Physiology & Behavior 89 (2006) 133 – 138 Effects of prenatal stress on sexual partner preference in mice Leslie R. Meek a,⁎, Kalynn M. Schulz b , Courtney A. Keith c a Division of Social Sciences, University of Minnesota, Morris, 600 E. 4th St., Morris, MN 56267, United States b Michigan State University, E. Lansing, MI, 48824, United States c University of South Dakota, Medical School, Vermillion, SD, 57069, United States Received 23 March 2005; received in revised form 28 April 2006; accepted 1 May 2006 Abstract Three-month old, male Swiss Webster mice were born to either control dams or dams who had been prenatally stressed with light, heat, noise and handling during the last week of gestation. As adults, male offspring were tested on sexual partner preference and sexual behavior (mounting, intromissions and lordosis) with a sexually experienced male stimulus animal and a stimulus estrous female. In comparison to males born to control dams, prenatally stressed males showed a sexual partner preference for the sexually active male as demonstrated by a negative partner preference score, more and longer visits to the male's compartment, fewer and shorter visits to the female's compartment and longer latencies to and lower frequencies of mounts and intromissions of females. In addition, stressed males showed a greater frequency of lordosis and a higher lordosis quotient than did control males. This study is the first to investigate the effects of prenatal stress alone, without hormonal manipulation, on sexual partner preference using both a partner preference paradigm and measures of sexual behavior such as mounting, intromissions and lordosis. These findings support the suggestion that prenatal stress alone is enough to significantly affect sexual partner preference in male mice. © 2006 Elsevier Inc. All rights reserved. Keywords: Mice; Sexual partner preference; Sexual orientation; Homosexuality; Prenatal stress It has been unequivocally demonstrated that prenatal stress during the last week of gestation in rats results in marked deficits in adult male mating behavior, with males showing fewer attempts to copulate, fewer successful ejaculations and longer latencies to mounts, intromissions and ejaculations (e.g., [1–3]), without accompanying alterations in sexual organs [4]. It was subsequently determined that stress during gestation shifts the typically high concentration of testosterone released on gestational day 17 to gestational days 18 and 19 [5]. This shift in the timing of the peak of testosterone is positively correlated with decreased steroid aromatase activity in the brain during gestational days 17–21 [6] and a markedly smaller adult SDN-POA [7,8]. The SDN-POA of adult male rats is typically significantly larger than that of females [9,10]. In sexually naïve male rats, bilateral lesions of the SDN-POA have been shown to ⁎ Corresponding author. Tel.: +1 320 589 6213; fax: +1 320 589 6117. E-mail address: meeklesr@morris.umn.edu (L.R. Meek). 0031-9384/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.physbeh.2006.05.006 decrease the number of animals ejaculating and to increase latencies to mounts, intromissions and ejaculations [11]. However, others have found no relationship between the size of the SDN-POA and adult ejaculatory performance [8]. Instead, it is possible that the volume of the SDN-POA is related to the probability that an adult male rat will demonstrate lordosis when mounted. Thus, a smaller nucleus predicts the increased likelihood of lordosis when a male is mounted by another male [8,12,13]. The conclusion that has been drawn from the above evidence is that by shifting the timing of the peak of testosterone, a critical period of brain sexual differentiation is missed so that brain nuclei involved in the expression of male sexual behavior are demasculinized, resulting in a decrease in the expression of male-typical sexual behaviors as an adult. Concomitantly, there is a failure to defeminize central mechanisms, resulting in an increase in the probability of the expression of female-typical sexual behaviors in the adult male (e.g., [1,4,14]). Since these discoveries have been made, many have questioned whether the shift in the timing of the gestational 134 L.R. Meek et al. / Physiology & Behavior 89 (2006) 133–138 testosterone surge and the resultant effects on central control mechanisms might not also affect sexual partner preference in adult male rodents. Studies have indicated that in a sexual partner preference test, adult males, even with no prior sexual experience, show a