ABSTRACT THE EFFECTS OF EARLY CASTRATION AND TESTOSTERONE PROPIONATE TREATMENT ON THE DEVELOPMENT AND DISPLAY OF BEHAVIOR PATTERNS BY MALE RHEUSS MONKEYS (MACACA MULATTA) By Craig F. Bielert The aim of this thesis was an evaluation of the role gonadal hormones play in the development and display of social and sexual behavior patterns by male rhesus monkeys. Consideration was also given to the role social factors may play in moderating the frequency of certain hormone stimulated behavior patterns. The development of male rhesus monkeys castrated on either day of birth or at three months of age, was compared to the development of similarly reared intact males. Dur- ing the first two years of life, castrates resembled the intact males in their display of dimorphic behavior patterns which significantly distinguish intact males from females. During the first two years of life there was no evidence of defic- ient masculine behavior development on the part of the cas- trate males. In order to assess the effects of increased postpubertal gonadal hormone levels comparisons were made between castrates and intact males for the first four years of life. Two behavior patterns, erection and yawning mar- kedly increased postpubertally for intact males; similar increases were not observed for castrates. Intact and castrate males did not differ in the measure of brief contact and rough and tumble play suggesting that the decline in this behavior occurs independently of the gonadal events of puberty. Both castrates and intact males exhibited mounts, intromissions, and ejaculatory responses. Gonadal status appeared to affect the level of aggression exhibited during periods of social instability with intact males -exhibiting higher levels at such times than the castrates. In order to assess the activational potential of the major male gonadal hormone, testosterone, testosterone prOpionate injections were given to several groups of castrate male rhesus of two age classes: 1) fourteen and oneAhalf months; 2) fifty months. The effects of androgen were assessed under social as well as in pair test situations. Results of the testosterone treatment period were compared to the pre- hormone test period. For the social group testing castrates differed from intact males in erections, yawning, groom solicitations, and genital examinations received. No increas- es in mounting or aggression were obtained. The group of older castrates increased in the measures of yawn, erection, and auto-genital manipulation after testosterone treatment. Though the older castrates exhibited higher prehormone treatment levels of mounting and aggression than the younger castrates, these measures failed to increase during the testosterone treatment period. During pair testing young castrates increased their levels of approach, contact, and mount while under testosterone treatment. These measures differed significantly between castrates and intact males during the TP treatment period, but not the prehormone treatment period. Two young castrates exhibited intromission while receiving testosterone therapy and one of these two also displayed ejaculation. The old castrates also increased their sexual behavior scores, but varied individually. In another experiment the mating behavior of adult intact and castrate males which were raised with peers were compared. These two groups were also compared to a group of adult intact wild-caught males. Laboratory-reared males had a lower percentage of responders for several measures of sexual behavior. The castrate group had the smallest number of responders. Contact and mount frequencies of the intact laboratory—reared males were significantly lower than those of the wild-caught males. Female stimulus partners behaved differently with the three groups of males and measures of proximity and spontaneous were at low frequencies during tests with castrates. A dominance paradigm was followed in order to categorize the pair tests as one in which the male was dominant to the female or one in which the female was dominant to the male. The dominance interactions between the pair were shown to affect the level. with which certain male and female sexual behavior patterns were shown. Testosterone injections modified the dominance relations between the adult ' castrate males and their stimulus female partners. THE EFFECTS OF EARLY CASTRATION AND TESTOSTERONE PROPIONATE TREATMENT ON THE DEVELOPMENT AND DISPLAY OF BEHAVIOR PATTERNS BY MALE RHESUS MONKEYS (MACACA MULATTA) By Craig F. Bielert A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 197A "For you have learned. One school is finished, and the time has come for another to begin." Richard Bach, Jonathan Livingston Seagull ii ACKNOWLEDGMENTS Though scientific inquiry can be viewed as the employ- ment of the scientific method in the investigation of a specific problem, the progress of science is dependent upon much more than mere inquiry. To the extent that the results of this thesis may aid tothe'body of basic knowledge which human- ity draws upon for scientific progress grateful thanks and credit are extended to the following individuals and groups of individuals. To Drs. Lynwood Clemens and Robert Goy appreciation is extended for their encouragement, criticism, acceptance and support. To Mike Jaeger, John DWyer, Archie Vomachka, Linda Coniglio, John Czaja, Jeff Slimp, Jane Cords, and Mary Collins special thanks are offered for sincere interpersonal inter- actions. They have all shared in the occassional pleasure and pain which my day to day existence has revolved upon and they have all made important contributions to the several years of a human lifetime that I have spent in pursuit of my academic goals. I have no regrets about the time I have spent and I only hape that I have been able to add some dimension to the lives of those with whom I have worked. I have been lucky for such friends, it is these peOple who have ensured that my education has remained the part of living that I honestly believe it should be. ' I would be remiss if I failed to mention that numerous others along the course of my thesis project have made many tasks possible which otherwise would have remained unwhaldly. iii Staffs at Michigan State University as well as the Wisconsin Regional Primate Research Center have always been coOperative and helpful. Financial support for my time in graduate studies has come from the following sources: General Medical Science Grant l-TOl-GMO 1751 to the Department of Zoology, Michigan State University and Grant RR-00167 to the Wisconsin Regional Primate Research Center both from the National Institutes of Health and Grant MH-21312 to Dr. Robert Goy from the National Institute of Mental Health. iv 3. 4. TABLE OF CONTENTS Introduction Longitudinal comparisons of the social and sexual development of prepuberally cas- trated and intact laborato reared male rhesus (Macaca mulattg¥ 2.1 Introduction 2.2 Methods 2.3 Results 2.4 Discussion Testosterone treatments of castrate male rhesus of varying ages in group and pair testing situations 3.1 Introduction 3.2 Methods for Part I: Testosterone stimulation and group testing 3.3 Results for Part I 3.4 Methods for Part II: Testosterone stimulation and pair testing 3.5 Results for Part II 3.6 Discussion Sexual behavior of adult laboratory reared male rhesus monkeys: Comparisons with wild-caught adult males and laboratoryéreared prepuberally castrated adult males Page 23 23 24 31 6O 76 76 80 90 111 114 117 127 Contents (Continued) Page 4.1 Introduction 127 4.2 Methods 129 4.3 Results 135 4.4 Discussion ‘ 168 5. Summary and Conclusions 179 Bibliography 183 vi Table Table Table Table Table Table Table Table Table 2.1 2.2 2.3 2.4 3.1 3.2 3-3 3-4 3.5 LIST OF TABLES Social peer groups of castrate and in- tact male rhesus Results for the three way analysis of variance for repeated measures compar- ing males, castrate males, and females across their first three social test- ing runs Results for the three way analysis of variance for repeated measures comp paring prepuberally castrated and intact males across their first five social testing runs Results for the three way analysis of variance for repeated measures compar- ing males, castrate males and females across their first five social test- ing runs Social peer groups of castrate male rhesus Penis measurements for young male rhesus prior to and during an ex- perimental treatment period Behaviors of young male rhesus prior to and during an experimental treatment period. Behaviors of adult neonatally cas- trated male rhesus prior to and during a testosterone propionate therapy period Behaviors for young male, young castrate male, and adult neonatal- ly castrated male rhesus during a fifty-day test period vii Page 28 33 34 36 81-82 94 97 99 105 [I‘ll [It‘ll 9% .ll’l‘llf a 'Irl‘l. [1 . [I‘ll [fir List of tables (Continued) Table Table Table Table Table Table Table 3.6 3.7 3.8 4.1 4.2 4.3 4-4 Page Behavior scores for two castrate rhesus prior to and following a dominance shift during testosterone treatment 108 Behavior scores for young castrate male rhesus prior to, during, and after tes- tosterone prOpionate treatment 110 Sex behavior scores for young male and young castrate male rhesus prior to and during experimental treatment 116 Sex behavior scores for tests of castrate peer-reared, intact peer-reared, and intact wild-caught adult male rhesus monkeys 138 Results of the statistical analyses of male sexual behavior during pair tests for castrate peer-reared, intact peer- reared, and intact wild-caught adult male rhesus monkeys 140 Results of the statistical analyses of female sexual behavior during pair tests with castrate peer-reared, intact peer-reared, and intact wild-caught adult male rhesus monkeys ' 142 Trends of change shown during a second testing with the same group of stimulus females by castrate peer- reared, intact peer-reared, and intact wild-caught adult male rhesus monkeys 164 viii I .llf'11 fivl>l|lllll ‘rlllr.l\nl|||\’l\.lfjlllllllll ul- inlII‘tli 1 Il‘.