PHOTOPERIODIC REGULATION OF COPULATORY BEHAVIOUR IN THE MALE HAMSTER
L. P. MORIN
AND
I. ZUCKER
Department of Psychology, University of California, Berkeley, California 94720, U.S.A.
(Received 26 August 1977) SUMMARY
The influence of daylength on copulatory behaviour was assessed by comparing male hamsters exposed to long or short photoperiods (14 or 2 h light/24 h). Copulation declined in animals transferred from long to short days; 13 out of 14 hamsters ceased to ejaculate within 9 weeks and many no longer intromitted in tests with sexually receptive female hamsters. The decline in copulation in hamsters experiencing short days was associated with atrophy of the gonads and flank glands. Behavioural changes in these animals were far more gradual than those observed in hamsters after surgical castration. There was significantly more mating behaviour in tests during the subjective night of the hamsters than during their
subjective day. Exogenous testosterone was more effective in restoring copulation in castrated hamsters exposed to long days than in castrated animals experiencing short days. This suggests that in short days the substrate for copulatory behaviour is relatively refractory to androgens. Photoperiodically mediated changes in behaviour, physiology and morphology may each contribute directly to the reproductive quiescence presumed to occur in the field during the short days of autumn and winter. INTRODUCTION
During the annual reproductive cycle of the hamster, short photoperiods act through the visual system and brain to initiate and maintain testicular regression (Reiter, 1974; Rusak & Morin, 1976). Testicular atrophy is associated with an 80% reduction in the level of testo¬ sterone and cessation of spermatogenesis (Berndtson & Desjardins, 1974). Although changes in the reproductive endocrinology of hamsters exposed to short photoperiods have been extensively recorded, the regulation of sexual behaviour under these conditions has not received experimental attention. Fewer litters of hamsters are delivered during the months of short daylength than at other times of the year (Czyba, 1968; Reiter, 1973-74), but it is un¬ clear whether this represents an effect of reduced gametogenesis or also involves diminished copulatory behaviour. The adult photostimulated male hamster secretes about 5 ng testosterone/ml plasma (Berndtson & Desjardins, 1974). Termination of this androgen secretion by castration results in the cessation of copulatory behaviour; the great majority of hamsters will no longer intromit or ejaculate 3-4 weeks after castration (Whalen & Debold, 1974). Administration of testosterone can restore copulatory behaviour. Exposure to short photoperiods may, in effect, produce a physiological castration by reducing the level of testosterone in the plasma to about 1 ng/ml; this decrease in the secretion of the hormone apparently precedes the * Present address : U.S.A.
Department of Psychology, Dartmouth College, Hanover, New Hampshire 03755,
cessation of spermatogenesis (Berndtson & Desjardins, 1974). The level of hormone from atrophied testes is inadequate to maintain the accessory sex tissues (Reiter, 1969), but its effectiveness in relation to sexual behaviour is unknown. Resolution of this issue is of interest for two reasons. First, if the reduced levels of testosterone in hamsters exposed to short days are incompatible with normal copulation, a behavioural change could contribute to the seasonal termination of reproduction. Secondly, the overwhelming majority of studies have used the unphysiological procedure of castration to assess testicular influences on behaviour. It is appropriate, where possible, to study behavioural changes under physiological con¬ ditions that more nearly approximate those of the animal in the natural state (i.e. photo¬ periodically induced regression). In this case, the pattern of hormone withdrawal and the basal levels of hormones attained are quite different from those observed after surgical removal of the gonads. The present experiments assessed the copulatory behaviour of male hamsters during exposure to short daylengths. Sensitivity to testosterone among hamsters exposed to different photoperiods and tested at different times of day was also determined. MATERIALS AND METHODS
Animals naive male hamsters were Sexually purchased from the Lakeview Hamster Colony, Newfield, New Jersey, U.S.A. Upon arrival in the laboratory, all animals were exposed to a lighting régime of 14 h light (L):10 h darkness (D) (lights on 21.00-11.00 h Pacific Standard Time). Hamsters were housed singly in wire-bottomed cages with access to Simonsen rat pellets (maintenance diet) and water ad libitum. These conditions prevailed for at least 1 month before the screening sessions at which time the hamsters were 3-4 months old. To facilitate selection of experienced male copulators, a series of screening tests was initiated approximately 1 month before any experiment. A male hamster was placed with a receptive female hamster for 5jmin or until an ejaculation occurred. Ovariectomized hamsters bearing a chronic subcutaneous implant of a Silastic capsule containing oestradiol benzoate were rendered sexually receptive by a subcutaneous injection of 1 mg progesterone 2-5 h before the test. Each male hamster received between eight and ten of these sessions. The pool ofanimals showing regular ejaculation during the screening tests was then given five pre¬ liminary sex tests at 2-3 day intervals; these tests were also 5 min long. The latencies to ejaculation and first mount were measured. In general, only animals which ejaculated in four out of the five tests (including the last one) within 5 min of being placed with a receptive female hamster were allocated to the experimental groups. The preliminary sex tests occurred between 13.00 and 15.00 h.
Photoperiods Upon completion of the preliminary sex tests, animals were either retained in 14L : 10D transferred to 2L : 22D. In each case, the room lights went off at 11.00 h.
or
Laparotomies The effect of the photoperiod on gonadal activity was assessed during laparotomy. The left testis was exposed through a mid-line incision rostral to the penis. The index of testicular activity [(length width)/body weight] (Rusak & Morin, 1976), which is highly correlated with spermatogenesis and the production of testosterone (Berndtson & Desjardins, 1974), was
recorded.