strong preference for an estrous female over a non-estrous female or an intact, adult male [15–17]. Studies using a partner preference paradigm, where animals have a choice between an intact, sexually experienced male and an intact (or hormonally induced) estrous female have shown that male rats who have been treated pre- and/or post-natally with anti-estrogens or anti-androgens or who receive a bilateral lesion of the MPOA–AH changed their partner preference from that of an estrous female to an intact, sexually experienced male [18–23] lending support to this suggestion. To date, while a number of studies have investigated the effects of pre- and/or post-natal anti-androgens and antiestrogens on sexual partner preference (e.g., [21,22,24,25] there have been no investigations of the effects of prenatal stress on sexual partner preference in gonadally intact male rodents without manipulating hormones during development or adulthood. In addition, unlike other partner preference studies, this study evaluated direct measures of sexual behavior such as mounting, intromissions and lordosis along with standard partner preference measures. 1. Methods 1.1. Animals Three-month old, male, Swiss Webster mice were used in this experiment. All animals were bred at the University of Minnesota, Morris, from breeding stock obtained from Taconic Farms, Inc. Food (Purina Mouse Chow # 5015, Ralston-Purina Co., St. Louis, MO) and water was available ad lib. All animals were maintained in a reversed light–dark cycle of 14 h light and 10 h dark, with a temperature of 20° ± 2 °C. Animals were cared for according to the standards published by the University of Minnesota Animal Care and Use Committee and in accordance with federal regulations as detailed in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. 1.2. Procedure 1.2.1. Mating and stress Swiss Webster dams were paired with sexually experienced males for 1 week. During this time, animals were videotaped to determine the date and time that mating occurred. On day 12 after mating, dams were randomly assigned to either a stressed (n = 20) or a non-stressed (control; n = 20) group (gestation length is 19 d in Swiss Webster mice). Dams were singly housed in 45 × 26.5 × 20 cm opaque polycarbonate cages. Stressed dams were exposed to stress 3 times daily for 45 min each time during the last week of pregnancy. All stress occurred during the light portion of the light–dark cycle. Stress occurred at random times so that dams could not predict when stress would occur [26]. A record of when stress occurred each day was kept to ensure that stress did not occur at the same time each day and that it occurred roughly equally across the 14 h light period. Stress sessions occurred a minimum of 2 h apart. Stress consisted of handling (being picked up by the tail), noise (being moved to a different room on a noisy cart), light (180.5 L/ft2) and increased temperatures (34° ± 2 °C). All stress procedures occurred in the same room. Food and water were not available during stress. Non-stressed dams remained in their cages in a separate room while stressed animals were stressed in another, sound-proofed room. Stressed and non-stressed animals were housed in separate rooms. Following birth, pups were housed with their birth-dams and littermates in 45 × 26.5 × 20 cm cages until weaning at 21 d of age. They were then singly housed in 27 × 17 × 13 cm cages until testing at the age of 3 months. All animals were sexually naïve at the time of testing. One pup from each litter was used in sexual partner preference testing (n = 20 in each group) and one pup from each litter was used in lordosis testing to eliminate litter effects (n = 20 in each group; 27). 1.2.2. Sexual partner preference testing For 4 d prior to sexual partner preference testing, males were habituated to the testing arena for 30 min each day. During habituation, the same stimulus animals that were used during testing for each individual experimental animal were present (a 3-month old sexually intact, experienced male and a 3-month old sexually experienced behaviorally estrous female) but were confined to wire-mesh cages on opposite sides of the arena. Their position in the testing chamber was randomly alternated during each habituation session. On the fifth day, the experimental animal was simultaneously exposed to the two stimulus animals, each of which was tethered at opposite ends of the arena. The tether consisted of a plastic tie (Radio Shack) as a collar with a fishing swivel attached. A 36 cm length of flexible cord was attached by