Sex Ratios of Eels Reared Under Two Temperature Regimes

Canadian Technical Report of Fisheries and Aquatic Sciences 2135 November 1996 Sex Ratios of Eels Reared Under Two Temperature Regimes by R. H. Peterson, T. J. Benfeyl, S. A. McGeachy, M. Rommens, K. Richards l and P. Harmon Fisheries and Oceans Canada Biological Station St. Andrews, New Brunswick EOG 2XO Canada IDepartment of Biology, University of New Brunswick, Fredericton, N. B. E3B 6El This is the two hundred and twenty-third Technical Report from the Biological Station, St. Andrews, N. B.

ii Minister of Supply and Services Canada 1996 Cat. No. Fs 97-612135 ISSN 0706-6457 Correct citation for this publication: Peterson, R. H., T. J. Benfey, S. A. McGeachy, M. Rommens, K. Richards, and P. Harmon. 1996. Sex ratios of eels reared under two temperature regimes. Can. Tech. Rep. Fish. Aquat. Sci. 2135: iii + 13 p.

iii ABSTRACT Peterson, R. H., T. J. Benfey, S. A. McGeachy, M. Rommens, K. Richards. and P. Hannon. 1996. Sex ratios of eels reared under two temperature regimes. Can. Tech. Rep. Fish. Aquat. Sci. 2135: iii + 13 p. American eels (Anguilla rostrata (LeSueur» were cultured to a length of 30 cm at a continual high temperature (23 C) and at 23 C, but with a 3-mo exposure to 5 C, with triplicate tanks for each treatment. Survival to 30 cm was low, ranging from 1-25% for the six tanks. The period of cold exposure did not significantly increase the proportion of females (2 of 35 for cold exposure vs 0 of 22 for wann exposure). One eel from each treatment was sexually undifferentiated. Of 40 wild eels obtained from a local stream, 27 (67.5%) were females. were females. All wild eels greater than 37 cm Peterson, R. H., T. J. Benfey, S. A. McGeachy, M. Rommens, K. Richards. and P. Hannon. 1996. Sex ratios of eels reared under two temperature regimes. Can. Tech. Rep. Fish. Aquat. Sci. 2135: iii + 13 p. Les chercheurs ont eleve des anguilles americaines (Anguilla rostrata (LeSueur» jusqu'a ce qu'elles atteignent une longueur de 30 cm. Un premier groupe a ete expose de fayon continue a une temperature elevee (23 C et, pendant trois mois, SoC). Chacun des traitements a ete realise dans trois aquariums. Seul un faible pourcentage des anguilles ont survecu jusqu'a la taille de 30 cm (l a 25% dans les six aquariums). La periode d'exposition au froid n'a entraine une augmentation significative du pourcentage de femelles (2 femelles sur 35 anguilles pour Ie groupe expose a une basse temperature, comparativement e 0 femelle sur 22 anguilles pour Ie groupe expose de fayon continue a une temperature elevee). Dans chacun des groupes, on a observe un specimen sexuellement non differencie. Parmi 40 anguilles sauvages capturees dans un cours d'eau local, 27 (67,5%) etaient des femelles. Toutes les anguilles sauvages mesurant plus de 37 cm de longueur etaient des femelles.

