m ‘— —' ’ " '“’ ‘ W " MWo'-V‘O‘-““WW.‘:‘V 4—‘ - -_ w A _ ,' w m . .-w.fim1mfiu'.w‘ ‘ v.. ., _ Pesncma RESiDUES m — . c030 SALMON EGGS 1 _ o THEIR, RELATEONSHIP T0 ' ' AND FRY mumm- ._ Thesis for the Degree of My 8. MiCHlGAN STATE UNIVERSITY 7 LES HENRY PECOR 1972 . L -; '{m 0‘ -' - V r .-... . . h t ‘ . ‘ I o's’~ fl - -g“ ' _ _ . . ‘9' . - c. ‘ . .,' .. . . . 1 ’ 1 . _. ' ' ' . v i; O -- Ayl- ' 0" >0 ‘ . ' O ‘I ~.|.' ' . - 1* a. ‘ fl 0. ‘ 0 ~ . _ :-"l ““ ' . 1 --- r . :3. (V . 7'..- v I I I . ' .‘- 1.‘ ac.» _ , '_ ' ‘ ' 3"; . ‘ I a f I ( - . .- ’42. I ~. ‘ ‘ . I . " ' I I. ‘ -‘.- ' .‘.' ‘ ‘t ‘ . ' ' ' ' _ I "“ c ' ' ’ a -o/.-: J " ',, w ' ' _ . . . 1 v ‘ ' - W v — | ‘2‘; - . - ' ' _ r . l ' l ' ’ 'A v . . o ' w ' . ' : ' - '. J - "'~ " “a" . .. .‘ 4 -_n I. - . . (—5 1 I" . . . f l I . J n-. ’ . . 7v!) .4! - ‘3}: i x. _ . 0. . .' ' ‘ '. . ' ' ' I . . r " l. . 4.m_ . . l P f ' . . . . ' ' . ’.. ,L . ' J. I 3' a f a) 1 i; "I.' l,“ r .o r' c . ._ u ’0’ ’l f “6"” .|r a . ' " .‘.’\ g I”, I. O .‘ 5” v I' ( -~ $g.‘ 7 .t ' ' 'l’I "' 'r >v' ' 1‘ 'I'. '- V'9 1". .' " a! ‘ 7 ‘ '-‘f "'5 4‘..¢a ! ' 0 I ‘- """ ' l .h'.’ " .. '1 401 1- .v-“WC . f .7 v; v". * u ' g.‘ ’1' 3'! 5., 5' h’ ‘1 l, 0"” 6'. . D. I ,0. ‘3’ '3‘“ IQ...“ ' I v'J‘. ' ' 'I I ' ' '. ' . ' I. — — 3‘ 1 ' ':. v, I . . a . . 0.. ’, . . «l 2. .,.1.. p‘. . .".'l. '.o‘ -R:."‘ -,‘, .l‘ 1,1,. .g a. I-{f ‘ . .3; . 1 ., 01.1’ ‘...I . '- I . . A; ,' ' .. J 3 £373. :.'.{ ’3’, "-f. "kfa-l .? I \ 0 ' . ' _ . , - v n; '0‘ O. ..o 1‘» / ..".-','.."~ I. I . p . , 1" -"; 4.":'\' 3"”. 'l . v . , u a . .. o '0' . _ . 1- J ; "flur’fid'ft a: o . 'J-V'f' ' ‘1' 1‘ ’ ' ' .H..foo00;h-~. ~ J' '1 " 14:. ‘0' ‘ ' . :IP a_ f" ’ ‘-'-' .. 2. .. 03,- Ana.» “va‘f-‘V-o‘éb'air'rfi' ' I V. I O IIIIIIIIIIIIIIIII IllTlllllllllllllfillllllillllllllltllSIHIYIIILI'flITfiliI'iIW ! mm, 3 1293 10642 4868 Michigan Safe .. University p.20 0 51239 ’1 ABSTRACT PESTICIDE RESIDUES IN COHO SALMON EGGS AND THEIR RELATIONSHIP TO EGG AND FRY MORTALITY BY Charles Henry Pecor During the fall and winter of 1968-69, the pesti- cide residues in eggs of coho salmon and the mortality of eggs and fry were investigated. Fertilized egg samples were collected from 104 individual female salmon from four Lake Michigan streams, two Lake Superior streams and one Oregon stream. An additional 96 egg samples were ob— tained for pesticide analysis only. Four major pesticide residues were identified and quantified in the salmon eggs: p,p'-DDT, p,p'-DDD, p,p'—DDE and dieldrin. The total concentration of these four pesticide residues in Lake Michigan eggs was approxi— mately 6 times higher than those in Lake Superior eggs and approximately 55 times higher than in the eggs from Oregon. Among the three systems, the mortality of fry in Lake Michigan groups was higher than in Lake Superior and Oregon groups. The Lake Michigan fry mortality was n? Vb v!” .al r unerir snara ml. he 5 f4. 4| W L Nd.- x M LL 3. E mug. “C. c \F H Y5 ”5 WA «nu :4 v; e by A c .x; .1; n9 .ph Y. _Q. .n a e by . L Nu. m Irun‘uflznlw W... 5 _- w Esti. U o h Charles Henry Pecor characterized by loss of equilibrium, erratic swimming and prolonged convulsions in response to a disturbance. The symptoms and mortality appeared abruptly during the final stage of yolk-sac absorption. Mortality among Lake Superior and Oregon fry occurred at an earlier age and did not show these symptoms. Statistical analysis of the mortality data on the eight fry groups within Lake Michigan did not show a correlation between pesticide residue concentration in the eggs and mortality of fry with the exception of one group. However, pesticide residue content and mortality among Lake Michigan fry were both significantly lower in samples collected later in the spawning run. No statistical relationship was found between pesticide residue concentration in the eggs and amount of fat in the eggs, parent fish length or egg mortality. Although the data did not show a statistical correlation between pesticide residue concentrations and fry mortality, it did show substantial circumstantial evidence to support the relationship and the results also suggested that other toxic materials or unknown factors may be involved in the fry mortality. PESTICIDE RESIDUES IN COHO SALMON EGGS AND THEIR RELATIONSHIP TO EGG AND FRY MORTALITY BY Charles Henry Pecor A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Fisheries and Wildlife 1972 i1 I am- 4 Dr. H. StUdY. ESPeci tribut vided atiOn '71. H‘e :‘d‘ 512.4 1 C7 ACKNOWLEDGMENTS I would like to express my special thanks to Dr. H. Johnson for his advice and guidance during this study. I am also grateful to the graduate students, especially Jerry Hamelink and Ronald Waybrant, who con- tributed to this study by discussions and suggestions. Recognition and thanks is given to the super- visors and personnel of the Michigan Department of Natural Resource's hatcheries and egg-taking stations for their efforts and assistance in the collecting of eggs from coho salmon. The financial support for this study was pro- vided by: The William A. Angell Foundation in cooper- ation with the Sport Fisheries Research Foundation, The Michigan Department of Natural Resources, Research and Development Division and FWPCA Training Grant - 5T1- WP-109. ii TABLE OF CONTENTS INTRODUCTION . . . . . . . . . General . . . . . . . . . History of Michigan Salmon . . . Objectives. . . . . . . . . MATERIALS AND METHODS . . . . . . Field Collections . . . . . . Sampling locations and schedules Sampling procedures. . . . . Hatchery Procedures . . . . . Incubation of eggs and early fry Fry rearing . . . . . . . Analytical Methods . . . . . . Extraction and clean-up of pesticide dues in salmon eggs. . . . . Gas chromatographic methods . . Proximate determination for dry and fat weights . . . . . . . . Procedures for pesticide residue identi- fication . . . . . . . . Statistical methods. . . . . iii Page 10 10 10 15 16 16 18 21 21 24 26 27 29 RESULTS AND DISCUSSION . . . . . . . . . . Mortality Study . . . . . . . . . . . Egg and sac-fry mortality . . . . . . . Fry mortality . . . . . . . . . . . Residue Identification . . . . . . . . . Gas chromatography . . . . . . . . . Thin layer chromatography . . . . . . . Pesticide Residue Concentrations in Coho Salmon Eggs 0 O O O O O O O O C C O O O I General . . . . . . . . . . . . . Relationship of pesticide residues to percent fat in the eggs . . . . . . . . . . Comparison of pesticide residue levels in coho salmon eggs collected during 1967 and 1968 from Lake Michigan, Lake Superior and Oregon . . . . . . . . . . . . . Differences in pesticide residue levels in salmon eggs collected from various streams tributary to Lake Michigan during 1968. . . Relationship between fish length and pesti- cide residues in the eggs . . . . . . . Comparison of the pesticide residue levels in the eggs of residual stream fish and lake run fish . . . . . . . . . . Relationship between pesticide residues in coho salmon eggs and sampling dates. . . . Relationship of pesticide residues in coho salmon eggs to the egg and fry mortalities . SUMMARY 0 O O O O O C O O O O O O O O LITEMTURE CITED 0 O O C C O O O O O C C APPENDICES O O O O O O O I O O O O O 0 iv Page 31 31 31 38 53 53 6O 61 61 65 68 71 73 75 76 80 87 90 93 Table LIST OF TABLES Sampling dates and locations with the number and type of samples collected during the 1968 run of coho salmon . . . . . . . Water analysis data for the carbon filter discharge (August 1968) and reservoir water (February 1969). . . . . . . . Average percentage cumulative egg and sac-fry mortalities of coho salmon for the dates each of the streams were sampled . . . . Average percentage mortalities of coho salmon fry for the dates each of the streams were sampled . . . . . . . . . . . . The relative retention times (p,p'-DDE = 1.0) and identification of the residue peaks present in the ether extract of coho salmon eggs . . . . . . . . . . . Correlation coefficient (r) data for the relationships between the four pesticide residues quantified . . . . . . . . Correlation coefficient (r) data for the relationships between pesticide residue concentrations based on wet weight, dry weight and fat weight. . . . . . . . The average percentage fat based on dry weight and correlation coefficients for percent fat and total DDT (DDT, DDD and DDE) in ug/egg . . . . . . . . . Average pesticide residues in coho salmon eggs collected during 1967 and 1968 (1967 data taken from Johnson and Pecor, 1969) . Page 14 19 32 39 56 64 64 67 69 Table Page 10. The average fork length of the fish sampled and the correlation coefficients (r) for fish length and average DDT concentration in their eggs . . . . . . . . . . 77 ll. Correlation coefficients (r) for the relationship of pesticide residues with sampling dates. . . . . . . . . . 77 12. Correlation coefficient (r) data for the relationship between DDT residues (ppm wet weight) in the eggs and fry mortality in samples from each stream . . . . . . 82 13. Correlation coefficient (r) data for the relationship between DDT, dieldrin and total DDT (DDT, DDD and DDE) expressed as ppm wet weight and ug/egg and the total fry mortalities of the groups from Lake Michigan . . . . . . . . . . 82 vi Figure 1. 2. LIST OF FIGURES Page A map of Michigan showing the location of all the streams sampled during the 1968 spawning migration of coho salmon in Michigan . . . . . . . . . . . . 12 Average cumulative mortalities of eggs and sac-fry, from fertilization to swim-up, of coho salmon collected at the various streams tributary to Lake Michigan and Lake Superior. T.C., Thompson Creek, (1) 9-17-68, (2) 10-13-68; B.C., Bear Creek, (1) 10-24-68, (2) 12-13-68; L.M., Little Manistee River, (1) 11-7—68, (2) 12-13-68; P.L., Platte River, (1) 10-17-68, (2) 11-14-68, (3) l-4-69; and Lake Superior, (1) ll-l-68. . . . . 37 The average weekly and cumulative fry mor- tality during the nine-week study for the two groups of Thompson Creek samples. . . 44 The average weekly and cumulative fry mor— tality during the nine-week study for the two groups of Bear Creek samples . . . . 44 The average weekly and cumulative fry mor- tality during the nine-week study for the two groups of Little Manistee River samples 0 O O I O O O O C O I O 46 The average weekly and cumulative fry mor- tality during the nine-week study for the two groups of Platte River samples . . . 46 The average weekly and cumulative fry mor- tality during the nine—week study for the two groups of Lake Superior samples . . . 50 vii Figure Page 8. The average weekly and cumulative fry mor- tality during the nine-week study for the single group of Oregon samples . . . . . 50 9. The average weekly fry mortality during the nine-week study for Lake Michigan, Lake Superior and Oregon samples . . . . . . 52 10. A typical chromatogram of the first eluate fraction of coho salmon eggs using the Micro-Tek 220 gas chromatograph. Peaks numbering 5,7,8 and 10 correspond to the pesticides p,p'-DDE, p,p'-DDD, o,p'-DDT and p,p'-DDT, respective. Peak number 4 corresponds to both MDE and o,p'-DDE . . . 55 11. A typical chromatogram of the second eluate fraction of coho salmon eggs using the Aerograph 660 gas chromatograph. Peaks numbered 16 and 17 correspond to DDE and dieldrin respectively . . . . . . . . 59 12. Scatter diagram and slope of regression line for the average DDT residue concentrations in coho salmon eggs from individual streams compared with sampling dates. . . . . . 79 viii INTRODUCTION General Effects of pesticides on reproduction in fish through chronic exposure to sublethal concentrations have been shown to occur in nature. Burdick e: 31. (1964) concluded the mortality of the lake trout fry (Salvelinus namaycush) in hatcheries from several New York State lakes occurred when the ether extract of eggs contained 2.9 ppm or more DDT based on the wet weight of the fry. Cuerrier gt il' (1967) found that when levels of DDT and metabolites exceeded 400 ppb in trout eggs, fry mortality ranged from 30 to 90 percent in the 60-day period follow- ing the swim-up stage. Kleinert (1967) indicated a possible association between DDT levels in the eggs and the mortality of walleye (Stizostedion vitreum) eggs from several Wisconsin lakes, and Johnson and Pecor (1969) reported DDT residues were a possible cause of high mortalities of coho salmon fry in Michigan hatcheries during 1968. Anderson and Everhart (1966) also reported high DDT residues in landlocked salmon (Salmo salar) in Lake Sebago, Maine and a failure of recruitment, but did not observe an abnormal mortality of hatchery reared fry. :- “? 