( .\{E{xri"l}l‘l [iv 1.0!I I [if Figure Figure Figure Figure Figure Figure Figure Figure Figure 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 LIST OF FIGURES Page Dimorphic behavior patterns for male, castrate male and female rhesus dur- ing their first three social testing runs 39 Aggression by male, castrate male, and female rhesus for their first three social testing runs 43 Mounts by male, castrate male, and female rhesus for their first three social testing runs 46 Brief contact play and rough and tumble play by prepuberally castrated and intact male rhesus for their first five social testing runs 48 Aggression by prepuberally castrated and intact male rhesus for their first five social testing runs 50 Erections by prepuberally castrated and intact male rhesus for their first five social testing runs 52 Mounting behavior by prepuberally castrated and intact male rhesus monkeys for their first five social testing runs 55 Measures of grooming behavior for male, castrate male, and female rhesus for their first five social testing runs 57 Yawning by male, castrate male, and female rhesus for their first five social testing runs 59 ix List of figures (Continued) Page Figure 2JL> Aggression shown by intact peer-reared and intact socialized peer-reared males during four social testing runs 74 Figure 3.1 Castrate S-55 exhibiting the prone body display described as groom solicit while under testosterone prOpionate therapy 89 Figure 3.2 Castrate 5-55 exhibiting the seated body presentation described as groom solicit while under testosterone prOpionate therapy 89 Figure 3.3 Weight gains by young male rhesus monkeys prior to, during, and following an experimental treatment period 92 Figure 3.4 Castrate male 4874 showing autogenital man. ipulation while under testosterone propionate therapy 102 Figure 4.1 Percentages of castrate peer-reared intact peer-reared, and intact wild-caught males exhibiting various measures of male sexual behavior 147. Figure 4.2 Percentages of males and females show- ing agonistic behavior during their sex tests. 150 Figure 4.3 Dominance relationships between a pair for castrate peer-reared, intact peer-reared, and intact wild-caught male rhesus monkeys 153 Figure 4.4 Dominance relationship between a pair and its effects on male sexual behav- ior measures 155 Figure 4.5 Dominance relationship between a pair and its effects on female behavior measures 158 List of figures (Continued) Figure 4.6 Figure 4.7 Page Dominance relationship between a pair and its effects on yawning and grooming 161 Percentage of tests dominant for castrate peer-reared, intact peer- reared and intact wild-caught males on repeated testings with the same female partners 167 xi INTRODUCTION Many nonhuman primate Species exhibit physiological, social and behavioral similarities with man. These similari— ties make certain primates appropriate for use as models in the examination of human biomedical problems. One area of research in which the rhesus monkey (Macaca mulatta) has been extensively employed is that of sexual behavior. The rhesus appears sensitive to many of the same general social and endocrine factors that affect humans. Experimental manipulation of humans is impossible, and for this reason clinical cases and observations on unusual syndromes become an important source of basic information from which rele- vant hypotheses are generated. A critical examination of hypotheses relevant to a particular clinical condition does require experimentation and it is in this regard that data from nonhuman primate experiments take on a very important role. These data allow one to assess the relative merits of a particular hypothesis as generated by work with the mon- key and as it may be pertinent to the human. The concern of this thesis was the fashion in which gonadal hormones affect theubehavior of male rhesus monkeys. It is hOped that the projects carried out with these monkeys may aid in the understanding of the fashion in which gonadal hormones influence human behavior. It is with this in mind that a review will now be given to the experimental data which are already in the literature and may be relevant to a consider- 1 2 ation of the ways in which gonadal hormones affect behavior in a wide variety of species. For several years data have accumulated which are con- sistent with the hypothesis that hormones of testicular ori— gin can permanently alter mammalian behavior patterns if present at the appropriate period during embryological devel- opment. In 1959 workers from the laboratory of Dr. W.C. Young published a paper dealing with the organizing action of prenatally administered testosterone propionate on the tissues which mediate mating behavior in the female guinea pig (Phoenix, Goy, Gerall and Young, 1959). It was in this paper that attention was first drawn to the effects which early gonadal hormone exposure might have on a sexually dimorphic behavior, although the possibility of behavioral modification had been suggested by earlier work (Dantachakoff, 1938a, b, c; Raynaud, 1938; Wilson, Young, and Hamilton, 1940). Follows ing publication of the work on the guinea pig similar effects in behavioral modification were reported for the rat (Harris and Levine, 1965). Experiments assessing the effects of hormonal manipulation on the mating behavior of the genetic female have also been carried out with the mouse (Edwards and Burge, 1971), the Syrian golden hamster (Swanson and Crossley, 1971), and the dog (Beach and Kuehn, 1970; Beach, Kuehn, Sprague, and Anisko, 1972). Many excellent reviews have come out on the general t0pic area and they offer a breadth of interpretation (Valenstein, 1968; Whalen, 1968; Beach, 1971; 3 Brain, 1971; Gorski, 1971; Money and Ehrhardt, 1971; Davidson and Levine, 1972; Money and Ehrhardt, 1972a; Goy and Goldfoot, 1973; Steinbeck and Neumann, 1973). In general, the hypothesis as it currently stands is that fetal gonadal hormones act as inducers if present during embryological develOpment. In other words, nervous system tissue is permanently modifiable if appropriate hormonal exposure occurs during the period of neural organization. The period of neural organization and hormone sensitivity is neonatal in some species such as the hamster (Eaton, 1970; Carter, Clemens, and Hoekema, 1972), while for the rat it appears to include both uterine and postparturient periods (Gerall and Ward, 1966; Gerall, Hendricks, Johnson, and Bounds, 1967; Stern, 1969). For the guinea pig and rhesus monkey this sensitive period occurs during the time period inbutgrg (Goy, Bridson and Young, 1964; Resko, 1970). The genetic sex of the individual appears irrelevant to behavioral modifica- tion; neural tissues of females as well as males appear equal- ly modifiable, although genetic sex may yet be shown to exert some type of limitation in this regard. For the species studied, it is only the genetic male which is exposed to a high concentration of gonadal hormone during early development; the fetal ovaries show little secretory activity (Resko, Feder, and Coy, 1968; Resko, 1970; Resko, Malley, Begley, and Hess, 1973). There has however, been the suggestion that the ovary may serve in a potentiating capacity for the female (Gerall, 4 1973; Gerall, Dunlap and Hendricks, 1973), but this potentia- tion phenomenon doesn't appear to involve any central nervous system alteration. Experimental manipulations of gonadal hormones have been determined by the sex of the individual animal as well as by the nature of the species typical gestation for that particu- lar animal. Genetic females can be exposed to hormones by means of direct injection, hormone implantation, or indirect- ly while 22:22252 through injections of their mothers. The genetic male's hormonal environment can also be manipulated at several levels (Jost, 1971/72). One experimental approach is through interference with the glandular processes of hor- mone synthesis and release. A second and perhaps the most common manipulative technique is by castration. A third level at which hormone action can be manipulated in the genetic male is through interference with hormone receptors, hormone conversion, and gross cellular activity. It should be rem- embered