1 INTRODUCTION The eel (Anguilla spp.) is currently cultured in many countries, but principally in Europe and Asia. The annual production amounts to over 100,000 tonnes (Gousset 1992). Production of eels in culture presently is totally reliant on collection of wild elvers. Countries intensively culturing eels are periodically faced with elver shortages. Consequently, although the elvers of the American eel (Anguilla rostrata (LeSueur)) are not highly regarded by Asian eel farmers (due to poor adaptation to Asian temperatures and pathogens), the demand for North American elvers is increasing. One of the problems confronting the eel farmer is the highly skewed sex ratios of cultured eels in favor of males, which achieve a smaller size prior to maturation (=final size) than do females (Beullens et al. 1991; Usui 1991; Ingraham 1992; Chiba et al. 1993; Kushnirov and Degani 1995). In contrast, sex ratios of wild eel populations of eastern Canada are strongly biased toward females (Gray and Andrews 1970; Dolan and Power 1977; Ingraham 1992). The contrast in sex ratios between wild and cultured eel populations suggests an environmental control of sex determination, rather than genetic. Kuhlmann (1975), however, suggested that sex is determined prior to the freshwater stage. Large elvers tended to become females, while smaller elvers were generally males. This relationship may be true for wild populations, where the size of elvers entering streams increases latitudinally from south to north. Northern eel populations are predominantly female. The relationship between elver size and eventual sex cannot pertain to cultured eel populations, where northern elvers still produce a sex ratio greatly favoring males. There are many factors which differ between culture and wild environments, such as stocking densities, annual temperature regime, diet, and the physical nature of the habitat. Tesch (1977) observed that males were found primarily in densely populated habitats, and females in sparsely populated areas. Also, Degani and Kushnirov (1992) demonstrated that density appeared to influence sex ratio. In European eels reared communally, 77% developed as males, while 60% of the population became females when reared in isolation from each other. Temperature has been shown to influence sex determination in the Atlantic silverside (Menidia menidia) (Conover and Kynard 1981), and hence is a possible determinant in the eel as well. In nature, northern eels experience seasonal periods of low (near 00 C) temperature, while cultured eels experience continual high temperature. In the present study, the influence of a period of low temperature during the first year in culture on sex determination in eels was investigated. The sex ratio of a sample of wild eels from a local population was also determined for comparison with the results from the cultured eels. A number of sex steroids have been used successfully to manipulate sex in cultured eels, including 1713-estradiol (Degani and Kushnirov 1992; Lin et al. 1992; Chiba et al. 1993; Kushnirov and Degani 1995; Andersen et al. 1996), 17aethylnylestradiol (Colombo and Grandi 1992; 1995), sodium diethylstilbestrol (Satoh et al. 1992) and testosterone (Lin et al. 1992; Colombo and Grandi 1992; 1995). However, direct use of steroids on eels destined for the marketplace may lead to problems with consumer acceptance. Environmental control of sex determination would obviate this problem. Identification of the sex of eels is not a straightforward matter. When eels reach about 10 cm in length the gonad usually begins to differentiate morphologically into either a frilled "organ of Mondini", which is synonomous with ovary, or a lobed "organ of Syrski", which several (particularly earlier workers) have identified as the testis. However, identification of maleness, based on gross morphology of the gonad, is not always reliable (e.g. Dolan and Power 1977). In early gonadal development, both oocytes and spermatogonia may be present. When the decision is made with regard to sexual expression, one cell type proliferates and the other is supressed. Our conclusions as to sexual identity are based on histological examination of fixed and stained sections. MATERIALS AND METHODS TANK DESCRIPTION The experimental tanks were of fiberglass, 1 m x 1 m x 30 cm high (Fig. 1). The tanks were divided into three compartments: a smaller compartment (15 cm x 1 m) where the water entered and was aerated, a main compartment occupied by the eels and a third, effluent compartment. Six tunnels of PVC pipe embedded in a PVC support (Fig. 1) in the main compartment provided shelter for the eels. All water flow through the main compartment was via the tunnels, which exited into the effluent compartment. Caps on the effluent end of the tunnels could be unscrewed to examine the tunnel interiors for