1‘71 Several laboratory studies have documented effects of pesticides on fish reproduction. Allison 33 El. (1964) in long-term tests with cutthroat trout (Salmo clarki) found the mortality of developing fry increased in lots where the females were exposed to higher doses of DDT. Macek (1968) reported fry from brook trout fed higher doses of DDT suffered a greater mortality during the eight weeks following hatching. Johnson (1967) found abnormalities developed in medaka (Oryzias latipes) embryos when the females were exposed to 0.03 ppb endrin or greater. Other laboratory studies, with live bearers (viviporous species), have also shown effects of pesti- cides on reproduction. Mount (1962) stated that endrin in concentrations as low as 0.5 ppb curtailed repro- duction in guppies (Lebistes reticulata). King (1962) found DDT did not prevent reproduction in guppies but many were born dead or died within several hours. Boyd (l964) also found different insecticides may cause mosquitofish (Gambusia affinis) to abort. A number of workers have reported there is a critical period during the development of the fry when mortality occurs (Allison EE.2£‘.11962? Burdick 32 21., 1964; Currier et 31., 1967; Johnson, 1967; Kleinert, 1967; Macek, 1968). The critical stage varies with species, temperature and possibly other parameters not yet determined but the mechanism is assumed to be the same. It has been hypothesized that DDT and other persistent chlorinated insecticides, because of their high solu- bility in lipids, are concentrated in the conspicuous oil globules of fish eggs. It is further hypothesized that organochlorine pesticides are released from the yolk to the fry during the last stages of development. Smith (1957) reported that the triglyceride lipids are metabolized at this stage of development. It is during a similar critical period or stage of development that the coho fry from Lake Michigan experienced a high mor- tality (Johnson and Pecor, 1969). History of Michigan Salmon In the fall of 1964 the Michigan Department of Natural Resources received its first shipment of eyed coho salmon (Oncorhynchus kisutch) eggs collected from the Columbia River at the Bonneville Dam, Oregon. The fry were reared in Michigan hatcheries until the spring of 1966 when 650,000 were released in two streams (Platte River, Benzie Co. and Bear Creek, Manistee Co.) in the Lake Michigan drainage and 200,000 in the Big Huron River (Baraga Co.) in the Lake Superior drainage. In the spring of 1967 2.2 million smolts reared from coho eggs collected at the Cascade River, Oregon and the Toutle River, Washington, were released in four Lake Michigan tributaries (Thompson Creek, Schoolcraft I Co.; Platte River, Benzie Co.; Little Manistee and Bear Creek, Manistee Co.) and one Lake Superior tributary (Big Huron River, Baraga Co.). The fall of 1967 produced the first successful spawning run of mature coho salmon in Michigan. Approx- imately eight million fertilized eggs were collected at the Platte River, Bear Creek and Big Huron River egg- taking stations and distributed to hatcheries throughout Michigan. In 1968, 1.95 million smolts were released in 19 Lake Michigan streams, 8 Lake Superior streams and 4 Lake Huron streams. In the fall of 1968 approximately eight million fertilized eggs were collected from the second mature run of coho salmon and distributed to Michigan hatcheries. The coho salmon has a three—year life cycle. Juvenile coho salmon are released as smolts when they are approximately 18 months old and four to six inches in length. Mature adult salmon after 18 months in Lake Michigan ranged in size from seven to eleven pounds and two to four pounds in Lake Superior. In 1969 attention was focused on the existing pesticide residue levels in Lake Michigan fishes when the Federal Food and Drug Administration confiscated a large shipment of coho salmon because of high total DDT residues. Recent studies in selected areas of the Great Lakes have shown the presence of pesticide (John ‘nu REinE residues in all vertebrate, invertebrate and sediment samples tested (Hickey e£_al., 1966; Federal Water Pollution Control Administration, 1968; Brown and Hughes, 1969). Carr and Reinert (1968) found DDT residues in the flesh of all species of Great Lakes fishes with the residue levels in Lake Michigan fish two to four times higher than those from the other lakes. The eggs of coho salmon from Lake Michigan have total p,p'—DDT, DDD and DDE residues ranging from 3.5 to 7.3 ppm and p,p'-DDT residues of 1.0 to 2.5 ppm (Johnson and Pecor, 1969; Carr and Reinert, 1968; Reinert, 1969). The recent discovery of the presence of poly- chlorinated biphenyls (PCB's) in tissues of aquatic organisms and wildlife from various locations around the world (Koeman 32 31., 1969; Holmes 35 21., 1967; Reynolds, 1969) has led to some confusion regarding the identification of pesticide residues. PCB's may inter- fere with gas-liquid chromatographic analysis of chlori- nated hydrocarbon pesticides by producing residue peaks with retention times exactly the same as pesticide residue peaks. Lichtenstein et 31. (1969) has also demonstrated the potential toxic interaction of PCB's with DDT and dieldrin in insects. Veith (1970) has shown that chlorophenyl compounds are present in fish from the Milwaukee River and Lake Michigan with concentrations as high as 405 ug/gm body weight. Studies are currently underway at Michigan State Uni- versity, Pesticide Research Center, to develop methods of analyzing for the compounds and to evaluate the effects of these compounds in the environment. In 1967 the Michigan Department of Natural Resources reported an abnormally high mortality of coho salmon fry in their hatcheries. The mortality was restricted to fry from Lake Michigan sources. Similar mortalities were reported in other states where Lake Michigan salmon eggs were reared. Total losses in Michigan during this period accounted for 680,000 fry or approximately eleven percent of the original number of eggs collected (Michigan Department of Natural Resources, 1968). The fry mortality occurred one to four weeks after absorption of the yolk-sac, depending upon the rearing temperature (Michigan Department of Natural Resources, 1968). The mortality commenced during the fifth week after hatching, increased to peak numbers during the sixth and seventh weeks and decreased by the end of the eighth week. In each case the mortalities followed the same pattern relative to the development stage of the fry. The mortalities occurred during the period when the fry were undergoing a transition from dependence on yolk nutrition to hatchery diet. The fry mortalities were characterized by erratic swimming, loss of equilibrium, hypersensitivity and cessation of feeding. Many of the affected fish turned dark but this was not considered to be a specific symptom because many light-colored fish were also affected. Death usually followed one ro five days after the onset of symptoms. The affected fry gradually weakened, sank to the bottom, many in peculiar flexed positions, and died. There were no external or internal lesions observed although some did show degeneration of liver and kidney tissues but with no apparent correlation with fry mortality. Fry reared from Oregon and Lake Superior eggs did not suffer unusual mortalities and no evidence of symptoms were observed in these groups even when they were reared in the same hatcheries with Lake Michigan fry. Samples of affected and non-affected fry were examined by pathologists at the U.S. Fish and Wildlife Service, Eastern Fish Disease Laboratory in Leetown, West Virginia. No evidence of an infectious disease was found in the samples examined (Dr. Kenneth Wolf, personal communication). Additional tests by fish pathologists of the Michigan Department of Natural Resources failed to show any specific pathogen associ- ated with the mortality. The absence of any apparent bacterial or viral diseases led to speculation that insecticide contamination was a possible cause of the fry mortality. Objectives During the fall and winter of 1967-68 a study was undertaken at Michigan State University to examine the pesticide residues accumulated in the eggs of the first mature run of coho salmon in Michigan and the possible effects these residues might have on the off- spring (Johnson and Pecor, 1969). Results from this study showed the presence of pesticide residues in all egg samples and a possible correlation between pesticide residues and the mortality of the fry. This study was regarded as preliminary and exploratory. The present study was initiated in June 1968 and continued through July 1969. The major objectives of this research were: (1) Determine the identity and concentration of specific pesticide residues in coho salmon eggs from Lake Michigan, Lake Superior and Oregon stocks; (2) Determine the relationship of pesticide residues in the eggs of individual coho salmon to: (a) home stream or location sampled (b) size and age of parent fish (c) sampling date (d) mortality of eggs and fry MATERIALS AND METHODS Field Collections Sampling locations and schedules This study was based on egg samples collected from coho salmon in Michigan during the fall and winter of 1968-69. Samples of coho salmon eggs were obtained from all major streams tributary to Lake Michigan and Lake Superior in which mature spawning runs were expected. This included the Platte River, Bear Creek, Little Manistee River and Thompson Creek in the Lake Michigan watershed; and the Big Huron River and Cherry Creek in the Lake Superior watershed (Figure l). A shipment of coho salmon eggs from the state of Oregon was received by the Michigan Department of Natural Resources to supplement Michigan plants of salmon. Samples of these eggs were obtained from the Oden State Hatchery at Oden, Michigan to serve as controls for this study. Egg samples were collected at hatchery or egg- taking stations on each of the streams in conjunction with routine spawning Operations by state hatchery personnel. With the exception of the eggs from the 10 11 Figure l. A map of Michigan showing the location of all the streams sampled during the 1968 spawning migration of coho salmon in Michigan. 09° lAKE SUPERIOR RIVER CHERRY CREEK lAKE MICHIGAN 12 (:3 CANADA momnou Q CREEK PLAVIE RIVER IEAR CREEK MANISIEE RIVER lIIYlE MANIS‘IEE RIVER MICHIGAN lAKE HURON 54.. w .«.u a » VA r. r)». I E .(l . u .(x. U: . . .3 .C 3 a a .1. I e . Q C u at L . .9 . 1.4 a E 8 3|. 8 C . Hun. t E t» rt #L st. DA nku 13 Big Huron River and Oregon, all the samples consisted of fertilized eggs from individual females for comparison of hatching success and fry survival with corresponding subsamples taken for pesticide analysis. Approximately 500 fertilized eggs were obtained from each female. Samples of unfertilized eggs were also collected from additional females for pesticide analysis. The Big Huron River and Oregon samples consisted of single samples of approximately 2,000 eyed eggs each from a large number of individual females (Table 1). Four subsamples of approximately 50 eyed eggs were collected from each for pesticide analysis and the remaining eggs were reared in the laboratory. An attempt was made to sample each stream during the early, peak and late periods of the spawning runs. However, the unpredictable timings of the runs prevented such a sampling schedule. Samples were obtained at the beginning of the runs and again every 3 to 5 weeks until the spawning migrations were complete. The dates samples were collected are shown in Table 1. Additional samples were collected from the Platte River for pesticide analysis only. Unfertilized egg samples were obtained weekly for the duration of the coho salmon run in that river, extending from September 19 to November 14, 1968 (Table l). I. I Id. I II F F (C a. i WQHQEMW MHIOH ablhdlm UQHQEMW 0063 VNAMZUHVJ COWQnZOATH Zflm sousm mam ocm xmmnu muumcu monmmDm mxdq OH. OH OH. OH m m 6 OH. as m k w m nmnomaaoo 4 4 44 44 44 4 44 44 44 44 mwamfimm mouvna 4auaa Nauaa muaa vmuoa sauoa canes suoa mmnm moumaum wmamewm mama Hm>flm muumam OH. OH NH. OH NH wmuomaaoo 4 4 mmamfimm mfluma 4N-OH moumnoa cmHmEMm mums xmmuo Hmmm OH. 0H m. m k. m 6 ma omuomaaoo 4 4 4 mmamfimm mauma s-aa vmnoa omum monsum umHQEMm mums .m mmumflamz mauuflq HA. OH ma. OH cmpomaaoo 4 4 mmamfimm mauoa mousanm amHQEMm mama xmmuo semmfiona zmonUHE mm¢q .coEHmm onoo mo can mmma may mcfiuso omuomaaoo mmamsmm mo mama can Hogans mzu nuw3 mcoflumooH paw mmumo msHHQEMm .H mamas Ed' ta 6. 1.. +5 15 Sampling procedures Eggs from individual "ripe" females were gently stripped into a porcelain pan. A subsample of eggs for pesticide analysis was taken at this time. The remaining eggs were fertilized with milt from two to three males, rinsed several times and placed into two-quart glass jars filled with river water. The jars were then packed in insulated boxes for transit to the laboratory. In each case, samples from individual females were main— tained separately. Eggs fertilized in the field from streams in the lower peninsula of Michigan were placed in the incubator within 6 hours after fertilization and those from the upper peninsula were placed in the incubator approxi- mately 12 hours after fertilization. The samples for pesticide analysis consisted of 100 to 150 unfertilized eggs from each female. The sample was procured from the posterior end of one of the two ovaries if the eggs were "green" (eggs attached to ovary) or by a random sample of both ovaries if the eggs were "ripe" (eggs lying free in body cavity). The eggs were placed in sealed polyethylene bags with caution to avoid water and ovarian fluid, and imme- diately frozen over dry ice. Samples were maintained in frozen condition until analysis. 16 The data recorded for each female included its fork length and its total weight. A scale sample from each female was also obtained from the region between the anterior edge of the dorsal fin and the lateral line. The age of each female fish samples was deter- mined from the seasonal annuli formation of the scales (Rounsefell and Everhart, 1953). Four or five scales from each fish were washed and mounted between two microsc0pe slides. The scales were then projected using a micro-projector and a 43X objective. Hatchery Procedures Incubation of eggs and early :9: The samples of fertilized eggs collected in the field were reared in the laboratory at Michigan State University in a l6-tray salmon egg incubator. Each tray was subdivided into four equal compartments with a fiber- glass screen material. Thus the incubator had a capacity to contain egg samples from 64 individual females. The incubator, supported over an insulated 715- liter capacity reservoir, comprised a semi-closed recycling water system. Two submersible pumps delivered water to the incubators at approximately 7.6 liters/minute. The water entered the top of the incubators traversed down through all the trays and emptied back into the ti rd. .3 la Ea 17 reservoir. The recycled water in the reservoir was con- tinuously recharged with fresh water at a rate of approximately 26.5 liter/hour. The water supplied to the incubator system was treated East Lansing municipal well water which was passed through an activated charcoal filter to remove residual chlorine and sediments. The water in the incubator system was maintained at 10.2 : 0.l°C by a portable refrigeration unit. Fertilized egg samples received at the laboratory were acclimated to the lO.2°C water temperature of the incubator at a maximum rate of 2.0°C/half-hour. After acclimatization to the water temperature, individual samples were gently poured into separate compartments in the incubators. The number of eggs in the samples usually covered the bottom of the compartments with one layer of eggs. The incubator was in continuous operation from early September 1968 to early April 1969 when the last of the sac-fry were transferred to rearing tanks. The dissolved oxygen content of the water for this period ranged from 8.8 to 10.8 ppm as determined by weekly oxygen analysis. There was no measurable difference between the dissolved oxygen content of the water enter- ing at the top of the incubator and the water leaving the incubator at the bottom. Analysis of a water sample .