one end to the fishing swivel and by the other end to the top of the aquarium. This apparatus allowed the stimulus animals to move freely and engage in sexual behavior, but confined them within a small portion of their compartment. The arena consisted of a 75 × 30 × 28 cm glass aquarium that was visually divided into 3 equal areas with tape. The test male was placed into the middle compartment and separated from the stimulus male and estrous female by partitions for a 15-minute period. Once the partitions were removed, a 60-minute preference test ensued. All animals received exposure to one of 20 stimulus males and one of 20 stimulus females. Each stimulus animal was used in a test with a stressed male and a non-stressed male. All habituation and testing occurred at random times during the dark portion of the reversed day–night cycle and under red light. Each test was videotaped with a Panasonic Low-Light Videorecorder and no experimenters were present during testing. Behavioral estrous in the female stimulus animal was determined by daily vaginal smears, stained with thionin and examined microscopically to determine whether a smear was classified as proestrous. Smears were rated as proestrous if most of the cells were nucleated. Only females rated as being in proestrous were used as stimulus females. L.R. Meek et al. / Physiology & Behavior 89 (2006) 133–138 Table 1 Partner preference variables Stressed versus non-stressed males Behavior Frequency in female compartment Frequency in male compartment Time spent in female's compartment (sec) Time spent in male's compartment (sec) Preference score Latency to mount female (sec) Frequency of mounting female Latency to intromissions with female (sec) Frequency of intromissions Stressed males Non-stressed males 64.45 ± 4.0 131.8 ± 5.35 153.2 ± 3.68 81.65 ± 4.22 1535.85 ± 62.88 2692.4 ± 66.75 1843.45 ± 68.07 596.0 ± 31.27 − 307.01 ± 128.36 +2096.4 ± 95.41 3163 + 128.77 2165.0 + 83.40 0.66 + 0.54 3449.84 + 58.56 1.0 + 0.35 23.38 + 1.60 2175.47 + 47.87 15.55 + 1.19 t statistic t = 9.95, P < 0.0001 t = 15.42, P < 0.0001 t = 13.15, P < 0.0001 t = 18.75, P < 0.0001 t = 16.25, P < 0.0001 t = 5.92, P < 0.0001 t = 11.78, P < 0.0001 t = 11.81, P < 0.0001 t = 10.74, P < 0.0001 All values expressed as mean ± standard error of the mean. P ≤ 0.05 considered significant. There were 20 animals in each group that were tested on each dependent variable. The following behaviors were scored from the tape for sexual partner preference testing: 1. Frequency and duration of visits to the female's or male's compartment: A visit was scored as occurring when the test male's entire body (excluding tail) had crossed the tape marker and was inside the stimulus animal's compartment. 2. Latency to first mount and frequency of mounts by the experimental animal: A mount of a stimulus animal was recorded when the test male oriented himself on top of the caudal part of the stimulus animal's body with his hind feet positioned on the flanks of the stimulus animal. 3. Latency to first intromission and frequency of intromissions: An intromission was scored if the test male inserted his penis into the vagina of the stimulus female or attempted to insert his penis into the stimulus male. 4. Preference score: Time spent in the compartment containing the sexually active male was subtracted from the time spent in the compartment containing the estrous female; a positive score indicates a preference for the estrous female; a negative score indicates a preference for the sexually active male [18,22,28]. 1.2.3. Lordosis testing Forty, 3-month old, Swiss Webster mice (one from each litter; not the same animals as used in sexual partner preference testing) were tested for lordosis behavior for 1 h or until 10 mounts had occurred, whichever occurred first. Frequency of lordosis and lordosis quotient (number of lordosis responses / number of mounts × 100) were both measured; 29). Each animal 135 was habituated to the testing arena as above and then placed in the testing arena with an untethered, sexually active male. Lordosis was defined as the experimental animal demonstrating a concave back, tail deviation and neck extension in response to being mounted by the stimulus animal [29]. 1.3. Statistics Student's t-tests were used for all comparisons between stressed and control animals. Variability was expressed as standard error of the mean and P ≤ 0.05 was considered significant. 