Hormone administration Non-esterified testosterone was administered via subcutaneous Silastic implants. Sections of Silastic tubing (Dow-Corning ; internal diameter 1 -93 mm ; external diameter 3-175 mm) were
packed with crystalline testosterone. Each end of a capsule was sealed with silicone adhesive (Dow-Corning Type A) and the effective length of tubing was 50 mm. During implantation, the animals were anaesthetized with sodium pentobarbitone (Nembutal, 80 mg/kg body weight). In several instances, castration was performed at the same time or separately. Serum testosterone analysis Animals were anaesthetized with ether and about 1 ml blood was withdrawn by cardiac puncture. Animals in Groups 3 and 5 were bled after the tests at week 12 (Table 1). After the blood had been allowed to clot overnight in a refrigerator, serum was obtained by centrifu¬ gation and the samples were stored frozen at —20 °C. Radioimmunoassay of androgens in the serum (largely testosterone and dihydrotestosterone) was performed according to the procedures described by Tamarkin, Hutchinson & Goldman (1976). Behavioural tests Male hamsters were pre-adapted to a 20 23 34 cm3 glass aquarium for 3-5 min before the introduction of the receptive female hamsters. Tests were terminated when an ejaculation occurred or after 7 min had elapsed. Generally two animals were tested simultaneously, in separate aquaria, under dim white light (approximately 50 lx at floor level). Hamsters were tested in random order by a technician. The number of mounts and intromissions and the mount and ejaculation latencies were recorded. A composite sex score was also employed in which a score of 1 was assigned to an animal that mounted at least once, 2 if mounting and at least one intromission occurred and 3 if the animal mounted, intromitted and ejaculated. A score of 3 was the maximum possible on any given test. A summary of the experimental groups, photoperiod and other conditions is presented in Table 1. Table 1.
Experimental details and photoperiodic conditions
Group
n*
Photoperiod light (h) : darkness (h)
1
13(11)
2 3 4
12(11) 9 (8) 12
Time of testosterone
14:10
Time of castration Week 1
2:22 2:22 14:10
Weekl Week 1 Week 1
Week 12 Weeks 1 and 12 Weeks 1 and 12
implantation Week 12
Weeks tested
3,t 6,t 9,t 12,t 15,t 16,î 17, 18, 19, 22, 26, 30 As Group 1 As Group 1 3,t 6,t 9,t 12, 16, 17, 19, 22
3,6,13 3,6,9, 12,13§ * Number of animals at the start of the experiment; number of animals tested at week 17 (for Groups 1-3 only) given in parentheses. t Tested between 13.00 and 15.00 h; during weeks without superscripts, tests occurred both at that time and 5 6
7 14
14:10 2:22
—
—
—
—
between 23.00 and 01.00 h. X Tested between 23.00 and 01.00 h. § Tested between 05.00 and 07.00 and 17.00 and 19.00 h. Hamsters were exposed to photoperiods of 14 h light : 10 h darkness (14L : 10D) or 2 h light : 22 h darkness (2L : 22D) and various components of sexual behaviour were tested under various conditions. For animals stably entrained to the 2L : 22D photoperiod, tests between 23.00 and 01.00 and 13.00 and 15.00 h occurred during the subjective night and subjective day of the animal respectively. For animals exposed to the 14L : 10D photoperiod subjective-night tests occurred between 13.00 and 15.00 h and subjtxtive-day tests between 23.00 and 01.00 h. See Materials and Methods for further details.
Subjective-day and subjective-night tests of the experiment it became apparent that the time of testing relative to the photoperiod influenced copulation. In the hamster, the 24 h day has been divided into an approximately 12 h subjective day (inactive phase) and a 12 h subjective night (active phase). The positioning of these two phases depends, among other variables, upon the duration of In the
course
the daily photoperiod. Detailed discussion of this point is beyond the scope of the present article. Nevertheless, it should be recognized that for animals housed in the 2L : 22D photo¬ period, subjective night would normally begin at 20.00 h and end at 08.00 h, for the animals in the 14L : 10D photoperiod subjective night would begin at 10.30 h and end at 22.30 h. These values are calculated from estimates provided by Elliott (1976). Thus behavioural tests conducted between 13.00 and 15.00 h fall during the subjective night and subjective day respectively, for animals in the 14L : 10D and 2L : 22D photoperiods. The converse relation applies to tests conducted between 23.00 and 01.00 h. RESULTS
Effects of castration By week 3 after castration, hamsters no longer displayed the behaviour pattern of ejaculation (Fig. 1, Groups 1 and 2). The frequencies of intromissions and mounts declined to low levels by 9 and 12 weeks respectively, after castration. The decline was more rapid in castrated animals exposed to short days (Group 2) than in those exposed to long days (Group 1 ; P Group 2; P< 0-01 for each behavioural component). The 2 h interval during which the tests took place (13.00-15.00 h) coincided with the subjective night of Group 1 animals and with the subjective day of Group 2 hamsters (see Materials and Methods). When these hamsters were re-tested in week 16 between 23.00 and 01.00 h, the groups did not differ significantly with respect to any of the three parameters recorded. At this point we suspected that merely testing animals during the dark phase of their daily illumination cycle was insufficient: to assess the influence of photoperiod on the hormonal activation of behaviour accurately, it seemed important that all animals be tested at the same subjective time. Accordingly, during week 17 each animal was first tested during its expected subjective day and then 12 h later during its subjective night. This procedure demonstrated that the efficacy of testosterone in restoring copulation was affected by the photoperiod. Intromission and ejaculation com¬ ponents were displayed more frequently (P< 0-005 for each parameter) by hamsters exposed to long days (Group 1) than by those experiencing short days (Group 2) in tests conducted during their respective subjective nights (Fig. 2). The groups did not differ significantly in the frequency of any copulation component during subjective-day tests; the performance of all groups was generally poor. The frequency of ejaculation was higher in intact control hamsters than in either of the groups receiving testosterone replacement therapy and the frequency of intromission was not restored to the levels found before castration in animals in Group 2. The differences detected between Groups 1 and 2 during week 17 were no longer detected during week 19 or in subsequent subjective-night tests. During these tests, per¬ formance of all groups approximately equalled that of Groups 1 and 3 obtained at week 17. In subjective-night tests during week 17, animals in Group 3 (implanted with testosterone capsules at the time of castration ; Table 1) copulated more frequently than hamsters exposed to short days and not implanted with testosterone capsules until week 12 (Group 2). The performance of Group 3 animals was not significantly worse than that of castrated hamsters
experiencing long photoperiods (Group 1, Fig. 2). Day-night differences There was significantly more copulation during the subjective-night than during subjectiveday tests conducted during week 17 (P< 0-025 for each behavioural component, for data from all animals).