-s-.-—. u-n. . collected pH of 7.95 was not de Re were main 57 days 0 first 16 sensitive aPpearamc ized eggs GEVelopm Swim-Up 1 the Comp bECauSe hatching EatiOn v“ Fr!) r W 18 collected from the reservoir in February 1969 showed a pH of 7.95 and a total alkalinity of 320 ppm. Chlorine was not detectable (Table 2). Records of mortality for each of the egg samples were maintained on a biweekly schedule throughout the 57 days of development in the incubator, except for the first 16 days after fertilization when the eggs were too sensitive to examine. Dead eggs were recognized by the appearance of white coagulated yolk material. Unfertil- ized eggs were characterized by a lack of embryonic development. The mortalities after hatching and before swim-up were recorded as dead sac-fry. Early mortality records were not available for the composite samples from Oregon and the Big Huron River because the eggs were received only a few days before hatching. Complete records of mortality during incu- bation were obtained for all other samples collected. Fryirearing The facilities for rearing coho salmon fry con- sisted of four 2.5m X 0.6m X 0.6m insulated tanks. Each tank was subdivided into 20 chambers by three rows of baskets. Each basket measured 25cm X 18cm X 30cm and was made of 0.32mm mesh nylon netting. The baskets were suspended in the tanks by rods traversing the width of the tanks. Rectangular glass-rod frames were 19 TABLE 2. Water analysis data for the carbon filter dis- charge (August l968) and reservoir water (February 1969). Concentration* Filter Reservoir pH 7.5 7.9 Total alkalinity 310 320 Hardness 362 380 Redox potential ~50 --- Dissolved oxygen 1.0 8.8-10.8 Free CO2 18.0 7.0 C1 ND < 40 ppb ND < 40 ppb PO4 total 0.9 1.0 Temperature 11.5 10.2 * All units in ppm with exception of pH, redox potential (mv at 20°C) and temperature (°C). ND = Not detected < 40 ppb placed desire 20 placed in the bottom of the baskets to maintain the desired position and shape. The water supplied to the tanks was the same as that supplied to the incubators (Table 2). Polyvinyl chloride (PVC) plastic pipe was used to construct a water distribution system that allowed a controlled flow of water to be jeted into each of the baskets. A total flow of approximately 6 liters/min. was maintained through each of the tanks. The dissolved oxygen con- tent of the water remained nearly constant at 6.5 ppm 1 0.5 ppm and the temperature ranged from 11.5°C to 12.5°C. Two weeks after hatching, approximately 57 days after fertilization, subsamples of 100 fry selected at random from each sample were transferred from the incubator to the rearing baskets. The fry at this time had a visible yolk-sac and had not yet started to feed. The fry were offered ground beef liver 3 times daily for the first week until they were actively feed— ing. At this time the liver diet was replaced with an Oregon Moist pellet (3/32 in.) which was ground to a fine consistency. The Oregon Moist diet was supplemented with ground liver one to three times weekly throughout the rearing period. The condition and behavior of the fry in each basket were observed every two days and the mortalities were recorded. The fry were reared under uniform 21 conditions so peculiar behavior or mortality of the fry could be identified. All dead and dying fry were pre- served in 10 percent formalin. Fry remaining at the end of the experiment also were preserved. All eggs and fry were handled similarly and exposed to the same hatchery procedures and water temperature regime. The single exception to these conditions was the first ten samples of Thompson Creek fry. These fry were reared in ten gallon aquaria divided by a plastic screen so that two groups could be accom- modated in one aquarium. The water temperature of these groups ranged from 13°C to 15°C. This difference from the other rearing groups was reduced by reporting mor- tality on the basis of "degree day" (degree day or temperature unit = [degrees F — 32] x days). Analytical Methods Extraction and cleanup of pesticide residues in salmon eggs All organic solvents used in the extraction, cleanup and analysis were redistilled to remove inter- fering artifacts (Appendix I). Glassware was washed in hot water and detergent, rinsed once with distilled water and twice with acetone. The egg samples were thawed and a subsample of approximately 20 unbroken eggs was removed, blotted dry .Hu 3 l :1) my my» 22 and weighed to the nearest 0.0001 gram. The eggs were mixed with approximately 20 grams of anhydrous sodium sulfate and 2 grams of clean ignited sand, and ground to a dry powder with a mortar and pestle. After thorough grinding the sample was extracted with four 20 ml portions of 94:6 petroleum ether:diethy1 ether and the fractions were collected in a 125 ml erlenmeyer flask. A standard Florisil column (Mills, 1959) was used in the clean-up procedure. The Florisil was reactivated at 150°C for at least 24 hours prior to use. Approximately 6 grams of activated Florisil was poured into a 25 mm diameter chromatograph column followed by 3 grams of anhydrous sodium sulfate. The column was prewetted and rinsed with 50 ml of petroleum ether which was discarded. The total extract was added to the Florisil column. The flask which contained the extract was rinsed twice with 10 m1 portions of petroleum ether which were added to the column. The column was eluted with 100 m1 of 94:6 petroleum ether:diethy1 ether and the eluate (total approximately 200 ml) was collected in a round bottom flask as the first fraction. The column was then eluted with 300 m1 of 85:15 petroleum etherzdiethyl ether (the diethyl ether contained 2 per- cent ethyl alcohol) and this second fraction was collected in a separate round bottom flask. The flow 23 through the column was maintained at approximately 6 ml/minute. Each fraction was then evaporated to 3 or 4 ml on a rotary evaporator and quantitatively transferred to graduated centrifuge tubes. The first eluate (6 percent EE/PE), which contained all the pesti- cide residues except dieldrin, aldrin and endrin, was suitable for gas-liquid chromatography (GLC) without further clean-up. The second eluate fraction (15 percent EE/PE) which contained the remaining pesticide residues was not suitable for GLC without additional clean-up. The second fraction was saponified with 20 percent alcoholic (95 percent EtOH) KOH and then partitioned with petroleum ether (Mills, 1961). Approximately 5 ml of alcoholic KOH was added to the sample and then the centrifuge tube and contents were placed in a hot water bath (80°C) for about 45 minutes. The contents were then transferred to a 125 m1 separatory funnel containing 50 ml of distilled water and 20 m1 of petroleum ether, and gently shaken for one minute and allowed to stand until the two phases separated. The aqueous layer was drained into a second separatory funnel containing 20 m1 of petroleum ether and extracted a second time. The petroleum ether from each extraction was combined and washed three times with 50 m1 portions of 50 percent alcohol. The ether layer was dried by passing it through a filter funnel containing afiwdrc quantil tube. this c Procec 98 to dield; taken 24 anhydrous sodium sulfate, evaporated to 3 or 4 ml and quantitatively transferred to a graduated centrifuge tube. The sample was suitable for GLC analysis after this clean-up procedure. The efficiency of the extraction and clean-up procedure using fortified reagent blanks ranged from 98 to 101 percent for p,p'-DDT, p,p'-DDD, p,p'—DDE and dieldrin. A second group of efficiency tests were under- taken using approximately 3.5 grams of Oregon coho salmon fry fortified with 2.47 mg of DDT, 0.93 ug of DDD, 6.19 ug of DDE and 2.95 ug of dieldrin. Analyses of unfortified fry samples showed only a trace of DDE. The efficiencies for these tests were 89 i 2 percent for p,p'-DDT, 91 i 3 percent for p,p'-DDD, 93 i 3 percent for p,p'-DDE and 94 i 4 percent for dieldrin. Pesticide residue concentrations reported in this study are uncor- rected for these percentages. Analysis of three random subsamples from single ovaries showed a coefficient of variability of approxi- mately 5 percent. Gas chromatographic methods All samples for pesticide analysis were analyzed by gas-liquid chromatography. Two gas chromatographs, a Wilkens Aerograph Model 660 and a Micro-Tek Model MT-220 were used to identify and quantify pesticide residues in the samples. 25 The Aerograph instrument was equipped with a 183 cm X 0.32 cm coiled glass column packed with 3 per- cent QF-l on 60/80 mesh Gas Chrom Q. The electron capture detector was of concentric tube design with a tritium foil source. The Operating temperatures used throughout the study were: column 186°C, detector 188°C and injection port 200°C. The carrier gas was commercial purified nitrogen adjusted to a flow of 35 ml/minute. The instrument developed 1,150 theoretical plates for dieldrin. The MT-220 gas chromatograph was equipped with a 183 cm X 0.64 cm U-shaped glass column packed with 3 percent SE-30 on 60/80 mesh Gas Chrom Q. The column was connected to a parallel-plate electron capture detector with a tritium source. The detector utilized a 50 volt pulse mode power supply, which had a pulse rate of 100 and a width of one microsecond. A purge gas of 95%-argon—5% methane was maintained at 6 ml/minute through the detector. Commercial purified nitrogen served as a carrier gas at a flow of 70 ml/minute. The operating temperatures were 180°C for the column and detector and 220°C for the injection port. The chromato- graph developed over 3,200 theoretical plates for p,p'-DDT. Samples were made up to volume in graduated cylinders that would permit sample injections between U. a. nu. Tnty ~ 26 l and 3 ul. Serial dilutions of pesticide standards were run each day to determine linearity curves and a basis for quantitative calculations. Care was taken to insure that the injected samples were within the linearity range of the detectors. Results were calculated from a linear regression of peak height (mm) versus picograms of insecticide. If the 95 percent confidence interval about the Y of the standard regression for a set of analyses exceeded i 10 percent, the results were deemed unacceptable and the samples reanalyzed by GLC. Results were reported on the basis of ug/sample, ug/egg, ppm wet weight, ppm dry weight and ppm lipid weight (ether extractable lipids). Proximate determination for dry and fat weights A subsample of eggs from each sample was used to determine dry weights and percent fat. These values were then used to calculate the concentration of the pesticides in the original sample. A subsample of from 50 to 100 unbroken eggs was counted, blotted dry and weighed to the nearest 0.0001 grams. The eggs were placed in an oven at 50°C for 24 hours and then trans- ferred to a desiccator for 48 hours. After this period the sample was weighed and the amount of the residue remaining was calculated as the dry weight. The lipid weight was determined by continuous extraction of the DI EV. 6 hro resi with c iie V 27 previously dried sample in a Soxhlet extraction apparatus with 100 percent diethyl ether for 7 to 9 hours. The ether extractable weight or lipid weight of the sample was the difference between the dry weight and the weight of the residue after ether extraction. Procedures forppesticide residue identification Pesticide residues in the coho salmon eggs from Lake Michigan were identified by GLC and thin layer chromatography (TLC). GLC was used to tentatively identify the pesticide residues by comparison of sample residue retention times with those of authenic pesticide standards. Confirmation of the pesticide residue identity was made by separate analyses on a second column and by exchange of samples with two other laboratories. Pesti- cide standards were made up in benzene solutions from purified standards (98 percent plus) obtained from the Pesticide Repository, Perrine, Florida. Thin layer chromatography was used to verify the identity of the pesticide residues present in the samples. The procedure used was described by Kovacs (1963) as useful for the chromatography of all chlorin— ated hydrocarbon pesticides. Additional clean-up of samples was necessary after the initial ex— traction and clean-up before the samples could be ‘J 28 spotted on TLC plates. The samples were partitioned twice with acetonitrile saturated with petroleum ether and then passed through two magnesium oxide-celite columns. The eluate was evaporated to 0.25 ml and spotted with appropriate standards on 20 cm square glass plates coated with aluminum oxide or silica gel G absorbents. The spotted plates were developed in 100 ml of either n-heptane or 20 percent ethyl ether in n-hexane for 20 to 30 minutes, air dried and sprayed with either 0.05 percent silver nitrate or Rhodamine B. The plates were placed under ultraviolet light for approximately 15 minutes to develop the spots and Rf values were calculated. The identity of the pesticide residues in the samples were further verified by removing the TLC absorbent containing the spots, extracting the residue with petroleum ether and injecting the sample into a gas chromatograph. Retention times were compared to those of pesticide standards. To evaluate the possible interference of other compounds, random subsamples of three to six egg samples from each stream were tested. The samples, after extraction for pesticide residues, were divided in half. One subsample was saponified with 20 percent alcoholic KOH and partitioned into an ether solution for injection into the gas chromatograph. The other 29 subsample was not saponified. The peaks which remained in the saponified portion of the samples with retention times equal to DDT and DDD were considered to be inter- fering residues. The percentage error due to the inter- fering compounds in the original analysis was determined by comparing the peak heights in the saponified and unsaponified portions of the samples. Statistical methods The analysis of variance formula for testing the hypothesis of no difference between means as presented by Sokal and Rohlf (1969) was used to test for differences in mean pesticide residues and mean fry mortalities between years, between sampling locations and between sampling dates. If ANOVA indicated significant dif- ferences within a group of means, then an "a posteriori" least significant range test (Sokal and Rohlf, 1969) was used to determine which means were significantly dif- ferent. An arcsine transformation was applied to all percentage data before analysis. \ Correlation coefficients were calculated for relationships between: all possible pairs of the four pesticide residues; DDT residues expressed as ppm wet weight and ppm dry weight, ppm wet weight and ppm lipid weight and ppm dry weight and ppm lipid weight; DDT residues and percentage lipid in the eggs; DDT residues and length of parent fish; pesticide residues and 3O sampling date; egg mortality and DDT, dieldrin and total DDT, DDD and DDE; and fry mortality and DDT, dieldrin and total DDT, DDD and DDE residues. The statistical probability of occurrence of the specific correlation coefficients was determined using tables from Snedecor (1956). Correlation coefficient values having a proba- bility of occurrence of 0.05 or less were accepted as indicating statistical relationship between the two variables under study. RESULTS AND DISCUSSION Mortality Study Egg and sac-fpy mortality The mortalities of the eggs and sac-fry for each of the groups of samples from individual streams were determined for each of four developmental periods (Table 3): start of incubation (6-18 hours), eye-up of the eggs (19 days or 360 temp. units), hatch (45 days or 810 temp. units) and swim-up of the sac-fry (57 days of 1,026 temp. units). The mortalities during the first developmental period, from fertilization to start of incubation (6-18 hours), consisted of both fertilized and unfertilized eggs. The mortality during this period was considered an indication of the success of the sampling and transpor- tation procedures. The low mortality, averaging 3.1 per- cent among all groups and ranging from 0.6 to 7.2 percent, suggested that the sampling procedures were adequate. The mortalities were independent of the sampling sites and there were no abnormalities observed during this early period. 