewage sludge and sewer effluent.
The mean concentration of cadmium in the sewage sludge of 22 communities in Massachusetts was found to be 12 ppm, with 10%b of the samples having concentrations in excess of 45 ppm (102). The Zn/Cd ratio in 54 samples from the 22 stations ranged from 50 to 150. In digested sewage sludges in the Chicago area, the concentration of cadmium and the Zn/Cd ratio were found to be 70 ppm and 48, respectively, at the Calumet treatment plant and 470 ppm and 14 ppm, respectively, at the Stickney plant (79). If it were assumed, somewhat arbitrarily, that 10%o of all sewage sludge had a concentration of 470 ppm, 30%o had a concentration of 70 ppm, and the remainder 12 ppm, the total cadmium flux in the sewage sludge would be of the order of 650 tons/ year. Since it had been earlier estimated that half of the total cadmium in sewage would be retained by the sludge, this corresponds to a flux of 1300 tons/year of cadmium. These calculations are intended only to indicate the order of magnitude of the cadmium content of sewage; reliable estimates of the cadmium content must be based on a more representative sampling of sewage sludges.
The concentration of cadmium in the effluent streams of some treatment plants are also available. At the Stickney plant referred to above, the cadmium concentration in the effluent of the preliminary treatment tank was 90 ppb, most of the cadmium was retained in sewage sludge obtained from a subsequent digestion unit, and the concentration in the final effluent was only 3 ppb (79). Concentrations of cadmium in the primary sewage effluent of the City and County of Los Angeles of 20 and 80 ppb have been measured (103). Based on a per capita waste water flux of 120 gal/day, the concentrations of 20 and 80 ppb extrapolate to national cadmium flows in sewer effluents of 600 and 2640 tons/yr, respectively. Although more measurements of cadmium concentration are needed before reliable estimates of cadmium flow in sewage may be obtained, the value is expected to be of the order of 1150 tons/yr.
Further evidence of the importance of sewage as a major route for cadmium input into the environment is the enrichment of the cadmium in the coastal sediments in area of 25-130 km2 in the vicinity of Southern California coastal sewage outfalls by factors of 20 to 100 relative to sediments far from the outfall areas (104,105).
River Transport: The cadmium concentration and fluxes in the Mississippi-Missouri and other selected rivers based on the measurements of Durum et al. (63) are shown in Table 24. Even after allowance is made for the problems of nonrepresentative sampling, the rate of cadmium flow and Zn/Cd ratios appear to be anomalous. The peak cadmium flow of 5690 tons/yr in the Mississippi-Missouri Rivers corresponds approximately to the total U.S. cadmium consumption (6020 tons/yr) and would appear to exclude losses from industrial operation as the major source of contamination. The cadmium flow increases rapidly in the mineralized area about the tristate Tennessee-Missouri-Kentucky border, suggesting that the increase in the cadmium flux results from weathering or mine drainage. But if this were the case, the Zn/Cd ratio should be close to that found in geological reserves. The anomaly remains unexplained, and may be a consequence of analytical difficulties (e.g., by selective removal of zinc or contamination by cadmium in the 0.45 ,u filters through which the samples were passed). In contrast to the above observation of anomalously low Zn/Cd ratios, analyses of sediments in Massachusetts streams (Table  25) show considerable contamination by cadmium and zinc without significant departure of the Zn/Cd ratio from the range found in geological reserves (102). (1 Ag/l.) 4 10/14-/70 60    in the open Atlantic Ocean adjacent to the British Isles but that they were significantly higher in the shoreline waters near industrial sources (more specifically, the east Irish Sea, where the most extensive measurements were made). Preston et al. furthermore found (56) a negligible difference between concentrations measured in 1961 and 1970 except for cadmium, the concentration of which decreased 24%o in the 9-yr time interval. Their study suggests that cadmium contamination is confined to shoreline waters. The failure to detect significant offshore contamination by the metals may be a consequence of deposition of most of the heavy metals in estuaries and near-shore sediment.
It is instructive to compare man's input of cadmium to the oceans with the natural fluxes. Bertine and Goldberg (21) estimate the global value of the sedimentary flux of cadmium to be 500 tons/yr by assuming that the cadmium concentration in the solids transported to the ocean (2.73 x 109 tons/ yr) due to weathering is that in igneous rocks (ca. 0.20 ppm). The fluxes from man's input into coastal waters are therefore much greater than that estimated for weather mobilization. It should be noted, however, that the present concentration of cadmium in the oceans ca. 0.11 ppb, is much lower than that which would be calculated from the cumulative natural input, suggesting that cadmium is being continuously deposited from ocean waters. From the concentration of cadmium in the oceans and the sedimentation rate it is estimated that the retention time for cadmium in the oceans is 500,000 yr (106).

Summary
The available data on ambient concentrations and emissions of cadmium in the environment are fragmentary and sometimes of questionable validity but, nevertheless, permit the determination of order of magnitude values for the major fluxes and reservoirs of cadmium in the environment. These estimates are summarized in Table 26 and Figure 5. For completeness, missing data have been estimated or guessed and the reader is directed to the earlier discussion of the various fluxes for an appreciation of the uncertainties in the figures cited.  Environmental Health Perspectives Air Contamination: From consideration of the overall flow, the major sources of contamination of the atmosphere are identified to be the emission from smelters, from furnaces reprocessing cadmium containing metals, and from coal-burning equipment. In all cases, the emission of cadmium could be greatly reduced by use of efficient particulate collection equipment and, in the case of smelters, by replacement of pyrometallurgical by hydrometallurgical processing. A substantial fraction of the atmospheric emissions is deposited in the vicinity of the sources, as evidenced by the accumulation of cadmium in the soil surrounding smelters ana boilers. Measured ambient concentrations of cadmium are in fair agreement with the values estimated from emissions, demonstrating the value of crude models of transport for obtaining preliminary estimates of exposure levels.
Water Contamination: The high fluxes of cadmium in the Mississippi River as it flows through the mineralized areas in Tennessee-Missouri-Kentucky, supported by similar observations in Conway, Wales, suggest that mine drainage may be the dominant source of water contamination. In addition, substantial amounts of cadmium are found in sewage, in amounts greater than would be expected for losses from cadmiumplating facilities. The fallout of atmospheric cadmium and subsequent entrainment in surface run-offs may account for the high sewage load. Most of the cadmium in surface flows is in solution. It should be noted that the cadmium solubility is strongly dependent on solution composition, particularly acidity, and that there is a potential danger that cadmium deposits that are presently immobilized may be dissolved if the acidity of rains, mine drainage, or surface run-offs increases. Cadmium in sewage is partly retained in the sludge and partly lost in the aqueous effluent (with the major fraction retained in the sludge in plants with secondary treatment). The cadmium losses to water streams appears to be ultimately deposited in the sediments in estuaries and along the shoreline, and may create a problem in these ecologically sensitive areas.
Soil Contamination: Cadmium buildup in soils surrounding smelters or other major sources has reached unacceptably high levels. The concentration of cadmium in the top layers of the soil may be estimated from the cumulative emission from the local sources. Additional cadmium is added to soils as contaminants in phosphate fertilizer and sewage sludge used for soil conditioning. Removal of cadmium from soils by plants, although significant as a path to man, does not provide a mechanism for appreciable depletion of the accumulated cadmium deposits. Recommendations 1. In order to facilitate the modelling of transport of cadmium through the environment, particle size distribution and chemical state should be measured in addition to mass emission at the sources of discharge into the environment.
2. The cadmium content of sewage, based on limited measurements, appears inexplicably high. Additional data are needed on the amounts and origin of cadmium in sewage.
3. The contribution of mine drainage to surface water cadmium contents needs to be quantitatively determined.
4. Material balances, albeit crude, and departures of Zn/Cd ratios from their natural values should be used, whenever possible, to test the consistency of data reported for ambient concentrations.

Effects on Man and Experimental Animals
The very extensive literature on the biologic effects of cadmium has been the subject of a number of recent reviews (13,30,107,108). Much information has come from studies of acute toxicity rather than from those directed to evaluating low level environmental effects.