2. Results 2.1. Sexual partner preference Stressed males made significantly fewer visits to the female's compartment (t = 9.95, P < 0.0001) and significantly more visits to the male's compartment (t = 15.42, P < 0.0001) than did control males. Stressed males also spent significantly less time in the female's compartment (t = 13.15, P < 0.0001) and significantly more time in the male's compartment (t = 18.75, P < 0.0001) than did control males. Stressed males had a significantly lower mean preference score than did control males. See Table 1. Stressed males showed a significantly longer mean latency to mount the estrous female (t = 5.92, P < 0.0001) and showed significantly fewer mounts than control males (t = 11.78, P < 0.0001). Similarly, stressed males had a significantly longer mean latency to intromissions with the female (t = 11.80, P < 0.0001) and showed significantly fewer mean intromissions than did control males (t = 10.74, P < 0.0001). See Table 1. 2.2. Lordosis frequency and quotient Stressed males showed significantly more mean incidences of lordosis than did control males (t = 4.50, P = 0.0002) when paired with an experienced male. See Fig. 1. For those animals Fig. 1. Mean frequency of lordosis (±SEM) in stressed and non-stressed males. ⁎ indicates significantly different from non-stressed animals, P < 0.05. 136 L.R. Meek et al. / Physiology & Behavior 89 (2006) 133–138 Fig. 2. Mean lordosis quotient (±SEM) of stressed and non-stressed males. ⁎ indicates significantly different from non-stressed animals, P < 0.05. that received 10 mounts, the lordosis quotient for stressed males was significantly higher than for control males (t = 5.16, P < 0.0001). See Fig. 2. Thirteen of twenty stressed males showed lordosis when paired with another male, while only 3 of 20 control males did. 3. Discussion Stressed males behaved significantly different than control males on several definitive measures of sexual partner preference. For example, when paired with a sexually active female, stressed males demonstrated longer latencies to mounts and intromissions and lower frequencies of both behaviors than control males. When paired with a sexually active male, 65% of stressed males exhibited lordosis, in contrast to 15% of control males. Stressed males who received 10 mounts had a lordosis quotient of 52.3% in comparison to that of 12.8% for control males. This sexual partner preference by stressed males for a sexually active male partner was corroborated and supported by their negative partner preference score (indicating a male partner preference) and the finding that they visited the male's compartment more frequently and spent more time there than did non-stressed control males. They also made fewer visits to the female's compartment and spent less time there than did control males. Overall, the above pattern of behavior on the part of prenatally stressed male mice shows a marked shift in sexual partner preference from estrous females to sexually active males. This pattern of behavior is astonishing in Swiss Webster mice, a highly aggressive species that is normally characterized by intense inter-male aggression. In this study, inter-male aggression was absent and appeared to be replaced by sexual interest and behavior. In addition to sexual interest, partner preference can also be motivated by social attraction, curiosity or aggression on the part of the experimental animal [30]. At least some of the behavior that is expressed when an animal meets an unfamiliar conspecific can be assumed to be motivated by curiosity. However, we believe that our lengthy habituation period before testing should have attenuated much of that curiosity, so that true sexual partner preferences were expressed by our animals. In addition, since Swiss Webster mice, unlike rats, are not a social species, few if any of the interactions between animals should be mediated by social attraction. Swiss Webster mice, in particular the males, are quite aggressive when they meet, and social interaction between males would normally consist of brief risk assessment behavior followed by fierce aggression. During our partner preference tests, however, we saw no aggressive interactions, so aggression can also be discounted as a motive for approaching another animal, particularly the male stimulus animal. Vasey [31] notes that in order to determine true partner preference, behaviors measured as dependent variables must be sexual, at least in part. One way of demonstrating that nonheterosexual behaviors are sexual is to show that they are similar to male–female sexual behaviors [31]. For this reason, we included sexual behaviors such as mounting, intromissions and lordosis in our study