0-8r
Ejaculating
Mounting
Intromitting
0-6 0-4 0-2
0: 12
3
12
3
12
3
Group number Fig. 2. Proportion of male hamsters ejaculating, intromitting or mounting in tests during week 17 conducted during the subjective night of each group after exposure to various photo-periods. Animals in Groups 1, 2 and 3 were treated as described in the legend for Fig. 1. All animals were implanted with Silastic capsules containing testosterone at week 12. Diurnal fluctuations in behaviour were further investigated in animals experiencing the 14L : 10D photoperiod (Group 4). These hamsters were castrated and implanted with testosterone capsules during week 1 ; the original capsule was then removed and replaced with a second testosterone implant during week 12 (Table 1). Both subjective-day and sub¬ jective-night tests were given during weeks 12, 16, 17, 19 and 22. The composite sex scores (Fig. 3) were significantly higher during the subjective-night tests. In week 12, the proportion of animals ejaculating, intromitting and mounting during the subjective-day tests were 0-17, 0-42 and 0-58 respectively, compared with a value of 0-92 for each component during the subjective-night tests. When blood was taken at the end of week 12, there was a significant difference in the level of androgens in the serum of the six hamsters of Group 4 bled during the subjective day and the levels in six others bled during the subjective night (9-14 + 0-32 (S.E.M.) and 6-65 + 0-38 ng/ml respectively, < 0-04, Student's i-test).
Fig. 3. Mean ( ± s.e.m.) copulation score for castrated hamsters experiencing 14 h light : 10 h darkness /day. Animals were initially castrated and implanted with testosterone capsules at week 1 and the original capsule was replaced at week 12 with a fresh one. Tests were conducted during the subjective night (·) or the subjective day (O).
During week 17, six hamsters were tested first during their subjective night and then during their subjective day; the remaining six animals were tested in the reverse order. This control procedure revealed that the higher copulation scores associated with subjectivenight tests were not a function of the order of testing. Diurnal differences were also analysed in intact hamsters (not castrated) housed in the 14L : 10D photoperiod throughout the experiment. The composite sex scores for these
during weeks 3, 6 and 13 failed to reveal a significant difference in copulation in subjective-day v. subjective-night tests. No fewer than six out of the seven hamsters ejaculated during each of the day and night tests. animals
Influence ofphotoperiod copulatory behaviour of the intact hamsters gradually declined in frequency with increasing duration of exposure to the short-day photoperiod (Morin, Fitzgerald, Rusak & Zucker, 1977). Animals were tested at a single fixed clock-time (13.00-15.00 h) both before and after the change from the long- to the short-day photoperiod. After we became aware of the significance of the subjective time in relation to copulatory performance, the experiment described by Morin et al. (1977) was repeated with 12 hamsters (Group 6). These animals were transferred from the 14L : 10D to the 2L : 22D photoperiod The
and their behaviour was assessed between 23.00 and 01.00 h as well as between 13.00 and 15.00 h during each test sequence. This procedure greatly increased the likelihood of sampling both subjective-day and subjective-night phases of the daily cycle even while re-entrainment of the circadian rhythms of the animals to the new photoperiod was in progress. Copulation declined during both the expected subjective-day and subjective-night tests (Fig. 4). The percentage of animals ejaculating decreased at a rapid rate between weeks 3 and 9 of exposure to short days (7% of the animals ejaculated during week 9). However, only 50% of the animals ceased to intromit by week 12. Tests conducted during week 13 at intermediate clock-times (05.00-07.00 h and 14.00-19.00 h) confirmed the generally poor copulatory per¬ formances established during tests at the two standard times.
Fig. 4. Mean (±s.e.m.) copulation score for intact male hamsters transferred from 14 h light : 10 h darkness to 2 h light : 22 h darkness at week 1. Tests were made during the expected subjective night (·) and subjective day (O). The testicular index (TI) of these animals assessed at week 13 was 1-09 + 0-97 (s.e.m.). This is indicative of gonads that have not fully regressed (see Table 2, short photoperiod control) and may account for the residual levels of copulation recorded in the tests during week 12. The main finding remains clear, however: short-day photoperiods inhibit copulation by male hamsters. discussion
The copulatory behaviour of many male rodents (Young, 1961), including hamsters (Tiefer, 1970; Whalen & Debold, 1974), is dependent on androgens secreted by the testis. Exposure to short daylengths effectively eliminates the mating behaviour of male hamsters, especially the ejaculatory pattern essential for successful reproduction. This loss of androgen-activated behaviour is probably influenced by the loss of gonadal androgens associated with gonadal atrophy induced by short photoperiods (Berndtson & Desjardins, 1974; Reiter, 1974). A
similar loss of male sexual behaviour associated with gonadal atrophy induced by short-photoperiods has been reported for the Japanese quail, Coturnlx coturnix (Sachs,
1969; Adkins, 1973).