2 cleaning and removal of mortalities. The effluent of each experimental tank exited to a separate 200-L plastic container filled with Koch rings and fitted with aeration and temperature control. This container thus functioned as a biofilter from which the water was pumped back to the inlet compartment of the experimental tank; however, the main function was to conserve heat, as the water was replaced every 4 h. The actual flow of water through the experimental tanks was maintained at 2000-3000 ml/min. EEL ACQillSITION In May 1994, 2000 elvers were obtained from a local supplier. They were reared at 15 C for the first 2 mo and then the temperature was increased gradually for 1 mo until a level of 23-24 C was achieved. Due to very high mortalities during this initial period, 130 elvers were retrieved from the wild in August 1994 by e1ectrofishing. In addition, 198 elvers were brought in from a local eel farm on November 1, 1994. Eight days later, the eels were assigned to one of six tanks and subjected to temperature treatments according to Table 1. In August 1995,42 wild eels were collected from Dennis Stream for sex determination. The sex ratio of these wild eels was then compared to those obtained under culture conditions. PROPHYLACTIC TREATMENTS Considerable aggressive behavior occurred in the tanks during the course of the experiments. This behavior seemed to increase after removal of large eels for sex determinations, possibly associated with establishment of new dominance hierarchies. Thus, it was periodically necessary to control invasion of fungus into lesions resulting from biting through application of salt and/or methylene blue. Salt (through addition of seawater) was used at 10 ppt, while methylene blue was used at 1 ppm. TEMPERATURE REGIMES AND DIETS Upon transfer into the tanks, water temperature was appoximately 23 C. On November 11, 1994, the water temperature in the cold temperature treatment tanks was lowered by 1 per day until reaching ambient levels on November 24, 1994 (Fig. 2, 3). This temperature regime was held constant until February 21, 1995. At this time, temperatures in the cold treatment tanks were increased by 1 per day until reaching the target temperature of 23 C on March 24, 1995. Temperatures In the hot treatment tanks were maintained at 23 C. (The elvers were fed a diet consisting of herring roe, frozen artemia, herring pieces, and commercial feeds - Moore-Clark Royal Response and Moore-Clark 2.0 mm salmon pellets.) PREPARATION OF EELS FOR SEXING Eels were kept continually under visual assessment as to whether they had achieved adequate size for sexing (30 cm). When eels were deemed to have attained sufficient length, they were removed from the tank and exposed to a sufficient concentration of MS222 to induce anaesthesia in 5 min. Their weight and length were then measured. If they had attained 30 cm in length, they were lethally exposed to MS222, slit ventrally to expose internal organs, and preserved in 10% formalin. One gonad was removed from each eel in its entirety, classified as to being frilled or lobed, and cut into segments for processing. Three segments of each gonad (anterior, middle and posterior) were examined. Each gonad segment was placed in a biopsy bag, then enclosed in a labelled cassette, which was stored in 10% formalin. The gonad segments were infiltrated with paraffin and embedded using standard Tissue Tek techniques. The paraffin blocks were sectioned at 7 /lffi, and Haupt's adhesive was used to fasten the sections to the slides. The sections were de-paraffinized and stained with Harris' hematoxylin and eosin + orange G. The stained sections were examined by light microscopy at 100-400x. The shape, morphology and cellular appearance of the gonads were used to confirm the sex of the fish (Fig. 4, 5) as described in the literature (Gray and Andrews 1970; Wenner and Musick 1974; Dolan and Power 1977; Colombo et al. 1984; Chiba et al. 1993; Colombo and Grandi 1995, 1996; Kushnirov and Degani 1995). SURVIVAL RESULTS Survival was high for the first 3 mo of the experiment (Table 1), ranging from 72.8-99.0% for the six groups. The groups exposed to a period of low temperature had higher survival than those maintained at 23 C. This higher survival probably was not a direct effect of temperature, but indirect through a slower metabolic rate in cold-exposed groups, resulting in slower growth, less size variation and less aggressive behavior. As the size range increased with time, due to variation in growth rates,

3 mortality due to aggression and cannibalism resulted in poorer survival, averaging about 17% by November 1, 1995. The specific growth rates for the three groups at 23 C for this time period (tanks 2, 3, 5) were 0.29, 0.6, and 0.43% body weight/d, respectively. The cold-exposed eels lost weight while at low temperature. CHRONOLOGY The chronology of removal of eels which had attained the 30-cm. criterion (Table 2) indicates that, as expected, the eels reared at constant high temperature attained a 30 cm length earlier than did those exposed to a period oflow temperature. Ten of the 19 eels surviving to 30 cm from the constant high temperature treatments were sacrificed between February 1995 and July 1995. The two collected on February 23, 1995 illustrate the widely varying growth rates among groups of cultured eels. They had attained a weight of nearly 70 g as compared to the group mean of 10.4 g (Table 1). In contrast, those eels subjected to low temperature treatment were harvested mostly from September 1995 until February 1996. SEX DETERMINATION Only two of the 57 cultured eels reared to 30 em and sexed histologically were identified as female (Table 3). These two eels had frilled "organs of Mondini" which were unequivocally ovarian when examined histologically (Fig. 4). Both were from Tank 6 which received low temperature treatment; however two of a total of 35 eels from the low temperature treatments is not sufficient to warrant further statistical consideration. A total of 22 eels from constant high temperature treatments were sexed, none of which were unequivocal females. Two eels from low temperature treatments and one from a high temperature tank had lobed "organs of Syrski" that had some apparent oocytes in the histological sections, but appeared to be composed mostly of testicular tissue. The gonads from the remaining eels appeared to be males (Fig. 5), although many were not well differentiated. Of the 40 wild eels successfully sexed, 27 were females and 13 were males. Those sexed as females averaged larger (mean length = 42.8 cm, sd = 8.4 cm) than those sexed as males (mean length = 34.1 cm, sd = 1.6 cm). All eels >37 em were female. DISCUSSION The prevalence of males in populations of cultured eels is well documented for European eels (Beullens et al. 1991; Usui 1991), in contrast to the situation with wild populations of northern latitudes. That temperature influenced sex differentiation in eels seemed a reasonable hypothesis for two reasons. Firstly, cultured eels do not experience the periodic low temperatures ofwild eels. Secondly, temperature has been shown to be influential in determining sex ratios in some other fish species (e.g. Conover and Kynard 1981; Baroiller et al. 1995; Craig et al. 1996; Romer and Beisenherz 1996; Striissmann et al. 1996). However, the experiments described in this report provide no support for the hypothesis. It may be that other environmental conditions unique to culture are influencing sex determination instead of temperature, or perhaps interact with temperature. It is also possible that our choice of temperature or timing of cold exposure was not optimal for feminization. There is evidence that eel densities in wild populations may influence sex ratios, with higher proportions of males in denser populations (Tesch 1977). The high densities employed in culture systems may well be of influence. Degani and Kushnirov (1992) found that 60% of eels reared in isolation became females, as opposed to 77% males when reared in groups. They postulate that stress associated with high stocking density results in increased corticosteroid secretion, which, in tum, may affect sex determination. Any inferences made from the results of these experiments are conditional upon correctly sexing eels. Smith and Saunders (1955) found that they could not reliably distinguish between the organ of Syrski and undifferentiated gonads for eels less than 45 cm in length, on a macroscopic basis. The relatively undifferentiated state of the gonads we presume to be testes, and the prescence of some oocytes in some of these, lends some uncertainty to the determinations. Degani and Kushnirov (1992) found that injection with human chorionic gonadotropin stimulated differentiation, and its use in these experiments might have resulted in more well defined testes and ovaries. The high proportion of presumed males in the sample of wild eels taken from Dennis Stream (a small river just to the west of S1. Andrews) may be due to high densities. The influence of population density and stream size on sex ratios in wild populations should be systematically examined.