Metabolism of Cadmium
Although there is little doubt that man and other mammals absorb cadmium through May 1974 the lung and by mouth, there have been very few studies designed to estimate quantitatively human uptake from the environment. Tipton and Stewart (109) described a balance study in three normal subjects over periods from 140 to 347 days. Using atomic absorption methods, the amount of cadmium and a number of other metals were determined in the diet, and the amount excreted in the urine and feces was measured.
The subjects were reported to have ingested an average of 170 jug Cd/day from the diet and to have excreted about 42 pg/day in the feces and 94 jug/day in the urine. Friberg et al. (30) point out that these results indicated an absorption of 75%o of the intake, much higher than reported by others. The urinary excretion levels also seemed very high and it seems possible that analytical errors resulted from sodium chloride interference (S0).
Bostrom and Webster (110) studied the intake in a human for short periods of 5 days. They found cadmium intake levels of 12 jug/day in each of two periods and fecal excretion of about 5 pg/day. The short duration of the study would make interpretation of absorption uncertain.
Rahola et al. (111) recently reported on work in progress in which five human subjects were given orally 115 Cd(NO)2 mixed with a kidney suspension. The dose used was 100 jug as Cd, containing about 5 ,ug Ci l1smCd, and body retention was studied by whole body counting. Urine and fecal excretion were also followed. About 6%o of the dose appeared to have been retained at two weeks. The subsequent rate of decrease in the amount retained in the body was extremely slow. The half-time for body retention was in excess of 100 days. Therefore total absorption of the oral dose of cadmium probably was not much greater than the 6%9 found to be retained after 2 weeks (Figs.  6 and 7).
Various animal studies also suggest that the absorption of cadmium from the gastrointestinal tract is poor. For example, Decker et al. (112) found 2.6%o of a single oral dose of 115Cd in the liver and kidneys of rats three days later. At 7 and 15 days after administration 2.0% of the dose was present in these organs. Thus, at least 2.6% of the dose had been absorbed, but probably not much more than in view of the fact that little or no radioactivity could be detected in muscle, lung, bone, spleen, and urine.
As part of a chronic toxicity study in rats, Decker et al. (118) measured cadmium levels in the kidney and liver. These contained about 0.30.5%b of the dose of Cd ingested in one year. Since 50-75% of the body burden will be in these organs, it is clear that the overall retention probably was only 1%b or less of the dose. Durbin et al. Another method of estimating absorption in humans is to determine body burden at autopsy and estimate total intake over a lifetime. Though subject to many errors, such calculations are compatible with an absorption rate of about 3-8% (30).
Although it is quite evident from its acute and chronic effect in occupational studies that cadmium is absorbed from the human respiratory tract, the precise degree and conditions governing pulmonary deposition, clearance, and absorption are unknown. Perhaps the earliest indication of respiratory absorption came from a report by Stephens (116) in 1920. A 67-year-old man who had worked in a zinc smelter for many years was alleged to have had lead poisoning, but his symptoms were atypical. At autopsy, no lead was found in the liver, but 120 ppm Cd and similar amounts of zinc were present. Eight other similar cases were noted. Subsequent -studies of autopsy material from persons with occupational exposure by Friberg (117), Kazantizis, et al. (118), Bonnell (119), and Smith, et al. (120) showed both liver and kidneys to have higher concentrations than normal. Lung concentrations were also higher, especially when the exposure was to insoluble pigments (118). Increased blood and urine levels have also been noted in exposed workers (30), but the relation to exposure and pulmonary uptake or deposition cannot be determined.
Generalizations concerning the retention of inhaled cadmium are difficult to make because of the likely variations in range of particle size and solubilities of cadmium compounds from different atmospheric sources. There is strong evidence that cigarette smoking contributes substantially to the cadmium body burden of smokers (121,122). Rough calculations suggest that the retention of cadmium inhaled in cigarette smoke is substantial. The following calculations lead to such a conclusion.
The body burden of cadmium due to cig- In industry the major hazard from cadmium is from inhaling cadmium oxide fumes.
These are probably particles having mass median diameters of 0.5 u or less. Cadmium oxide also is classed as being very poorly soluble. Using the International Commission on Radiological Protection (ICRP) lung model (124), one would expect roughly 40% deposition and less than 10%'o retention of inhaled fume particles, mainly because of predicted retrograde movement of particles to the pharynx with ultimate swallowing. There is no experimental confirmation reported for the specific case of cadmium oxide. Thus, the fate of inhaled cadmium is poorly known. In any event, ambient air is only a minor source of cadmium intake for the general population, cigarette smokers expected.
Estimates of cadmium retention in animals have been made. Prodan (125) studied the accumulation of cadmium in cats exposed to the fume, oxide, and sulfide for short periods. The largest percentages were found in the lungs, liver, and kidney. Prodan estimated the retention to vary from about 17 to 30%o. In the case of the sulfide exposure, virtually all of the cadmium was found in the lung. Friberg (117) exposed rabbits to a mixture of cadmium and iron oxide dust and found the principal amounts in the lung, liver, and kidney and an es-timated pulmonary absorption of about 30%0.
Little attention has been given to skin absorption. Skog and Walberg (126) applied a 115CdCL2 isotope to guinea pig skin and found 1.8%l absorbed in 5 hr. It seems likely that skin absorption is relatively insignificant.
The various studies (SO, 127) on absorption and distribution after parenteral injection will not be reviewed here. Cadmium is absorbed well from the injection sites and, again, seems to be stored mainly in the liver and kidney.
The exact mechanism by which cadmium is transported through intentinal mucosa is unknown although there has been some speculation that it may involve mechanisms similar to those of copper and iron transport. There now is substantial evidence that once it reaches the liver its presence stimulates the formation of an unusual protein of low molecular weight, discovered by Margoshes and Vallee (128) in 1957 and named metallothionein. It was originally found in equine kidney, but now is known to be present in the liver and kidney of many mammalian species. It usually contains about equal molar concentrations of Cd and Zn and has a molecular weight of about 10,500. It may contain as much as 5.9%b Cd and also has the ability to bind mercury.
Recent work by Nordberg et al. (129) has resulted in the finding of 60% of 109Cd attached to a similar protein in mouse blood after repeated subcutaneous injections over a 6-month period. The protein was attached to the red cell, but clearly separable from hemoglobin. The plasma levels were too low to allow separation and identification of the binding proteiii. Some portion of plasma binding was thought to be on a low molecular weight protein. This fraction may be filtered in the glomerulus and thus perhaps represents the pathways for tubular reabsorp6 tion, storage, and urinary excretion.
The excretion of cadmium in the urine has been reported to be approximately 1-2 Ag/day in the general adult population (180)(181)(182). Little is known concerning the relative importance of urinary and fecal excretion in man. The only study that provides any information indicates that 30 days following oral administration of an oral dose of 115Cd, the rate of fecal excretion was approximately half the rate of urinary excretion (111). There also is some evidence for excretion through the intestinal tract into the feces following parenteral injection in rats (112). In man there is a slight but statistically significant increase in cadmium excretion in the urine with age (180). No correlation with arterial pressure was found.
Hair normally contains about 1 or 2 ppm Cd (30). Although it is possible that this level may vary somewhat with exposure or intake, it does not seem to be a quantitatively important route of excretion. Attempts have been made to correlate hair levels with environmental exposure, but more work appears to be needed to evaluate the technique.
The levels of cadmium in blood and certain organs in the general population appear to be fairly well established and have been reviewed in detail (18,80). Some of the values for blood, liver, and kidney (80, 132-186) are given in Table 27. The correlation between the concentration in blood and other organs has not been established for man. Thus, blood cadmium has not as yet been shown to be a reliable index of exposure. Concentrations in other organs are considerably lower than in either liver or kidney.
The body burden of Cd in an adult is estimated to be about 30 mg. The newborn is said to contain only about 1 jug of Cd, so there is a gradual increase with age. Of the total body burden, 50-75%b will be in the liver and kidneys, about one-third of it in the kidney. Normally the kidney concentrations are 5-20 times those of liver. This may not be the case, however, in exposed persons. From the scattered data collected thus far, it would appear that the estimated body burden in humans varies with age from virtually zero in the newborn to a maximum of 10-18 mg in adults (80). Autopsy data suggest that this maximum is attained in middle life. There is some evi-Environmental Health Perspectives dence that the kidney levels may decrease slightly after that. Friberg (80) has speculated that there may be some loss from the kidney in older age groups but an alternate possible explanation is proposed by Hammer et al. (187), suggesting that the present older generation may have had lower exposures during their lifetime than the current younger generation. The fact that cadmium accumulates to a large extent in the liver and kidney is better established than the range of total body burden and its variation with age.
The concentration of cadmium in the kidneys is of special interest. Industrial experience suggests that the kidneys are es--pecially sensitive to the toxic effects of cadmium. Comparison of data from various sources is complicated by the fact that some report concentrations in outer cortex and others report values for whole kidney tissue.
The approximate concentration ratio cortex/ medulla is 2.
The threshold concentration of cadmium in kidney above which renal damage is likely to occur is estimated to be 200 ppm in outer cortex (30). This is approximately four times the concentration reported for adults in the general population in one study (186) and about six times the concentration reported for the category of moderate smokers of advanced age in another study (122). The margin of difference might be considerably smaller for moderate smokers in middle age, since the concentration of cadmium is known to fall in advanced age. Normal Human Intake of Cadmium from Environmental Sources The principal source of cadmium would Friberg et al. (80) have reviewed data on cadmium in food in various countries, and there seems to be some general agreement that food averages about 0.05 ppm Cd (wet weight) with, of course, wide variations depending on the source. There have been relatively few comprehensive studies of the total human intake via foods, but the available data would put the average of 50 jug/day or less with considerable variation. The best estimates of the U.S. food intake are by Schroeder and Balassa (186), Murthy et al. (139), and Duggan and Lipscomb (140) (Table 29). Further evidence of the approximate correctness of the figure of 50 jug/day is provided by data on daily fecal excretion of cadmium in the general population. Thus, Tsuchiya (141) reported that daily fecal excretion in four nonoccupationally exposed men was 57 Ag, and Essing et al. (142) reported daily fecal excretion to be 31 ,ug.
Two studies of water supplies would indicate that except for unusual instances of contamination, the intake via water is probably negligible (148,144). Table 30 summarizes data from the latter study on com- munity water supplies. The average intake from drinking water is about 1 or 2 jug/day. Little seems to be known as to the contribution of particulate content of water to the total cadmium content.
Airborne particulates or aerosols provide an additional source of Cd to the body. A large amount of data from 35 stations gave an average airborne Cd concentration (145) of 0.002 ,ug/m3. There was considerable variation, and Kneip et al. (27) found higher levels in urban air than in suburban air. For example, lower Manhattan averaged 0.023 /ig/m8 while suburban areas were 0.003 /ig/ms. Lee et al (28) studied the particle size distribution of metals in urban air, using cascade impactors. The mass median diameter was 3.1,u in Cincinnati, at a concentration of 0.08 Ag/ms. In suburban Fairfax, the concentration of cadmium was lower (0.02 ,ug/m8), and the mass median diameter was about 10 A. The mass median diameter for cadmium seems to have been considerably larger than those measured for lead and chromium in the same areas. On the average, the intake would seem very low, probably not more than a few micrograms per day.  (121) suggests that this may be an important item. Autopsies were performed on 172 adults, including 45 male smokers, whose approximate cigarette consumption was known, and 23 nonsmoking males. The mean age at death for each group was 60 years.
Cadmium levels in lungs, liver, and kidney were determined by using atomic absorption. The estimated body burden of cadmium in nonsmokers averaged 6.63 mg and was about double, 15.8 mg, in the smokers. Lewis et al. estimate that their data point to a nonsmoker retention of 1 jg or less per day compared with about 2.5 jug per day for smokers.
Estimates of total body burden from the data would give nonsmokers about 12 mg and smokers about 30 mg at age 60.
Szadkowski, Schulze, et al. (146) reported that about 1.4 ,fg Cd was found in a cigarette and estimated that 0.1 jg Cd per cigarette would be in the particulate phase, 0.03 ,Ag Cd in the gaseous phase. About 0.1-0.13 jug might therefore be inhaled per cigarette. The respiratory intake from two packs per day would be about 4-6 ,ug, or 10-20 times the intake from reported levels in the air of lower Manhattan.
Menden et al. (128) also estimated that the amount of cadmium inhaled from mainstream smoke was about 0.1 jug per cigarette. They point out, however, that the sidestream smoke contains considerably more total cadmium than the mainstream smoke. It is difficult to estimate how much sidestream smoke is inhaled by the smoker or by others in his vicinity. The figure is no doubt quite variable, depending on whether the smoker and others are out of doors or are in an enclosed room.