31 32 TABLE 3. Average percentage cumulative egg and sac-fry mortalities salmon for the dates each of the streams were sampled. Sample Incubator Mortality Source Date Initial Eye-up Hatch Transfer 0 TU* 360 TU 810 TU 1026 TU 0 days 19 days 45 days 57 days LAKE MICHIGAN Thompson Creek 9-17-68 2.9 21.4 25.3 28.6 10-13-68 6.7 35.5 43.1 49.1 Bear Creek 10-24-68 7.2 26.0 41.9 42.1 12-13-68 1.5 9.4 32.9 33.6 Little Manistee ** River 11-7-68 1.8 33.0 65.5 71.8 12-13-68 0.6 11.2 65.4 66.5 Platte River 10—17-68 1.8 32.3 43.6 44.8 11-14-68 1.5 20.9 60.2 --- 1-4-69*** 12.5 85.0 96.8 97.5 Average 3.1 24.5 47.4 48.0 LAKE SUPERIOR Cherry Creek 11-1—68 0.9 11.4 38.0 41.4 * TU = Temperature units - [(degrees F - 32) x days] ** Include the result of two samples collected 10-24-68 *** Not included in average 33 During the second period of development, from start of incubation to eye-up of the eggs (visible eyes within eggs) there was an average mortality of 21.4 per- cent among all groups. The average cumulative mortality of 24.5 percent to eye-up was within the range of cumu- lative mortalities (7 to 30 percent) reported in Michigan hatcheries for the same period (Michigan Department of Natural Resources, 1968). The mortality consisted of both fertilized and unfertilized eggs. Also included in the mortality for this period was a relatively large number of unfertilized "blank" eggs that were removed after the eye-up stage. A little over half (11.1 per- cent) of the mortality for this period was accounted for by unfertilized "blank" eggs. There is no way of knowing what percentage of the dead eggs removed before eye-up were unfertilized eggs because an initial estimate of the percent fertilization was not obtained. However, a maximum of 88.9 percent fertilization was realized if the unfertilized "blank" eggs removed after eye-up are assumed to be the only unfertilized eggs. On the other hand, a minimum of 75.5 percent fertilization was achieved if the entire mortality up to eye-up, including the unfertilized "blank" eggs, is considered to be unfertilized eggs. The Michigan Department of Natural Resources (1968) reported a range of 86 to 89 percent fertilization for the eggs from the spawning run of coho salmon. 34 The eggs during the third period of development, from eye-up to hatch (26 days or 450 temp. units) sus- tained an average mortality of 22.9 percent. The average cumulative mortality up to the end of this period was 47 percent. The relatively high mortality of embryonated eggs was associated with a breakdown of the chorion just prior to hatch. The chorion in many eggs became very thin approximately 30 days after fertilization and ruptured with resulting protrusions of yolk material. Some of the embryos hatched early and survived but the condition was fatal to a large percentage of the embryos. The condition was first observed in the second group of samples collected at Thompson Creek. This group of samples was the second group to be placed in the incubator and every subsequent group, whether from Lake Michigan or Lake Superior streams exhibited chorion breakdown. There were no reports of this con- dition exhisting in Michigan hatcheries during 1968-69. The fourth period of development, from hatch to transfer (12 days or 216 temp. units), accounted for less than one percent of the total cumulative incubator mor- tality. The mortality consisted mainly of deformed embryos. The Michigan Department of Natural Resources (1968) reported losses ranging from 0.6 to 3.0 percent for the same period of development and a total cumulative mortality ranging from 30 to 45 percent. The average 35 cumulative incubator mortality of eggs and sac-fry at the end of the fourth mortality period in this study was 48 percent. The majority of the losses can be attributed to premature hatching and chorion breakdown of the eggs. The pattern of the egg and sac-fry mortalities were very similar with respect to developmental periods for both Lake Michigan and Lake Superior streams (Figure 2). The mortalities during the first and fourth developmental periods were very low but the second and third periods mortalities ranged between 8 and 54.2 percent. The last group of samples from the Platte River did not adhere to the general mortality pattern. These samples sustained the highest average cumulative mor- tality of all samples collected with an 85 percent loss by the end of the second development period. The egg quality of these late groups was considered poor because of low fertilization and low survival to hatching. The Michigan Department of Natural Resources (1968) also reported a higher than average mortality for eggs col- lected at the same time from the Platte River. This group of samples was not considered representative and was excluded from the analysis of egg mortality data. The average cumulative mortality for samples from the Little Manistee River was significantly greater than the mortalities of samples from other streams 36 .mmuauaa lac .uofiummsm mxmq ..m.q 6cm “monqua Ame .mwnvanaa ANV .monanOH AHL .um>flm muumam .q.m .mo-ma-ma Ame .mous-HH lav .Hm>flm ompmucmz maupuq ..z.q “mmnmauma Ame .mouvmuoa has .xmmuo Hmmm ..o.m “monmanOH Ame .monsanm lav .xmmuo sommEonB ..U.B .Hoflummsm mxmq ocm cmmfinoflz mxmq on mumpsnfiuu mammuum msowum> mcp um monomaaoo coEHmm 0:00 we .msnfia3m on cofiumN IAHHUHTM Eoum .MHMIUMm cam mmmm mo meUHHmpHoE m>flumasaso mmmum>< .m mudmflm 37 cotangt: .33.. «>01 u! _ ... 32 03 con v.22: £2”.— a... me o. 3.3 \\...‘.\.\..*..§\Lv \\\\\M..u\...\..V\\\ \ \\\ \o ooooooo o\\\ x \\\\..».\k.. \ \ \\ \\\\\ ooooooo o\ \ \\\ \ ON \\ ...... . \ \\\. \ \‘ \\\\ ....... V\ \\\\ X\ \ A: Uh \\\\\ oooooooo \ \\\\\\\ \ .5 un IIIIIIIIIIIII .\\ ............ \\ \\\ \ .\\ \\ \\\\\\\ \ coo-cocoooooooooooooloh\\\\UH\ .\. is anwunnulwfluhhuhh\\\\ \ \. .\ \ 2:...\ \ \ . \ \ ANVUH ANVCE\ o\ \o \ \ .\. \. .\ \ o o\ \ '0‘ .«EE..I.I...I:..I.u.\J\\ . \ \. \.\ \ A: mid . 3:... \ \ coo INIDUSd ALI'IVIIOW 38 sampled in the Lake Michigan and Lake Superior watersheds. The differences in the average cumulative mortalities of samples from Bear Creek, Platte River, Thompson Creek (Lake Michigan) and Cherry Creek (Lake Superior) were not significantly different (p < 0.05). The egg and sac-fry mortalities observed in this laboratory study were similar to the egg and sac-fry mortalities in Michigan hatcheries during 1968. The mortalities in each case were considered high when com- pared to that in West Coast hatcheries (Michigan Depart- ment of Natural Resources, 1968). primortality Observations of fry mortality were initiated at the time the sac-fry were transferred from the incubators to the rearing tanks, approximately 57 days (1,036 temp. units) after fertilization, and continued for 63 days (1,371 temp. units) after transfer. The average cumu- lative mortalities of fry from each of the streams sampled are shown in Table 4. Mortalities of fry from individual females showed a wide variation, ranging from 0 to as high as 90 percent. In recording the daily mortality in each group the symp- toms of the dying fry were carefully noted. Specific symptoms which were common to a large percentage of the dying fry were: loss of equilibrium, erratic swimming, spasmodic and convulsive movements when disturbed, long 39 TABLE 4. Average percentage mortalities of coho salmon fry for the date each of the streams were sampled. Sampling No. of Fry Mortality Date Groups 1371 TU (63 days) Source With Total symptoms LAKE MICHIGAN Thompson Creek 9-17-68 10 23.4 18.3 10-3-68 8 44.5 39.1 Bear Creek 10-24-68 10 37.6 31.6 12-13-68 8 2.5 < 1.0 Lt. Manistee River 11-7-68 9 22.0 18.6 12-13-68 7 10.0 7.0 Platte River 10-17-68 10 27.4 21.6 11-14-68 1 41.0 38.0 1-4-69 4 32.7 26.4 LAKE SUPERIOR Cherry Creek ll-l-68 10 55.6 < 1.0 Big Huron River 1-20-69 14 6.1 0.0 OREGON Oregon 1-4-69 10 11.3 0.0 * Symptoms described in text 40 periods of inactivity while laying on the bottom and cessation of feeding. The affected fry gradually weakened and died within three to six days after noticeable symptoms. This syndrome was characteristic of a major percentage of the mortalities in the samples from Lake Michigan. A very low percentage of the fry from Lake Superior and none of the fry from Oregon showed these symptoms (Table 4). Identical symptoms of dying fry were observed in Michigan hatcheries during 1968 (Michigan Department of Natural Resources, 1968) and in a preliminary study on salmon development (Johnson and Pecor, 1969). Examination of the affected fry showed no external lesions. A small percentage had a clouding of the lens on one or both eyes. Microscopic examination of tissues from healthy and sick fry showed no major differences between the two with the exception of the frequency of the presence of food in the gut. Food, in most cases, was absent from the stomachs of affected fry. A clear yellowish mucus Was present in the stomach and intestine of affected fry whether food was observed in the stomach or not. A clear oily liquid was also observed in the body cavity of many fry, both affected and nonaffected. Average cumulative fry losses through the ninth week after transfer from the incubator ranged from 2.5 to 41 44.5 percent in individual sample groups from Lake Michigan streams. The average cumulative mortalities associated with the specific mortality syndrome described above ranged from less than one percent to 39.1 percent for the same groups of samples. There was no significant difference in the average cumulative mortalities of fry from eggs taken at different times from the same streams with the exception of Bear Creek samples. The second group of samples from Bear Creek (12-13-68) had signifi- cantly lower mortality (p < 0.05) than those from the earlier sample from this stream. Less than one percent of these mortalities were associated with the specific symptoms. The mortalities of the Lake Michigan fry exhibiting the mortality syndrome were very specific with respect to time. The mortality generally increased and peaked abruptly during the sixth week (1,938 temp. units) after transfer from the incubator, although some fry with symptoms were observed during the fifth week and some mortality of fry with symptoms extended into the seventh week. The peak mortality occurred 99 days after fertilization during early feeding and final yolk absorption stage. The mortality of fry between the fifth week (1,797 temp. units) and the eighth week (2,221 temp. units) after transfer accounted for 81 percent of the total mortality observed in the Lake Michigan fry 42 samples. The Michigan Department of Natural Resources (1968) reported a similar mortality pattern in Michigan hatcheries, although the time period to initiation and peak mortalities with symptoms was different for hatcheries with different water temperatures. The period of mortality for each of the groups of samples collected from each of the four Lake Michigan streams, Thompson Creek, Bear Creek, Little Manistee River and Platte River, were almost identical to the general mortality pattern previously discussed (Figures 3, 4, 5 and 6). Only minor differences existed within streams and between streams. The second group of samples from Thompson Creek exhibited a low mortality during the sixth week but a high mortality occurred for the next two consecutive weeks, the seventh (2,080 temp. units) and eighth (2,221 temp. units) weeks after transfer (Figure 3). The second group of samples from Bear Creek sustained the lowest mortality of all sample groups and specific symptoms were virtually absent (Figure 4). The peak period of mortality in the first group of samples from the Platte River occurred during the seventh week (2,080 temp. units) but in the second group of samples from this stream the mortality began to increase during the fifth week (1,797 temp. units) and peaked during the sixth (Figure 6). 43 Figure 3. The average weekly and cumulative fry mortality during the nine~week study for the two groups of Thompson Creek samples. Figure 4. The average weekly and cumulative fry mortality during the nine-week study for the two groups of Bear Creek samples. MORTALITY PERCENT MORTALITY PERCENT 44 50 I I o// u momvsou CREEK 5’ 6’" o°/’ - . I —— I (9 I7 6.) l/ ---- 2 (IO-3'6.) [I 30 I I / I I, cumuhflvo / zo / I I / wooIIIL___\ an." \ /’ // \\\\ Io / s ‘I I, I / ,’ '...