4 REFERENCES Andersen, D., 1. Boetius, L. O. Larsen, and P. H. Seidler. 1996. Effects of oestradiolenriched diet and feeding with porcine testicular tissue on macroscopic gonadal sex in European eels. J. Fish BioI. 48: 484-492. Baroiller, J. F., D. Chourant, A. Fostier, and B. Jolabert. 1995. Temperature and sex chromosomes govern sex ratios of the mouth breeding cich1id fish Oreochromis niloticus. J. Exp. Zooi. 273: 216-223. Beullens, K., W. Beeckman, and F. Ollevier. 1991. Influence of 17-~ estradiol on the gonadal sex of the European eel (Anguilla anguilla L.). Gen. Compo Endocrinoi. 82: 240 (Abs.). Chiba, H., K. Iwatsuki, K. Hayami, and K. Yamauchi. 1993. Effects of dietary estradiol-17~ on feminization, growth and body composition in the Japanese eel (Anguilla japonica). Compo Biochem. Physioi. 106A: 367-371. Colombo, G., and G. Grandi. 1992. Further experiments on the effects of sex steroids on the gonad sex differentiation of European eel. Irish Fish. Invest. Ser. A 36: 126-129. Colombo, G., and G. Grandi. 1995. Sex differentiation in the European eel: histological analysis of the effects of sex steroids on the gonad. J. Fish BioI. 47: 394 413. Colombo, G., and G. Grandi. 1996. Histological study of the development and sex differentiation of the gonad in the European eel. J. Fish BioI. 48: 493-512. Colombo, G., G. Grandi, and R. Rossi. 1984. Gonad differentiation and body growth in Anguilla anguilla L. J. Fish BioI. 24: 215-228. Conover, D.O., and B. E. Kynard. 1981. Environmental sex determination: interaction of temperature and genotype in a fish. Science 213: 577-579. Craig, J. K., C. J. Foote, and C. C. Wood. 1996. Evidence for temperature-dependent sex determination in sockeye salmon (Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. 53: 141-147. Degani, G., and D. Kushnirov. 1992. Effects of 17~estradiol and grouping on sex determination of European eels. Prog. Fish-Cult. 54: 88 91. Dolan, J. A., and G. Power. 1977. Sex ratio of American eels, Anguilla rostrata, from the Matarnek River system, Quebec, with remarks on problems in sexual identification. J. Fish. Res. Board Can. 34: 294-299. Gousset, B. 1992. Eel culture in Japan. Bull. Inst. Oceanogr. (Monaco) 10: 1-128. Gray, R. W., and C. W. Andrews. 1970. Sex ratio of the American eel (Anguilla rostrata (LeSueur)) in Newfoundland waters. Can. J. Zool. 48: 483-487. Ingraham, D. L. 1992. Sex ratios and sexual differentiation in a Saint John River population of the American eel, Anguilla rostrata. Honors Thesis, Dept. BioI., University of New Brunswick, Fredericton, N.B. 34p. Kuhlmann, H. 1975. Der einflub von temperature, futter, grobe und herkurft auf die sexuelle differengierung von glasaalen (Anguilla anguilla). Helgoland. wiss. Meeresunters. 27: 139-155. Kushnirov, D., and G. Degani. 1995. Sexual dimorphism in yellow European eels, Anguilla anguilla (L.). Aquacult. Res. 26: 409-414. Lin, H.-R., M.-1. Zhang, L. Chen, G. Van Der Kraak, and R. E. Peter. 1992. Effects of sex steroids on gonadotropin synthesis and secretion as well as ovarian development in female Japanese silver eel Anguilla japonica. Irish Fish. Invest. Ser. A 36: 142. Romer, U., and W. Beisenherz. 1996. Environmental determination of sex in Apistogramma (Cichlidae) and two other freshwater fishes (Teleostei). J. Fish BioI. 48: 714-725. Satoh, H., Y. Nimura, and T. Hibiya. 1992. Sex control of the Japanese eel by an estrogen