Environmental Health Perspectives
The amount of cadmium in dustfall in 77 midwestern cities has been presented by Hunt et al. (147). The geometricmeans varied from 0.040 mg/M2 per month in residential areas to 0.075 mgm/m2 per month in industrial areas. These levels would not appear to contribute significantly to human intake via soil or water contamination. However, localized increases are found in some studies in the vicinity of cadmium-emitting metallurgial operations (30).
In summary, it can be stated that-the intake for man under ordinary circumstances is principally from food, and most estimates would put this at about 20-50 JLg/day. Due to poor absorption from the intestinal tract, it is probable that only about 2 pg/day or less actually is assimilated (0.06 x 35-g) The intake from drinking water is presumably 1 or 2 pg/day on the average; due to poor absorption, probably only about 0.1 ,tg/day is assimilated. The intake from ambient air is also probably very low. Assuming 0.003 jug Cd/mi3, 40% retention of inhaled cadmium, and the inhalation of 18m3 of air per day, the daily assimilation of cadmium from ambient air would be approximately 0.02 Jpg. Air would not be a significant source of cadmium even at the levels reported for lower Manhattan (0.023 /Lg/m3).
For the nonsmoking, nonindustrially exposed U.S. adult, the likely daily assimilation can be summarized as follows: from air, 0.02 fLg; from food, 2.0 pg; from water, 0.1 Fg, for a total of 2.12 ,Ag.
This value approximates the reported daily excretion of cadmium, suggesting that adults in the general population are approximately in cadmium balance. This conclusion may be unwarranted, because available data on urinary excretion of cadmium do not distinguish between smokers and nonsmokers, whereas the estimate of daily assimilation for smokers and nonsmokers combined is probably appreciably higher than 2.12 pg and, therefore, probably higher than urinary excretion or perhaps even than urinary and fecal excretion combined.

Toxic Effects in Man and Animals
There are a number of well documented acute and chronic effects of cadmium in man and animal studies indicate a variety of toxic effects, the significance of which has not yet been demonstrated for man. Many animal experiments have been carried out using relatively large doses parenterally, but because of specific interest in some aspect of cadmium metabolism or toxicity, many of these experiments shed little light on potential, low-level environmental effects. Effects in Man: The effects in man can be classified as follows: acute oral, chronic oral, acute inhalation, and chronic inhalation effects. The acute oral effects are similar to those of zinc, and there have been many well documented epidemics of acute gastroenteritis from the ingestion of acid-type foods that have been stored in cadmiumlined containers (148,149). The dose causing the symptoms has been estimated (150) to be as low as 15-30 mg. The symptoms come on almost instantly after ingestion of the contaminated food, and the acute nausea and vomiting may in some cases be followed by a severe gastroenteritis. The actual oral lethal dose in man has not been established, but estimates have been made that it is probably in the neighborhood of several hundred milligrams (150).
As far as is known at present, no chronic toxic effects (except for a Japanese incident) have ever been reported from the oral ingestion of cadmium by human subjects. It is of course possible that some of the chronic toxic effects by inhalation may have been due to swallowing of material being cleared from the lung.
The second important acute toxic effect in man is that of acute pulmonary edema induced by inhalation of metallic fumes or cadmium oxide dust (13,30). It has been estimated that fatalities have occurred from a 5 hr exposure at about 8 mg/ma although in one instance recovery was reported after exposure to 11 mg/ms for 2 hr. In one brief account of a few intermittent exposures to cadmium fumes during silver soldering, a nonfatal acute pneumonitis was noted at concentrations estimated to have varied be-tween 0.5 and 2.5 mg/mrn over a 3-day period (151).
The acute pulmonary changes seen in man have been reproduced in experimental animals (129). Evidence does not indicate that acute renal injury is a part of the picture of acute inhalation toxicity with cadmium, nor is it known whether repeated acute exposures would subsequently result in chronic renal or lung injury.
The chronic exposure to cadmium through the respiratory tract produces a number of toxic effects, the most important of which is the chronic emphysema first described in the classical report by Friberg (117). Along with the emphysema, a peculiar renal disturbance has been noted, with excretion of relatively low molecular weight proteins in the urine together with some increase in amino acids and at times glucose and calcium. These findings have now been confirmed in a number of other publications and have been reviewed in detail by Friberg et al. (30). Among the more important of these studies are those by Kazantzis (154,155), and Princi (156).
It is apparent that chronic cadmium emphysema appears only after a period of exposure averaging about 20 years. It was not possible in most of these studies to arrive at the exact exposure levels since in some cases these may have occurred many years previously. However, from Friberg's initial study it would appear that exposure to cadmium oxide dust at levels of about 3-15 mg/ms may have been principally responsible. In a more recent study (157) of 11 subjects (eight of whom were smokers) exposed to cadmium oxide in the course of extracting cadmium from master alloys, no physiologic disturbances compatible with emphysema were noticed in the group. However, the length of exposure varied between 7 and 11 years, and the air concentrations were lower than those described in Friberg's original work (1.21-2.70 mg CdO/m8). Elevated excretions of cadmium in the urine were noted, varying from about 3 to 65 /Lg/24 hr. Although this study included an extremely thorough analysis of all aspects of respiratory function, it is unfortunate that no renal function or urinary protein levels were determined.
The renal effects have been particularly thoroughly studied by Kazantzis et al. (118), by Piscator (154,155), and by Smith et al. (120,153). From autopsy material it appears that the principal lesions were in the tubules, but in general the anatomical lesions were not very pronounced except in the most severe cases.
Kazantzis measured cadmium in the lung, liver, and kidney of one case (152) and compared it with data presented by Bonnell (119), Friberg (30), and Smith et al. (120,153). It is of interest that these combined data on these heavily exposed individuals indicated that the concentrations in the liver might actually have been higher than those in the kidneys and somewhat in contrast to normal individuals.
Another finding of interest in the Kazantzis work is that there was evidence in some individuals of increased output of calcium in the urine.
There may also have been some increase in incidence of renal stones in those with long exposure. Ahlmark et al. (158) also reported a high incidence of renal stones with long exposure to cadmium dust. Nicaud et al. (159) described x-ray changes characteristic of pseudo fractures in certain workers exposed to cadmium oxide dust. Findings of this type were not noted in later studies made by Friberg. Kennedy (160) studied serum calcium levels in rabbits given repeated injections of cadmium. There was a very slight fall in serum calcium which was interpreted as being due to possibly increased renal excretion.
The nature of the protein in the urine has been the subject of a number of studies, particularly by Piscator and his associates (154,155) Electrophoresis separations show that these have a molecular weight of about 20,000 to 30,00 and that they consist mainly of low molecular weight globulins. Some albumins may be present in minor quantities and these are sometimes Environmental Health Perspectives called minialbumins. There may be some increased excretion also of amino acids but, in general, the renal disability is not particularly severe. The renal changes, however, usually precede the development of emphysema and the proteinuria is usually accompanied by an increased cadmium output from the kidney. Friberg (30) interprets this as being due to renal damage.
The form in which cadmium is excreted in the urine is not known although it is possible that it may be in the form of a metallothionein complex.
Slight anemia has been noted in some subjects and it is thought possible that this might be related in some way to interference with zinc, copper, or iron metabolism. At present no explanation is available for the production of anemia. Anosmia has been reported in a group of alkaline battery workmen who were exposed to both cadmium and nickel dust (161).
In 1955 a somewhat unique disease was described in the vicinity of a mine in Toyama prefecture, Japan. The disease was epidemic among elderly women who had borne many children (average of 6). The outstanding features of the disease were lumbar pain and niyalgia, spontaneous fractures with skeletal deformation. Pain was readily elicited from pressure applied to bones. Extensive epidemiological studies were instituted after it was demonstrated that the water, rice, and fish in the endemic area were found to contain high concentrations of cadmium and other metals, probably due to contamination of the local river by the effluent from a zinc-lead-cadmium smelter. These studies continue and have been extended to include other areas in Japan where similar mining operations exist. The results of these investigations have been summarized by Friberg et al. (3a, 162).
The evidence available to date strongly indicates that this syndrome, termed itaiitai, is due to long-term cadmium exposure. It is the first likely instance of cadmium poisoning in man due to general environmental contamination. The characteristics skeletal changes found in the older women are not usually observed in industrial cadmium poisoning. They are ascribed to the interplay of cadmium exposure and certain other factors not usually encountered in industrially exposed groups, such as old age, low nutritional status, and multiparous motherhood. The neuromuscular signs and skeletal defects described in itai-itai, however, have also been observed in a series of cases of cadmium poisoning in France during World War II. Four women and two men were affected. All had been exposed for at least 8 years (159). Other cases of this type in industrial cadmium poisoning also have been reported (119,163,164). Thus, the musculoskeletal features of itaiitai are far from unique as manifestations of excessive cadmium exposure. Itai-itai is not solely a musculoskeletal disease. It is accompanied by the more classical renal effects of cadmium seen in industrial poisoning. Proteinuria was always found in clinical cases of itai-itai. Glucosuria and aminoaciduria also usually were present (30). Further, the incidence of proteinuria and glucosuria was much higher among older women and men in the endemic area then elsewhere in Toyama Prefecture. The urinary excretion of cadmium also was three times greater among people in the endemic area than in the nonendemic area of Toyama Prefecture. The prevalence of proteinuria and glucosuria was only somewhat greater among women than among men in the endemic area. Thus, while the musculoskeletal manifestations of itai-itai are seen almost exclusively among older women, the renal manifestations do not appear to be sex-related.
It has not been possible to develop from the available data a dose-response curve for itai-itai. Data on dietary intake of cadmium in the endemic area have been available only for recent years. The recent daily oral intake has been estimated to be 600 /ug, ten times the estimated intake for the general population in Japan (165). It seems likely that exposure of that order or higher occurred for at least 20 years and that such long-term exposure may have been neces-sary to attain the recent incidence of proteinuria among older women in the endemic area. Women born in another area but residing in the endemic area for 20 years or more have an incidence of proteinuria not quite as great as those born and living in the area all their lives (30). The statistical significance of the difference in the incidence of proteinuria in these two groups has not been reported.
More recent studies conducted in Japan indicate that excessive exposure to cadmium may be more widespread in that country than had been previously thought (162). Unfortunately, due to faulty experimental design, these studies have not provided any significant new information in regard to dose-response relationships.
The possible relation of cadmium to hypertension first proposed by Schroeder et al. (166) and by Perry (167) has been the subject of much discussion. The relationship is based primarily on the fact that some data suggest that humans with essential hypertension have more cadmium and a higher Cd/Zn ratio in their kidneys than those without hypertension. Experimental animals such as rats or rabbits are also reported to develop hypertension following cadmium intake by mouth or by injection. However, no relationship between cadmium levels in the kidney and cardiovascular disease has been found. For example, Morgan (168) determined the cadmium content of liver and kidney tissue from 80 individuals at postmortem and could find no significant difference between a control group and those with either hypertensive or other types of cardiovascular disease. There was also no significant difference in the cadmium-zinc ratio in these individuals. Szadkowski et al. (130) measured urinary excretion of cadmium in a large series of individuals and could find no relation between hypertension and cadmium excretion in the urine.
Epidemiological studies of people industrially exposed to cadmium do not support the idea that cadmium is a significant factor in hypertension (30). People occupationally exposed to cadmium differ from those not occupationally exposed in some respects that may invalidate the comparison. As an example, Hammer (137) found no rise in blood pressure in cadmium-exposed workers. As is often the case among workers exposed to cadmium, these men also were exposed to zinc. In view of the fact that zinc antagonizes certain vascular effects of cadmium, concurrent levels of zinc exposure may have an important bearing on the question of hypertension due to cadmium. Further research on this important topic is clearly needed.
Other Effects Demonstrated in Animal Studies: -One of the most interesting effects of cadmium in animal studies has been its ability to cause an acute necrosis of the rat testis following either parental injections or relatively large sublethal oral doses. This matter has been reviewed recently (13,30). Studies by Parizek (169) first established the extraordinary protective effect of zinc treatment against the acute atrophy of the testis resulting from cadmium injections. This has also been studied by Gunn and his coworkers (170). Cysteine and selenium also will protect against the cadmium testicular atrophy. By use of 109Cd, it was shown that none of the protective agents actually lowered the amount of cadmium reaching the testis. In the case of selenium there was actually an increased level of cadmium in the testis. Anatomical studies have demonstrated (171) that in all probability this acute cadmium testicular damage results from a toxic effect on the unique vascular system of the testis.
Because of the well-known effect of cadmium on the testis, a number of investigators have looked at the effect of cadmium on the course of pregnancy and fertility. Most of these studies have been based on cadmium injection. Parizek (172) showed that there could be fairly rapid destruction of the fetal placenta with only slight effects on the maternal placenta. When selenium salts were simultaneously injected, no placental effects could be produced. A teratogenic effect of cadmium has been demonstrated in the hamster (173). This effect was inhibited by zinc.