--n—_-- . ’v‘p’. - ’—_’_,J o ‘._____—-_- 'v— —-'-"- r I DAYS 64 7. 92 I06 I20 TJI. I23I ISIJ I707 20.0 2363 TIME days chef fonIIIunIon 401 BEAR CREEK 30' —— I (IO-24-6.) ——-— 2(I2-I3-bl) 20 I0 cumuIlflVL ______._ —-—""’-"‘ —-——--—-----;.—.k" _____ o — ‘ ' '1 l _ —-—-——-—-———-- ———————————— DAYS 64 7. 92 I06 I20 T.I.I. I23I ISIO I797 20.0 2363 TIME days Cher IonIlemlon 45 Figure 5. The average weekly and cumulative fry mortality during the nine-week study for the two groups of Little Manistee River samples. Figure 6. The average weekly and cumulative fry mortality during the nine-week study for the two groups of Platte River samples. PERCENT MORTALITY PERCENT MORTALITY m4 DAYS ‘. u 0 40¢ 304 201 46 LITTLE MANISTEE RIVER —— I (II-7-00) ----- 2 (12-13-60) 1. 0““ ‘0. I”’ ¢mu|QTIV1 ____/” /”_--- I ’l” ”’\\ w 1’ I” \\ ..k' I -—-’ ’ ‘ ’ ------—-—-’-- l” \ I 64 70 02 106 I20 I23I I514 I797 2000 2860 TIME days of!" “Huh-flan / I PLATTE awn ,z’ 10 ’l’ «9‘38” —— I (Io-mu) 9'” -_-...- 2 (1.4.59) l/’ 1” I’ I I / / l ’l / // Il’ I ,’ \ ”I/ [I ‘42:”? I I ~-~— -—-/’ ” ", «“' ’7 64 70 92 I06 I20 I20I 15!! I791 2000 2303 T l M E days char IQnIIquIon 47 The extent and pattern of mortality among fry from Lake Superior and Oregon was distinctly different from that observed in the Lake Michigan groups. The losses occurred at an earlier age and were not charac- terized by the symptoms observed in Lake Michigan samples. The Michigan Department of Natural Resources (1968) also reported that symptoms associated with the Lake Michigan fry mortality were not observed in the fry from Lake Superior sources. The average cumulative mortality of fry from Big Huron River, Lake Superior was 6.1 percent, which was generally lower than those from Lake Michigan. This group of fry sustained a low constant mortality up to the ninth week when a slight increase in mortality was recorded (Figure 7). No specific symptoms were observed. A very high loss of fry in the Cherry Creek samples (average cumulative mortality of 55.6 percent) was believed to be due to starvation shortly after yolk- sac absorption. The fry at swim-up averaged only 0.29 grams/fry as compared with 0.69 grams/fry for those from Lake Michigan. The small size probably contributed to their rapid emaciation and failure to feed. The mor- tality occurred one to two weeks earlier than was typi- cally observed in Lake Michigan samples (Figure 7). Most of the fry had developed a "pinhead" condition when death occurred. 48 The cumulative losses of Oregon fry averaged 11.3 percent during the nine-week period. Lesions were observed on most of the dead and dying fry. The moribund fry showed discoloration and gradual necrosis of the skin and flesh around the dorsal fin and caudal peduncle. The losses appeared to be due to a systemic bacterial infection of low virulence which was fatal during the first three weeks in the rearing tanks (from 1,231 temp. units to 1,514 temp. units), but then gradually disappeared (Figure 8). The average weekly mortality of the fry from each of the three major systems, Lake Michigan, Lake Superior and Oregon, showed a different pattern (Figure 9). Oregon fry samples suffered a peak mortality during the first three weeks after transfer from the incubator and a loss of less than one percent during the rest of the nine-week period (Figure 9). Lake Superior fry experienced a peak mortality during the fifth week after transfer and a second during the ninth week (Figure 9). The Lake Michigan fry samples experienced the highest average total cumulative mortality and the highest peak mortality which occurred during the sixth week (Figure 9). The absence of the Lake Michigan mortality syn- drome in Lake Superior and Oregon fry reared in the same tanks with Lake Michigan fry indicated the cause of death was not the same. There was no evidence of disease being a factor in the Lake Michigan fry mortalities. 49 Figure 7. The average weekly and cumulative fry mortality during the nine-week study for the two groups of Lake Superior samples. Figure 8. The average weekly and cumulative fry mortality during the nine—week study for the single group of Oregon samples. MORTALITY PERCENT MORTALITY PERCENT “I 50 LAKE SUPERIOR 50 d 1 (II-F60) “'0 NTI“ ----- (I - 20- 69) 40 1 2 30 I 20 10 W..k’ o . n I “‘11:- -' '1— --:_:'_ ------------------- DAYS 64 70 92 106 T. U. 1201 1514 1797 2000 TIM E days cflor IcflIIIzotIon I2 10 0 O R E GON 6 I-4-09 ----- weekly \ ‘ \\ —— tunnel-11v. \ \ \ \ \ \ 2 d \\ \ \ \ \ k - ‘ — ‘ ~ -- 0h— ‘~“~ --------- —~~“‘- U I DAYS 64 72 92 106 T. U. 1231 1514 1797 2000 TIME (Icy: aflor fonIllzcflon 51 Figure 9. The average weekly fry mortality during the nine-week study for Lake Michigan, Lake Superior and Oregon samples. I3 II MORTALITY PERCENT 52 LAKE MICHIGAN LAKE SUPERIOR ..——-.._—— -— ‘— .._...—————-" OREGON \ \ I, \\ I, I \ l 3 ’I \ / / I \\ [I I \\ / I 0 \ I, ’ / \\ / \ I I V I, o 1 I J, / I: ’ ‘— ‘I' .— 6 / . >/ 64 18 92 106 120 1231 I 514 1797 2030 2363 TIME days char Iorflllzoflon 53 Residue Identification Gas chromatography Gas chromatographic analysis of the ether extract of coho salmon eggs showed 14 to 17 residue peaks in every egg sample. Six of these peaks were identified as pesticide residues, two were artifacts from the clean-up procedure and the identity of the remaining peaks was not determined. Five of the pesticide residues were DDT and related isomers and one was dieldrin, a chlorinated cyclodiene. Analysis of the first elution fraction of the egg extracts with the Micro-Tek 220 gas chromatograph usually showed 13 residue peaks (Figure 10). The retention times of these 13 peaks, relative to p,p'-DDE (Table 5), were compared with retention times of pesti- cide standards for initial identification. The retention times of endrin, aldrin, methoxychlor, heptachlor and lindane pesticide standards were not related to any of the 13 peaks in the samples. One peak was identical to an anhydrous sodium sulfate residue and was apparently the result of the extraction and clean-up procedure. Samples of polychlorinated biphenyl mixtures (Arochlors 1221, 1232, 1242, 1253 and 1260--Monsanto) were injected into the gas chromatograph and the resulting chromatograms compared to those of typical egg extracts. The polychlorinated biphenyls (PCB's) Figure 10. 54 A typical chromatogram of the first eluate fraction of coho salmon eggs using the Micro Tek 220 gas chromato- graph. Peaks numbering 5, 7, 8 and 10 correspond to the pesticides p,p'-DDE, p,p'-DDD, o,p'-DDT and p,p'-DDT, respectively. Peak number 4 cor- responds to both MDE and o,p'-DDE. TABLE 5. The and 56 relative retention times (p,p'-DDE = 1.0) identification of residue peaks present in the ether extract of coho salmon eggs. Chromatograph P325 Re;:;:i°n Residue Micro-Tek 220 l 0.49 NaSO4 2 0.64 unidentified 3 0.74 unidentified 4 0.81 o,p'-DDE MDE 5 1.00 p,p'-DDE 6 1.11 unidentified 7 1.27 p,p'-DDD 8 1.35 o,p'-DDT 9 1.50 unidentified 10 1.77 p,p'-DDT 11 2.11 unidentified 12 2.59 unidentified 13 3.13 unidentified Aerograph 660 14 0.32 unidentified 15 0.79 unidentified 16 1.00 p,p'-DDE 17 1.32 dieldrin 18 4.04 Florisil artifact 57 produced from 30 to 35 residue peaks of which at least one peak matched almost every peak in the egg sample extracts. The saponification of PCB's and pesticide standards showed that DDT and DDD were destroyed but the PCB mixture was not affected. Comparison of saponified and unsaponified portions of individual egg extracts indicated that an average of 19.8 percent (n = 41, standard deviation = $5.95) of the DDT residue was not degraded upon saponification, indicating an inferfering residue was present. There appeared to be no difference in the per- centage of the interfering residue present in the samples from the four Lake Michigan streams or between Lake Michigan and Lake Superior streams. Inferferences of this type were not present or they were not identified for DDD and DDE although PCB's have been shown to inter- fere with both DDD and DDE (Koeman ep’al., 1968). No positive identification was made of the interfering residue and all DDT residue concentrations reported in this study are uncorrected for the average 19.8 percent interfering residue. Analysis of the second elution fraction of the egg extracts with the Wilkens Aerograph 660 gas chroma- tograph usually showed the presence of five peaks in every egg sample (Figure 11). A trace of p,p'-DDE was also present in this fraction. The retention times of Figure 11. 58 A typical chromatogram of the second eluate fraction of coho salmon eggs using the Aerograph 660 gas chromato- graph. Peaks numbered 16 and 17 cor- respond to DDE and dieldrin, respec- tively. AAAAAAAAAAAA 60 the peaks are recorded in Table 5. Two peaks were identified as p,p'-DDE and dieldrin, and one peak was identified as an artifact from the Florisil used in the clean-up procedure. Two peaks were not identified by comparison to pesticide standards, although the retention time of one was very close to that of lindane. Tests with PCB standards showed these were not found in the second elution fraction and thus were not potential inferfering compounds in this fraction. Thin layer chromatography Thin layer chromatography confirmed the presence of p,p'-DDT, p,p'-DDD, p,p'-DDE and dieldrin in the egg samples. These four pesticides were considered to be the most significant with respect to the quantity present and their toxic properties. Samples spotted on both aluminum oxide and silica gel G plates produced spots with Rf values which corresponded to Rf values of respective pesticide standards. The spots were scraped into vials, extracted with petroleum ether and an aliquot reinjected into a gas chromatograph. Single peaks were usually detected with retention times analogous to those of the appropriate pesticide standards. PCB standards (Arochlor 1221, 1232, 1242, 1254, 1260--Monsanto) were also spotted on thin layer plates along with sample extracts and pesticide standards. The PCB standards moved with the solvent front and had an 61 Rf value similar only to DDE. The spots associated with DDE were extracted and injected into a gas chromatograph. The chromatograms showed a series of peaks with retention times similar to those in a chromatogram of the whole egg extract. These data are further evidence that PCB compounds may be present in the salmon eggs. Pesticide Residue Concentrations in Coho Salmon Eggs General The concentrations of p,p'-DDT, p,p'-DDD, p,p'-DDE and dieldrin were determined for 200 samples of coho salmon eggs collected during 1968 from Lake Michigan, Lake Superior and Oregon. The average con- centrations of pesticide residues for each of the samples are tabulated in Appendix II according to sampling date on the basis of ppm (parts per million) wet weight, dry weight, fat weight and in Appendix III on the basis of ug/egg. In the salmon eggs from Lake Michigan and Lake Superior, DDT, DDD and DDE constituted approximately 27.1, 5.4 and 67.5 percent, respectively, of the total DDT complex and dieldrin accounted for approximately one percent of the total residues quantified. In the eggs from Oregon DDT, DDD and DDE constituted 17.9, 15.4 and 66.7 percent, respectively, of the total DDT complex while dieldrin accounted for approximately four 62 percent of the total residues quantified. Statistically there was not a significant difference (p < 0.05) between the percentages of the individual residues in Lake Michigan and Lake Superior eggs or between the per- centages of DDE in Lake Michigan, Lake Superior and Oregon eggs. The percentages of DDT, DDD and dieldrin in Oregon eggs were significantly different (p < 0.05) from the respective percentages in Lake Michigan and Lake Superior eggs. The percentages of the four residues relative to the total were very uniform within the three major "lake" systems. Of the four pesticide residues which were quan- tified in the study, DDT and dieldrin are generally con- sidered to be the most toxic (Henderson, Pickering and Tarzwell, 1959; Katz, 1961). DDD is considered less toxic to vertebrates than DDT (O'Brien, 1967) but the effects of DDD on fish have not been adequately studied. DDE, generally, has a much lower toxicity to fish than DDT or DDD. Burdick (1964) found no relationship between DDE concentrations in lake trout eggs and mortality of the fry. The concentration of DDT in the coho salmon eggs from each sampling site was significantly correlated (p < 0.05) with the concentration of DDD and DDE (Table 6). In considering the relationship of these compounds to the parameters evaluated in this study, 63 DDD and DDE were not considered separately. In the following discussion only p,p'-DDT values are noted except in some instances where the total DDT residues (DDT, DDD and DDE) are considered. Dieldrin is approximately 2 to 3 times more toxic to fish than DDT (Henderson, Pickering and Tarz- well, 1959; Katz, 1961). Dieldrin was not significantly correlated (p < 0.05) with DDT, but there was a signifi- cant positive correlation with DDD and DDE although a very low percentage of the variation was common to diel- drin and DDD or DDE. Dieldrin residues were considered separately in the analysis and interpretation of the data. Some discussion has occurred as to which expression of the pesticide residue concentration, ppm wet weight, ppm dry weight or ppm fat weight, presents the most useful and meaningful information for comparisons. It is my opinion that for discussion pur- poses it is adequate to express the pesticide residues as ppm wet weight in the salmon eggs in this study. Wet weight showed a significant positive correlation with both dry weight and fat weight as did dry weight with fat weight (Table 7). Expressing the pesticide residues as ppm wet weight accounts for a maximum of variation between the three expressions of pesticide residue concentration (Table 7). 