Model ecosystem studies of the environmental fate of six organochlorine pesticides.

Despite more than 20 years of intensive use there are major uncertainties about the environmental distribution and degradative fate of the various organochlorine pesticides in food web organisms. The problem is both controversial and momentous, as 88,641,000 pounds of cyclodienes and toxaphene and 59,316,000 pounds of DDT were produced in the United States in 1970 (1). The laboratory model ecosystem devised in this laboratory (2) has been used to estimate the comparative environmental properties of DDT, methoxychlor, and other DDT analogs (3-5). These investigations have demonstrated environmental degradative pathways and have provided quantitative data on ecological magnification and biodegradability index (6). This paper presents data on six additional organochlorine pesticides which should be of value in judging their overall effects on environmental quality. Furthermore, the data provide a realistic background of the environmental toxicology of standard substances against which screening data for new candidate pesticides can be weighed.

Despite more than 20 years of intensive use there are major uncertainties about the environmental distribution and degradative fate of the various organochlorine pesticides in food web organisms. The problem is both controversial and momentous, as 88,641,000 pounds of cyclodienes and toxaphene and 59,316,000 pounds of DDT were produced in the United States in 1970 (1). The laboratory model ecosystem devised in this laboratory (2) has been used to estimate the comparative environmental properties of DDT, methoxychlor, and other DDT analogs (3)(4)(5). These investigations have demonstrated environmental degradative pathways and have provided quantitative data on ecological magnification and biodegradability index (6). This paper presents data on six additional organochlorine pesticides which should be of value in judging their overall effects on environmental quality. Furthermore, the data provide a realistic background of the environmental toxicology of standard substances against which screening data for new candidate pesticides can be weighed. *Department of Entomology, School of Life Sciences, University of ilinois, Urbana-Champaign, ilinois 61801.
The model ecosystem evaluation was carried out in a small glass aquarium with a sloping terrestrial-aquatic interface of pure white sand exactly as described previously (2). A 5-mg portion of radiolabeled pesticide June 1973 was applied to sorghum seedlings grown on the terrestrial portion from an acetone solution by using a micropipet. The dosage is equivalent to 1.0 lb/A (1.1 kg/hc). Fourth instar salt marsh caterpillars (Estigmene acrea) were fed on the leaves until these were consumed, and their fecal products and the larvae themselves contaminated the aquatic portion of the system. The radiolabeled products were transferred through several food chains, e.g., alga Oedogonium cardiacum snail Physa; planktonwater flea Daphnia magnamosquito Culex pipiens quinquefasciatusfish Gambusia affinis. After 33 days in an environmental plant growth chamber at 80 ± 10 F and a 12-hr photoperiod, the experiment was terminated, and the amount and nature of the 14C determined by homogenization of the organisms, extraction with acetonitrile, TLC autoradiography, and liquid scintillation counting (2)(3)(4). TLC was carried out on silica gel containing fluorescent marker (E. Merck GF-254) with the developing mixtures shown in Tables 1-6, and autoradiography was performed on no-screen x-ray film (Eastman Kodak). Wherever possible, identity of degradation products was determined by cochromatography with known standards. Liquid scintillation counting was done in Cocktail D (7 g PPO, 100 g naphthalene in dioxane to make one liter), and counts were corrected to dpm by using channel ratioquenching corrections.
The results are expressed as equivalent ppm values in Tables 1-6 for the pesticides and their degradation products. The ecological magnification (EM), defined as ppm parent compound in organism/parent compound in water, and biodegradability index (BI), defined as ppm polar products in organism/nonpolar products (6), are used to provide quantitative characterization of the environmental behavior of these organochlorine pesticides.