Environmental Health Perspectives
There is no evidence at present that the testicular effects found by injection in experimental animals are to be found in man. Favino et al. (174) investigated the fertility of ten cadmium workers and could find no evidence of infertility. However, one. individual stated that he was impotent and this individual had somewhat low testosterone levels in the blood. Friberg et al. (30) quote Cvetkova (175) as having reported somewhat lower weights in male and female children born to women working in a cadmium accumulator factory with relatively high levels of exposure to cadmium. Obviously many more studies of this sort are necessary before any conclusions can be drawn as to human effects of this type.
Studies of possible carcinogenic effects of cadmium have been primarily in experimental animals. The tumors found mostly have been of the sarcomatous type and have mainly been localized at the site of injection (176)(177)(178)(179)(180). In two of these studies metastatic tumors occurred in the regional lymph nodes and in the lungs (179,180). Interstitial cell neoplasms have been shown to develop in the testes of rats, both after direct injection into the testicle and after subcutaneous administration (181,182). Long-term feeding of cadmium also has been shown to increase the incidence of neoplasms in rats but not in mice (183)(184)(185). Kazantzis (personal communication reported that long-term feeding studies in rats nearing completion have not yet revealed any evidence of carcinogenic activity. Other studies reviewed by Shubik and Hartwell (186) and studies by Decker et al. (113) and Anwar et al. (187) did not reveal evidence of cadmium-related cancer formation in animals.
In view of the frequent demonstration of a carcinogenic effect in experimental animals, the implications for a carcinogenic effect in man need to be explored thoroughly. Indeed, some limited epidemiologial studies have been conducted in industrially-exposed populations. Potts (188) reported a high incidence of cancer among 74 men exposed to cadmium for at least 10 years.
Three of the eight men who died had prostatic cancer and two had other forms of cancer. The incidence of cancer of the prosstate was also reported to be high in another study of industrially exposed workers (four cases versus an expected number of 0.58) (189). Malcolm (190) concluded that cadmium does not cause cancer of the prostate in man in spite of the suggestive evidence.
In view of the relatively high concentration of cadmium in cigarette smoke, its possible role in bronchogenic carcinoma has received some consideration (30) but a causal role has not yet been demonstrated.
Finally, one of the most interesting aspects of the cadmium toxicity is the discovery by Terhaar et al. (191) that pretreatment with very small oral doses of cadmium would protect against the effects of subsequent very large or fatal doses of cadmium chloride. Doses as low as 10 jug/kg given orally 24 hr before a subsequent dose of 100 mg/kg completely protected against the massive testicular atrophy. Others have since found similar protective effects of pretreatment with small doses against a variety of lesions (192,193).
It now seems almost certain that the nature of this remarkable protective effect is due to the induction of metallothionein in the liver by subcutaneous or oral administration as demonstrated by Shaikh and Lucis (194). Metallothionein induction by small doses of cadmium has also been confirmed by others (J. Piotrowski, personal communication). Ability of cadmium to induce the formation of such a specific protein has raised a question as to whether this is not a fundamental protective mechanism against toxicity from this and perhaps certain other metals. Although metallothionein has thus far only been isolated from mammalian tissues, MacLean et al. (195) reported recently on the uptake of 109CdCl2 and of 65ZnCl2 by various microorganisms. The organisms in- I to extract both cadmium and zinc from the solution. Protein separations indicated that both the zinc and the cadmium were bound to a protein with similar properties to the metallothionein isolated from mammalian tissues. Whether such proteins represent the form of cadmium in plants is unknown but this seems possible. In this regard a recent paper by Neathery et al. (196) indicated a difference in the disposition of zinc fed to animals as a salt compared to naturally-occurring zinc. In this study 05Zn was added to media in which corn plants were grown.
Calves that had been placed for a short period on a zinc-deficient diet (2.7 ppm of zinc) were then given 65Zn by capsule in pure form or an equivalent dose of naturally occurring zinc in corn. These studies showed that although the absorption and the fecal excretion of 65Zn were similar, there appeared to be a difference in the soft tissue distribution, especially in tissues metabolically active such as kidney and liver. Similar studies need to be carried out with cadmium, since it is clear from other studies of these workers that ruminants have a strong homeostatic mechanism allowing the absorption of zinc but rejecting that of cadmium. Efforts thus far to demonstrate some essential role for cadmium have been unsuccessful. However, the apparently nearly universal occurrence of metallothionein or similarproteins containing roughly equal mols of cadmium and zinc and the obvious "evolutionary antiquity" of this type of protein continue to make this an important question. A review (197) of cadmium in the metabolism of albumin suggests that perhaps cadmium ions may play an important physiological role in regulating the biosynthesis of albumin and perhaps other proteins including some enzymes.

Evaluation of Environmental Hazard
One approach to the estimation of environmental hazard to man (198) is to compare known toxic and apparently nontoxic exposure levels in animals and man. Table 31 lists some of the more important studies and attempts to convert exposure levels to intake in terms of mg/kg (= ppm) for various species. The term intake as used here means the quantity entering the respiratory tract or mouth and does not include any estimate of systemic absorption from either route. It is necessary to evaluate the oral or respiratory hazard separately since the hazardous intake levels may differ. Tables  32 and 33 summarize some of the estimated hazardous intake levels in comparison with estimated average intake levels for man from the environment. On the basis of these data it could be concluded that the present U.S. average daily intake (assuming a 50-100 jug intake and a 70 kg man) via food and water is some 10-20 fold smaller than the uncertain estimates of Japanese intakes causing human renal damage. It is about 1/380 of the lowest reported level causing minimal adverse effects in rats on lifetime feeding, about 1/70 of the lowest level known to cause very minor or negligible effects in the dog over four years, and about 1/35 of the level in dogs causing no demonstrable effect in four years.
The inhalation intake from ambient air in a nonsmoker (assuming an ambient air level of 0.02 jug/m3, as in lower Manhattan) would be about 1/130,000 the level causing no pathological changes in dogs breathing CdO or CdS dusts daily for 37 weeks. It is about 1/2600 of the current threshhold limit value (TLV) level for for CdO fume. Smoking two packs a day has been estimated to give an intake about 10-20 times the ambient air intake. Although we do not know the form of cadmium in ambient air, it seems reasonable to assume it is a CdO or CdS particle.
It would appear that although there is a considerable safety margin for the environmental intake from both air and food, the margin is narrower in the case of food. This is true even if one corrects the values for estimated retention via the two routes. It is also probable that the inhaled concentration levels may play a relatively important part in localized pulmonary effects. This should be borne in mind in evaluating any possible pulmonary effects from cadmium in smokers since the concentration of cadmium in the May 1974 smoke is about 10,000 times that in ambient air in cities. The approach to this problem by Friberg et al. (30) is of interest. The assumption is made that there is a direct relation between the concentration in the target organ and the toxic effect. Using available data for the absorption rate from the gut or lung and the excretory rate in the urine, it is possible to calculate the daily intake required to reach a given level in the kidney in a stated length of time.
Assuming a 5% intestinal absorption, it would be necessary to ingest about 130 ug/ day over a period of 50 yr to reach the threshold level of 200 ppm in the renal cortex, or 260 jug/day over 25 yr. In estimating the accumulation of cadmium as a result of inhalation, lung retention was assumed to be 25%o. Friberg estimates that inhalation Environmental Health Perspectives of air containing as little as 10 jug Cd/m8 could result in the accumulation of 200 ppm Cd in the kidney cortex in 15 yr of industrial exposure (8 hr/day, 225 days/yr) even if the daily excretion were 0.01%o of the body burden. Again, using the assumptions of 25%o lung retention, Friberg et al. (30) have estimated that smoking one pack of cigarettes per day from age 15 to 50 could cause a 80% higher renal cadmium level than in a nonsmoker.
Although the assumptions made seem valid as a working hypothesis, there are a number of uncertainties. First the critical level in the target organ is derived from relatively sparse data on concentrations in human kidneys and on animal data. Secondly, the unusual nature of the cadmium binding protein in the kidney makes it uncertain whether overall Cd levels relate to toxicity or whether other secondary factors are important. Finally, we have little or no information regarding the absorption of the chemical complexes of Cd which exist in food, and analytically valid long-term human balance studies of the sort done by Kehoe (208) on lead are not available.
The induction of metallothionein synthesis in the liver by small doses may be a protective mechanism, but the effects of this on long-term toxicity are not known.
Environmental studies such as those described by Hammer (137) represent another approach. These are difficult and require expert planning. Where possible, some form of biological monitoring of the hazard needs to be a part of such studies, and these may need to be carried out over many years. It is unfortunate that neither cadmium urine or blood levels seem at present to be useful as lead or mercury determinations in prediction of hazard. Research on this matter needs to be continued. Based on present knowledge, and there are many gaps; there is no solid evidence of a hazard to the general population from cadmium in food, air, or water. Contamination of food and water rather than air would seem to present more of a potential problem. This might not be the case with smoking or excessive occupational exposures. The increased biochemical, toxicological, and epidemiological programs now underway should give the needed data for the answers to these questions.