64 TABLE 6. Correlation coefficient (r) data for the relationships between the four pesticide residues quantified. Residues n r r2 X 100 DDT vs DDD 182 .753* 56.70 DDT vs DDE 182 .679* 46.10 DDD vs DDE 180 .563* 31.70 dieldrin vs DDT 115 .125 dieldrin vs DDD 115 .264* 6.97 dieldrin vs DDE 115 .375* 14.06 * Significant at the 0.01 level of significance TABLE 7. Correlation coefficient (r) data for the relationships between pesticide residue con- centrations based on wet weight, dry weight and fat weight. Weights n r r2 X 100 Wet vs Dry 182 .977* 95.48 * wet vs Fat 167 .811 65.77 * Dry vs Fat 167 .770 59.29 2' Significant at the 0.01 level of significance 65 Reinert (1969) stated concentrations based upon fat weight were the most useful in comparisons between fish species in the Great Lakes because chlorinated hydrocarbon pesticide residues were correlated with the amount of fat present. In cases where fat weight and dry weight are both highly variable or significantly different for different species or locations it would be more meaningful to express the residues as ppm fat weight and/or ppm dry weight. But in the situation where the percentage of fat and water are relatively constant, as in this study where water in the eggs averaged 56.72 percent (S.E. 10.121, n = 158) and fat averaged 10.6 per- cent (S.E. $0.126, n = 178) based on the dry weight of the eggs, there appeared to be no advantage in using either dry weight or fat weight over wet weight. On the basis of the points mentioned in the above dis- cussion, residue concentration based upon wet weight was used for the interpretation of the results of this study. Relationship of_pesticide residues to percent fat in the eggs The relationship between the total DDT, DDD and DDE as ug/egg and percentage ether extractable fat based on the dry wet of the eggs was inconsistent. The samples from the Little Manistee River were the only ones in which a significant positive correlation (p < 0.05) 66 between percent fat and pesticide residue was found (Table 8). The samples from the remaining four streams did not show a significant correlation (p < 0.05) between fat and pesticide residue concentration (Table 8). The results of this study indicate the general lack of cor- relation between pesticide residue concentration and amount of fat in the eggs. Reinert (1969) reported a direct relationship between the percent fat and the amount of pesticides present in lake trout from Lake Michigan and Lake Superior. Assuming the quantity of fats present and the exposure of the fish are major factors in the con- centration of pesticides, all species and sizes of fish exposed to similar pesticide residues should have the same concentration of pesticides in their fats. Reinert (1969) found this relationship existed in three out of four different species of fish from Lake Michigan. Although there were distinct species differences in the amount of fat, the concentration of DDT residues fell within the narrow range of 34 to 38 ppm. The coho salmon eggs from Lake Michigan in this study averaged 36 ppm DDT in the fat, although the coefficient of varibility (CV) ranged from 19 to 40 percent for the four Lake Michigan streams. A second important factor in the relationship between fats and pesticide residues is the concentration 67 TABLE 8. The average percentage fat based on dry weight and correlation coefficients for percent fat and total DDT (DDT, DDD and DDE) in ug/egg. River n Average Correlation percent fat coefficient Platte River 78 10.2 -0.017 * Little Manistee 44 10.9 0.589 Thompson Creek 10 12.1 -0.028 (Bear Creek 34 11.0 0.300 Cherry Creek (Lake Superior) 10 10.9 0.403 * Significant at the 0.05 level of significance 68 of pesticides in the environment. The rather uniform concentrations of pesticides found in the fat of dif- ferent fish species indicates an equilibrium with environmental concentrations. This equilibrium is con- siderably lower than saturation level. If maximum con- centrations are assumed to be an equilibrium and not a saturation, then lower pesticide residues in the environ- ment would result in lower concentrations in the fat of fish tissues. The lower pesticide residue concentrations in the eggs of coho salmon from Lake Superior and Oregon (Pacific Ocean) probably reflect the lower concentration of the residues in these environments. The percent fat in the eggs from all three systems were similar (Table 8). Comparison of pesticide residue levels IE coho salmon eggs collected during I967 and 1968 from Lake Michigan, Lake Superior and Oregon There were no significant differences (p < 0.05) in the average DDT concentrations (ppm wet weight) in salmon egg samples from corresponding streams during 1967 and 1968 (Table 9). The DDT residue concentration in samples from Oregon, Lake Superior and Bear Creek (Lake Michigan) were very similar for 1967 and 1968 (Table 9). The Platte River samples showed the greatest difference between years (Table 9). One possible explanation for this difference is that the coho salmon stocked in this stream were from two different sources. 69 40.0 4 zoommo 0004 040.0 4 004.0 040.0 4 000.0 000.0 4 040.0 000.0 4 040.0 4 200040 0004 u u s 004.0 4 000.0 0 40>44 004:4 04m 0004 004.0 4 004.4 004.0 4 000.0 440.0 4 000.0 000.0 4 000.0 04 40000 004040 400.0 4 000.0 440.0 4 000.0 000.0 4 000.0 000.0 4 000.0 0 40>44 004:4 040 40440000 0444 0004 u u n 04.0 4 00.4 04 00>44 000040 n u u 44.0 4 40.4 04 40040 4000 0004 044.0 4 400.0 004.0 4 000.0 000.0 4 000.0 000.0 4 004.4 00 40>44 000040 004.0 4 000.0 044.0 4 040.0 040.0 4 000.0 000.0 4 004.4 44 .4 00004002 040044 004.0 4 004.0 044.0 4 040.4 440.0 4 040.0 000.0 4 400.4 40 40040 4000 000.0 4 000.0 000.0 4 040.4 400.0 4 044.0 004.0 4 000.4 00 40040 00005040 20044042 0444 0004 40009 moo mom son 2 coflumooq 4000 .I0004 .40000 0:0 0000400 0040 00400 0000 00040 momH pom 0mm4 mafiuso owuomaaoo mmmm Goadmm 0:00 :0 mmsoflmmu mowowpmmm mmmum>< .m mqmde 70 The salmon which returned to the Platte River in 1967 were an Oregon strain averaging ten to twelve pounds as adults, but in 1968 a Washington strain returned which averaged seven to ten pounds. The same strains of fish were stocked in Bear Creek and Lake Superior both years. Dieldrin residues were not quantified for the small group of samples collected in 1967, but Reinert (1969) reported an average dieldrin residue level of 0.16 ppm in the eggs from four coho salmon collected at the Platte River during 1967. The average concentration of dieldrin residues determined in this study ranged between 0.086 to 0.095 ppm in the eggs from Lake Michigan tributaries during 1968. The difference between these values and Reinert's probably reflect different analyti- cal methods although it is possible that higher concen- trations were found in 1967 than in 1968. The concentration of pesticide residues in coho salmon eggs from each of the three major systems, Lake Michigan, Lake Superior and Oregon, were significantly different (p < 0.01). Lake Michigan samples had the highest average total pesticide residue of 5.83 ppm, including DDT, DDD, DDE and dieldrin. The average total pesticide residue in Lake Superior samples was 0.97 ppm, which was approximately one-sixth the level in the Lake Michigan samples. Oregon samples had the lowest average total pesticide residue content of 0.11 ppm, approximately 71 55 times lower than the Lake Michigan samples. Reinert (1969) reported that identical species of fish from Lake Superior had 4 to 7 times less DDT and 2 to 7 times less dieldrin than those from Lake Michigan. Differences in pesticide residue levels in salmon eggs collected from varibus streams tributary to Lake Michigan duringgl968 There were distinct differences in the pesticide residue content of egg samples from the four streams tributary to Lake Michigan (Thompson Creek, Bear Creek, Platte River and Little Manistee River) (Appendix II). The Thompson Creek samples had significantly higher average DDT concentrations (p < 0.05) than the samples from the other Lake Michigan streams. The Bear Creek samples showed the next highest average DDT concentrations although they were not significantly different (p < 0.05) from the average DDT residues in the Platte River and Little Manistee River samples. Platte River and Little .Manistee samples had almost identical average DDT con- centrations. There was no apparent relationship between the «average DDT residues present in the egg samples and the IKDrth-south location of the parent streams. Little is kilown about the migrations of salmon in Lake Michigan 1R1t.it is expected that they move extensively throughout the lake. It is possible that the different strains, 72 such as those stocked in Thompson Creek, follow a dif— ferent migration pattern and are thus exposed to dif- ferent pesticide levels during their growth. Reinert (1969) reported higher pesticide residues in fish from southern Lake Michigan with relatively lower residues in fish from the northern portions of the lake. It is known that numerous salmon are located in southern Lake Michigan during the spring and summer prior to their spawning runs. If salmon from different streams spend different periods of time in southern Lake Michigan or other areas of different pesticide levels, then they may accumulate different pesticide concentrations. There is some evidence to show a relationship between the genetic strains of salmon sampled and the average DDT concentrations found in the eggs. Thompson Creek fish were reared from an Alaskan strain, Bear Creek fish were reared from an Oregon strain and Platte River and Little Manistee fish were reared from a Wash- .ington strain of coho salmon. The three different strains of salmon all showed different average DDT con— centrations. However, the samples from the two streams which had a single genetic strain of salmon had very Silnilar average DDT concentrations. Thompson Creek and Beéar Creek samples exhibited higher but different con- Cetitrations and the Platte River and Little Manistee Saluples had lower but almost identical residue 73 concentrations. The differences in the pesticide residues of the three strains of coho salmon indicates a possible physiological difference in the accumulation of pesticide residues between genetic strains or a different migration routes. This could be related to differences in growth rates. Dieldrin residues were determined for samples from the Platte River, Bear Creek and Little Manistee River (Appendix II). Dieldrin residues were not deter- mined for Thompson Creek samples because early attempts to analyze for dieldrin were inadequate and resulted in the loss of the residue during the clean-up procedure. There was no pattern in the residue levels for the two strains analyzed. The lack of dieldrin values for Thomp- son Creek limits the evaluation of these data. Relationship between fish length and pesticide residues in the eggs The relationship between the fork length of the .adult fish and DDT residue levels in their eggs was inconsistent for the different streams sampled. There Vmas a significant positive correlation (p < 0.05) between fish length and DDT residue levels in the eggs of the PJuatte River samples, while the same comparison for Listtle Manistee River samples showed a significant nesaative correlation (p < 0.05) (Table 10). The fish 1eImgth and DDT residues in the eggs of the other streams 74 sampled, Thompson Creek, Bear Creek and Cherry Creek (Lake Superior) were not significantly correlated (Table 10). The lack of a significant relationship between fish length and DDT residues in three groups of samples and the opposing significant correlations for the other two groups of samples, suggests there was no relationship between fish length and DDT residue concentration in the eggs. Although it was not statistically significant at the 5 percent level, there was indicated a trend for a decrease in DDT residues with increased fish length. Other studies have shown both agreement and apparent conflict with the results concerning the relationship between female length and pesticide residue levels in the eggs. Kleinert (1967) stated pesticide residues in the eggs of walleye from Wisconsin lakes did not appear to be related to the length of the females. Reinert (1969) reported DDT and dieldrin residues in body tissues were directly related to the length of lake trout and walleyes collected from Lake Michigan. How- ever, it is likely that ovarian tissues are subject to Ciifferent physiological constraints than body tissues. CPhe fish analyzed by Reinert also varied over a large EBize range incorporating several year classes as opposed t<3 the restricted size range and single year class of tile present study. 75 Comparison of the pesticide residue levels Tn the eggs of residual stream fish andTake run fish One of the original objectives of this study was to compare the pesticide residue levels of eggs from However, scale analysis indi- salmon of different ages. A cated that all mature females were three years old. group of small fish collected at the Platte River, averaging two to five pounds, exhibited two annuli indicating a three-year-old fish. A frequency dis- tribution of the lengths of the fish collected at the Platte River was distinctly bimodal, indicating that two independent constraints were acting on the growth rates. One explanation proposed before age determination Was that the small fish were precocious females, but the fact that all the fish were three years old eliminated this possibility. A second proposal, which is most com- patible with the data, is that the small fish were residual stream fish (remaining in the stream). A comparison of the average pesticide residues in the large and small fish lends support to the residual stream theory. The group of small fish had higher average DDT residues (1.61 ppm) than the group of larger fish (1 . 