The decline in copulatory behaviour is far more rapid after surgical castration than after exposure of hamsters to short days (Morin et al. 1977). The abrupt withdrawal of androgenic hormones and possibly the trauma associated with surgery may contribute to this difference. Gradual withdrawal of hormones, as occurs during photoperiodically induced testicular regression, is presumably more representative of the natural testicular cycle in the field
(Reiter, 1975) and constitutes a more valid assessment of hormone-behaviour interactions than can be obtained from studies involving surgical castration. Several investigations with male hamsters (Turek, Elliott, Alvis & Menaker, 1975; Tamarkin et al. 1976 ; Zucker & Morin, 1977) have indicated that the mechanisms controlling the secretion of gonadotrophins may become more sensitive to androgen feedback during exposure to short photoperiods. The present experiments do not support the concept of a similar increase in behavioural sensitivity to the activational effects of testosterone induced by exposure to short photoperiods. On the contrary, the data establish that the regressed gonads are not sufficiently functional to maintain normal copulatory behaviour. This could reflect at least two separate photoperiodic influences : the amount of hormone secreted by the regressed gonad may be below threshold values for activating behaviour or the thresholds of the neural substrates upon which androgens act to activate copulation may have been raised. Exogenous testosterone, in amounts apparently equal to, or in excess of, that endogenously secreted, was more effective in restoring copulation in castrated hamsters experiencing long days than in those exposed to short days. There is evidence for a similar effect of photoperiod on a variety of animals such as the ewe (Raeside & McDonald, 1959; Reardon & Robinson, 1961; Fletcher & Lindsay, 1971), the red deer stag (Lincoln, Guiness & Short, 1972), the female canary (Steel & Hinde, 1972) and the male stickleback (Baggerman, 1966). Some of the possible mechanisms underlying this phenomenon have been discussed by Steel «fe Hinde (1972) and Hutchison (1976) and include the requirement of photoperiodically increased thresholds in the neural substrates for mating behaviour, altered secretion of pituitary hormones which permissively influence the interaction between androgens and their central nervous system (CNS) target tissues, non-copulatory behavioural changes which decrease the probability of mating behaviour (e.g. increased aggression) and altered metabolism of hormones within the CNS reflecting changed enzyme activity. At present it is not possible to choose among these and other plausible alternatives. Nor is it possible to specify why certain animals become behaviourally less sensitive to exogenous steroids during short photoperiods and others do not (e.g. male or female Japanese quail, Adkins & Nock, 1976 ; female hamster, Morin et al. 1977; female lemur, Reynolds & Van Horn, 1977). Long-day castrated hamsters exhibited copulatory performance at week 17 which was superior to that of short-day castrated hamsters, but this difference did not extend beyond that test time. Perhaps refractoriness of the hamster neuroendocrine system begins to dissi¬ pate at this time. Whether this process is spontaneous, as it is in intact hamsters (Turek et al. 1975; Zucker & Morin, 1977) or is accelerated by the testosterone implants is not clear. It does, however, parallel the time course of gonadal recovery in intact hamsters exposed to short days ; after about 20 weeks of short photoperiod the testes spontaneously recrudesce
(Reiter, 1972). In our experiments, implants of testosterone were not completely effective in restoring copulatory behaviour of castrated hamsters to preoperative levels. This phenomenon is reminiscent of the decline in behavioural sensitivity to androgens after castration in barbary doves (Hutchison, 1976). In the latter species, sensitivity of the neural substrate to androgens declines as a function of time after castration and may also be operative in the present case. This explanation could also account for the superior performance of hamsters experiencing
short days and implanted with testosterone pellets at the time of castration as compared with others first given exogenous hormone 12 weeks after the operation. The substantial differences in behavioural responsiveness between animals tested at different times in their circadian cycle raises important methodological considerations. We had initially and naively assumed that valid comparisons could be made between animals housed in the 2L : 22D and 14L : 10D photoperiods as long as tests were performed during the dark portion of the cycle of each animal. This assumption can now be rejected; instead one must equate the subjective times at which tests are made. In animals exposed to photo¬ periods providing 10 or less hours of light/day, subjective night will begin approximately 11-5 h after the onset of light (Elliott, 1974, 1976). A number of our earlier results may have been confounded by failure to compensate for this variable. Accordingly, we are reluctant to decide whether the decline in copulation is really more rapid in castrated hamsters exposed to short days than in those exposed to long days or whether this represents an artifact of testing at inappropriate times of day. Significant day-night differences were obtained in the majority of experiments that examined this phenomenon; copulation was more frequent during subjective-night tests. There are several precedents for such a finding (Beach & Levinson, 1949; Larsson, 1958; Dewsbury, 1968; Richter, 1970). However, in several experiments the expected differences did not materialize, most notably in intact hamsters exposed to the 14L : 10D photoperiod. These animals performed at high, nearly asymptotic, levels during all tests. The basis for this seemingly anomalous finding is unclear. Perhaps day-night differences are seen most clearly under conditions of less than maximum androgenic stimulation (e.g. in castrated hamsters experiencing short or even long days and treated with testosterone) rather than in intact hamsters experiencing long days and stimulated by a variety of endogenous androgens. An alternative explanation may again reflect the methodological difficulties of determining sub¬ jective days and nights when groups are repeatedly tested (and therein photostimulated) at two times a day. Ideally, such tests should be performed according to subjective daytimes measured against an ongoing circadian rhythm such as the hamster locomotor rhythm. This research was supported by grant no. HD-02982 from the National Institute of Child Health and Human Development (NICHD); L.P.M. was supported by a postdoctoral fellowship from NICHD and preparation of the manuscript was supported by NICHD grant no. HD-10740. We are grateful to Bruce Goldman of the University of Connecticut for performing the androgen assays and to Elise Ravel and Darlene Frost for their excellent technical assistance. REFERENCES
Adkins, E. K. (1973). Functional castration of the female Japanese quail. Physiology and Behavior 10, 619— 621.
Adkins, E. K. «fe Nock, B. (1976). Behavioural responses to sex steroids of gonadectomized and sexually regressed quail. Journal of Endocrinology 68, 49-55. Baggerman, B. (1966). On the endocrine control of reproductive behaviour in the male three-spined stickle¬ back (Gasterosteus aculeatus L.). Symposia of the Society for Experimental Biology 20, 427-456. Beach, F. A. «fe Levinson, G. (1949). Diurnal variation in the mating behavior of male rats. Proceedings of the Society for Experimental Biology and Medicine 72, 78-80. Berndtson, W. E. «fe Desjardins, C. (1974). Circulating LH and FSH levels and testicular function in hamsters during light deprivation and subsequent photoperiodic stimulation. Endocrinology 95, 195-205. Czyba, J. C. (1968). Les fluctuations de la fécundité chez le hamster doré (Mesocricetus auratus, Waterhouse) au cours de l'année. Comptes Rendus Hebdomadaires des Séances de la Société de Biologie 162,113-116. Dewsbury, D. A. (1968). Copulatory behavior of rats variations within the dark phase of the diurnal cycle. Communications in Behavioral Biology, Al, 373-377. Elliott, J. A. (1974) Photoperiodic regulation of testis function in the golden hamster : relation to the circadian system. Ph.D. Thesis, University of Texas at Austin. Elliott, J. A. (1976). Circadian rhythms and photoperiodic time measurement in mammals. Federation Proceedings. 35, 2339-2346. -
Fletcher, I. C.