Conclusions
This review has been concerned with an evaluation of the known toxic effects in man and of some of the other possible or alleged effects based on studies in experimental animals. Food is generally the most important source of cadmium in the environment, where it occurs in close association with zinc. The average intake from food is probably not more than 60 4g/day. Drinking water and ambient air contribute relatively little to the daily intake under normal circumstances. Airborne cadmium may become significant because of the probability of a higher percentage of respiratory absorption than of oral absorption. However, to date there is no evidence to suggest that ambient air concentrations of cadmium are presenting a hazard. This may not be true, however, if one considers the rather striking potential increase from cigarette smoking. It is evident that in the future epidemiologic studies will have to be much more carefully planned to rule out incidental effects such as smoking, and they should take into account all the known facts regarding cadmium metabolism and toxicity.
An estimate of the safe daily intake in the diet is at present very difficult because of a lack of appropriate information. This is especially true with regard to the nature of the chemical bond of cadmium in plants and meat and other foods. Predictions of toxic effects are also rendered difficult by the nearly universal occurrence of very much higher levels of zinc in combination with cadmium and the protective effect of zinc against cadmium toxicity.
The question of a buildup in the food chain similar to that seen with mercury compounds has been raised but there are reasons to believe this is unlikely. Fassett (13) has pointed out there are fundamental differences in the electronic structure of cadmi-um and that of mercury. Mercury is able to form extremely stable carbon-mercury bonds, probably because of the presence of an additional 32 electrons, as compared to cadmium (209). Although alkylcadmium compounds are known, they are extremely unstable and very unlikely to be formed or exist for very long in the natural environment. Furthermore, the methylmercury that is built up in the food chain is virtually completely absorbed from the intestinal tract, whereas the cadmium-protein complexes and other forms of cadmium are absorbed only to a very limited extent. In the unlikely event that an individual were to derive his sole source of protein from liver, kidney, or oysters, it is conceivable that the dietary cadmium intake might become hazardous. Until more is known about the absorption of such compounds, however, speculation is very tenuous. The accidental contamination of water supplies may present a very real hazard as well as that of the ambient air. The new data on smoking deserve a thorough investigation.
The role of metallothionein and its induction by small increases in cadmium intake is imperfectly understood. Cadmium bound as metallothionein may be biologically unavailable. If this is indeed the case, evaluation of the toxicological significance of cadmium levels in tissues (particularly the kidney) will require that the implications of binding as metallothionein be thoroughly explored.
Studies of interactions of metals in regard to absorption, transport, and toxic effects have been relatively few. In the case of cadmium, biological interactions with zinc and calcium have been shown. More studies are needed of interactions in order to provide a deeper understanding of the significance of epidemiological data regarding cadmium concentrations in biological systems. Thus, renal damage may depend not only on the cadmium level in the kidney but also on the concurrent level of zinc or other metals.
More effort must be devoted to definition of the kinetic constants which determine the rate of accumulation of cadmium in sensi-tive organs as a result of exposure via air, food, and water. The paucity of necessary information in this area was clearly revealed by Friberg (30) in his attempt to specify acceptable levels of long-term air exposure in industry.
There is no evidence to suggest that the general population is in imminent danger of excessive cadmium exposure. However, many people (particularly heavy cigarette smokers) probably have approximately one sixth to one fourth of the minimally toxic concentration of cadmium in their kidneys. This margin of safety may not be adequate since the minimally toxic concentration is estimated on the basis of very limited data. Further, there is a possibility that industrially exposed workers may have an abnormally high incidence of cancer of the prostate. Further research should be done to determine whether cadmium is carcinogenic under any circumstances of human exposure, and further efforts should be made to establish no effect levels for cadmium exposure in man.

Research Recommendations
The accumulation of cadmium among people in the general population is substantial. Knowledge concerning the margin of safety is inadequate. Such knowledge as we have is specifically largely limited to the margin of safety for adverse renal effects. It is estimated that current exposure is about one sixth to one fourth of the level necessary to produce minimal kidney damage. Essentially nothing is known concerning the margin of safety in the general population for other health effects of cadmium.
There is need for research in three general areas: (1) endpoints of toxicity; (2) disposition of cadmium in the body; (3) criteria of exposure.
Endpoints of Toxicity: Although much has already been learned concerning the renal effects of cadmium, present knowledge is still inadequate for the purpose of defining minimal toxic effects and the associated exposure. Proteinuria is currently considered to be the most sensitive index of toxicity.
More information is needed as to minimal exposure levels in man associated with this effect. Further work also should be done to explore the possibility that other renal effects may occur at even lower levels of cadmium exposure, such as effects on maximal tubular transport of organic compounds. More studies also are needed of the effects of cadmium on pulmonary function, again with the view of searching for effects occurring at relatively low exposure levels.
Since cadmium has been found to be carcinogenic in experimental animals, further efforts should be made to explore the implications of these observations. As an example, it should be established whether or not cadmium enhances the carcinogenicity of benzpyrene in experimental animals. Similarly, further epidemiological studies with industrially exposed and nonindustrially exposed populations should be conducted to establish whether there is any association between cancer and cadmium exposure. Disposition of Cadmium in the Body: It is essential to know the rate of input of cadmium necessary to attain minimally toxic concentrations in specific target organs. Kinetic studies of cadmium absorption, distribution, excretion, and accumulation are needed. Where possible, studies should be carried out on human volunteers. The technique of radioisotope dilution by use of stable isotopes of cadmium as the tracer is promising for gaining much essential knowledge about cadmium metabolism in man. Since a major concern is with the consequences of long-term accumulation, the influence of age on cadmium kinetics is important. Thus, while the daily rate of cadmium excretion may be 1% 7 of the body load in young adults, it may be 2%o in old people. It is also important to gain a better understanding of the factors which determine the biological availability or toxicity of cadmium in the body. The role of metallothionein in this regard needs to be explored. Similarly, the relationship of other metals to cadmium levels in tissues and to cadmium toxicity has not been adequately studied. Thus, reports on the incidence of hyperten-sion or other presumed effects of cadmium in industrially exposed people may not have much relevance to exposures in the general population because of differences in concurrent exposure to other metals, such as zinc, which influence cadmium toxicity.
Criteria of Exposure: Little is known about the relationship between the concentration of cadmium in readily procured biological samples and the concentration at sites of toxic effect in the body. The low concentrations of cadmium found in urine and blood as found in the range of exposure for the general population pose an analytical problem. More sensitive and precise methods for analysis of cadmium in biological specimens are needed. More effort also should be made to develop alternative approaches to measuring cadmium exposure, such as have been developed for lead exposure.

Cadmium in Plants and Animals Cadmium in Plants
Cadmium in low concentration is probably a normal constituent of all plant tissue, although it generally is not thought to be an essential micronutrient. The toxicity of cadmium to plants is believed to be much greater than that of zinc, an element with which it is commonly associated (210).
The concentration of cadmium in plant tissue is determined by the inherent ability of a plant species to absorb cadmium, and by the concentration of this element in the plant's environment. Where environmental cadmium levels are low, cadmium concentrations in plant tissue frequently vary more with species than with soil type (211). All soils analyzed in a recent geochemical study of Missouri (212) contained less than 1 ppm cadmium, yet the average cadmium content of ash from stems of hickory (Carya ovata) was 24 ppm, whereas the average content of ash from stems of white oak (Quercus alba) that grew with the hickory was only 4.3 ppm.
Where cadmium content of soils are higher than background amounts, the cadmium content of plant tissue tends to increase with increased concentrations of soil cadmium. For example, fescue grass (Festuca rubra) was found to contain 0.7 ppm cadmium in dry matter where growing on soil that contained 0.4 ppm cadmium, but contained 40 ppm cadmium where growing on soil that contained 26 ppm cadmium (47).
The cadmium contents of many species of plants reflect above-normal amounts of cadmium that are introduced into the environment both from natural sources, and from cadmium pollution of soils, water, and air. Cadmium from these sources may be absorbed by the plant through roots or leaves, or both, and thus be incorporated into the tissues, or airborne particulate matter containing cadmium may be deposited on the surface of leaves.
Certain peat bogs in Orleans County, New York, are enriched in cadmium by the entrance of ground water from dolomite beds that locally contain abnormal amounts of zinc, cadmium, and some other metals. After these bogs were drained for agricultural use, 15 vegetable samples from soil of the bogs were found to have a mean cadmium content of 0.12 ppm in dry material, with a range in concentration from 0.05 to 0.97 ppm (18).