that no matter what the sex of the individual or the manipulation employed the effects are to a great degree de- pendent upon ongoing developmental processes. Not all behav- ioral patterns exhibit the same period of sensitivity (Coniglio, Paup, and Clemens, 1973). During a particular behavioral pattern's period of sensitivity small manipulations of the hormonal environment can result in radical modification, however by starting at a later period somewhat similar chang- es may be induced by increasing the hormone do se level (Goy et al., 1964). Demonstrations of the limits of this physiological sensitivity have frequently utilized the rat. ‘Ward (1972) has shown that by stressing pregnant female rats a decrease in mounting by the male offspring can be detected In addition, Clemens and Coniglio (1971) have shown that the number of males present $£;2£2£g and their relative proximity to their female siblings appear related to the frequency of mounting eventually shown by these females. Partly because of the unidirectional nature of deve10pmental processes, there are time periods past which no amount of hormonal exposure will result in behavioral modification. Hormones during postnatal periods have what has been de- scribed as an activational action on behavior patterns, and the activational action of gonadal hormones on mating as well as social behaviors has been well reviewed (Guhl , 1961; Yuan; 1961; Hart, 1974). WOrk on the hormonal activation of sex behavior has dealt with the problem through two experimen- tal approaches: a) deveIOpmental studies of mating behavior; b) manipulations of gonadal condition in animals which have already exhibited complete mating behavior. For both males and females the elimination of gonadal hormones will result in the deterioration and decline of the complete mating pattern. Males as well as females gonadectomized prepuberally may also fail to deveIOp the complete mating pattern. Complete mating behavior can be restored in gonadectomized males and females of numerous species, through the administration of hormones of the appropriate variety and concentration. The hormone itself, however, remains passive in terms of its affect on mating be— 6 havior; hormones don't have the capacity to alter the Specific pattern shown or increase the maximal response rate though their level may affect certain measures of response within narrow limits (Grunt and Young, 1953; Peretz, 1968). It appears that it is experiential (Lehrman, 1962:)1971) as well as genetic factors (Valenstein, Riss and Young, 1954; Goy and Young, 1957; Goy and Jakway, 1962) which are critical in determining the behavioral responses which hormonal stimulation may elicit. It is largely as a result of the work which has gone on with rodent species that it has come to be accepted that gonad- al hormones can affect behavior patterns in two different and distinct fashions. These two actions have been categorized as ”organization” and "activational"; and can be described and distinguished as follows. 1. The organizational action of gonadal hormones for those mammalian species in which it has been investigated is of pro- tracted duration and limited to fetal and larval stages. Effects of hormonal action of this type are persistent and may or may not require adult sex hormone levels for their exhi- bition. 2. The activational action of gonadal hormones may be limited in one direction by an age boundary. Effects of hor- monal action of this type are transient or tonic depending upon the presence of the apprOpriate hormonal substrate and its level. Though the preponderance of research dealing with the effects of gonadal hormone on sexual behavior has utilized non- primate species, there are data on the rhesus monkey (Macaca mulatta) in this regard as well as clinical literature on the human. A recent review (Luttge, 1971) dealt very broadly with the tOpic for both males and females. The remainder of this literature review will focus on the male and the female pseudohermaphrodite and the affects which gonadal hormones have on their behavior. The psychological literature contains numerous references to the importance which genetic sex plays in determining be- havioral response. Dr. Harry Harlow and his associates at the University of Wisconsin were the first to demonstrate in a conclusive fashion that for the rhesus monkey these behavior- al differences can emerge in a situation devoid of maternal intervention. Dr. Leonard Rosenblum (1961) published work on young surrogate reared rhesus monkeys which were allowed only peer contact showing that under these conditions young male and female rhesus exhibited dimorphic patterns of play, sex and aggression. Werk carried out by Drs. William Young, Robert Goy and Charles Phoenix at the Oregon Regional Primate Research Center corroborated these findings. These researchers have enumerated five separate patterns of social interaction in which, (in terms of frequency of occurrence), juvenile males differ significantly from juvenile females. The patterns have been descriptively named: a) threat b) play initiation c) rough and tumble play; d) pursuit play; and e) mounting behavior. Under the peer rearing conditions employed, males differed 8 quantitatively from females at levels of from 5x to 10x de- pending upon the behavior pattern examined (Young, Goy and Phoenix, 1964; Phoenix, Goy and Resko, 1968; Coy, 1968, 1970; Goy andPhoenix, 1971; Goy and Resko, 1972). A sixth be- gavior pattern, that of presenting, was not found to be di- morphic under these rearing conditions. This is an interest- ing observation in light of the fact that presenting is the posture which is most analogous to that of the lordotic posture exhibited by receptive female rodents. In rodents, this be- havioral pattern is exhibited dimorphically and its display may be affected by early hormone exposure. In more recent work carried out at the Wisconsin Regional Primate Research Center by Dr. Goy and his colleagues, a rear- ing condition has been utilized in which adult females are housed with their infants in small social groups. Sex dimorphisms con- sistent with those previously observed have again been documented (Goy and Goldfoot, 1974). It may, in fact, be the case that un- der this social rearing condition, situations are such that they maximize the reinforcement which mothers may provide for those behavior patterns which are biologically determined. It is known that mother-infant interaction is important to the rhesus (Harlow and Harlow, 1969) and others have documented marked differences in mother-infant interaction.which are related to the sex of the infant for the pigtail macaque (Jensen, Bobbitt, and Gordon, 1968). The biological nature of sex differences in behavior for the rhesus has received further support from the work of Dr. Gene Sackett who has examined isolate reared rhesus monkeys and has found sex to be an important variable for a wide variety of measures including locomotor behavior, explor- ation time, and self-directed behavior (Sackett, 1972). The literature on humans also contains documentation of sex differences in motor and facial patterns for infant play (Brindley, Clarke, Hutt, Robinson, and Wethli, 1973). These differences appear to be culturally modifiable but hold up to the cross cultural test of biological validity (Blurton—Jones and Konner, 1973). The sex differences between young boys and girls are not limited to measures of play alone (Hutt, 1972). Differences exist in the tendency of a child to leave its mother, the frequency of returns to mother, and the interval of these returns. Genetic sex also affects the choice of play objects. In a study of exploratory behavior carried out by Hutt (1970) a novel "play object" was presented to young nursery school children. Fewer girls than boys approached the object and the treatment of the object also differed for the sexes. Young girls tended to treat the play object in a very conventional and utilitarian fashion, while young boys tended to be creative and innovative in terms of their employment of the object in their play interaction. The theory of the organizational action of gonadal hormones on behavior has been recently extended to the rhesus. Drs. Goy and Phoenix have demonstrated that it is possible through early androgen exposure to modify the exhibition of dimorphic behavior patterns by the genetic female rhesus monkey. Preg- 1O nant female rhesus were given daily injections of testosterone propionate from about day 40 to day 110 of gestation. The injection dosages varied as did the duration of treatment; complete details on the manipulations can be found in Goy (1970). The genetic female infants of mothers so treated exhibited marked anatomical masculinization at the time of birth. This masculinization was in good accord with that which Wells and Van'Wagenen (1954) had obtained. The young pseudohermaph- rodites had well deveIOped penes and scrotum with no vestiges of vaginae. The young pseudohermpahrodites also had secondary male sex organs as well as functional ovaries. Menarche for these females was retarded from that of normal females, how- ever (Goy, 1970; Goy and Phoenix, 1971; Goy and Resko, 1972). Once menarche had occurred, ovulatory ovarian cycles were docu— mented (Goy and Resko, 1972). The behavioral develOpment of these pseudohermaphrodites was characteristic of that of the genetic male. In all behavior categories for which sex differ- ences had been described the pseudohermaphrodites failed to differ from the males. Furthermore, even in the mother-infant rearing condition, a social setting under which rapid develOp- ment of mounting behavior occurs in genetic males, the pseudo- hermaphrodites followed the deve10pmental path which was typi- cal of genetic males raised under this condition. It appears clear from this work that the hypothesis of hormonal organiza- tion, which was formulated from research on rodent mating be- havior, is validly extendable to the primate and to the rich social repertoire of these animals. 