L. P. MORIN
AND
I. ZUCKER
Department of Psychology, University of California, Berkeley, California 94720, U.S.A.
(Received 26 August 1977) SUMMARY
The influence of daylength on copulatory behaviour was assessed by comparing male hamsters exposed to long or short photoperiods (14 or 2 h light/24 h). Copulation declined in animals transferred from long to short days; 13 out of 14 hamsters ceased to ejaculate within 9 weeks and many no longer intromitted in tests with sexually receptive female hamsters. The decline in copulation in hamsters experiencing short days was associated with atrophy of the gonads and flank glands. Behavioural changes in these animals were far more gradual than those observed in hamsters after surgical castration. There was significantly more mating behaviour in tests during the subjective night of the hamsters than during their
subjective day. Exogenous testosterone was more effective in restoring copulation in castrated hamsters exposed to long days than in castrated animals experiencing short days. This suggests that in short days the substrate for copulatory behaviour is relatively refractory to androgens. Photoperiodically mediated changes in behaviour, physiology and morphology may each contribute directly to the reproductive quiescence presumed to occur in the field during the short days of autumn and winter. INTRODUCTION
During the annual reproductive cycle of the hamster, short photoperiods act through the visual system and brain to initiate and maintain testicular regression (Reiter, 1974; Rusak & Morin, 1976). Testicular atrophy is associated with an 80% reduction in the level of testo¬ sterone and cessation of spermatogenesis (Berndtson & Desjardins, 1974). Although changes in the reproductive endocrinology of hamsters exposed to short photoperiods have been extensively recorded, the regulation of sexual behaviour under these conditions has not received experimental attention. Fewer litters of hamsters are delivered during the months of short daylength than at other times of the year (Czyba, 1968; Reiter, 1973-74), but it is un¬ clear whether this represents an effect of reduced gametogenesis or also involves diminished copulatory behaviour. The adult photostimulated male hamster secretes about 5 ng testosterone/ml plasma (Berndtson & Desjardins, 1974). Termination of this androgen secretion by castration results in the cessation of copulatory behaviour; the great majority of hamsters will no longer intromit or ejaculate 3-4 weeks after castration (Whalen & Debold, 1974). Administration of testosterone can restore copulatory behaviour. Exposure to short photoperiods may, in effect, produce a physiological castration by reducing the level of testosterone in the plasma to about 1 ng/ml; this decrease in the secretion of the hormone apparently precedes the * Present address : U.S.A.
Department of Psychology, Dartmouth College, Hanover, New Hampshire 03755,
cessation of spermatogenesis (Berndtson & Desjardins, 1974). The level of hormone from atrophied testes is inadequate to maintain the accessory sex tissues (Reiter, 1969), but its effectiveness in relation to sexual behaviour is unknown. Resolution of this issue is of interest for two reasons. First, if the reduced levels of testosterone in hamsters exposed to short days are incompatible with normal copulation, a behavioural change could contribute to the seasonal termination of reproduction. Secondly, the overwhelming majority of studies have used the unphysiological procedure of castration to assess testicular influences on behaviour. It is appropriate, where possible, to study behavioural changes under physiological con¬ ditions that more nearly approximate those of the animal in the natural state (i.e. photo¬ periodically induced regression). In this case, the pattern of hormone withdrawal and the basal levels of hormones attained are quite different from those observed after surgical removal of the gonads. The present experiments assessed the copulatory behaviour of male hamsters during exposure to short daylengths. Sensitivity to testosterone among hamsters exposed to different photoperiods and tested at different times of day was also determined. MATERIALS AND METHODS
Animals naive male hamsters were Sexually purchased from the Lakeview Hamster Colony, Newfield, New Jersey, U.S.A. Upon arrival in the laboratory, all animals were exposed to a lighting régime of 14 h light (L):10 h darkness (D) (lights on 21.00-11.00 h Pacific Standard Time). Hamsters were housed singly in wire-bottomed cages with access to Simonsen rat pellets (maintenance diet) and water ad libitum. These conditions prevailed for at least 1 month before the screening sessions at which time the hamsters were 3-4 months old. To facilitate selection of experienced male copulators, a series of screening tests was initiated approximately 1 month before any experiment. A male hamster was placed with a receptive female hamster for 5jmin or until an ejaculation occurred. Ovariectomized hamsters bearing a chronic subcutaneous implant of a Silastic capsule containing oestradiol benzoate were rendered sexually receptive by a subcutaneous injection of 1 mg progesterone 2-5 h before the test. Each male hamster received between eight and ten of these sessions. The pool ofanimals showing regular ejaculation during the screening tests was then given five pre¬ liminary sex tests at 2-3 day intervals; these tests were also 5 min long. The latencies to ejaculation and first mount were measured. In general, only animals which ejaculated in four out of the five tests (including the last one) within 5 min of being placed with a receptive female hamster were allocated to the experimental groups. The preliminary sex tests occurred between 13.00 and 15.00 h.
Photoperiods Upon completion of the preliminary sex tests, animals were either retained in 14L : 10D transferred to 2L : 22D. In each case, the room lights went off at 11.00 h.
or
Laparotomies The effect of the photoperiod on gonadal activity was assessed during laparotomy. The left testis was exposed through a mid-line incision rostral to the penis. The index of testicular activity [(length width)/body weight] (Rusak & Morin, 1976), which is highly correlated with spermatogenesis and the production of testosterone (Berndtson & Desjardins, 1974), was
recorded.