RPsults and Disussion
The overall quantitative degradation and accumulation of the six pesticides investigated is shown in Tables 1-6.
Aldrin (1,2,3,4,10,10-hexachloro-1,4,4a,5, 8,8a-hexahydro-1,4-endo,exo-5,8dimethanonaphthalene) is well known to be converted environmentally to the 6,7-epoxide dieldrin. The data of Table 1 indicate that this conversion occurred rapidly and nearly completely in the model ecosystem and that dieldrin was stored in the various organisms as 95.9% of the total 14C in the fish Gambusia, 91.6% in the snail Physa, and 85.7% in algae Oedogonium (Fig. 1). In Gambusia, at the top of the food chain, only 0.5% of the 14C was stored as aldrin. Three minor degradation products were clearly visible in the autoradiographs of TLC plates developed in ether-hexane 1:1, Rf 0.63, 0.45, and 0.34 ( Fig. 1). These three degradation products were also found in the model ecosystem run with 1 4 C dieldrin. The compound having Rf = 0.45 cochromatographed with an authentic sample of 9-hydroxydieldrin and that at Rf = 0.34 with an authentic sample of 9ketodieldrin (Shell Chemical Company). The unknown II, Rf = 0.08 is probably transdihydroxydihydroaldrin formed by hydration of the 6,7-epoxide of dieldrin. The biodegradability index (BI) of aldrin for fish is 0.00014 and for snail is 0.0017 and the ecological magnification (EM) in fish is 3140 and in snail is 44,600. Aldrin is clearly an undesirable micropollutant in its own right, apart from its easy conversion to dieldrin. The EM values for the other degradation products in Table 1 provide useful infonnation about their relative environmental hazards. For dieldrin, the EM in fish is 5957 and in snail is 11,149. The 3-OH dieldrin has EM values of 617 in fish and 327 in snail and 3-C=O dieldrin of 220 in fish and 542 in snail. Thus aldrin and dieldrin provide the major environmental hazards from bioconcentration and the other compounds including aldrin trans diol (unknown II) are on the degradation pathways and are slowly being eliminated.
The exceptional environmental stability of dieldrin is indicated by the BI values of 0.0018 in fish and 0.0009 in snail, and the EM values of 2700 in fish and 61,657 in snail. As with the aldrin experiment (Table 1), the degradation products had lower EM values, 777 in fish and 5666 in snail for 3-OH dieldrin.
Endrm or 1,2,3,4,10,10-hexachloro-6,7epoxy-1 ,4,4a,5,6,7,8,8a-octahydro-1,4-endo, endo-5, 8-dimethanonaphthalene is the endo,endo isomer of dieldrin. As-shown in Table 3, it is somewhat less stable in the model ecosystem than dieldrin and is stored as 75.8% of the total 1 4C in Gambusia, 82.8% in Physa, and 84.9% in Oedogonium. The major degradation products were unknowns I (Rf = 0.83) and II (Rf = 0.52) (Fig. 2). The latter compound did not cochromatograph with an authentic sample of A-ketoendrin (Shell Research). By analogy with dieldrin, unknown II is probably 9-hydroxyendrin, one of the chief degradation products found in the rat together with 9-ketoendrin (9) iand 5-hydroxyendrin (M. K. Baldwin, Shell Research, Tunstall Laboratory, Sittingboume, Kent, England; unpublished data). Reference samples of these compounds were unfortunately not available for comparison. The BI for endrin was 0.009 in fish and 0.0124 in snail, and the EM values were 1335 in fish and 49,218 in snail. The EM value for unknown II was 270 in fish and 1701 in snail, supporting the conclusion that it is a hydroxylated degradation product with a lower lipid/H20 partition coefficient.
Biological observations on the model system experiment with endrin were particularly informative. The compound was highly toxic to the salt marsh caterpillar, which had difficulty consuming it. As the level in the water rose to 0.06 ppm the Daphnia and mosquito larvae in the aquarium were repeatedly killed and had to be reinfested. Most imnportantly, the water phase was incredibly toxic to Gambusia, which began violent convulsions within a few minutes of being added to the aquarium and died within a few hours. The high toxicity of the water persisted for more than 60 days from the beginning of the experiment and was associated with water concentrations of endrin ranging from 1 to 2 ppb. This toxicity delayed the termination of the experiment for twice the usual 33 days and provided a striking parallel to the Mississippi River fish kils associated with the leaching of endrin wastes (10). This experiment demonstrated the substantial predictive value of biological observations during the model ecosystem evaluation. Mirex (1,2,3,4,5,5,6,7,8,9,10,10-dodecachloro-octahydro-1,3,,4-metheno-2Hcyclobuta-[c,d] -pentalene) is one of the most stable compounds yet evaluated in the model system. Autoradiographs of the TLC plates containing extracts of organisms showed only unchanged mirex and traces of poiar material at the origin of the chromatograms (Fig. 2).  (11), where it was stored and slowly excreted virtually unchanged. These data suggest that mirex is a highly persistent micropollutant.
The data of preparation of lindane before it was repurifled. This degradation product cochromatographed with y-pentachlorocyclohexene prepared from lindane by reaction with sodium   (11), which is the key degradation product of lindane in the rat (12). Gambusia contained only lindane and a substantial amount of polar radioactivity. The four unknowns of decreasing Rf found in the water phase (Fig. 3) may be trichlorothiophenols as found by Grover and Sims (13) to be further degradation products offlindane in the rat.
The BI values for lindane were 0.091 in fish and 0.052 in snail, and the EM values were 560 in fish and 456 in snail. Thus lindane is considerably more biodegradable than the cyclodiene compounds.
Hexachlorobenzene has been used as a substitute fungicide and seed dressing to replace the alkylmercury compounds. Abbott et al. (14) have found residues in human adipose tissues in Great Britain ranging up to 0.29 ppm. It is evidently a persistant environmental micropollutant. In the model system it was found in substantial quantities in the various organisms with little evidence of degradation products except highly polar materials and conjugates (Table 6). Hexachlorobenzene comprised 85.1% of the total radioactivity in Oedogonium, 90.8% in the snail, 87.2% in Daphnia, 58.3% in Culex, and 27.2% in Gambusia. The water phase, however, contained an appreciable quantity of pentachlorophenol (Rf 0.5), identified by cochromatography. This compound was not found in free form in any of the organisms of the system. Hydrolysis of the polar products in the water, with HCL, showed a family of related compounds with Rf values of 0.27, 0.22, 0.17, and 0.10 ( Fig. 3) which are other chlorinated phenols. Mehendale (15) has tentatively identified, 2,4,5-trichlorophenol along with pentachlorophenol as urinary degradation products of hexachlorobenzene in the rat. The BI values for hexachlorobenzene were 0.46 in fish and 0.10 in snail, and the EM values were 287 in fish and 1247 in snail.
Ecological magnification in organisms of the food chain appears to be the most pemicious environmental effect resulting from the general usage of the organochlorine pesticides. This phenomenon, which is too well-known to need documentation here, is clearly revealed in the values of Table 7, showing the bioconcentration of the organochlorine pesticides from water to Physa snail and Gambusia fish. Bioconcentration clearly results from two important properties of Hexachloro-benzene (Rf= 0.80)a 0.00298 these organochlorine micropollutants, their high lipid solubility and water insolubility, i.e., a large lipid/water partition coefficient, and their resistance to degradation by the multifunction oxidase enzymes through which organisms protect themselves against xenobiotic pollutants. All of the organochlorine pesticides studied here have a high lipid/ water partition coefficient and this factor does not relate convincingly to values for ecological magnification which cover three orders of magnitude ( Table 7). As suggested by Hamelink et al. (16), the water insolubility of these highly lipid soluble compounds is likely to be the key driving force in producing lipid storage, e.g., in fish from successive partitionings from water to blood and blood to tissue lipids. This relationship holds surprisingly well for the organochlorine pesticides studied here, together with DDT, DDE, DDD, methoxychlor, ethoxychlor, and methylchlor previously investigated (2)(3)(4). As shown in Figure 4, when the logarithm of ecological magnification in Gambusia from the model ecosystem was plotted against the logarithm of water solubility, a clearly discemable linear relationship was found. Water solubility values were taken from Gunther et al. (17) and those for which no literature values could be found (    The question arises as to the extent to which the ecological magnification found in the organisms of the ecosystem (Table 7) is the result of successive concentration through food chains or of direct absorption through cuticle and gills. Some indications may be gained from experiments to determine the relative rates of concentration occurring in Daphnia, Culex and Gambusia exposed for 24-72 hr in one liter of standard reference water (18) containing approximately 1-3 ppb of the radiolabeled pesticide. The results in Table 8 show the extent of biomagnification of the pesticide in the three organisms. The various compounds are clearly concentrated manyfold directly from water into the organism and the degree of concentration varies with the individual pesticide. Bioconcentration is greatest with Daphnia, which has the largest ratio of surface to volume. There is a substantial trend toward increased accumulation with time, but some of the results show an apparent decrease with time suggesting degradation and excretion. The data with Gambusia suggest that a 3-day exposure to levels of pesticide contamination of the same general order of magnitude as found in the model ecosystem experiments does not produce levels of pesticide storage as high as those from ingestion of food chain organisms containing the pesticide. For the fish Gambusia, comparison of the values in Tables 7  and 8 suggests the following ratios for ecological magnification through food to biomagnifi-cation by direct uptake from water: aldrin 6.7X, endrin 2X, mirex 0.41X, lindane 3.3X, hexachlorobenzene 3.1X, DDT 250X, and DDE 125X. 'Biodegradation is, of course, materially enhanced by successive passage through food chain organisms.