48 ppm), but the difference was not significant (p < 0.05) . One would expect that fish living in the ri‘rer for all of their lives would be exposed to higher 76 pesticide levels and more restricted conditions for growth than the fish in the Great Lakes. Relationship between pesticide resicfies in coho salmon eggs and samplinldates DDT residues in coho salmon eggs, considering the saInples from all the Lake Michigan streams together, were significantly correlated (p < 0.05) with sampling dates of the 1968 salmon spawning migrations (Table 11) . The general trend was for DDT concentrations to decrease With later sampling periods (Figure 12) . Samples were cellected over a period of approximately 120 days cover- ing’ the entire spawning run. Dieldrin residues expressed as ppm were not significantly correlated with sampling dates (Table 11), possibly because dieldrin residues were not determined for the early samples. DDT residue concentrations in the eggs collected weekly at the Platte River for the duration of the run Were not significantly correlated (p < 0.05) with sainpling date (Table 11) . However, the samples did not necessarily represent fish that returned to the stream du-:"—"ing the week the samples were collected. The move- ment of the salmon into the Platte River and up to the e9g“taking station at Honor, Michigan was controlled by a harvest weir at the mouth of the river. Large nu“fibers of salmon were allowed past the harvest weir only two or three times during the run, so the samples 77 TABLE 10. The average fork length of the fish sampled and correlation coefficients (r) for fish length and average DDT concentration in their eggs. Correlation River N Average length Coefficient (Cm) (r) Platte River 66 66.4 -o.317* Little Manistee River 45 67.7 0.416* Thompson Creek 26 58.9 -0.209 Bear Creek 34 67.8 -0.048 Cherry Creek (L.S.) 10 49.7 -0.521 * Significant at the 0.05 level of significance TABLE 11. Correlation coefficients (r) for the relation- ship of pesticide residues with sampling dates. Residue N r All Lake Michigan Samples Together DDT vs. sampling date 168 -0.267* DDD vs. sampling date 168 -0.291 DDE vs. sampling date 168 -0.161* Dieldrin vs. sampling date 106 -0.150 Platte River Samples Only DDT vs. sampling date 63 -O.l96 * Significant (p < 0.05) 78 .mmumo mafiamEMm £003 omummfioo mammnum Hmsp0> swosw Eonm mmmm GOEHMm 0:00 :0 mcowumnusmocoo mSpHmmH Boo mmmum>m map How 0004 cofimmmummu Mo mmon was EMHOMHU Hmuumom .04 040040 79 0001 as: 00:»:- 075 «NI: 0:: 072 halo nous no 3.2a o«— 02 on 90 at ON 0 m»:— [ [ L I ca.— 0 09.— co.— X p000... I :05; u > On.— ..ch 300.25.. 300: . 00>... 0005.. o 30 0.0.3 .20- o 0.35 2000203 0 u D “I“ nouvumaauoa 80 collected at the egg-taking station were representative of only two or three groups which had been held in the river for varying amounts of time. Fish length and percentage fat were not cor- related with sampling date, so the relationship between DDT residues and sampling date was not a secondary function of either. A possible biological explanation for the inverse relationship between DDT residues and sampling date is that the later fish remained in the streams where pesticide levels are probably lower during this time of year and have ceased feeding for a longer period of time. The decrease or elimination of the two major sources of pesticides would allow a decrease in the total body burden of pesticides. Relationship of pesticide residues in coho salmon eggs to the egg and'fry mortalities Egg mortalities did not appear to be related to DDT or dieldrin residue concentrations (ppm wet weight). Correlation coefficients for DDT (-0.128) and dieldrin (-0.225) compared to egg mortalities of all the samples from Lake Michigan streams were not significant (p < 0.05). Ten samples from Lake Superior were the controls available during this portion of the study and the mortality of these eggs was similar to the lots of Lake Michigan eggs. No unusual mortality occurred that could not be accounted for. Allison gt 31. (1964) reported that egg lots from 81 females dosed with various concentrations of DDT had similar survival rates. However, Macek (1968) indicated eggs from females dosed with DDT had a higher mortality than eggs from control fish. The lack of adequate con- trols in this present study hindered the interpretation of the results. To test the hypothesis that higher mortalities of coho fry are the result of higher pesticide residue concentrations in the eggs, a correlation analysis of DDT concentrations (ppm wet weight) and fry mortality was calculated (Table 12). No significant correlations (p < 0.05) were found between DDT concentrations and fry mortalities in the samples from each of the streams, with the exception of the second group of samples from Thompson Creek. Higher mortalities of these fry were associated with higher pesticide residues in the eggs. Lake Superior fry mortalities also were not significantly correlated to DDT residues (ppm wet weight). A comparison of the fry mortalities of all sample groups from Lake Michigan to DDT, total DDT (DDT, DDD and DDE) and dieldrin residue concentrations (ppm wet weight) in the eggs showed negative correlations that were not significant (p < 0.05) (Table 13). The same comparisons on the basis of pg of DDT, dieldrin and total DDT (DDT, DDD and DDE) in the eggs also showed no significant cor- relations (Table 13). 82 TABLE 12. Correlation coefficient (r) data for the relationship between DDT residues (ppm wet weight) in the eggs and fry mortality in samples from each stream. Source Sample date N r Thompson Creek 9-17-68 10 -0.492 10—13-68 8 0.692* Bear Creek 10-24-68 10 -0.128 12-13-68 8 -O.223 Little Manistee River ll-7-68 9 -0.211 12-13—68 7 0.559 Platte River 10-17-68 10 -0.254 1-4-69 4 -O.844 Lake Superior 11-1-68 10 0.550 * Significant at the 0.05 level of significance TABLE 13. Correlation coefficient (r) data for the relationship between DDT, dieldrin and total DDT (DDT, DDD and DDE) expressed as ppm wet weight and ug/egg and the total fry mor- talities of groups from Lake Michigan. n r ppm DDT vs. fry mortality 67 -0.l48 ug/egg DDT vs. fry mortality 57 0.056 ppm total DDT vs. fry mortality 67 -0.010 ug/egg total DDT vs. fry mortality 57 0.142 ppm dieldrin vs. fry mortality 45 -0.154 ug/egg dieldrin vs. fry mortality 45 -0.068 *No significant correlations (p < 0.05) 83 On the basis of these correlation analyses there does not appear to be a direct relationship between DDT, dieldrin or total DDT (DDT, DDD and DDE) residues in the eggs and the mortality of the fry. However, these analy- ses alone are not adequate to assess the relationship between pesticide residues and fry mortality. Further consideration must be given to the interaction between specific pesticide residues in the eggs and to the remaining unidentified residues which were found in the eggs. It would also be interesting to compare the effects of pesticides on specific genetic strains of coho salmon reared under similar conditions. There are many factors which may influence the effect of pesticide residues on the development of fry. Burdick 33 El. (1964) in a study of lake trout fry mortality from a number of lakes determined that a mortality syndrome occurred in fry groups reared from eggs which had a total DDT concentration of 4.75 ppm wet weight (combined ether and methanol extracts) or higher based on the Schechter Haller method. The com- bined DDT and DDD residue concentrations (ether extract) in the eggs of the present study are generally less than half this critical value, which would suggest no mortality resulting from pesticide contamination. Cuerrier gt_§l. (1967) reported DDT and metabolites (p,p'-DDT, o,p'-DDT, DDD and DDE) concentrations of grea mor1 val1 eve: the cal The Cid on cri Stl Clc of De de re Of (11' 01' b] t1 S1 84 greater than 400 ppb wet weight in eggs resulted in fry mortalities ranging from 30 to 90 percent. The critical value of 400 ppb is below the total DDT concentrations of even Lake Superior eggs which had approximately one-sixth the residue concentrations of Lake Michigan eggs. Criti- cal values are probably not reliable as a general rule. The number of parameters affecting the toxicity of pesti- cide residues, such as fish species, temperature, stress on the fish, resistance to pesticides and general water quality, are too variable in individual systems for critical values to be determined and applied to separate studies. The lack of a direct relationship between pesti- cide residues and fry mortality suggest that the majority of Lake Michigan coho salmon have not exceeded a critical pesticide residue level, but based on circumstantial evi- dence some of the fry mortalities are probably the result of pesticide residues. The timing and symptoms of the fry mortality are of particular interest. Bur- dick gt_al. (1964) reported fry mortalities due to pesti- cide contamination were characterized by distended air bladders and air in the intestinal tract. Although these symptoms were not major characteristic in this study, they were present. Burdick et 31. (1964), Allison gt 2£° (1964), Cuerrier st 31. (1967) and Macek (1968) reported that 85 fry mortalities occurred during final stages of yolk absorption at a time when the fry were just starting to feed. The mortalities in this study occurred at a similar time in development. At the time of this mor- tality, little or no external evidence of yolk-sac was visible in the fry, but yolk and lipid material were still present within the gut. Separate analyses of composite samples of eviscerated fry and gut tissues showed that gut tissues had approximately 6 to 12 times higher concentrations of DDT than the eviscerated fry. It was hypothesized that pesticide residues are stored in the triglyceride lipids which are not utilized by the fry until the final yolk absorption (Smith, 1957). The metabolism of these lipids could result in toxic levels of pesticides entering the circulatory system. This would explain the abrupt onset of mortality just prior to the initial feeding stage. The results of this study do not provide a satis- factory explanation for the cause of mortality among the Lake Michigan coho salmon fry. A comparison of pesticide residue concentrations in eggs of salmon from Lake Michi- gan, Lake Superior and Oregon indicates the mortality is associated with the higher pesticide residues, but the relationship does not hold for samples taken within the Lake Michigan system alone. After assessing the data collected in this study it is apparent that the factor(s) 86 causing the mortality is more than a simple critical threshold level of one specific pesticide. Thus the interaction of pesticide residues, possible PCB residues or other unidentified toxic materials may be important. SUMMARY Samples of coho salmon eggs were collected during 1967 and 1968 from four Lake Michigan streams, Bear Creek, Little Manistee River, Platte River and Thompson Creek; two Lake Superior streams, Big Huron River and Cherry Creek and one Oregon source. Pesticide residues in coho salmon eggs were determined and tested for their relationship to the length of parent fish, sampling dates, location of sampling sites, percentage fat in the eggs and egg and fry mortality. Four pesticide residues were identified and quantified in each sample: p,p'-DDT, p,p'-DDD, p,p'-DDE and dieldrin. Compounds were found that interferred with the analysis of DDT in samples from all three sources. The interference resulted an average error of 19.8 percent in the quantification of DDT residues. These compounds were believed to be polychlorinated biphenyls (PCB's). 87 88 Lake Michigan coho eggs had the highest average total pesticide concentration of 5.83 ppm. The average total pesticide concentration in Lake Superior eggs was 0.97 ppm, which was approxi- mately one-sixth the level in Lake Michigan eggs. Oregon coho eggs had the lowest average total pesticide residue content of 0.11 ppm, approxi- mately 55 times lower than Lake Michigan eggs. Among the Lake Michigan streams, the Little Manistee River and Platte River egg samples had almost identical concentrations of pesticide residues, Bear Creek samples were slightly higher and Thompson Creek samples had the highest pesticide residue concentrations. The type and concentration of pesticide residues found in 1968 samples were very similar to those found in the 1967 samples. There was no relationship between the length of the parent fish or the amount of fat present in the eggs and the pesticide residue concentration in the eggs. Egg mortality was similar for both Lake Michigan and Lake Superior, and no relationship with pesticide residues was found. 10. 89 The mortality of the fry from Lake Michigan, Lake Superior and Oregon, in actual numbers, was similar; but the symptoms, timing and causative agent of the mortalities were dif- ferent. One group of samples from Thompson Creek, Lake Michigan, was the only group to show a significant positive relationship between fry mortality and pesticide residues. Circumstantial evidence, including the timing of the mortality, specific symptoms associated with the mortality of the Lake Michigan fry and the higher concentration of pesticide residues in the Lake Michigan coho eggs, indicates the mortality may be associated with pesticide toxicity. LITERATURE CITED LITERATURE CITED Allison, D., B. J. Kallman, O. B. Cope and C. C. Van Valen. 1964. Some chronic effects of DDT on cutthroat trout. U.S. Bureau Sport. Fish and Wildl., Res. Rept. 64 30p. Anderson, R. B. and W. H. Everhart. 1966. Concen- trations of DDT in landlocked salmon (Salmo salar) at Sebago Lake, Maine. Trans. Am. Fish Soc. 95:160-164. Boyd, C. E. 1964. Insecticides cause mosquitofish to abort. Progve. Fish. Cult. 26, 138. Brown, J. R. and H. Hughes. 1969. Pesticide levels in wildlife located in the region of the Great Lakes. 12th Conf. on Great Lakes Res. (Abstracts) Ann Arbor, Michigan. Burdick, G. E., E. J. Harris, H. J. Dean, T. M. Walker, J. Skea and D. Colby. 1964. The accumulation of DDT in lake trout and the effect on reproduction. Trans. Am. Fish. Soc. 93(2):127-l36. Carr, John F. and Robert E. Reinert. 1968. DDT and dieldrin in Great Lakes fish. 11th. Conf. on Great Lakes Res. Milwaukee, Wisconsin. Currier, J. P., J. A. Keith, and E. Stone. 1967. Prob- lems with DDT in fish culture operations. Le Naturaliste Canadian, Vol. 94:315-320. Federal Water Pollution Control Administration, Great Lakes Region. 1968. Water pollution problems of Lake Michigan and tributaries, Report of Conference (Jan. 1968) 74pp. Henderson, Croswell, O. H. Pickering, and C. M. Tarywell. 1959. The toxicity of organic phosphorus and chlorinated hydrocarbon insecticides to fish. Trans. 2nd Seminar on Biol. Prob. in Water Poll. U.S. Public Health Serv., Robert A. Taft Sanitary Engineering Center. Cincinnati 26, Ohio. 90 91 Hickey, J. J., J. A. Keith and F. B. Coon. 1966. An exploration of pesticides in a Lake Michigan ecosyslem. J. Appl. Ecol. 3:141-154. Holmes, D. C., J. H. Simmons, J. O'G. Tatton. 1967. Chlorinated hydrocarbons in British Wildlife. Nature. 216:227-229. Johnson, H. E. 1967. Effects of endrin on reproduction in a freshwater fish (Oryzias latipes). Ph.D. Thesis. Univ. of Wash. Seattle, Washington. 149p. Johnson, H. E. and C. Pecor. 