in addition to relying on partner preference score to indicate choice of sexual partner. When partner preference data is combined with data indicating that stressed males are mounted by non-stressed stimulus males and show lordosis in response, a stronger argument for same-sex sexual partner preference can be made. Previous studies have suggested that inexperienced males show no preference for an estrous female over another stimulus animal such as a sexually active male [32,33], but others have found that even inexperienced males will show a preference for an estrous female [15–17]. One author has suggested that habituation to the testing arena prior to a sexual partner preference test allows an inexperienced animal to express its sexual attraction to an estrous female [17]. Since our study allowed for adequate habituation to the testing arena and to the stimulus animals, our sexually inexperienced animals can be assumed to have expressed their real sexual partner preferences. Stressed males almost never mounted stimulus males or females. In contrast, as evidenced by the lordosis data, they were mounted by the stimulus males. This suggests that stressed males were not perceived by the stimulus males as a typical male. When male Swiss Webster mice meet, aggressive fighting usually occurs quickly. In contrast, in this study, stimulus males treated the stressed males as if they were receptive females and mounted them. It would be interesting to determine whether stressed males were behaviorally mimicking estrous female posture in any way, thus disarming the aggression response and triggering sexual behavior on the part of the stimulus male. Previous studies have demonstrated that even without the use of exogenous hormones during adulthood, some male rats will demonstrate lordosis as adults in response to pairing with another male or manual stimulation by an experimenter [34,35]. In our study, 15% of control males displayed lordosis, indicating that some inexperienced male mice that have not been exposed to any manipulation will demonstrate female-typical sexual behavior as adults and show lordosis when paired with a sexually active adult male. It is possible that this is a facultative response, seen in some young, inexperienced males in response to being paired with another L.R. Meek et al. / Physiology & Behavior 89 (2006) 133–138 male. This response may occur as part of a constellation of submissive behaviors that forestalls aggressive interactions and functions to protect the submissive mouse from injury or death. Thus, it appears as though there may be two functions of male– male mounting, one of which occurs to forestall aggression and one that actually indicates male–male sexual interest, as in our study. If so, this suggests that each may arise from different developmental events, but result in the same behavioral endpoint. While the majority of stressed males exhibited lordosis in response to being paired with a sexually active male, none of them displayed female-typical proceptive behaviors such as dart-hopping or ear-wiggling. It is possible that different aspects of female-typical mating behavior differentiate during different critical perinatal periods, and had we disrupted testosterone beyond the day of birth, at least some males would have exhibited the full range of female mating behavior, as has been demonstrated in rats (e.g., [4,22]). The evidence of the partner preference data plus the sexual behavior data indicates that 3-month old prenatally stressed males showed a marked sexual partner preference for same-sex partners. However, it remains possible that stress merely delayed the development of the expression of full-blown male sexual behavior. Had we tested animals at an older age, this sexual preference may have been attenuated or reversed. Further research needs to be done to determine whether age is a factor in the expression of sexual partner preference following prenatal stress. Thus, future studies could be conducted in which adult stressed males could be tested for sexual behavior with non-stressed males and females separately to determine the pattern of sexual behavior that occurs with each stimulus animal. It would also be interesting to place two prenatally-stressed males together in an arena to determine the pattern of behavior they might express towards one another. It seems unlikely that they would express either aggressive or sexual behavior towards one another, since neither will be perceived by the other to be a sexually active heterosexual male. However, it would be interesting