5 (DES-Na) in feed. Nippon Suisan Gakkaishi 58: 1211-1218. Smith, M. W., and J. W. Saunders. 1955. The American eel in certain fresh waters of the Maritime Provinces of Canada. J. Fish. Res. Board Can. 12: 238-269. Striissmann, C. A., S. Moriyama, E. F. Hanke, J. C. C. Cota, and F. Takashima. 1996. Evidence of thermolabile sex determination m pejerrey. J. Fish BioI. 48: 643-651. Greenwood (ed.). English eddition. Chapman and Hall, London. 434 p. Usui, A. 1991. Eel Culture, 3rd Edition. Fishing News Books, The University Press, Cambridge. 148 p. Wenner, C. A., and J. A. Musick. 1974. Fecundity and gonad observations of the American eel, Anguilla rostrata, migrating from Chesapeake Bay, Virginia. J. Fish. Res. Board Can. 31: 1387-1391. Tesch, F.-W. 1977. The eel: biology and management of anguilloid eels, P. H.

6 Table 1. Numbers (n), lengths (1), weights (wt) and condition factors (cf) ofeels placed in the experimental tanks at the beginning of the experiments (1994), and at two subsequent dates. * - eels were obtained from a local eel farmer. Length in cm, weight in g, condition factor = wtje x 1000, % s = percent survival. Eels in tanks 1, 5, 6 received cold treatment. November 1, November 9,1994 Februarv 23, 1995 1995 Tank # n I (sd) wt (sd) cf n I wt cf %s n %s 1 70 10.7(1.4) 2.5(1.0) 2.04 64 10.9(1.5) 2.0 (0.9) 1.54 91.4 7 10.0 5 71 10.9(1.7) 2.7(1.3) 2.08 61 11.1(1.7) 2.1 (1.1) 1.54 85.9 15 21.1 6* 98 15.0(2.5) 6.8(3.9) 2.02 97 15.2(2.6) 6.0 (3.4) 1.71 99.0 34 35.1 2 70 11.1 (1.6) 3.0(1.2) 2.19 51 13.4(2.6) 4.1 (2.6) 1.70 72.8 3 "4.3 3 70 10.9(1.3) 2.8(1.0) 2.16 52 14.0(3.2) 5.3 (6.1) 1.93 74.3 19 27.1 4* 100 14.6(2.9) 6.6(4.6) 2.12 85 16.8(4.6) 10.4(12.4) 2.19 85.0 7 7.0