Hormone administration Non-esterified testosterone was administered via subcutaneous Silastic implants. Sections of Silastic tubing (Dow-Corning ; internal diameter 1 -93 mm ; external diameter 3-175 mm) were
packed with crystalline testosterone. Each end of a capsule was sealed with silicone adhesive (Dow-Corning Type A) and the effective length of tubing was 50 mm. During implantation, the animals were anaesthetized with sodium pentobarbitone (Nembutal, 80 mg/kg body weight). In several instances, castration was performed at the same time or separately. Serum testosterone analysis Animals were anaesthetized with ether and about 1 ml blood was withdrawn by cardiac puncture. Animals in Groups 3 and 5 were bled after the tests at week 12 (Table 1). After the blood had been allowed to clot overnight in a refrigerator, serum was obtained by centrifu¬ gation and the samples were stored frozen at —20 °C. Radioimmunoassay of androgens in the serum (largely testosterone and dihydrotestosterone) was performed according to the procedures described by Tamarkin, Hutchinson & Goldman (1976). Behavioural tests Male hamsters were pre-adapted to a 20 23 34 cm3 glass aquarium for 3-5 min before the introduction of the receptive female hamsters. Tests were terminated when an ejaculation occurred or after 7 min had elapsed. Generally two animals were tested simultaneously, in separate aquaria, under dim white light (approximately 50 lx at floor level). Hamsters were tested in random order by a technician. The number of mounts and intromissions and the mount and ejaculation latencies were recorded. A composite sex score was also employed in which a score of 1 was assigned to an animal that mounted at least once, 2 if mounting and at least one intromission occurred and 3 if the animal mounted, intromitted and ejaculated. A score of 3 was the maximum possible on any given test. A summary of the experimental groups, photoperiod and other conditions is presented in Table 1. Table 1.
Experimental details and photoperiodic conditions
Group
n*
Photoperiod light (h) : darkness (h)
1
13(11)
2 3 4
12(11) 9 (8) 12
Time of testosterone
14:10
Time of castration Week 1
2:22 2:22 14:10
Weekl Week 1 Week 1
Week 12 Weeks 1 and 12 Weeks 1 and 12
implantation Week 12
Weeks tested
3,t 6,t 9,t 12,t 15,t 16,î 17, 18, 19, 22, 26, 30 As Group 1 As Group 1 3,t 6,t 9,t 12, 16, 17, 19, 22
3,6,13 3,6,9, 12,13§ * Number of animals at the start of the experiment; number of animals tested at week 17 (for Groups 1-3 only) given in parentheses. t Tested between 13.00 and 15.00 h; during weeks without superscripts, tests occurred both at that time and 5 6
7 14
14:10 2:22
—
—
—
—
between 23.00 and 01.00 h. X Tested between 23.00 and 01.00 h. § Tested between 05.00 and 07.00 and 17.00 and 19.00 h. Hamsters were exposed to photoperiods of 14 h light : 10 h darkness (14L : 10D) or 2 h light : 22 h darkness (2L : 22D) and various components of sexual behaviour were tested under various conditions. For animals stably entrained to the 2L : 22D photoperiod, tests between 23.00 and 01.00 and 13.00 and 15.00 h occurred during the subjective night and subjective day of the animal respectively. For animals exposed to the 14L : 10D photoperiod subjective-night tests occurred between 13.00 and 15.00 h and subjtxtive-day tests between 23.00 and 01.00 h. See Materials and Methods for further details.
Subjective-day and subjective-night tests of the experiment it became apparent that the time of testing relative to the photoperiod influenced copulation. In the hamster, the 24 h day has been divided into an approximately 12 h subjective day (inactive phase) and a 12 h subjective night (active phase). The positioning of these two phases depends, among other variables, upon the duration of In the
course
the daily photoperiod. Detailed discussion of this point is beyond the scope of the present article. Nevertheless, it should be recognized that for animals housed in the 2L : 22D photo¬ period, subjective night would normally begin at 20.00 h and end at 08.00 h, for the animals in the 14L : 10D photoperiod subjective night would begin at 10.30 h and end at 22.30 h. These values are calculated from estimates provided by Elliott (1976). Thus behavioural tests conducted between 13.00 and 15.00 h fall during the subjective night and subjective day respectively, for animals in the 14L : 10D and 2L : 22D photoperiods. The converse relation applies to tests conducted between 23.00 and 01.00 h. RESULTS
Effects of castration By week 3 after castration, hamsters no longer displayed the behaviour pattern of ejaculation (Fig. 1, Groups 1 and 2). The frequencies of intromissions and mounts declined to low levels by 9 and 12 weeks respectively, after castration. The decline was more rapid in castrated animals exposed to short days (Group 2) than in those exposed to long days (Group 1 ; P Group 2; P< 0-01 for each behavioural component). The 2 h interval during which the tests took place (13.00-15.00 h) coincided with the subjective night of Group 1 animals and with the subjective day of Group 2 hamsters (see Materials and Methods). When these hamsters were re-tested in week 16 between 23.00 and 01.00 h, the groups did not differ significantly with respect to any of the three parameters recorded. At this point we suspected that merely testing animals during the dark phase of their daily illumination cycle was insufficient: to assess the influence of photoperiod on the hormonal activation of behaviour accurately, it seemed important that all animals be tested at the same subjective time. Accordingly, during week 17 each animal was first tested during its expected subjective day and then 12 h later during its subjective night. This procedure demonstrated that the efficacy of testosterone in restoring copulation was affected by the photoperiod. Intromission and ejaculation com¬ ponents were displayed more frequently (P< 0-005 for each parameter) by hamsters exposed to long days (Group 1) than by those experiencing short days (Group 2) in tests conducted during their respective subjective nights (Fig. 2). The groups did not differ significantly in the frequency of any copulation component during subjective-day tests; the performance of all groups was generally poor. The frequency of ejaculation was higher in intact control hamsters than in either of the groups receiving testosterone replacement therapy and the frequency of intromission was not restored to the levels found before castration in animals in Group 2. The differences detected between Groups 1 and 2 during week 17 were no longer detected during week 19 or in subsequent subjective-night tests. During these tests, per¬ formance of all groups approximately equalled that of Groups 1 and 3 obtained at week 17. In subjective-night tests during week 17, animals in Group 3 (implanted with testosterone capsules at the time of castration ; Table 1) copulated more frequently than hamsters exposed to short days and not implanted with testosterone capsules until week 12 (Group 2). The performance of Group 3 animals was not significantly worse than that of castrated hamsters
experiencing long photoperiods (Group 1, Fig. 2). Day-night differences There was significantly more copulation during the subjective-night than during subjectiveday tests conducted during week 17 (P< 0-025 for each behavioural component, for data from all animals).