11 Additional support for the extension of the original theory comes from human clinical literature (as reviewed by Money and Ehrhardt, 1972b). There is evidence from cases of progestin- induced hermaphroditism as well as from cases of the andreno- genital syndrome which suggest that early exposure to andro- genic substances will modify both the anatomy and the behavior of genetic female humans (Ehrhardt and Meney, 1967; Ehrhardt, Evers, and Money, 1968). It should be added that the modifi- cations of this hormonal exposure in the human are consistent with those of the rhesus. These individuals possess a penis and scrotum as well as functional ovaries. Further as was the case for the rhesus, there is evidence of delayed menarche for these girls (Jones and Verkauf, 1971). It has been reported that the young girls which received such early androgen exposure exhibited marked tomboyism in their play patterns. This has been suggested to be analogous to the rough and tumble play pattern which male and pseudohermaphrodite rhesus monkeys exhibit. These young girls also exhibited toy preferences which were of a more masculine character. In marked contrast to these propensities, young males afflicted with the syndrome of androgen insensitivity show very little masculine behavior although in rare cases some degree of anatomical masculini- zation may have occurred (Money, 1972, 1973; Money and Ehrhardt, 1972b). It appears that data from both the rhesus monkey and the human are consistent with the hypothesis that early hormone action in the primate can be characterized as organizational 12 in regard to its effects on behavior patterns which can normally be classed as dimorphic. Rodent sexual behavior, though fre— quently employed as an endpoint in examinations of the or- ganizational action of gonadal hormones on behavior fails in these Species to appear until puberty and cannot really be considered hormone-independent, may not be the best test of the theory on the organization action of gonadal hormones on behavioral patterns peg gs. In addition to the work on the rhesus monkey work with the effects of early gonadal hormones on the urinary postures of the dog has added convincing evi- dence that hormone independent behavior patterns can be permanently altered (Beach, 1970, 1974; Sprague and Anisko, 1973). The best evidence for the theory of hormonal organiza- tion of behavior, however still remains the experimental work on the rhesus for the rich social repertoire of this species offers an unrivalled potential. Though early reports exist on the effects of castration and subsequent androgen therapy on the sexual behavior of nonhuman primates (Thorek, 1924; Zuckerman and Parkes, 1938, 1939; Clark, 1945), these reports are far from complete or systematic. Field reports following the effects of prepu- beral castration in the rhesus also exist (Sade, 1968; Wilson and Vessey, 1968; Loy, 1971), but an assessment of the reported decline in mating behavior is difficult to make. In recent years, however, the effects of castration in the adult male rhesus monkey have been systematically examined under labora- tory conditions (Michael, 1972; Michael, Wilson and Plant, 1973? Phoenix, Slob, and Goy, 1973; Michael and Wilson, 1974). 13 Following castration, peripheral androgen levels drop signif- icantly within hours (Resko and Phoenix, 1972; Michael 22 al., 1973). The sexual behavior of the castrated animals shows a decline over time, but the rate of decline seems highly in- dividualistic. In addition to sexual behavior, a marked de— cline in frequency of occurrence is observed for yawning and spontaneous erection. It appears that the rate of decline for intromission and mounting may diverge markedly in some animals with a significant decrease in mounting rate occurring only over a very extended period. Treatments with testosterone propionate have been shown to be effective in the restoration of precastration levels of sexual behavior (Michael, 1972; Michael gt_al,, 1973; Phoenix, 1973; Phoenix.gt.al,, 1973; Michael and Wilson, 1974). Testosterone was equally effective in the restoration of yawning (Phoenix, 1973; Phoenix 23.31., 1973). a behavior pattern which has been suggested to be androgen stimulated (Goy and Resko, 1972). Dihydrotestosterone propionate has also been shown to be capable of restoring sexual behavior as well as yawning (Phoenix, 1973. 1974). It appears that age may affect the decline in sexual behavior as well as the effec- tiveness of testosterone as a replacement therapy (Michael, 1972; Michael e£_§l,, 1973). The results which have been reported for the castrate talapoin are consistent with those which have been reported for the rhesus (Dixson, Everitt, Herbert, Rugman, and Scruton, 1973). The castrate talapoin study is interesting in that it clearly demonstrated that social 14 position can modify the expression of androgen-stimulated sexual behavior. The activational potential which testosterone has for male primate sexual behavior is certainly clear from the data pre- viously mentioned for the rhesus. An interesting sidelight in this.regard comes from work on ovariectomized female pseudohermaphrodites. Recently it was demonstrated that adult pseudohermaphroditic rhesus which had been ovariecto— mized and given testosterone propionate treatments can exhibit the complete male sexual pattern, including ejacula- tion (Eaton, Goy and Phoenix, 1973). This observation ex- tends the data which had already been obtained from work on rats and guinea pigs (Goy, Phoenix and Meidinger, 1967; Whalen and Robertson, 1968) and demonstrates the level of sexual behavior modification which can be‘achieved through the organizational action of gonadal hormones. The human clinical literature on the effects of castra- tion was reviewed by Tauber (1940). By and large the data from the human is of the case observation variety (Daniels and Tauber, 1941; Miller, Hubert, and Hamilton, 1938; Feinier and Rothman, 1937), and often includes observations on individuals afflicted with syndromes such as Klinefelter's syndrome (Beumont, Bancroft, Beardwood, and Russell, 1972). Even with the paucity of human data on the role of gonadal hormones in sexual behavior expression the human appears consistent with the rhesus in many regards. Follow- ing castration peripheral testosterone values in most cases 15 have declined rapidly by ten days post operation (Sciarra, Sorcini, DiSilverio, and Gagliardi, 1973). Spontaneous erections for castrates become infrequent for both man (Tauber, 1940) and the rhesus (Michael and Wilson, 1974) though they do still occur. In this regard, it appears that sexual per- formance declines even though some capability for erection is maintained. Miller 33.31. (1938) reported that their subjects described intercourse as "arduous work" partially with reference to the difficulty they encountered in main- taining an erection. Michael and Wilson (1974) have suggested that erectile capability may be a critical factor in the main- tenance of rhesus sexual behavior. Though the individual differences in level of sex behavior and the occurrence of spontaneous erection were not thought to be related to adrenal androgens (Hamilton, 1943) recent work utilizing radioimmuno- assay has suggested that the adrenal may be responsible for low levels of testosterone peripherally in at least some cases (Sciarra gt.gl., 1973). Work employing the antiandrogen cypro- terone acetate has also resulted in data consistent with the proposition that gonadal hormones are critical to the full expression of primate sexual behavior (Michael, Plant, and Wilson, 1973; Mbrse, Leach, Rawley, and Heller, 1973). Cyproterone acetate treatments result in depressed peripheral testosterone levels and concurrent decreases in sexual behav- ior. Cessation of drug treatment is followed by recovery of sexual drive as well as resumption of normal testicular function and testosterone titer. There is some reason to 16 believe that for the monkey complete behavioral restoration may not occur (Michael, Plant and Wilson, 1973) but revers- ability for the human has been documented (COOper, Ismail, Phanjoo, and Love, 1972). Testosterone propionate appears capable