to observe what behaviors they do engage in and compare that to intact males interacting, intact females interacting and a male–female dyad interacting. This study supports the previous findings in rats that prenatal stress alone can result in deficits in male sexual behavior and the expression of female sexual behavior in adult rodents (e.g., [1– 5,7]) and extends those findings to show that prenatal stress alone can also alter sexual partner preference in mice. In addition, since the litter effect was controlled for in this study, it also disputes the previous suggestion that prenatal stress effects are mainly due to a litter effect [27]. In summary, this study suggests that prenatal stress alone, without post-natal stress or hormonal manipulations can significantly change sexual partner preference in mice, so that males prefer sexually active males to an estrous female. These findings have implications for sexual orientation studies in non-humans. A number of studies have documented that some species (e.g., rams) typically have a small number of males who are unequivocally uninterested in females and 137 show male-typical behavior directed towards other males [36]. This study suggests that maternal stress might mediate such sexual partner preferences. Vasey [31] has suggested that trying to fit non-human male–male preferences into an adaptationist framework is not feasible for every species, and that for some species, it is reasonable to suppose that a shift in sexual partner preference is maladaptive or functionally neutral, while for others it may actually serve an adaptive function. Such determinations must be made based on the evolutionary history and current ecology of each individual species [31]. Thus far, human studies have not been able to document any association between prenatal stress and human male homosexuality (e.g., [37–40]), suggesting that rodent and human responses to prenatal stress are not similar. Further investigations may determine whether the rodent response to prenatal stress is functional or maladaptive, although at least one recent study suggests that this response is adaptive in that it prepares offspring for the environmental and social conditions they will encounter as adults [41]. References [1] Ward IL. Prenatal stress feminizes and demasculinizes the behavior of males. Science 1972;175:82–4. [2] Crump CJ, Chevins PFD. Prenatal stress reduces fertility of male offspring in mice, without affecting their adult testosterone levels. Horm Behav 1989;23:333–43. [3] Gerardin DCC, Pereira OCM, Kempinas WG, Florio JC, Moreira EG, Bernardi MM. Sexual behavior, neuroendocrine, and neurochemical aspects in male rats exposed prenatally to stress. Physiol Behav 2005; 84:97–104. [4] Ward IL. Exogenous androgen activates female sexual behavior in noncopulating prenatally stressed males. J Comp Physiol Psychol 1977;91:465–71. [5] Ward IL, Weisz J. Maternal stress alters plasma testosterone in fetal males. Science 1980;207:328–9. [6] Weisz J, Brown BL, Ward IL. Maternal stress decreases steroid aromatase activity in brains of male and female rat fetuses. Neuroendocrinology 1982;35:374–9. [7] Anderson DK, Rhees RW, Fleming D. Effects of prenatal stress on differentiation of the sexually dimorphic nucleus of the preoptic areas (SDN-POA) of the rat brain. Brain Res 1985;332:113–8. [8] Kerchner M, Ward IL. SDN-POA volume in male rats is decreased by prenatal stress, but is not related to ejaculatory behavior. Brain Res 1992;581:244–51. [9] Gorski RA, Gordon JH, Shryne JE, Southam AM. Evidence for a morphological sex difference within the medial preoptic area of the rat brain. Brain Res 1978;148:333–46. [10] Gorski RA, Harlan RE, Jacobson CD, Shryne JE, Southam AM. Evidence for the existence of a sexually dimorphic nucleus in the preoptic area of the rat. J Comp Neurol 1980;193:529–39. [11] De Jonge FH, Louwerse AL, Ooms MP, Evers P, Endert E, Van De Poll NE. Lesions of the SDN-POA inhibit sexual behavior of male Wistar rats. Brain Res Bull 1989;23:483–92. [12] Van de Poll NE, Van Dis H. The effect of medial preoptic-anterior hypothalamic lesions on bisexual behavior of the male rats. Brain Res Bull 1979;4(4):505–11. [13] Hennessey AC, Wallen K, Edwards DA. Preoptic lesions increase the display of lordosis by male rats. Brain Res 1986;370:21–8. [14] Meisel RL, Dohanich GP, Ward IL. Effects of prenatal stress on avoidance acquisition, open-field performance and lordotic behavior in male rats. Physiol Behav 1979;22:527–30. [15] Kagan J. Differential reward value of incomplete and complete sexual behavior. J Comp Physiol Psychol 1955;48:59–64. 