0-8r
Ejaculating
Mounting
Intromitting
0-6 0-4 0-2
0: 12
3
12
3
12
3
Group number Fig. 2. Proportion of male hamsters ejaculating, intromitting or mounting in tests during week 17 conducted during the subjective night of each group after exposure to various photo-periods. Animals in Groups 1, 2 and 3 were treated as described in the legend for Fig. 1. All animals were implanted with Silastic capsules containing testosterone at week 12. Diurnal fluctuations in behaviour were further investigated in animals experiencing the 14L : 10D photoperiod (Group 4). These hamsters were castrated and implanted with testosterone capsules during week 1 ; the original capsule was then removed and replaced with a second testosterone implant during week 12 (Table 1). Both subjective-day and sub¬ jective-night tests were given during weeks 12, 16, 17, 19 and 22. The composite sex scores (Fig. 3) were significantly higher during the subjective-night tests. In week 12, the proportion of animals ejaculating, intromitting and mounting during the subjective-day tests were 0-17, 0-42 and 0-58 respectively, compared with a value of 0-92 for each component during the subjective-night tests. When blood was taken at the end of week 12, there was a significant difference in the level of androgens in the serum of the six hamsters of Group 4 bled during the subjective day and the levels in six others bled during the subjective night (9-14 + 0-32 (S.E.M.) and 6-65 + 0-38 ng/ml respectively, < 0-04, Student's i-test).
Fig. 3. Mean ( ± s.e.m.) copulation score for castrated hamsters experiencing 14 h light : 10 h darkness /day. Animals were initially castrated and implanted with testosterone capsules at week 1 and the original capsule was replaced at week 12 with a fresh one. Tests were conducted during the subjective night (·) or the subjective day (O).
During week 17, six hamsters were tested first during their subjective night and then during their subjective day; the remaining six animals were tested in the reverse order. This control procedure revealed that the higher copulation scores associated with subjectivenight tests were not a function of the order of testing. Diurnal differences were also analysed in intact hamsters (not castrated) housed in the 14L : 10D photoperiod throughout the experiment. The composite sex scores for these
during weeks 3, 6 and 13 failed to reveal a significant difference in copulation in subjective-day v. subjective-night tests. No fewer than six out of the seven hamsters ejaculated during each of the day and night tests. animals
Influence ofphotoperiod copulatory behaviour of the intact hamsters gradually declined in frequency with increasing duration of exposure to the short-day photoperiod (Morin, Fitzgerald, Rusak & Zucker, 1977). Animals were tested at a single fixed clock-time (13.00-15.00 h) both before and after the change from the long- to the short-day photoperiod. After we became aware of the significance of the subjective time in relation to copulatory performance, the experiment described by Morin et al. (1977) was repeated with 12 hamsters (Group 6). These animals were transferred from the 14L : 10D to the 2L : 22D photoperiod The
and their behaviour was assessed between 23.00 and 01.00 h as well as between 13.00 and 15.00 h during each test sequence. This procedure greatly increased the likelihood of sampling both subjective-day and subjective-night phases of the daily cycle even while re-entrainment of the circadian rhythms of the animals to the new photoperiod was in progress. Copulation declined during both the expected subjective-day and subjective-night tests (Fig. 4). The percentage of animals ejaculating decreased at a rapid rate between weeks 3 and 9 of exposure to short days (7% of the animals ejaculated during week 9). However, only 50% of the animals ceased to intromit by week 12. Tests conducted during week 13 at intermediate clock-times (05.00-07.00 h and 14.00-19.00 h) confirmed the generally poor copulatory per¬ formances established during tests at the two standard times.
Fig. 4. Mean (±s.e.m.) copulation score for intact male hamsters transferred from 14 h light : 10 h darkness to 2 h light : 22 h darkness at week 1. Tests were made during the expected subjective night (·) and subjective day (O). The testicular index (TI) of these animals assessed at week 13 was 1-09 + 0-97 (s.e.m.). This is indicative of gonads that have not fully regressed (see Table 2, short photoperiod control) and may account for the residual levels of copulation recorded in the tests during week 12. The main finding remains clear, however: short-day photoperiods inhibit copulation by male hamsters. discussion
The copulatory behaviour of many male rodents (Young, 1961), including hamsters (Tiefer, 1970; Whalen & Debold, 1974), is dependent on androgens secreted by the testis. Exposure to short daylengths effectively eliminates the mating behaviour of male hamsters, especially the ejaculatory pattern essential for successful reproduction. This loss of androgen-activated behaviour is probably influenced by the loss of gonadal androgens associated with gonadal atrophy induced by short photoperiods (Berndtson & Desjardins, 1974; Reiter, 1974). A
similar loss of male sexual behaviour associated with gonadal atrophy induced by short-photoperiods has been reported for the Japanese quail, Coturnlx coturnix (Sachs,
1969; Adkins, 1973).