of restoring erection capability (Hamilton, 1937a, b; Miller 32 al., 1938) and has been said to increase libido (Daniels and Tauber, 1941; Beumont 22 al., 1972). An interesting observation in this regard is that certain cases of human impotence are accompanied by significantly lower urinary testosterone levels than controls, suggesting decreased production as a possible factor for the condition (COOper, Ismail, Smith and Loraine, 1970). It would appear that testosterone is in some way necessary for the sexual behavior of the male primate. One effect of testosterone may be the maintenance of penile integrity, since it has recently been demonstrated that penile feedback of some variety is necessary for the maintenance of the complete mating pattern in the male rhesus (Herbert, 1973). It might be important to comment in this regard that the penile sensory systems for both the male rhesus and the human have certain anatomical similarities such as the presence of complicated claw-like nerve endings and a high incidence of typical genital corpuscles (Malinovsky and Sommerova, 1972 a, b). There are, however, certain data which are inconsistent with a functionally significant role for testosterone and male sexual behavior. There is a report for humans which suggests that even prepuberal levels of testosterone are sufficient for an active sex life (Raboch and Starka, 1973). 17 It also appears clear that testosterone level pgr‘ge and sexual vigor are not closely related. Failure had been re- ported for clinical improvement of chronically impotent men despite a doubling of endogenous testosterone levels through the use of clomiphene (COOper, Ismail, Harding and Love, 1972). In a full year study of sexual behavior and testos- terone level for a group of eight rhesus monkeys, Plant (1974) failed to find any consistent relation between hormone level and sex performance though inconsistencies were reported. As previously mentioned, castrate humans have difficulty in maintaining an erection. Erections do, however, occur (Hamilton, 1943). Human infants are also capable of spon- taneous erection (Halverson, 1940; Korner, 1969). In the case of infants this behavior may be at least indirectly related to the testosterone levels during this period (Forest, Cathiard, and Bertrand, 1973). It is well documented for the rhesus (Hines, 1942; Boy, 1966; Phoenix, Goy and Young, 1967; Goy, 1968) as well as for some other primate species (Cercgpithecus aethiOps, Struhsaker, 1967; Macaca radiata, Rahaman and Parthasarathy, 1969; Macaca fuscata, Hanby and Brown, 1974; 222$g_gynocephalus, Anthoney , 1968; Pan troglodytes, Van Lawick-Goodall, 1968) that young males exhibit mounting attempts, pelvic thrusting and erections. The occurrence of intromissions has been re- ported for juvenile chimpanzees (van Lawick-Goodall, 1968) as well as for baboons (Hall and DeVore, 1965; Hammer, 1968). Documentation of the testosterone levels for animals of these 18 ages for species other than theztesus are lacking. In the case of the rhesus, peripheral testosterone values don't reach an appreciable level until the age of about 2 yrs. and 10 mos. (Resko, 1967). Puberty, as measured by mating tests and testicular biOpsies, has been reported to occur in laboratory male rhesus at ages of 2 yrs. 11 mos. (van Wagenan and Simpson, 1954) to 3 yrs. 6 mos. (Hartman, 1932). Re- ports on free ranging male rhesus suggest that puberty has taken place by 3 yrs., 6 mos. (Conaway and Sade, 1965). The data are consistent with the hypothesis that the high andro- gen levels which accompany puberty aren't necessary for the develOpment of apprOpriate patterns of mature mounting in the primate. Further evidence of a minimal role for gonadal hormones in early male sexual development comes from data on the deve10pment of young castrate male rhesus. These young monkeys were frequently observed to deveIOp erections as well as to exhibit mounting at levels equivalent to those of similar- ly aged intact males (Goy, 1966, 1968; Phoenix, et al., 1967); thoweven:n:reports have appeared on the occurrence of intro- missions or the ejaculatory pattern for prepuberally castrated rhesus. In fact, certain reports from the literature suggest a critical role for the pubertal production of androgen in this regard. Loy and Loy (1974) in a recent paper on two year-old male rhesus stated, "In no instance was a series mounting observed to end in the motor pattern characteristic of ejaculation (Loy, 1971) or in the female clutching reac- tion (Zumpe and Michael, 1968). No female was seen with a 19 vaginal plug of coagulated semen. These results were not surprising, since Conaway and Sade (1965) reported that free- ranging rhesus males on Cayo Santiago do not reach maturity until about three and one-half years of age." One recent study has followed the deveIOpment of the intromission pattern and ejaculation and correlated the onset of these behaviors with physiological changes characteristic of puberty (Michael and Wilson, 1973). This study might be taken as suggestive evidence that pubertal hormone levels are necessary for the occurrence of the intromission and ejaculatory responses characteristic of adult sexual behavior (Bielert and Goy, 1973). The activational action of gonadal hormones has also been examined through the manipulation of sexually immature animals. These attempts have been carried out in order to assess the role which chronological age may play in determining the effects of hormonal action on behavior. In this regard, the experimental induction of precocious sexual behavior through exogenous testosterone propionate treatments has been achieved in the rat and the guinea pig. Beach (1942), following up some work done by Stone (1940), was able to stimulate complete male sexual behavior from his experimental animals thirty days in advance of his control animals. Beach stated that although it was generally accepted that c0pulatory behavior would not be shown before puberty, thebehavioral system for c0pulation was well organized and capable of functioning before this time. Further, the gonadal hormone wasn't an active organizer, it merely lowered a threshold for the behavior pattern. Pre- 20 cocious sexual behavior in the guinea pig was elicited through exogenous TP treatment but the difference between the experi- mental and controls was only 3-6 days (Riss, Valenstein, Sinks and Young, 1955; Gerall, 1958; 1963). A clinical condition known as "pubertus praecox" exists in the human and has profound consequences. For example, "we find in the records that Salamis, the son of Euthmenes, grew to 4'6" in his third year; he walked slowly and had a bass voice, and was carried off by a sudden attack of paralysis when he turned three." (Pliny, Natural History). Experimenta- tion on this condition has been done and numerous studies are available on the neuroendocrine basis for precocious ovula- tion and spermatogenesis. An adequate review of the area is Donovan and van der'Werff ten Bosch (1965). The psychologi- cal ramifications of precocious puberty are little known, al- though Money and c0dworkers have reported data on the psycho- sexual state of some two dozen cases of idiOpathic sexual precocity (Money and Alexander, 1969; Mbney and Walker, 1971). This work has suggested that the psychosexual develOpment of boys that were studied was clearly not correlated with the precocious appearance of the gonadal hormone. The body be- came sexually mature first and the intellectual and psycho- logical maturity followed. The boys exhibited a tendency to high I.Q. early occurrence of the capacity for frankly sexual imagery in dreams and daydreams, and early establishment of the capacity for erotic and sexual arousal in relation to visual imagery and visual perception as well as tactile sen- 21 sation. At present, no study has attempted through hormonal manipulation to induce sexual behavior in prepuberal rhesus monkeys. The activational action which gonadal hormones might have on social behaviors other thanthose related to sex has been a tOpic of considerable research effort. Field reports have documented increases in aggressive interaction for free-rang- ing rhesus during the mating season (Kaufman, 1965; Southwick, Bed, and Siddiqi, 1965; Vandenbergh and Vessey, 1968). The suggestion has been made that since recrudescence of the testis has occurred by this time (Sade, 1964), these increases in aggression may be androgen-related (Wilson and Boelkins, 1970). Relevant to these observations are the reports that individual male's peripheral androgen levels are correlated with their social positions as well as with their levels of aggression (Rose, Holaday, and Bernstein, 1971), and that in humans testosterone level may correlate with aggression (Persky, Smith, and Basu, 1971). Further work has suggested that social interactions connected with defeat may depress peripheral androgen levels (Rose, Gordon, and Bernstein, 1972). Numerous laboratory studies (Clark and.Birch, 1945; Birch and.Clark, 1946; Mirsky, 1955; Kling, 1968; Green, Whalen, Rutley, and Battie, 1972; Joslyn, 1973) have attempt- ed to answer questions about how androgens might affect social interactions in group situations. By and large, these attempts have been unsuccessful in altering social interactions (Mirsky 22 1955; Kling, 1968; Green g£_al,, 1972) but a recent study has reported social change when young females were given daily androgen injections (Joslyn, 1973). In the light of the preceding literature review, it seems clear that the rhesus monkey is an ideal animal with which to pursue certain basic questions about gonadal hormones and their effects on male primate behavior patterns. The thesis project which was carried out attempted to gain basic data on the following questions: 1) How does neonatal castration affect the deveIOpment of social and sexual behavior patterns in the male rhesus? 