1969. Coho salmon mor- tality and DDT in Lake Michigan. Proc. 34th N. Am. Wildl. and Nat. Res. Conf. 8p. Katz, Max. 1961. Acute toxicity of some organic insec- ticide to three species of salmonids and to the threespine stickleback. Trans. Am. Fish. Soc. 90:264-268. King, G. T. 1962. Some effects of DDT on the guppy and the brown trout. Bur. Sport. Fish and Wildl. Spec. Sci. Rept., Fish. No. 399. Kleinert, Stanton J. 1967. Survival of walleyes from eggs of known DDT and dieldrin residues in three southeastern Wisconsin lakes. Wis. Dept. of Nat. Res., Tech. Bull. No..21..9p. Koeman, J. H., M. C. TenNoever DeBruw, R. H. DeVos. 1969. Chlorinated biphenyls in fish, mussels and birds from the River Rhine and the Netherlands coastal areas. Nature Vol. 221. Kovacs, M. F. Jr. 1963. Thin-layer chromatography for chlorinated pesticide residue analysis. J.A.O.A.C. 46:884. Lichtenstein, E. P., K. R. Schultz, T. W. Fuhremann, and T. T. Liang. 1969. Biological interaction between plasticizers and insecticides. J. of Econ. Ent. 62(4):?61-765. Macek, K. J. 1968. Reproduction in brook trout (Sal- velinus fontinalis) fed. sublethal concentrations of DDT. J. FiSh. Res. Bd., Can. 25(9):1787-1796. Michigan Department of Natural Resources. 1968. Inter- office communications. Unpublished Hatchery Reports. 92 Mills, P. A. 1959. Detection and Semiquantitative estimation of chlorinated organic pesticide residues in foods by paper chromatography. J.A.O.A.C. 42:734‘739. Mills, P. A. 1961. Collaborative study of certain chlorinated organic pesticides in dairy products. J.A.O.A.C. 44:171. Mount, Donald I. 1962. Chronic effects of endrin on bluntnose minnows and guppies. Fish. Wildl. Serv., Bur. Sport Fish. Wildl., Res. Rept. 58:1-38. O'Brien, R. D. 1967. Insecticides--Action and Metabolism. Acad. Press. New York, N.Y. 332pp. Reinert, R. E. 1969. Summary of available information from the Bureau of Commercial Fisheries, Ann Arbor Biological Laboratory on pesticide levels in Great Lakes fish. Unpublished manuscript. Reynolds, Lincoln M. 1969. Polychlorobiphenyls (PCB's) and their interference with pesticide residue analysis. Bulletin of Environmental Contami- nation and Toxicology. 4(3):128-143. Rounsefell, George A. and W. R. Everhart. 1953. Fishery Science, Its Methods and Application. John Wiley and Son Inc., New York. 444pp. Smith, S. 1957. Early development and hatching. Brown, M. E. (ed.). In V01. I, p. 323-360. The Physiology of Fishes, Acad. Press, New York, N.Y. Snedecor, George W. 1956. Statistical Methods. State College Press, Ames, Iowa. 534pp. Sokal, Robert R. and F. James Rohlf. 1969. Biometry. W. H. Freeman and Co., San Francisco, Calif. 776pp. Iowa Veith, G. D. 1970. Environmental chemistry of the chlorobiphenyls in the Milwaukee River. Ph.D. Thesis. Univ. of Wis., Madison, Wis. 180p. APPENDICES APPENDIX I Reagent Purification Procedures: l. Petroleum ether Fisher brand, reagent grade, 30-60°C boiling range petroleum ether was used throughout most of the study. Several other brands were tried, but they frequently smelled foul and contained artifacts which could not be removed. The petroleum ether was purified by refluxing 3 liters of the solvent with 10 grams of sodium-lead granules (Dri-Na) for 3 to 4 hours. The solvent was slowly distilled in an all glass system, discarding the first 50 ml of distillate and collecting the remaining fractions up to 50°C. Ethyl ether Fisher brand, reagent grade ethyl ether was twice redistilled in an all glass system prior to use. It was observed that the addition of 2 percent by volume ethyl alcohol greatly facilitated the elution of dieldrin from Florisil columns. 93 94 Benzene Fisher brand, reagent grade benzene was refluxed over freshly cut sodium, decanted and slowly dis- tilled at 80°C in an all glass system. Acetonitrile Fisher brand, reagent grade acetonitrile was purified weekly as required. The solvent was purified by refluxing for 4 hours with 85 percent H3PO4 and P205, added in the proportions of 1 ml acid and 30 grams pentoxide to 4 liters of acetoni- trile slowly distilling and collecting the fraction distilling at 81 to 82°C. 95 000.0 4 440.0 000.0 4 000.0 000.0 4 040.0 444 00-4-4 200000 000.0 4 000.0 000.0 4 000.0 000.0 4 000.0 400 00-00-4 40344 00400 040 040.0 4 000.0 400.0 4 000.0 000.0 4 000.0 04 00-4-44 40040 044000 40440000 0404 000.4 4 040.40 000.0 4 404.0 000.0 4 004.4 00 0004030 400.0 4 400.00 000.0 4 400.0 000.0 4 040.4 04 400-44-44 004.4 4 000.00 044.0 4 000.0 040.0 4 044.4 04 00-4-4 000.4 4 000.00 000.0 4 000.0 004.0 4 404.4 04 00-44-44 000.0 4 004.40 000.0 4 004.0 000.0 4 004.4 04 00-0-44 00 40-04 004.4 4 000.44 000.0 4 404.0 444.0 4 400.4 04 00-04-04 400.4 4 000.00 004.0 4 040.0 000.0 4 400.4 00 00-04-04 00 04-0 40>44 000040 000.0 4 000.40 4000 004.0 4 000.0 000.0 4 004.4 44 0004030 040.4 4 400.40 004.0 4 404.0 000.0 4 040.4 04 00-04-04 040.4 4 000.00 000.0 4 004.0 404.0 4 040.4 0 00-0-44 400.0 4 400.00 000.0 4 000.0 400.0 4 000.4 0 00-40-04 000.0 4 004.40 000.0 4 000.0 004.0 4 004.4 0 00-00-0 000.0 4 044.00 400 000.0 4 000.0 044.0 4 440.4 04 00-0-0 40340 mmflmflfimz mHflflfiQ 004.4 4 400.40 044.0 4 000.0 000.0 4 400.4 40 0004030 000.4 4 400.00 004.0 4 004.0 000.0 4 004.4 04 00-04-04 400.0 4 004.00 004.0 4 000.0 000.0 4 000.4 04 00-40-04 040.0 4 004.00 040.0 4 000.0 000.0 4 000.4 04 00-0-04 40040 4000 000.0 4 000.00 4040 000.0 4 000.4 004.0 4 000.4 00 0004030 000.0 4 000.00 004.0 4 000.4 404.0 4 040.4 44 00-04-04 004.4 4 000.04 400 004.0 4 040.0 004.0 4 000.0 44 00-04-0 40040 00000000 Z004004s 0404 000403 000 420 000403 040 000403 003 z 0mm + NV 2040040s0osoo 0504000 Boo 00aoa0m 0000 00000000 .0000500 0403 0500400 0:0 00 0000 00000 000 400 000403 00004 Ema 0:0 .000003 040 Boo .000003 003 Boo m0 000004ox0 soEH0m 0:00 00 0000 000 :0 0:000040000 -000 0004004 04404040 000 4000 .000 .0000 000 40000 .000 .000 .000-.0.0 000 .HH XHQZMQQd 96 040.0 4 044.0 000.0 4 000.0 000.0 4 040.0 444 00-4-4 200000 000.0 4 040.4 000.0 4 400.0 000.0 4 000.0 440 00-00-4 40340 00400 040 000.0 4 000.4 000.0 4 004.0 440.0 4 000.0 04 00-4-44 00040 044000 00400000 0000 440.0 4 000.0 400.0 4 400.0 000.0 4 000.0 00 0004030 444.0 4 000.0 040.0 4 000.0 040.0 4 040.0 04 400-44-44 004.0 4 000.0 040.0 4 000.0 000.0 4 000.0 04 00-4-4 400.0 4 400.4 000.0 4 040.0 000.0 4 000.0 04 00-44-44 000.0 4 000.0 040.0 4 040.0 040.0 4 000.0 04 00-0-44 00 40-04 000.0 4 044.0 000.0 4 000.0 400.0 4 400.0 04 00-04-04 044.0 4 004.0 000.0 4 000.0 040.0 4 000.0 00 00-04-04 00 04-0 40340 000040 400.0 4 000.0 0000 000.0 4 040.0 040.0 4 000.0 0004030 400.0 4 000.0 000.0 4 000.0 040.0 4 000.0 04 00-04-04 000.0 4 040.0 400.0 4 040.0 400.0 4 000.0 0 00-0-44 000.0 4 040.0 000.0 4 400.0 000.0 4 000.0 0 00-40-04 004.0 4 000.0 000.0 4 000.0 000.0 4 000.0 0 00-00-0 000.0 4 000.0 000.0 4 000.4 000.0 4 044.0 04 00-0-0 40340 00004002 040040 400.0 4 000.0 000.0 4 040.0 440.0 4 040.0 40 0004030 000.0 4 000.0 040.0 4 000.0 040.0 4 000.0 04 00-04-04 400.0 4 040.0 040.0 4 000.0 400.0 4 400.0 04 00-40-04 000.0 4 000.0 000.0 4 000.0 040.0 4 000.0 04 00-0-04 00040 4000 004.4 4 000.04 400.0 4 000.4 400.0 4 044.0 00 0004030 440.4 4 000.44 004.0 4 000.4 040.0 4 000.0 44 00-04-04 400.0 4 040.0 000 400.0 4 000.0 400.0 4 000.0 44 00-04-0 00040 00000004 20040042 0000 000403 000 020 000403 040 000403 003 2 0040800 0000 :0000004 0mm 4 My coflumnucmonoo 0560000 can . 0.00000 .44 04020000 7 9 NHH.O H NOO.H NNO.O H NON.O OH0.0 H NNO.O HHO OOnOnH zoommo OOO.O H NNO.NH HOO.O H OOO.H OH0.0 H Omm.O HHN OOnONnH Hm>Hm cousm mHm mOO.N H OON.OH mON.O H OmO.N OOH.O H OON.O OH OOanHH xmmHo NHHmnu monmmOm mMNH HHO.m H OOH.OO NON.O H HmO.O OOH.O H ONO.N ON mmmum>4 HON.m H OOH.NO OOO.O H NO0.0 OmN.O H OOO.m OH OwsvHuHH HNN.O H OmO.ON OHm.O H OOH.O HOH.O H HNm.m OH H mouvnH NOO.m H HON.mO HHm.O H NOO.O ONN.O H OON.m OH OOaOHnHH HOO.O H OmH.OO NNO.O H NON.O NON.O H OOO.O NH OOuOuHH oH ONnOH OOm.O H HOO.OOH NOm.O H HON.O mmN.O H NOO.m OH OquHuOH mmv.O H ONO.OO OOm.O H ONH.O OON.O H OOO.m ON OO-OH-OH oH OHaO Hm>Hm mHHmHm ONO.N H OOO.NO HOOO NON.O H mOm.O OOH.O H mHO.m HO mmmHm>4 mmO.m H OON.HO mHm.O H NON.O ONH.O H Omm.m OH OOannNH OOm.O H OOm.OO mmm.O H Omm.m mOm.O H ON0.0 O OOaOuHH Om0.0 H NHO.HO OOO.O H NON.O NON.O H NON.m O OOnONnOH NOO.O H OOO.OO ONO.O H OmO.O NON.O H NNO.m O OOnONuO OHO.HH H ONO.NO OHO.O H NOO.O mOO.O H OON.N NH OonOuO Hm>Hm mmgmflcmz wHQHHA mmN.m H mO0.00 NHm.O H mO0.0 NOH.O H NON.O Om momnm>¢ OOO.m H ONO.NO ONO.O H OOO.O OOH.O H ONO.N OH OOumHuNH OOO.O H OHO.OO OOO.O H OOO.OH OHN.O H OOO.O NH OO-ON-OH OOO.m H NO0.00 NOm.O H NNO.O NON.O H OHO.O NH OOnmnOH HmmHu Hmmm OOO.O H HOO.OOHAOHO mOO.O H mOm.HH OmN.O H mHm.O mN mmmuw>¢ NNO.NH H NOO.HHH OON.H H OOO.NH ONO.O H OOO.m HH OOan-OH NNO.NH H mHm.OO AmO mmm.O H Omm.OH OHN.O H NOO.O OH OOnNHum xmmuu commEosa z¢uHmuH2 mmdq usmflm3 umm sz unmflm3 who unmflmz umz z vmamemm myma coaumuoq Amm H av cowumnucmocou mscflmmm man .HH XHQmemd 98 OOO.O H mmO.N OHO.O H OOO.O NH0.0 H OOH.O HHO OOanH zoommo ONO.O H HON.mN mO0.0 H OOm.N HN0.0 H NOO.O HHN OOnONnH Hm>Hm cousm mam NOO.m H NOO.ON ONN.O H mOO.N NmH.O H OOH.H OH OoanHH HmmHu NHHmnu monmmOm mMOH mOO.O H NHm.ONH mOm.O H OOO.NH OOH.O H HNm.m ON mmmHm>¢ NO0.0 H OOm.mHH mOO.O H NOO.NH HHN.O H ONO.m mH HOOaOHuHH HHN.m H OON.NHH OOO.O H Nmm.NH OOH.O H OOH.m OH OmnHuH NOO.m H OOm.ONH mmO.O H HOm.NH Omm.O H Hmm.m OH OOuOHnHH OHm.O H OOH.HNH OOO.O H OOO.NH HO0.0 H HNN.m NH OOuOuHH oH ONnOH NHH.OH H OOH.OmH NOO.H H OOO.HH NOm.O H NOO.m OH OOnOHIOH NNO.O H mmN.HHH ONO.O H Hmv.mH mmm.O H HOO.m ON OOnOHIOH 0» OHaO Hm>Hm mHHmHO HNO.m H OHN.HNH AOmO HOO.O H OOO.NH NOH.O H OOm.m OH mmmHm>4 mmO.m H ONH.OHH OOO.O H OOH.NH NOH.O H mNH.m OH OOanuNH HNO.O H mmm.mNH mmH.H H OOO.mH OOO.O H NHO.m O OOnOuHH HMO.O H NON.ONH mOO.O H HmO.OH mOm.O H ONN.O N OmnqNLOH HO0.0 H HmO.HmH OHO.O H NOO.NH NH0.0 H OOO.m O OOuONnO HON.NH H mmO.OOH HOV HmO.H H mOO.mH OO0.0 H OOO.m NH OOnOnO mwumwcmz mHHHHH mmm.O H OOO.ONH HOO.O H NHN.OH OOH.O H OOH.O Hm mmmHm>m HOO.m H HNO.HNH OHO.O H NHO.NH NNN.O H mom.m OH OOannNH ONm.O H mON.OHH ONO.O H HOH.mH HOm.O H ONm.O NH OO-HN-OH OOO.O H NOH.ONH ONO.O H ONO.HH mOm.O H OH0.0 NH OOnmnOH HmmHu Hmmm OOO.mH H HOH.OmH HOHO HOO.O H HOH.OH Nmm.O H ONO.O mN HOHHm>4 NNO.OH H ONN.HOH HON.H H OOO.OH mOO.O H ONm.O HH OOumHnOH ONO.OH H OHO.OmH HOV OOO.H H OHO.OH OHm.O H HOO.O OH OOnOHnO xmmHo commsona zgonon mMHH usmflmz umm sz unmflmz who unmwmz #03 z vamEmm mama cowumooq Hum H HO coHHHHHcmocou msOHmmm man can can Baa Hmuoe A.#GOUV HH xHDmem4 99 mHmEMm uflmomEoo mco mo mmmMHmc¢ HH nmwm Emmuum Hmsvflmmu mo msono .1 omo.o H HHH.o moo.o H mao.o ooo.o H voo.o «He mmlvla zommmo ooo.o H Hom.o ooo.o H wmo.o ooo.o H moo.o HHN molomla H0>Hm cousm mHm «HH.o H mmm.o HH0.0 H mmo.o moo.o H mmo.o 0H monauaa xmmuu mnumno mOHmmmDm mMm mmm.o H mHn.H mHo.o H mwa.o hoo.o H mno.o m «mmnvalaa HvH.o H mno.m «Ho.o H NNN.o moo.o H mmo.o OH mmlvna mma.o H mmm.a mHo.o H mna.o woo.o H mno.o oa wmlvalaa mNH.o H mmn.a oao.o H mma.o voo.o H wwo.o HH mmlmlaa ou vmloa Hmm.o H Nam.m omo.o H ~om.o moo.o H mmo.o 0H mmlnanoa mom.o H mmn.m mmo.o H «mm.o oao.o H mHH.o Hm moloaloa 0H mHIm Hm>Hm muHmHm NHH.O H OHO.N HH0.0 H OOH.O mO0.0 H OO0.0 NN HOHHm>4 mma.o H moo.m OHo.o H oa~.o moo.o H vmo.o 0H mmlmalma mma.o H voa.m mmo.o H Hom.o «Ho.o H mmo.o m monnlaa Hm~.o H Nmm.a oao.o H mmH.o noo.o H «no.0 O monvmloa Hm>Hm _ mmuchmz mHuHHA mO0.0 H mOO.N OO0.0 H mHN.O HO0.0 H MO0.0 NN mmmum>¢ oma.o H Omm.H mao.o H mom.o moo.o H mmo.o OH mommanma mmH.o H moo.m mao.o H vmm.o moo.o H nmo.o NH mmlvmloa xwmuo Hmwm meMHucmsw Ho: Han ucwmmum mm xmmuo acmmfiona z¢onqu mmmq HamHmz Hmm unmwmz who unmflm3 umS z Umamamm mama GOHHMUOA Amm H MO GOHHmuuamoaou mswwmmm CHHUHmHQ H.Haouv .HH xHQZMmmd 100 0.0 H OHO.O 0.0 H OO0.0 0.0 H OO0.0 HHO OOnOnH zowmmo HO0.0 H HO0.0 0.0 H OO0.0 OO0.0 H OO0.0 HHN OOnONnH Hm>Hm :oHsm OHm OO0.0 H OOH.O 0.0 H OOO.O HOO.O H OOO.O OH OOaHnHH xmmHu OHHmOu monmmam mmmH ON0.0 H OH0.0 HO0.0 H OO0.0 OO0.0 H OON.O OO momHm>m OO0.0 H OH0.0 OO0.0 H NOO.O OH0.0 H OON.O OH HOO-OHnHH OOO.O H OOO.O HOO.O H OOO.O OHO.O H HON.O OH OOuOuH OOO.O H OOO.O OOO.O H OOO.O HH0.0 H OON.O OH OOuOHnHH OO0.0 H ONO.O OOO.O H OO0.0 OH0.0 H OO0.0 NH OOuOnHH oH ONuOH HOO.O H OOO.O HOO.O H OOO.O OH0.0 H OON.O OH OOuOHaOa OOO.O H OHO.O NOO.O H NOO.O HHO.O H OON.O ON OOnOHnOH oH OHnO Hm>Hm mHHmHO NOO.O H OOO.O NOO.O H OOO.O HHO.O H OON.O NO mOmHm>« OOO.O H OOO.O HO0.0 H OOO.O HH0.0 H OON.O OH OOumHnNH HOO.O H HO0.0 OO0.0 H OOO.O NNO.O H OO0.0 O OOnOnHH OOO.O H OOO.O OOO.O H OO0.0 ON0.0 H HON.O O OOIONuOH OO0.0 H OH0.0 HO0.0 H OO0.0 ON0.0 H OON.O O OOnONuO Hm>Hm . mmHchmz mHHHHH OO0.0 H ONO.O NO0.0 H OO0.0 HH0.0 H OH0.0 Om mOmHm>< NO0.0 H OOO.O OO0.0 H NOO.O OH0.0 H OO0.0 OH OOnOHnNH OO0.0 H OOO.O OO0.0 H OO0.0 OH0.0 H ONm.O NH OOIONIOH OO0.0 H OOO.O HOO.O H OOO.O OH0.0 H HNO.O NH OOIOuOH xmmHo Hmmm OO0.0 H OOO.O OOO.O H HO0.0 OO0.0 H OO0.0 OH mmmHo>m OOH.O H OO0.0 HH0.0 H OOH.O OO0.0 H OOO.O HH OOanuOH OOO.O H OO0.0 HO0.0 H OO0.0 ONO.O H OOO.O O OOIOHuO HmmHu commsons zmonqu mmmq man can son 2 mmaafimm muma cowumooq Amm xv mGOHumHucmocoo mswfimmm msmHmmH cHHmHmHm mam Anna .aao .HQQO Boa HmHoH .maa .aaa Hana .m.m may .mmaafimm mum3 mammuum map mo nomm mmumm map How mmm\mn mm Ummmmuaxm cosamm 0:00 mo mmmm may cw mGOHHmHHamoaoo .HHH XHQmemfl 101 mHaEmm muHmanoo mao mo mmmmama< HH zmwmfimmupm HmDUOmmH mo asouw .1 OOO.O H HOO.O HOO.O H ONO.O HHO OOuOuH zoummo OOO.O H HOO.O OOO.O H OHH.O HHN OOnONnH Hm>Hm coHsm OHm OOO.O H OOO.O NH0.0 H OOH.O OH OOnHuHH HmmHu OHHmnu monmOOO mmaq HOO.O H OHO.O AOOO OOO.O H OOO.H OO mOmHm>4 HOO.O H HHO.O HOV OOO.O H OO0.0 OH HOOnOHnHH HOO.O H ONO.O OOO.O H OOH.H OH OOnOuH HOO.O H OHO.O OOO.O H OHH.H OH OOuOHuHH HOO.O H OHO.O AHHO OOO.O H OOH.H NH OOumuHH oH ONuOH HOO.O H OHO.O OOO.O H OOO.H OH OOnOHnOH NOO.O H ONO.O AHNO HOO.O H OOO.H ON OOnOHnOH oH OHnO Hm>Hm HHHHHO HOO.O H OHO.O ANNO OOO.O H NOH.H NO mOme>< HOO.O H OHO.O OOO.O H OOH.H OH OOnOHnNH HOO.O H OHO.O AOO OOO.O H OOH.H O OOuOuHH HOO.O H OHO.O OHH.O H HOO.O O OOuONuOH . OOO.O H OOO.H O OO-ON-O Hm>Hm mmHchmz mHHHHH HOO.O H OHO.O HO0.0 H OON.H OO mOmHm>4 HOO.O H OHO.O NOO.O H OOH.H OH OOIOHsNH HOO.O H OHO.O OOO.O H OON.H NH OOIONIOH n OOO.O H NHN.H NH OO-O-OH HOOHU Hmmm . ONH.O H OHO.H OH HOOHO>4 . OOH.O H OHO.H HH OOuOHnOH . HO0.0 H OOO.H O OOnOHnO xmmHo commeona ZHOHmqu HHOH OHHOHOHO sz HOHoa z mmaafimm mHma GOHHmooa Ham H xv mQOHumHHcmoaou mswflmmm A.uGOUV .HHH NHszmmfi “7'1'ifiilfiflfljflmflfifllmflflfl Milli! ITS