The decline in copulatory behaviour is far more rapid after surgical castration than after exposure of hamsters to short days (Morin et al. 1977). The abrupt withdrawal of androgenic hormones and possibly the trauma associated with surgery may contribute to this difference. Gradual withdrawal of hormones, as occurs during photoperiodically induced testicular regression, is presumably more representative of the natural testicular cycle in the field
(Reiter, 1975) and constitutes a more valid assessment of hormone-behaviour interactions than can be obtained from studies involving surgical castration. Several investigations with male hamsters (Turek, Elliott, Alvis & Menaker, 1975; Tamarkin et al. 1976 ; Zucker & Morin, 1977) have indicated that the mechanisms controlling the secretion of gonadotrophins may become more sensitive to androgen feedback during exposure to short photoperiods. The present experiments do not support the concept of a similar increase in behavioural sensitivity to the activational effects of testosterone induced by exposure to short photoperiods. On the contrary, the data establish that the regressed gonads are not sufficiently functional to maintain normal copulatory behaviour. This could reflect at least two separate photoperiodic influences : the amount of hormone secreted by the regressed gonad may be below threshold values for activating behaviour or the thresholds of the neural substrates upon which androgens act to activate copulation may have been raised. Exogenous testosterone, in amounts apparently equal to, or in excess of, that endogenously secreted, was more effective in restoring copulation in castrated hamsters experiencing long days than in those exposed to short days. There is evidence for a similar effect of photoperiod on a variety of animals such as the ewe (Raeside & McDonald, 1959; Reardon & Robinson, 1961; Fletcher & Lindsay, 1971), the red deer stag (Lincoln, Guiness & Short, 1972), the female canary (Steel & Hinde, 1972) and the male stickleback (Baggerman, 1966). Some of the possible mechanisms underlying this phenomenon have been discussed by Steel «fe Hinde (1972) and Hutchison (1976) and include the requirement of photoperiodically increased thresholds in the neural substrates for mating behaviour, altered secretion of pituitary hormones which permissively influence the interaction between androgens and their central nervous system (CNS) target tissues, non-copulatory behavioural changes which decrease the probability of mating behaviour (e.g. increased aggression) and altered metabolism of hormones within the CNS reflecting changed enzyme activity. At present it is not possible to choose among these and other plausible alternatives. Nor is it possible to specify why certain animals become behaviourally less sensitive to exogenous steroids during short photoperiods and others do not (e.g. male or female Japanese quail, Adkins & Nock, 1976 ; female hamster, Morin et al. 1977; female lemur, Reynolds & Van Horn, 1977). Long-day castrated hamsters exhibited copulatory performance at week 17 which was superior to that of short-day castrated hamsters, but this difference did not extend beyond that test time. Perhaps refractoriness of the hamster neuroendocrine system begins to dissi¬ pate at this time. Whether this process is spontaneous, as it is in intact hamsters (Turek et al. 1975; Zucker & Morin, 1977) or is accelerated by the testosterone implants is not clear. It does, however, parallel the time course of gonadal recovery in intact hamsters exposed to short days ; after about 20 weeks of short photoperiod the testes spontaneously recrudesce
(Reiter, 1972). In our experiments, implants of testosterone were not completely effective in restoring copulatory behaviour of castrated hamsters to preoperative levels. This phenomenon is reminiscent of the decline in behavioural sensitivity to androgens after castration in barbary doves (Hutchison, 1976). In the latter species, sensitivity of the neural substrate to androgens declines as a function of time after castration and may also be operative in the present case. This explanation could also account for the superior performance of hamsters experiencing
short days and implanted with testosterone pellets at the time of castration as compared with others first given exogenous hormone 12 weeks after the operation. The substantial differences in behavioural responsiveness between animals tested at different times in their circadian cycle raises important methodological considerations. We had initially and naively assumed that valid comparisons could be made between animals housed in the 2L : 22D and 14L : 10D photoperiods as long as tests were performed during the dark portion of the cycle of each animal. This assumption can now be rejected; instead one must equate the subjective times at which tests are made. In animals exposed to photo¬ periods providing 10 or less hours of light/day, subjective night will begin approximately 11-5 h after the onset of light (Elliott, 1974, 1976). A number of our earlier results may have been confounded by failure to compensate for this variable. Accordingly, we are reluctant to decide whether the decline in copulation is really more rapid in castrated hamsters exposed to short days than in those exposed to long days or whether this represents an artifact of testing at inappropriate times of day. Significant day-night differences were obtained in the majority of experiments that examined this phenomenon; copulation was more frequent during subjective-night tests. There are several precedents for such a finding (Beach & Levinson, 1949; Larsson, 1958; Dewsbury, 1968; Richter, 1970). However, in several experiments the expected differences did not materialize, most notably in intact hamsters exposed to the 14L : 10D photoperiod. These animals performed at high, nearly asymptotic, levels during all tests. The basis for this seemingly anomalous finding is unclear. Perhaps day-night differences are seen most clearly under conditions of less than maximum androgenic stimulation (e.g. in castrated hamsters experiencing short or even long days and treated with testosterone) rather than in intact hamsters experiencing long days and stimulated by a variety of endogenous androgens. An alternative explanation may again reflect the methodological difficulties of determining sub¬ jective days and nights when groups are repeatedly tested (and therein photostimulated) at two times a day. Ideally, such tests should be performed according to subjective daytimes measured against an ongoing circadian rhythm such as the hamster locomotor rhythm. This research was supported by grant no. HD-02982 from the National Institute of Child Health and Human Development (NICHD); L.P.M. was supported by a postdoctoral fellowship from NICHD and preparation of the manuscript was supported by NICHD grant no. HD-10740. We are grateful to Bruce Goldman of the University of Connecticut for performing the androgen assays and to Elise Ravel and Darlene Frost for their excellent technical assistance. REFERENCES
Adkins, E. K. (1973). Functional castration of the female Japanese quail. Physiology and Behavior 10, 619— 621.
Adkins, E. K. «fe Nock, B. (1976). Behavioural responses to sex steroids of gonadectomized and sexually regressed quail. Journal of Endocrinology 68, 49-55. Baggerman, B. (1966). On the endocrine control of reproductive behaviour in the male three-spined stickle¬ back (Gasterosteus aculeatus L.). Symposia of the Society for Experimental Biology 20, 427-456. Beach, F. A. «fe Levinson, G. (1949). Diurnal variation in the mating behavior of male rats. Proceedings of the Society for Experimental Biology and Medicine 72, 78-80. Berndtson, W. E. «fe Desjardins, C. (1974). Circulating LH and FSH levels and testicular function in hamsters during light deprivation and subsequent photoperiodic stimulation. Endocrinology 95, 195-205. Czyba, J. C. (1968). Les fluctuations de la fécundité chez le hamster doré (Mesocricetus auratus, Waterhouse) au cours de l'année. Comptes Rendus Hebdomadaires des Séances de la Société de Biologie 162,113-116. Dewsbury, D. A. (1968). Copulatory behavior of rats variations within the dark phase of the diurnal cycle. Communications in Behavioral Biology, Al, 373-377. Elliott, J. A. (1974) Photoperiodic regulation of testis function in the golden hamster : relation to the circadian system. Ph.D. Thesis, University of Texas at Austin. Elliott, J. A. (1976). Circadian rhythms and photoperiodic time measurement in mammals. Federation Proceedings. 35, 2339-2346. -
Fletcher, I. C.