7 Table 2. Length (cm)/weight (g) data for individual eels sampled for histological sex determination. Six eels were not weighed or measured prior to histology. Number of eels in Tank 1-6 was 5,1,11,7,5, and 25, respectively. Sampling date Tank 1* Tank 2 Tank 3 Tank 4 Tank 5* Tank 6* Feb. 23/95 31.4/68.7 30.2/68.7 Apr. 5/95 29.8/55.4 31.3/68.4 Apr. 26/95 32.4/73.9 30.1/60.4 30.5/60.0 June 14/95 30.5/57.6 July 10/95 29.5/42.7 30.4/60.8 July 17/95 29.4/48.9 Sept. 18/95 35.3/96.9 35.31107.8 30.5/69.8 32.0/68.7 31.7/74.7 31.7/76.3 36.6/112.6 33.3/76.9 30.5/66.6 33.2/85.1 32.9/84.2** 32.8/77.8 31.3/66.1 32.9/69.8 33.4/92.6 30.5/52.7 30.6/72.2 Nov. 1/95 30.1/60.9 30.1/53.2 29.5/56.3 31.9/72.0 29.9/54.4 31.3/73.0 34.1/92.3 32.5/83.9 30.8/59.2 31.7/78.1 31.0/67.3** 33.0/84.3 Jan. 8/96 34.2/80.5 34.3/94.1 34.1/97.7 34.0/83.4 35.5/104.5 36.4/107.3 32.3/78.5 32.3/75.1 34.7/98.5 Feb. 6/96 30.0/57.2 32.1/83.2 31.1/58.6 32.9/76.3 31.3/67.1 *Cold treatment tanks. **Female.

8 Table 3. The sexes of eels reared to 30 cm from each of the experimental tanks, and of 40 wild eels collected from Dennis Stream Tank Males Females Undifferentiated Unknown Cold treatment 1 5 0 0 0 5 5 0 0 0 6 20 2 1 2 Total 30 2 1 2 Warm treatment 2 0 0 1 0 3 12 0 0 0 4 9 0 0 0 Total 21 0 1 0 Wild 13 27 0 0

9 Fig. 1. Photograph of the experimental eel tanks. Compartment to the left was aeration and mixing. The water then passed through the perforated partition to the rearing chamber. The dark black in the rearing chamber contained six PVC tubes to provide shelter for eels. Three tubes exited into the black effluent chamber on the far right. Note flanges around the rim of the rearing chamber to prevent escape ofeels. A plexiglass lid is omitted from the photograph.

-+-AVGLOW -ll-avg HIGH Fig. 2. Average temperatures for eels reared at continuous high temperatures, and for eels exposed for 3 rna at 5 C. Daily temperatures for tanks 2, 3 and 4 are averaged for the high temperatures, and for tanks 1,5 and 6 for the low. -I 30 I I ';;"-'1; 25 20-0 0-CD.... ::I 16 15 CD Co f-' o E CD I- 10-5 It'~~/?.~,n~;;il~t:~":;: ':'1;'~:,,;: ::i~~'~, NoM ~';: " I i I, o! 11111111111111111111111' 11111111' 1111111111 fit 11111111111' 111111111 n 1rt 1f 1111111111111 j 111111111111'1 11111111111111 i I1I1I1 f 1'11111111111111111111111111 1-Nov 1-Dec 1-Jan 1-Feb 1-Mar Date

. ~ ~=~==1 30 I I -+-TANKI """-TANK2 ~TANK3 ~TANK4 -'-TANK5 -'-TANK6 I-' I-' 25 -- --- -.----- -~~-- ----- ----ac O - 20 0-2..,..! ::I 16 15 ~ Q) c. E Q) I- 10, I 5 0 '<t '<t '<t 10 10 10 10 10 10 10 10 10 10 10 10 <0 <0 <0 OJ OJ OJ OJ, OJ OJ OJ OJ q> OJ OJ OJ OJ OJ q> q> OJ OJ.b.!..!. U > ~ 0 Z Cl..., u. ~ «...,..., ~ ~. 0.. U >.b.!. <= <= :; a. 0 <= 0 ro Q) ro a. =' =' Q) 0 Q) ro Q) ro Date «(J) 0 z Cl..., u. ~ Fig. 3. Monthly mean temperatures for an experimental eel tanks for the duration of the experiment.

12 A o f /./ Fig. 4. Histological sections demonstrating progressive ovarian development in the American eel. A, B: 32.9 cm cultured fish from tank 6; C,D: 52.5 cm wild eel; E,F: 60.0 cm wild eel; Bar = 250 J.IlI1 (A, C, E) or 100 J.IlI1 (B, D, F).

13 ::A -=~..,.,...-~----::::----~-~-----... G E Fig. 5. Histological sections demonstrating progressive testicular development in the American eel. A, B: 30.1 cm cultured eel from tank 1; C, D: 29.5 cm cultured eel from tank 4; E, F: 35.0 cm wild eel; Bar = 250 /llil (A, C, E) or 100 /llil (B, D, F).