Samples of cedar (Juniperus virginiana
L.) from a roadside site in Missouri where the soils were thought to have been contaminated by lead ore from passing ore trucks were reported (213) to contain an average of 9.3 ppm cadmium in the ash from leaf and branch samples, whereas similar samples from trees that grew at the same site but at greater distances from the road contained an average of 2.8 ppm cadmium.
Cadmium pollutioli of agricultural soils through irrigation water from mine drainage in Japan increased the cadmium content of cereal grains that grew on these soils. The polished rice contained an average of 0.49 ppm cadmium in the dry material, and wheat and barley contained several times as much as the rice (42).
Airborne cadmium that is deposited on soil can be absorbed by a plant and trans-ported throughout the plant. There is no effective method of removing this cadmium from the plant tissues. Goodman and Roberts (47) reported that if plants grew in areas where only a moderate amount of airborne pollution was present, no significant amount of cadmium could be removed by washing the plant samples, but where extreme pollution prevailed, washing removed as much as 45%o of this metal from grass samples. These results suggest that at locations where large amounts of airborne cadmium are present at least part of the total amount found by analysis of these plants occurs in surface deposits.
The area around Helena, Montana, is heavily contaminated by cadmium and other metals from smelter emissions. Miesch and Huffman (41) reported that concentrations of cadmium in the upper 10 cm of soil ranged from an average of 72 ppm at sites two-thirds of a mile from the smelter to 1.6 ppm at sites 7.5 miles from the smelter. The U.S. Environmental Protection Agency (133)  In studies of contamination from a zinc refinery in Japan, Kobayashi (42) found Environmental Health Perspectives  used as an index of airborne cadmium pollution in Sweden. The normal cadmium con-0.23 -tent of this moss was reported to range from 0.7 to 1.2 ppm in the dry plant. In the center of the Greater Stockholm area it contained as much as 7.5 ppm cadmium (215, ,rial by assuming 25%o 216). Oskarshamn, a city of about 25,000 population, has a storage battery factory that releases cadmium, chromium, nickel, and lead into the atmosphere.
In a study of airborne pollution in the area around Swansea, Wales, Goodman and Roberts (47) reported that the moss (Hypnum cupressiforme L.) contained as much as 9.5 ppm cadmium in dry matter where growing near centers of contamination (the moss could not survive in the areas of heaviest pollution), but only 2.0 ppm where growing in nonpolluted areas. In transplant experiments, samples of this moss contained in nylon mesh bags were suspended from trees at sites receiving various levels of airborne pollution. After 2 weeks of exposure, the cadmium content of the moss samples rose from the background value of 2.0 ppm to 337 ppm at locations of greatest airborne pollution.
A lichen, Clacdonia alpestris (L.) Rabenh., was sampled as an indicator of cadmium pollution from a recently constructed zinc refinery in Finland (59). The background concentration of cadmium in this plant is 0.1-0.2 ppm in the dry matter, but at locations near the refinery the lichen contained 1 ppm. These investigators stated, "Only six months of operation have increased the cadmium level fivefold at about 1 kilometers' distance from the zinc works, but no increase is yet noticeable at a distance of about four kilometers." Spanish moss (Tillanisia usneoides L.), a flowering plant that somewhat resembles true mosses, was sampled throughout its range in the Atlantic and Gulf coastal plains from North Carolina to Texas in a U.S. Geological Survey study (213). This plant has no roots, and grows as intertwined masses of stems and leaves on the branches of trees and other supports, and obtained all essential nutrients from airborne materials. Analyses of 122 samples showed that the lowest cadmium concentrations (2.0-3.8 ppm in ash) were in samples from nonmetropolitan or rural areas where the amounts of airborne cadmium are expected to be low, whereas the highest cadmium concentrations (20 ppm or more) were in samples from metropolitan areas that had concentrations of heavy industry and other potential sources of airborne cadmium. Table 34 summarizes data from the numerous reports of cadmium in plants or plant parts occurring in environments presumed to have the normal low levels of cadmium and in those from environments reported to have greater than normal levels of cadmium. Because most reports gave cadmium concentrations on a dry weight basis but did not give ash weights, the data have been converted to ppm in dry plant material. Some data based on ash weight, but without the ash yield being given, were converted to approximate concentrations in dry weight by assuming a reasonable ash yield for the particular kind of plant tissue that was analyzed. In instances of conflicting data, personal judgment was used in selecting the data to be included.
The values given in Table 34 must be used with caution in evaluating specific problems of cadmium in plants. The categories of plants and plant parts represented in this table were, of necessity, often very broad, and doubtless included certain species or varieties and plant parts whose normal cadmium contents differ greatly from the values given for the category. It should be noted that some of the cadmium values in the table were based on very few samples. Furthermore, for some kinds of plant materials, the cadmium concentrations in the environments where the plants grew were only loosely characterized, or were unknown. Cadmium in Terrestrial Animals Table 35 summarizes levels of cadmium in small samples of terrestrial animals analyzed by Schroeder and Balassa (136). With the exception of a few samples which may have been contaminated during processing, levels were generally low in the tissues of domestic animals raised for human consumption, and much higher in wild mammals and birds. Using kidney concentrations for comparison, those in domestic animals were of the order of one tenth of those in the wild animals samples, and of the order of one-hundredth of those in adult humans. Jaakola et al. (59) reported concentrations of 2.3, 5.5, and 51 ppm (dry weight) in the meat, liver, and kidney of an elk from south Finland. All the data quoted above were derived from very small samples and appear insufficient for statistical comparisons.
Schroeder and Balassa (186) determined cadmium in samples of leaves, twigs, and fruit of various trees used as food by deer. Levels ranged from not found to 0.43 ppm with a median of about 0.17 ppm, i.e., of the same order as the level of cadmium in human diets (see above). The levels of cadmium in the kidneys of deer (Table 35) were similar to those recorded in man in the first decade of life (186). Unfortunately, the ages of the sampled deer were not known, but 1-10 years is a likely range for wild deer, so it seems probable that deer concentrate cadmium at roughly the same rate as man. Table 36 summarizes levels of cadmium in soil, grass, and in the livers and kidneys of rodents (mostly mice and voles) at various sites in the polluted Helena Valley, Montana (138). With the exception of one group of samples collected within 1/2 mile of the smelter (sites 13, 19, and 20), cadmium levels in the kidneys (1.6-7.7 ppm, wet weight) were in the same range as those in the dried grass (1-7 ppm, dry weight). Cadmium levels in the livers (0.2-2.7 ppm, wet weight)  were 3-4 times smaller. However, there is no significant correlation between high levels in soils or grass and high levels in the rodents. The age of the rodents was not known, which is again a confounding factor in assessing the rate of uptake. Rabbits collected in the same area (mostly at station 6) had much higher levels of cadmium than the rodents (3.9-9.1 ppm, wet weight, in the livers; 19-61 ppm, wet weight, in the kidneys). However, in a feeding experiment, rabbits fed on a diet including 50-60% of lettuce from East Helena containing 5-12 ppm (dry weight) of cadmium accumulated no more than 1 ppm (wet weight) in 6 weeks (188). Hence the high levels found in wild rabbits of this area probably represent accumulation over one or more years. It might be expected that predatory animals would concentrate cadmium to higher levels than herbivorous mammals. However, the only measurement reported for a preda-  cadmium daily for 2 weeks resulted in concentrations of cadmium in the milk of less than 100 ppb (221). Table 37 summarizes the results of four studies in which cadmium concentrations were measured in several different organisms in the same area. Additional data on levels in marine animals are given in Tables  38 and 39. The cadmium levels summarized in these tables were measured and reported in various ways and are not necessarily quantitatively comparable.

Cadmium in Marine Animals
Where data are available on the distribution of cadmiurn within marine animals, the highest levels have generally found in the digestive and renal systems. Mullin and Riley (53) found that in mollusca levels of cadmium were generally less than 0.02 ppm in shells, of the order of 1.5 ppm in muscle, and up to 550 ppm in digestive glands and renal otgans. Brooks and Rumsby (227) found that in oysters cadmium was strongly concentrated in the gills, visceral mass' and heart. Schroeder and Balassa (136) found that in a lobster levels of cadmium were 14 times higher in the digestive gland than in muscle. Jaakkola et al. (59) found that in pike cadmium levels in the kidney  were 20-60 times those in the muscle (Table  37). In marine birds, levels in the liver were much higher than those in muscle or bone (Table 39).
On using levels in the whole body or muscle for comparison, most measurements in invertebrate animals in unpolluted areas fall in the range 0.4-2.6 ppm, dry weight (0.1-0.4 ppm, wet weight) (Table 38). Levels in fish (muscle) were generally considerably lower (less than 0.1 ppm, wet weight) (Tables 37 and 38). However, much higher levels were reported in zooplankton from the North Atlantic [mean 13 ppm, dry weight (220)] and in seabirds from the Antarctic [3-6 ppm, dry weight (225)]. The whales and seabirds are long-lived animals and may have accumulated cadmium over a number of years, but the mean life of the zooplankton is only of the order of months.
The higher levels in marine organisms listed in Tables 37-39 are generally associated with higher levels in the ambient water, usually the result of pollution. For example, at least three of the groups of seawater samples listed in Table 37 contained considerably more cadmium than the global average suggested earlier in this report, and even the fourth (the 1955 Irish Sea data) probably included some polluted samples. The apparent increase on cadmium levels in the Irish Sea between 1955 and 1969 is probably a reflection of additional pollution (228) and is reflected in higher levels in the sampled molluscs. The high levels in some samples of oysters from the U.S. reported in Table 38 (224) are probably also a reflection -of pollution. The fact that cadmium levels in petrels from California are roughly twice as high as those from ecologically equivalent species from the Antarctic (Table  39) may also be the result of pollution (225).
Tables 37 and 38 indicate that marine organisms concentrate cadmium to levels much higher than those in seawater. Concentration factors (ppm in fresh organism/ppm in seawater) appear to be of the order of 6000 for zooplankton [mainly copepods (222)], 104 for molluscs, 10W for echinoderms and crustacea, and 10-103 for fish. Higher concentration factors-up to 3 x 105 for oysters and 2 x 106 for scallops-have been recorded in laboratory experiments with very low levels of cadmium in water (227) but concentration factors higher than 105 have not been recorded in free-living molluscs, even in extreme individuals.
There is little evidence that cadmium is concentrated to a substantial extent in marine food chains. Martin's (222) figure of 6000 for the concentration factor in zooplankton is some seven times those recorded for marine phytoplankton (229). However, levels in plankton-eating birds (petrels) are no higher than those in zooplankton, despite the longevity of the birds (Table 39). Levels in fish are generally lower than those re-corded in marine invertebrates or even marine plants; levels in predatory fish (tuna and swordfish) are not markedly higher than those in fish low in food chains. Levels in the livers and kidneys of fish-eating birds (guillemots) were only of the order of 10 times those in fish in the same area (54). The only measurements reported for starfish (predators on molluscs) were lower than most measurements reported for molluscs (Table 37 and 38). However, in none of these examples were measurements made on successive species in a food chain.