138 L.R. Meek et al. / Physiology & Behavior 89 (2006) 133–138 [16] Hetta J, Meyerson BJ. Sexual motivation in the male rat. A methodological study of sex-specific orientation and the effects of gonadal hormones. Acta Physiol Scand Suppl 1978;453:1–67. [17] Agmo A. Unconditioned sexual incentive motivation in the male Norway rat (Rattus norvegicus). J Comp Psychol 2003;117(1):3–14. [18] Paredes RG, Tzschentke T, Nakach N. Lesions of the medial preoptic areas/anterior hypothalamus (MPOA/AH) modify partner preference in male rats. Brain Res 1988;813:1–8. [19] Matuszcyzk JV, Larsson K. Sexual preference and feminine and masculine sexual behavior of male rats prenatally exposed to antiandrogen or antiestrogen. Horm Behav 1995;29:191–206. [20] Houtsmuller EJ, Brand T, De Jonge FH, Joosten RNJMA, van de Poll NE, Slob AK. SDN_POA volume, sexual behavior, and partner preference of male rats affected by prenatal treatment with ATD. Physiol Behav 1994;56 (3):535–41. [21] Brand T, Kroonen J, Mos J, Slob AK. Adult partner preference and sexual behavior of male rats affected by prenatal endocrine manipulations. Horm Behav 1991;25:323–41. [22] Bakker J, Brand T, van Ophemert J, Slob AK. Hormonal regulation of adult partner preference behavior in neonatally ATD-treated male rats. Behav Neurosci 1993;107:480–7. [23] Bakker J, Ophemert JV, Slob AK. Postweaning housing conditions and partner preference and sexual behavior of neonatally ATD-treated male rats. Psychoneuroendocrinology 1995;20(3):229–310. [24] Johnson WA, Tiefer L. Sexual preferences in neonatally castrated male golden hamsters. Physiol Behav 1971;9:213–7. [25] Vega Matuszczyk J, Fernandez-Guasti A, Larsson K. Sexual orientation, proceptivity and receptivity in the male rat as a function of neonatal hormonal manipulation. Horm Behav 1988;22:362–78. [26] Fride E, Weinstock M. The effects of prenatal exposure to predictable or unpredictable stress on early development in the rat. Dev Psychobiol 1984;17:651–60. [27] Chapman RH, Stern JM. Failure of severe maternal stress or ACTH during pregnancy to affect emotionality of male rat. Dev Psychobiol 1979;12 (3):255–67. [28] Edwards DA, Pfeifle JK. Hormonal control of receptivity, proceptivity and sexual motivation. Physiol Behav 1978;30:437–43. [29] Carter CS. Neuroendocrinology of sexual behavior in the female. In: Becker JB, Breedlove SM, editors. Behavioral endocrinology. Cambridge (MA): MIT Press; 1992. p. 74. [30] Meisel RL, Sachs BD. The physiology of male reproduction. In: Knobil E, Neil JD, editors. The physiology of reproduction, vol. 2. New York: Raven Press; 1988. p. 3–105. [31] Vasey P. Same-sex sexual partner preference in hormonally and neurologically unmanipulated animals. Annu Rev Sex Res 2002;13: 141–79. [32] Matuszczyk JV, K. Larsson. Experience modulates the influence of gonadal hormones on sexual orientation of male rats. Physiol Behav 1994;55(3):527–31. [33] Vega Matuszczyk J, Appa RS, Larsson K. Postnatal testicular secretions partially restore the disturbances in reproductive activity caused by prenatal hormone manipulation. Physiol Behav 1997;62(1):137–42. [34] Dahlof L-G, Hard E, Larsson K. Influence of maternal stress on offspring sexual behaviour. Anim Behav 1977;25:958–63. [35] Sodersten P, Larsson K. Lordosis behavior in castrated male rats treated with estradiol benzoate or testosterone propionate in combination with an estrogen antagonist, MER-25, and in intact male rats. Horm Behav 1974;5:13–8. [36] Perkins A, Fitzgerald JA. Sexual orientation in domestic rams: some biological and social correlates. In: L. Ellis, L. Ebertz, eds. Sexual orientation: toward biological understanding. Westport, CT: Praeger Publishers/Greenwood Publishing Group, Inc.: pp 107–128. [37] Siegelman M. Parental backgrounds of homosexual and heterosexual men: a cross national replication. Arch Sex Behav 1981;10(6):505–13. [38] Ellis L, Ames MA, Peckham W, Burke D. Sexual orientation of human offspring may be altered by severe maternal stress during pregnancy. J Sex Res 1988;25(1):152–7. [39] Bailey JM, Willerman L, Parks C. A test of the maternal stress theory of human male homosexuality. Arch Sex Behav 1991;20(3):277–93. [40] Hines M, Johnston KJ, Golombok S, Rust J, Stevens M, Golding J, ALSPAC Study Team. Prenatal stress and gender role behavior in girls and boys: a longitudinal, population study. Horm Behav 2002;42:126–34. [41] Kaiser S, Sacher N. The effects of prenatal social stress on behaviour: mechanisms and function. Neurosci Biobehav Rev 2005;29(2):283–94.