Degradation in Salt Marsh Caterpillar
The contribution of the second element in the food chain, Estigmene acrea, to the degradation of each pesticide was measured by feeding individual larvae approximately 100 gg of radiolabeled pesticide and analyzing excreta and body homogenate by TLC (3). The organochlorine pesticides are very resistant to metabolism in the insect body as shown by the radioautographs of excreta and body homogenate shown in Figure 5

Condusions
The results of laboratory model ecosystem investigations with radiolabeled organochlorine pesticides support the value of this technique for study and measurement of the environmental toxicology of micropollutants. The organochlorine pesticides were shown to accumulate in the tissues of fish and snail to levels from 200 to 84,000 times greater than those in the water of the model system. There was a satisfactory inverse correlation between levels of ecological magnification and the water solubility of the compounds investigated. The model ecosystem studies demonstrated the chemical pathways for the environmental degradation of the pesticides and provided estimates of the acute toxicological hazards of the pesticides and their transformation products to the organisms of the ecosystem. The high values for ecological magnification and the low biodegradability indices for the organochlorine pesticides are in marked contrast to the low ecological magnification values and high biodegradability indices found with common organophosphorus and carbamate insecticides. Therefore, the laboratory model ecosystem provides a useful screening device for evaluating the environmental toxicology of new candidate pesticides.