Cadmium in Fresh-Water Animals
In a survey of 406 fish from 49 waters in New York State (230), 68.5%b of the fish contained 0.020 ppm or less (whole body residues, wet weight) and 23.5%o contained 0.020-0.100 ppm. Only eight fish contained more than 0.1 ppm cadmium, but these in-Hudson River. Somewhat higher levels of cadmium were found in fish from the Adirondack mountain area than elsewhere, probably reflecting higher natural levels of cadmium in the mountain lakes. No association between cadmium level and age was found in a sample of lake trout from Cayuga Lake.
Freshwater mussels samples in Massachusetts contained 1.7-3.7 ppm cadmium, wet weight (Massachusetts Department of Public Health, unpublished data), but the mussels had been transplanted as part of a pesticide monitoring program (P. Palermo, personal communication), and it is not clear whether these relatively high concentrations of cadmium (Tables 37 and 38) reflected background levels of local pollution in the rivers to which they had been transplanted.

Summary
Plants: Cadmium in low concentrations appears to be a normal constituent in all plant tissues. The concentration in the tissue is determined by the inherent ability of a plant species to absorb cadmium and by the cadmium concentration in the environment. Where the cadmium levels in the soils are low, the cadmium contents of plants vary more with species than with cadmium concentrations of the soils in which they grow. Beyond certain background amounts of cadmium in soils, the cadmium contents of plant tissue tend to increase with increased concentrations of cadmium in the soil.
Airborne cadmium, originating in the emissions from the combustion of hydrocarbons or from certain industrial processes, may enter the soils and be absorbed by plant roots, may be absorbed by the leaves, or may be deposited on the surface of plants in particulate matter, until very high levels of cadmium are accumulated by the plants.
There appears to be no natural means by which cadmium is eliminated from plant tissue, and no cultural practice has been found effective in reducing or preventing the absorption of cadmium.
Animals: Only scattered data are available on levels of cadmium in wild or domestic animals. Except for one study on trout, no data appear to have been published on levels of cadmium in animals of known age.
Among small samples of terrestrial animals, cadmium levels in domestic animals were of the order of 1/10 those recorded in wild animals, and of the order of 1/100 those in adult humans. No clear geographical correlation has been demonstrated between levels in herbivorous animals and levels in vegetation.
In marine animals, the highest concentration of cadmium recorded have been in pelagic zooplankton (13 ppm, dry weight), molluscs (locally up to 73 ppm, wet weight) and plankton-eating birds (20-53 ppm, wet weight, in livers). The highest levels in molluscs and birds appear to reflect human pollution. Invertebrates concentrate cadmium to levels 10W-104 or more above those in the ambient water; fish to levels 10-103 times higher. There is no evidence for concentration of cadmium in marine food chains.

Recommendations
A systematic, planned survey of the distribution of cadmium in natural environ-Environmental Health Perspectives ments is required. Data available at the present are fragmented among many reports and cannot be integrated into a uniform format because of differences in analytical procedures, various (or unreported) methods of sampling and preparation, and often poorlydefined samples. Reports should include measurements of moisture content and ash weight of samples, to permit interconversion to wet-weight, dry-weight, and ash-weight measurements.
The survey should be planned to serve three functions: (a) to provide baseline information on the distribution of cadmium in relatively unpolluted systems; (b) to investigate effects of locally high ambient levels of cadmium; (c) to identify locally high concentrations which may present a hazard to sensitive wildlife or the risk of high human intake. The survey should be combined with measurements of cadmium levels in soils and waters to provide information on uptake by plants and animals, and should be combined with controlled studies of uptake and retention.
The survey should be based on a small set of indicator species, preferably widespread, common species representing several types of environment and several trophic levels.
For plants, the selection should include grasses, trees, algae, and plants used for human food and feed for domestic animals. For animals, it is important to select species in which the age of collected specimens can be determined, either from anatomical characteristics or previous tagging. A list of possible candidates might include deer, voles, trout, crayfish, mussels, seagulls, and petrels.
The list should include at least one terrestrial predator (e.g., domestic cats), one predator on molluscs, and one plankton-feeder.
On the largest scale of sampling, the survey network might include one or two sampling sites per state. This should be supplemented with local studies in areas where significant gradients in ambient levels are known to occur, e.g., near smelters, highways, and polluted estuaries.
-Experimental studies of uptake over the lifetime of experimental animals are required for a number of representative species.
At least one food chain study should be made in each of three environments: terrestrial, fresh-water, and marine. Model ecosystems (microcosms) might be the most appropriate systems for these studies.
Field studies and laboratory experiments should be designed to determine the relative effectiveness of airborne and soil-held cadmium in influencing the content of this element in plants; that is, foliar versus root absorption. In addition, the properties of soils that affect uptake of cadmium by plants should be studied, with a view to developing cultural methods for reducing cadmium absorption in food or forage crops.
For convenience and economy, all the above studies could profitably be combined with surveys of other heavy metals.

Ecological Effects Effects on Plants
Reports of cadmium toxicity symptoms in plants grown under field conditions have not been found. The most complete publication on the toxicity of metallic elements and compounds to plants (231) does not mention cadmium as a possible phytotoxin. A bibliography of references to cadmium in soils and plants up to 1966 (282) includes no references to harmful effects of cadmium on plants. An older, much quoted reference (238) stated, "Very little is known of the effect of cadmium in plants or of its occurrence in soils. " John, Chuah, and Van Laerhoven (48) grew oat plants on soils contaminated by cadmium from a battery smelter, and reported, "Oats grown on the contaminated soils contained very high amounts of cadmium in the roots, with smaller amounts in the above-ground portions. Soil treatments affected the cadmium content of roots significantly, but did not affect the cadmium content of tops." In the experience of the writer, plants that were found by analysis to contain anomalous concentrations of cadmium in their tissues exhibited no symptoms of injury that could be attributed to this element. Reports found in the literature on cadmium concentrations in plants make no mention of toxic effects, except a Czechoslovakian paper (210) which stated, "All authors agree that cadmium ions are definitely more poisonous [to plants] than zinc ions," and that doses greater than 160 mg cadmium per plant reduced the inorganic phosporus content of tobacco, and produced necrotic spots on the leaves. The results are hardly applicable to plants growing under field conditions, because of the high concentrations of cadmium used, and because necrosis was produced only when the cadmium solutions were sprayed on the leaves. The same conclusion applies to the data of John, Van Laerhoven, and Chuah (234), who found that addition of 50 mg cadmium as chloride to 500 g of soil reduced the yields of radish and lettuce.

Effects on Animals
No systematic studies of the effects of environmental levels of cadmium on wild animals have been reported. Kobayashi (42) mentioned without detailed description damage to silkworms feeding on mulberry leaves containing 3-17 ppm cadmium (dry weight). In the Helena Valley, Montana (235), horses were reported to be markedly more susceptible to environmental toxicants than other species of farm animals. Chronically impaired horses had higher concentrations of lead (up to 35 ppm) and cadmium (up to 9 ppm) in their mane hair. One horse that died had very high levels of cadmium and lead in the kidney and liver, but these were not reflected by high levels in the mane (235). Goodman and Roberts (47) reported a similar case in Wales, in which a horse died with lethal levels of both lead and cadmium in the kidneys after feeding on grass and hay containing 7.6-9.9 ppm cadmium (dry weight).
The 24-hr tolerance limits for various fish' species under various conditions lie between Cd levels of 0.67 and 88 mg/l. (236). However, Ball (237) found that the toxicity of cadmium to rainbow trout was cumulative, the 7-day TLm being as low as 8-10 ug/l. Pickering and Gast (238) found adverse effects of chronic exposure to cadmium on the reproduction of fathead minnows: reduced survival of embryos was observed at levels as low as 57 jug/l. D. I. Mount (personal communication) reports that similar tests with other species of fish have shown adverse effects of cadmium in reproduction at levels below 10 jug/l. Sangalang and O'Halloran (239) report testicular damage in brook trout exposed to cadmium levels of 10 and 25 Jug/l. (as chloride), and gave evidence from in vitro experiments that the damage was associated with alterations in androgen synthesis. The level of 10/ug/l. associated with reproductive effects in these experiments overlaps with levels observed in natural waters (63, 64) but it is not clear whether results of experiments in aquarium conditions can be directly extrapolated to the more complex physical and chemical environment of natural waters.