2) Can the action of testosterone be considered acti- vational in terms of its effects on the social and sexual behavior patterns of the male rhesus? 3) To what extent does an individual's social position as well as his experiential state determine the behavioral effects of testosterone stimulation? This thesis will present the data from three seperate experimental evaluations of behavior by castrate and intact male rhesus monkeys. These evaluations will be presented as sections 2, 3, and 4. These sections will be presented as discrete units with their own introductions, results, and discussions. It is hoped that each of these sections can‘be pub- lished in a seperate paper sometime in the future. Section 2. Longitudinal comparisons of the social and sexual development of prepuberally castrated and intact laboratory-reared male rhesus (Macaca mulatta). 2.1 INTRODUCTION The rhesus monkey (Macaca mulatta) is a comparatively long— lived animal and shares with many primates, including man, a period of extended social development preceding adolescence. The contributions gonadal hormones may make to male primate be- havior patterns are most frequently associated with the be- havioral changes of puberty, but early hormone exposure also has been shown to have profound consequences (Goy and Phoenix, 1971; Goy and Resko, 1972; Money, 1972, 1973). In a recent review of the literature on the role of gonadal hormones in the sexual behavior of the rhesus and the human (Luttge, 1971), it was stated that, ". . . the effects of castration in both species should be further examined [p. 83]." Since then, several papers have come out on the role of gonadal hormones in adult male rhesus sexual behavior (Michael, 1972; Resko and Phoenix, 1972; Michael, Wilson, and Plant, 1973; Phoenix, 1973; Phoenix, Slob, and Goy, 1973; Michael and Wilson, 1974). Reports also exist on the effects of prepuberal castration in the rhesus (Goy and Dodsworth, 1962; Wilson and Vessey, 1968; Phoenix 33 al., 1967; Goy, 1968; Phoenix, 1973). but no study has yet attempted to deal in a longitudinal fashion with the 1 effects which prepuberal castration may have on the rhesus. 23 24 A longitudinal study which extends through puberty allows the assessment of both the "organizational" and "activational" actions of the gonadal hormones (Beach, 1970; Goy and Goldfoot, 1973) and may help in the determination of gonadally depen- dent and gonadally independent behavior patterns. As the literature stands today it is primarily from work primates and canids that dimorphic behavior patterns have been described which appear to be hormonally independent (Young, Goy, and Phoenix, 1964; Beach, 1970, 1974; Goy and Phoenix, 1971; Goy and Resko, 1972; Sprague and Anisko, 1973). A recent report (Michael and Wilson, 1973) has documented changes in sexual behavior of male rhesus at puberty, but the tests were "pair" sex tests, and few data were offered on social behaviors other than those normally associated with mating. Evaluation of behavioral changes which occur within a constant social grouping over several years may add important dimensions to our understanding of behavioral events associated with puber- ty in the male primate. 2.2 METHODS Subjects. The animals selected for this analysis were all born of wild-caught female rhesus who were part of the Oregon Regional Primate Research Center's breeding colony. Experimental animals lived with their mothers until the age of three months, at which time they were separated and caged individually. Social peer groups were made up of similarly aged 25 animals, once a peer group was assembled its membership re- mained constant, barring death. The sex composition and size of the peer groups varied, but for this analysis a match- ing was made between castrate and intact groups in terms of the total number of animals in the group and the ratio of genetic males to genetic females. Table 2.1 lists the animals and groups used in this analysis. All castrations were carried out either the day of birth (day 1) or at approximately three months of age. The castra- tions were carried out under general anaesthesia with a subumbilical incision in a fashion consistent with sterile surgical practice. Rearing procedures and observations. Within controllable limits, all groups were treated similarly. Following the designation of a group of animals born roughly within the Span of one month as a social peer group, all animals remained caged individually. Beginning when the mean age of the ani- mals designated as a peer group was 3 months, daily one-half hour group tests were carried out five days a week for 100 test days during the first year of life. During a group test the entire peer group was assembled and released into a small testing room 8 x 7 x 5 ft. which contained a stainless steel ramp and platform. Individual animals were scored on a rotating basis by one observer for five minutes per animal. During the same testing period, a second observer kept a record of the occurrence of a wide variety of sociosexual behaviors, noting the interaction and the animals involved. 26 Following the 100-day test run, animals remained caged in— dividually. These same social groups were again brought together for social test runs when the mean age of the an- imals in a group reached 12, 24, and 36 and 48 months. For these later test runs, the groups were tested five days a week for a total of 50 test days. The same scoring pro- cedures that had been used during the first year of social testing were used on these later runs. On the later test runs, the animals were tested in a slightly larger testing room 10 x 7 x 6 ft. Animal care and maintenance procedures re- mained relatively constant for all groups across all the runs and no extraneous manipulations were carried out on the animals of any of the groups. At various times and for vary- ing periods of duration, the social groups were given oppor- tunities for 24 hour continuous living experience. The duration and occurrence of these constant living periods are as follows: group 38: one month between runs 2 and 3, four months between runs 4 and 5; group 41: two months between runs 2 and 3, one month between runs 4 and 5; group 43: four months between runs 2 and 3, one month between runs 4 and 5; group 36: one month between runs zzand 3, one month between runs 4 and 5; group 30: three months between runs 1 and 2, three months between runs 2 and 3, one month between runs 3 and 4; group 37: one month between runs 2 and 3. 27 Data analysis. Data from all groups were used when available. Group 41 was a group which was of the compo- sition listed in Table 2.1 only for social test runs numbers 3, 4, and 5. Originally, group 41 had been composed of animals #2321, 2329, 2324, 2320 and another castrate and second intact female. The castrate and second intact female died after the end of run 2. Female 2344 was added to group 41 prior to social test run 3; she was of similar age and had received similar social testing and early rear— ing. Some data for males #1638, 1558, and 1618 and females #1642 and 1551 were incomplete for the first social run and in consideration of the behaviors for which this was the case, it was necessary to omit these animals from the re- peated measures analysis of variance. All figures carry no- tations as to the numbers of animals used in the final analyses. When.ma1es 1636, 1558 and 1618 were omitted from an analysis, females #1642 and 1551 were also deleted. For purposes of the repeated measure design, only the first 50 test days of the 100—day social test run number 1 were used in the comparisons. All analyses of variance considered three factors: 1) sex or condition; 2) year of test run; and 3) blocks of tests within thesocial test run. When comparisons of means between sex or condition within years were made, the NewmanéKuels test (Winer, 1962) was employed. The minimal level of statistical significance acceptable throughout this paper is p‘<.05. 28 owe mo 9.285 n we 8333.39 as uses sagas no use no coaumapmmu. a asmn a amms a HmmN o NomN a some a Hmmn a vsmu a momN a mamN xom s ammm a gnome a Nwoa a ONmN a anmN a ome new n mmmN e Beams 4 mass 4 seNnN e mmmN a aamN .oz Hmaae< e ommN a pause e mmma a pmNnN e NmmN e «camu e «mam a games e ones » sa~m~ e ammu » «sham me on 44 an an a asses .msmofi eds: 8.35 use 3833 we 3595 soon 330w Em manna. 