Conclusions
Our ignorance of the effects of cadmium in natural or polluted systems is almost total. There is inadequately documented evidence of toxic effects on horses and silkworms in highly polluted areas. Laboratory tests suggest the possibility of adverse effects on fish reproduction even in lightly polluted areas. Some wild birds have been reported with kidney and liver concentrations near to those suspected of causing adverse effects in man.
On the basis of the scanty data available on the uptake, concentration, and effects of cadmium in natural systems, four main types of effect appear likely to occur, at least locally: (1) direct toxicity to plants and animals in polluted areas, especially near smelters, but also perhaps in areas with high natural levels of cadmium; (2) cumulative toxicity to predatoryf animals which eat the kidneys of their vertebrate prey (such effects should be sought for additionally in lightly polluted areas such as roadsides); (3) cumulative toxicity to animals which feed regularly on molluscs, such as starfish and seagulls (the possibility should be in-Environmental Health Perspectives 310 vestigated that such animals may have evolved biochemical immunity to cadmium toxicity); (4) cumulative toxicity and adverse effects on the reproduction of fish. The data of Ball (237), Mount, and Sangalang and O'Halloran (239), quoted above, suggest that such effects may occur with sensitive species even at natural levels of cadmium in fresh water, or in polluted sea water.
Further investigation is needed of these likely ecological effects. However, they are generally expected to -be manifested as rather subtle changes in natural systems, in particular replacement of sensitive species by tolerant species. Hence they may be difficult to detect without large-scale experimentation, and difficult to relate conclusively to cadmium in the presence of other pollut-

ants. Recommendations
Experimental studies are needed to determine levels of cadmium that are toxic to plants, and to define symptoms of toxicity. This may require only simple experimental designs and limited expenditures to determine that toxic thresholds are too high to have practical agricultural importance.
More information is needed on the toxic thresholds of cadmium in wild animals. At first, the species selected for testing should be those in which adverse effects are already suspected, or those in which very high levels are detected by sampling. Examples are silkworms, horses, grouse, and fish-eating birds. The tests should investigate longterm sublethal effects (e.g. on reproduction) rather than acute or chronic lethal toxicity.
In particular, toxic thresholds for effects on phytoplankton, zooplankton, and fish reproduction should be investigated. These experiments should utilize physical and chemical conditions appropriate to natural environments, rather than pure chemicals in clean, filtered water.
Another approach to the elucidation of possible effects of cadmium in natural systems is to compare species populations and community structure in polluted and unpol-luted systems. One suitable site for such a study would be an area around a smelter with marked gradients in ambient concentrations. These should be related to gradients in biological parameters such as the reproductive rate in voles, or the species composition of soil invertebrate communities. Other suitable comparisons would be between biological indicators in polluted and unpolluted estuaries, or polluted and unpolluted streams in the same area. Where differences are found, the causative influence of cadmium should be verified by experiments in model ecosystems.

Analysis for Traces of Cadmium
Trace concentrations of cadmium can be determined by many methods and there are many volumes devoted to the description of these techniques (e.g., [240][241][242][243]. Unfortunately, analyses for trace constituents are subject to many complications. A technique desirable for the determination of one element may be unsuitable for another element. The reliability of a method for a given element may vary with the nature of the particular sample being analyzed and with the proportions of the coexisting major and minor elements present. It would be most desirable to have a method of analysis for trace amounts of cadmium that would be sensitive, accurate, simple, reliable, inexpensive, rapid, require small samples, give reproducible results, and be unaffected by the presence of other elements and compounds. Such a method does not yet exist, and we must compromise by using the method most suitable for the type of material studied and the objective of the particular experiment. Some of the more common methods that can be used to determine trace amounts of cadmium are (1) spectrophotometric colorimetry (dithizone method), (2) emission spectroscopy, (3) atomic absorption, (4) neutron activation, (5) electrochemistry (anodic stripping voltammetry, polarography), (6) spark source mass spectrometry, and (7) isotope dilution. A selection of references to papers in which these methods have been used is given in Table 40.

Material
Method Reference Granite and diabase rocks Atomic absorption, mass spectrometry, neutron activation, Table 41 optical spectrography, polarography, spectrophotometry. Waters Atomic absorption, spectrophotometry, polarography, Dithizone, atomic absorption Table 42 These methods are briefly summarized below. It is not our purpose to give details of procedures or to discuss the necessary precautions required to obtain representative samples and avoid contamination, which are readily available in standard reference works. Tables 41 and 42 give an indication of the variation in analytical results obtained by different laboratories on samples of geological and biological material.
Further research on methods is clearly needed. It would be highly desirable also that many more interlaboratory comparisons be made on standard samples. Such samples have been available for some time of geological material, for example, the eight rocks distributed by -the U.S. Geological Survey, and several others (256). Until recently, the only standard of biological material has been the kale prepared by H. J. M. Bowen; the new standards prepared by the U.S. National Bureau of Standards of "orchard leaves" and beef liver should be very useful.

Spectrophotometric Colorimetry
The method is based on the measurement of the degree of absorption of light at a given wave length that is characteristic of a specific ion or complexion. Thus, the maximum absorption of light by a solution of cadmium dithizonate (the most commonly  ate complex is a necessary step, which can remove many interfering elements with proper adjustment of pH during the extraction. The method is less sensitive than some of the others, except where concentration is comparatively easy, as for sea water, in which levels of parts per billion can be determined (53). Special modifications, such as the use of ultraviolet spectrophotometry or fluorometric methods give promise of increased sensitivity of determination of cadmium. Emission Spectroscopy When elements are vaporized by means of a spark or an electric arc, the atoms present are energized to excited energy states; return of these energized atoms to the ground state is accompanied by the emission of light, the frequency of which is characteristic of each element. Resolution of these spectral lines and the determination of their intensities serve as the basis for estimating the concentrations of the trace elements present.
The method is most commonly used for direct analysis of solid samples, without pretreatment, because it makes possible the determination of many elements simultan-eously. The inherent variations due to varying composition of samples, variations of degree of volatility, etc. require careful standardization against known material of similar composition and the use of internal standards. The sensitivity for the direct method (generally 5-10 ppm) is not as good as in most of the methods discussed; it can be improved by concentration methods, followed by spark emission on solutions, or by arc emission of the evaporated dried solids. Such methods have not however, been used extensively.

Atomic Absorption Spectrophotometry
The basic principle of the method is that vaporized elements will absorb radiation of their characteristic frequencies by being activated from the ground energy state to a higher electronic energy state. The concentration of the atomized element is measured by the degree of absorption of its characteristic frequency of light.
Atomic absorption is one of the most widely used methods for determining traces of cadmium, because of its relative simplicity, speed, and sensitivity. It is commonly preceded by a concentration procedure, most often by dithizone extraction, to improve the sensitivity and to eliminate interferences, especially the serious interference caused by the presence of NaCl; some of the early data on cadmium in biological samples may be seriously in error because this source of error was not then known (30).
Recent modifications of the method, such as atomic fluorescence flame spectrometry (257) and flameless atomic absorption, give promise of lowering sensitivity limits for cadmium to well below 1 ppb Cd (258). Neutron Activation Many elements, when subjected to bombardment by neutrons in a reactor, form radioactive isotopes. The amount of a given isotope formed is proportional to the concentration in the original sample of the specific element, the neutron flux used, and the crosssection of the parent nuclide. Instrumental analysis of the energy of radiation and the decay curve is used to identify the desired radioactive isotope, the amount of which is determined by comparison with standards that are irradiated simultaneously with the unknowns and which have been carried through the identical separation and counting procedures (259). Rarely is direct counting possible; usually separations must be made from other radioactive isotopes that might interfere in the final counting procedure. The chemical separations are made after adding a known, generally much larger amount of the nonradioactive element after irradiation is completed. Although such separations may be quite complex, it is unnecessary to make quantitative recoveries of the element sought, because yields can be calculated from a knowledge of the amount of nonradioactive carrier added (250).
The neutron activation method is extremely sensitive. For cadmium the isotope 115Cd is generally measured, with sensitivity in the ppb range if chemical separations are made. The principal disadvantage is the need for a nuclear reactor and a "hot" laboratory. Electrochemical Methods (Polarography, Anodic Stripping Voltammetry) The polarographic method is based on the current voltage curves obtained by the elec-trolysis of solutions under special conditions, using a dropping mercury electrode. With proper choice of electrolyte, it is possible to obtain separate steps in the curves for each element present, the height of each step (current) being proportional to the concentration of -the ion, and the location (potential) being dependent on the ion and on the nature of the base electrolyte.
The method in its conventional form can be used to analyze solutions for cadmium at about 10-5M concentrations and with special techniques for concentrations down to 10-7M.
Anodic stripping voltammetry is essentially a polarographic method in which the element is slowly plated out of a small volume of solution on a small electrode (usually mercury-plated graphite) under carefully controlled conditions. After electrolysis is complete, a reversed voltage is applied, causing rapid dissolution of the plated element from the amalgam and thus producing a relatively large signal on the plot of current flow versus voltage.
The method is extremely sensitive and is especially useful for natural waters; its use may yield information on the nature of binding of cations in waters.

Spark Source Mass Spectrometry
Analyses for traces of metals by this method involves the volatilization and ionization of the material being analyzed by applying a radio-frequency spark, followed by measurement of the ions formed by their masses in a high-resolution mass spectrometer. The method is extremely sensitive, permits the simultaneous determination of many elements, and can be used for nearly all types of samples. The main problems are erratic variability of the emissions of ions from a spark source, which requires that an internal standard such as lutetium be used, and the formation of multiply charged ions. Precisions of about ± 10%o have been reported for rock samples (260); further research should improve this.

Isotope Dilution
The method of stable isotope dilution, ap-plicable to the determination of traces of any element composed of two or more stable isotopes, is based on the mass spectrometric determination of the proportions of two stable isotopes in a sample to which a known amount has been added of a "spike", i.e., of a sample of known isotopic composition enriched in one of the isotopes that is of low natural abundance (261). The procedure requires the complete solution of the sample, addition of the spike with thorough mixing, separation of the element from possible interfering elements, and determination of the isotopic composition. For cadmium, the spike might be enriched in 106Cd (normal abundance 1.22%) or 108Cd (normal abundance 0.88%o) and the ratio measured against 112Cd (24.07%o) or 114Cd (28.86%o). The method is extremely sensitive, with very high precision and accuracy. It has the advantage that quantitative recovery in separation is unnecessary; also the yield does not have to be determined, because a ratio, rather than an absolute amount, is measured. Contamination by reagents can be determined by parallel experiments. Although the method has not often been used for cadmium, probably because it is relatively slow and expensive, it is to be considered one of the ultimate means of monitoring faster and less expensive methods.