29 Definitions There were eighteen separate behavior categories on the 30=minute or group score sheet; the five minute or individual score sheet had a total of 29 categories. For purposes of this paper, only the behaviors to be discussed will be defined. Behaviors scored for thirty minutes per trial: Egan. Hyperextension of the mandible and retraction of the lips, exposing all of the teeth. The expression closely resembles that of the yawn stereotypic of human beings. Aggression. A vigorous biting of a portion of the body of another animal. Head and body shaking may accompany this contact; bleeding and laceration sometimes result from this interaction. Erection. Engorgement and subsequent turgidity of the penis sufficient to permit extension of theglans beyond the prepuce. This turgidity may or may not be accompanied by penis flips. Erection was scored when noticed such as dur- ing grooming bouts or as associated with some mount attempts as well as in seemingly unprovoked circumstances. Single foot or double foot clasp mount. An erect stance associated with coitus in the monkey, in which the hands are placed on the partner's hips, perineum or back. One or both feet of the male claSp one or both legs of the partner. Pelvic thrusting may occur. I‘ I |.-..I.Iu|l l I'll. I [III 1.. l I'll-III I 'lllll ' I'll Ill [I 3O Disoriented mount. An erect stance in which the hands are placed on the partner's back or head and appropriate hindquarter orientation is lacking, with thrusting directed either to the partner's side or head. Deficient mount. A stance in which full hind leg extension is never achieved and which thus is carried out in a crouched posture. gag. A facial emression characterized by full jaw opening without teeth exposure. Brow contraction is not evident nor is man- dible extension. This expression is often associated with play ini- tiation. .1323.“ A brief Open-mouth contact with another animal in which the point of contact becomes the focus of either the head or body's move- ments. Brief contact play. A vigorous upper torso involvement with ano- ther animal. The forelimbs and mouth are the principle contact areas with the hind limbs often involved in position or equilibrium main- tenance. Ro_ugh and tumble play. A vigorous whole body involvement with another animal in a complex and varied manner including components of biting, wrestling or tugging. Pursuit play. Vigorous chasing of another animal of sufficient duration or complexity to make the involvement rec0gnizab1e and not random. m. A stereotyped sorting or brushing aside of the fur of another animal. This behavior pattern may involve one or two hands and is often acconpanied by lip snacking. 31 Groom receive. A quiet and relaxed response to social grooming. The groomee may exhibit changes of posture or limb movement as if to accomodate the grooming partner. Groom solicit. A sustained presentation of head, throat, chest or side to another animal. Another posture considered a solicitation is the direct elevation of the hindquarters with accompanying tail deviation and collapse of the front quarters to a point of apparent rest on the cage floor. 2.3 RESULTS Tables 2.2, 2.3 and 2.4 give the results for the analyses of variance which were carried out. The tables give the an- alyses as they pertain to social runs numbers 1, 2 and 3 for males, prepuberally castrated males and females, and social runs 1, 2, 3, 4, and 5 for males and castrate males alone. The analyses for thebehavior categories of yawn, groom, groom received, and groom solicit were carried out for males, pre- puberally castrated males and females for social runs 1, 2, 3, 4, and 5. Figure 2.1 presents a composite array of those data for the behavior categories of gape, box, brief contact and rough and tumble play, and pursuit play. For all of these behaviors, the analysis of variance revealed significant effects of sex, year and block (Table 2.2). Post hoc comparisons, employ- ing the Newman-Kuels tests, were made for a consideration of the effect of sex or condition; males were compared to females, males were compared to castrates, and castrates were compared Table 2.2 32 Results for the three way analysis of variance for repeated measures comparing males, castrate males, and females across their first three social testing runs. 33 .3 v mes mo. v a. .9: RR .9: 04. . a .9: 3:59: Ame? uoom .9: om .9: ma. .9: moo v 03:8 w 035m .9: 32 .9: 3. 3:3: 30898 .9: as. v : 30. v a . . .8 .2 o 32.283 . . .9: m : um< u: .9: Hoo. v a mo. v a 838qu oWo: um omoflh m< . .9: RE .9: 3. . . a . v a moo. v a .83 33.5: .20: .9: m< mNo v Hoo .02 .9: 2 as: flea... s Seam . a .v : moo. v : gum and. moo v So on: 3:38 wows «09.: .9: u< . Q . v Q Hoo. v Q Xom «sum .9: me. So v So «05: .9: 9. . a .v : moo. v a use 2.0: .9: m< moo v Hoo 9836335 5: ammo. mo Moo:— .u 380:. we .80» .m 833:8 w :3 .< RMWS: N.N manna. Table 2.3 34 Results for the three way analysis of variance for repeated measures comparing prepuberally castrated and intact males across their first five social runs. 35 8.3.. mo. v9 e .9:u m4. .9:u< .9: moo. v o .9: mass: Raw Mono: .9:um .9:m< 3:338 H .m .9:u m4. .9:u< .9: mo. . o .9: mpfigowfimmwwoo Omocu m Om.:m< a“ e a «a u m < .9: u .o. .9: mo. v a no. Va 83005 6.: u m c m < «a u m 4. «e u < so. v o moo. v o .9: Samoa}. .3 u m .m.: m < .9: u m < .9: u .o. "Eu .9: 8o. v a .9: 93:3 98 fine: «a u m .9: m < o 939:3 moron 283883 was»... mo e83 .o 9:48: mo to» .o 833:8 .< E33: Quota} n.N manna. 36 Table 2.4 Results for the three way analysis of variance for repeated measures comparing males, castrate males, and females across their first five social testing runs. 37 8.1”: mo. v a . ceum<.m.: u< «cum. mé m4. .m.: .m.: 5.: fluflom 500.8 .m.: um<.m.c 04. :um. mé m< S. v Q So. v a .mé 958mm 58.5 .m.cum<.m.: u< .3099: m< .mé mo. v a .m.: 58.6 ¢«um<.m.a 3. tom :2 .9: m8. ya S... v a Ea» 5630335 :5 amok mfiumoe 33258 503mm Hauofifizom mo #0on .0 mo .80» .m .33 xom .< v. N 3an 38 Figure 2.1 Dimorphic behavior patterns for male, cas- trate male and female hesus during their first three social testing runs. A) the category of gape; B) the category of box; C) the category of brief contact and rough and tumble play; D) the category of pursuit play % 9 .. .0 ("3' d 0—. (N-IO no.3 24'27 Jane (M- 9 u.. IZ'IS no.2 I I I I I I I i -0- 39 24-27 no.3 l2345 ...2 3...‘ 5 .2349 IO TRIAL BLOCKS 2345l2345 I NO. I ”E IN MOS. 3.5'6 SOCIALTESTIWM o 8806 flimsy: adu! 2‘34 5 40 to females. For all four of these behavior patterns, statis— tically higher frequencies displayed by males on at least one of three social runs examined. During the test run in which the animals were aged from 3.5 - 6 months, differences between the sexes were obtained for the measure of gape, box and brief contact and rough and tumble play. For the cate— gory of pursuit play, there were differences in mean fre- quency for the two sexes with males showing higher frequen- cies than females as reference to Figure 2.1 (d) illustrates, but these differences failed to reach significance. No differences were statistically significant for any of these four behavior patterns in the comparisons made between males and castrates. During the Second test run when the animals were aged from 12-15 months, statistical differences were attained for the behavior frequencies of pursuit play, gape, and box on the comparison between males and females, but brief contact and rough and tumble play did not differ sig- nificantly between the sexes for this period of social test- ing, although reference to Figure 2.1 (c) makes it clear that the mean frequency with which males and females exhibited these play patterns was different. For the behavioral pat- terns for which sex differences had been statistically veri- fied, no differences were obtained between males and prepub- erally castrated males. For the behavioral category of gape, a statistically reliable difference did exist between cas- trates and females for the second testing period. For all 41 four of the behavior patterns illustrated in Figure 2.1, no statistical differences for sex or condition were con- firmed for the social testing run which was carried out when the animals were between 24-27 months of age. It is clear even during this run, however, that there were differences in the mean frequencies with which these behaviors were exhibited with the females consistently being the lowest level performers. The frequency of aggressive behavior was highest during the first social testing run and declined over subsequent testing periods (Figure 2.2). The effect of years was revealed to be significant by the analysis of variance, (Table 2.2). Analysis of variance also revealed a significant effect of sex or condition but post-hoc comparisons failed to establish any differences between any of the groups for any of the test runs. Females, however, consistently achieved the lowest mean frequency across all three social test runs. The frequency with which disoriented and deficient mounts occurred was high during the beginning of the first social test- ing run, but subsequently the mounting of the single or double foot clasp variety (Figure 2.3). Analysis of variance re- flected this through a significant interaction of condition or sex and year. Though the analysis of variance revealed sig- nificant effects of sex or condition and years on the mea- sure of disoriented and deficient mounts, only an effect of years was revealed for the single and double foot clasp mount category. The post—hoe analysis for the deficient and dis- 42 Figure 2.2 Aggression by male, castrate male and female rhesus for their first three social testing runs 43 AGGRESSION SOCIALTESI'ING RUN no I no.2 ~03 AGE u was 35-6 nz-s 24-27 44 oriented mount category failed to substantiate a difference between.males and castrates, or females although females exhibited the lowest level of response (Figure 2.3). The levels of play declined significantly from the first year across the next four social runs (Figure 2.4). The castrates did not differ significantly from intact males across all five years in terms of the frequency of aggression which they displayed, but social runs did have a significant effect across the five runs followed for these two groups, and aggression declined during the first three social runs and then showed some increases (Figure 2.5). This increase was particularly noticeable during the first block of runs four and five for intact males and a signif- icant interaction for condition, year and block was re- vealed (Table 2.3). The occurrence of erection was significantly affected by both condition and year of testing (Figure 2.6). Unpaired t tests (Winer, 1962) on the data for the castrates and in- tacts for all five social test runs, revealed no differences Ibetween the two groups for the social testing at 3.5 — 6 months of age (3 = 1.51, df = 14, 1-tailed test). However, one of the castrates, 2321, was markedly different from the eight others in frequency of erection during the first 10-day block of test run one; when he was excluded from a second analysis, there was a significant difference between the intact males and the castrates during the first social run (t a 2.25, p<.025, df = 13, 1-tailed 3 test). No 45 Figure 2.3 Mounts by male, castrate male and . , female rhesus for their first three social testing runs 0 MEAN FREQUENCY SCORED 30 MIN. PER TRIAL 20 46 DISORIENTED AND DEFICIENT MOUNTS SOCIAL TESTING RUN NO. I AGE IN MOS. 3.5-6 no.2 No.3 I2' I5 24'27 P P 1 O'H (Na?) d'O--o (Na-9) 9 ....... (N.|2) 9 \ \ \ \ \ \ \ \ ’-- -- 2345 I 2 3 4 5 I I0 TRIAL BLOCKS 6” 2345 47 Figure 2.h Brief contact play and rough and tumble play by pre- puberally castrated and intact male rhesus for their first five social testing runs. 48 BRIEF CONTACT PLAY AND ROUGH AND TUMBLE PLAY NJ "6 M no.2 no.3 no.4 I03 AGE IN N08. 3.5'6 E'IB 24'?! “'39 48"!“ HO‘IN-IO) r-afIN-OI g42 >538 2 :5 A “’ fixra’ E! ' §§2 ' \ ,« I I. I \ / \ ,0 )\ I’ 8 I4 I V ‘0’ w d" \( IO J °"' v’ ‘u G K r 2 L 1 L l A l 4 n l 1 1 1 1 A. 1 J 1 1 1 1 1 A l *Izsqslzsas 2345I2345I2345 49 Figure 2.5 Aggression by prepuberally castrated and intact male rhesus for their first five social testing runs. 50 AGGRESSION no.5 no.4 NO.3 24’27 1‘11 ddd‘d‘d d dddqfld‘ddddddd 2345 2345 2345 2345 2345 IOTRIALBLOCKS 51 Figure 2.6 Erections by epubeially castrated and intact male rhesus or their first five social testing runs. 52 MEAN FREQUENCY soonzo 30 MIN. PER TRIAL 24 22 83 ERECTION SOCIAL TESTING RUN . NO.I . no.3 No.4 No.5 AGE IN MONTHS 35-6 I245 24'26 36-39 48-5I I H (NIT) d' t .. " o--o (ll-9) d' ‘- 4 .L d qr ‘IUV'U‘ITfY'YU'fo‘T'V'VV'Y llllllllelllllllllllHlAll q 1 II: \\ 4:- at» A qbr'o-~'/ \ "q 1b b.’ \ A Jr A T?’ l A 7A: A ‘9 {’1‘ 4"...) a _1 2’ 1‘} j 1 a 4 I 2 3 4 5 I 2 3 4 5 I 2 3 4 5 I 2 3 4 5 I 2 3 4 IOTRIALBLDCKS 53 difference between the two groups for social test run two was revealed (3.: 1.27, df = 14, 1-tailed t test). Differences were significant for the two groups of males on social runs three, four and five (three: t = 2.27, p <(.025; four: t_= 2.01, p (.05; five: t_= 2.52, p (.025; for all runs, df = 14, 1-tailed t test). For both disoriented and deficient and single and double foot clasp mounts there was no statistical difference between the castrates and the intacts (Figure 2.7). For both categories, years was revealed to be a significant var— iable. However, the trends for these two mount categories were opposite to each other: the disoriented and deficient mount category showed a decrease from the first social run on, while there were increases in the frequency of single and double foot clasp mounts for social test runs numbers three, four, and five. There were no effects of sex or condition for measures of groom, groom receive and groom solicits (Figure 2.8, Table 2.4 ). However, for the measures of groom and groom receive, there was a significant increase across years. In addition, for the measure of groom receive a significant ef- fect of test blocks was also found. There was a significant and dramatic increase in the frequency with which intact males displayed yawning during the fifth social run (Figure 2.9). The yawning frequency dur- ing this run declined and this decline was reflected in a significant interaction between year and block (Table 2.4). 54 Figure 2.7 Mounting behavior by prepuberally castrated and in- tact male rhesus monkeys for their first five social testing runs. MEAN FREQUENCY SCORED 30 MIN. PER TRIAL SOCIAL TESTING RUN NO. I 55 DISORIENTED AND DEFICIENT MOUNTS AGE IN MOS. 3.5'6 3" '— lie.” 0.2 12'I5 no.3 24'27 NO. 4 36‘39 no.5 4B'5I ‘9 \‘--n SINGLE FOOT AND DOUBLE FOOT CLASP MOUNTS I2 II I0 —NU.0ONO@ If H (3' (94-7) o--o U IN-SI I2345 d I2345 J L L 1 1 l I 2 3 4 IO TRIAL BLOCKS 1L1; I2345 56 Figure 2.8 Measures of grooming behavior for male, castrate male, . and female rhesus for their first five social testing runs MEAN FREQUENCIS sconce 5 MIN. PER TRIAL SOCIAL l~ISS'TING RUN AGE IN M03 3.5'6 57 GROOM NO 2 NO 3 N04 NO 5 l2‘ l5 24'27 36'39 48'5I GROOM SOLICIT o-—-O d'IN'S) o——O O‘INIIOI " . ...... 4. 9 (N «L 58 Figure 2.9 Yawning by male, castrate male, and female rhesus for their first five social testing runs. 59 Ifll5 NOAI 36'39 4B'5I IK13 24'27 YAWN 1 0—0 0" (N-TI o--° 0‘ IN-SI .ym. 9 akin 4(=:.¢uu.z=anfinou¢00m >ozunou¢u.:mlz 2 3 4 5 2 3 4 5 I 2 3 4 5 I K)TRMM.BUDCK3 I 2 3 44 5 2 13 4» 5 60 The fact that it was only the intact males which yawned during run five was reflected in a significant interaction for year and condition. Prior to run five, yawning was in- frequently shown by any group of animals. The frequency with which females yawned was low for all five of the social runs and a similar situation existed for the castrate males. 2.4 DISCUSSION The sex differences for thebehavior patterns of box, gape, brief contact and rough and tumble play, pursuit play, and mounting behavior are consistent with data which have been previously reported (Goy, 1968, 1970; Goy and Phoenix, 1971). The lack of statistically significant differences between the castrate and intact males for these behavior patterns during the first three social runs suggests that gonadal hormones during this time period play little role in thedisplay of these behaviors. Further support for this position comes from a recent study (Joslyn, 1973) in which young female rhesus monkeys reared in a fashion similar to that described in this study were given testosterone injections in an attempt to modify the frequency with which these females exhibited the dimorphic behavior patterns already enumerated. The androgen treatments were ineffective in the modification of the frequencies with which these females exhibited any of these five patterns of social interactions (Joslyn, 1973). 61 Data from the present study and those collected on genetic female monkeys whose mothers were androgen-treated while pregnant (such androgen treatments were effective in the modification of the frequencies with which these fe- males exhibited these patterns of social interaction) are consistent with the hypothesis that prenatal gonadal hormone exposure in this Species modifies the expression of dimorphic behavior patterns in a direction characteristic of that of the genetic male. The diSplay of the behavior patterns de- scribed for the castrate males in this study appears to be independent of concurrent gonadal hormone stimulation and in this regard clearly differentiates the behaviors of box, gape, brief contact and rough and tumble play, pursuit play, and mounting behavior from those dimorphic patterns studied frequently in the rodents (Whalen, 1971). Recently, evidence similar to that which has been obtained with monkeys has been reported for the urination postures of dogs (Beach, 1970, 1974). ‘These postures are exhibited dimorphically by both males and prepuberally castrated males. In addition, genetic females which have been exposed to androgen while in 33253 exhibit the pattern typical of the male. The data from male castrate and female pseudohermaphrodite dogs in terms of the onset of this behavior pattern make it clear that this be- havior is independent of concurrent gonadal hormone stimulation. At the moment, the hormone-independent dimorphic behaviors of the dog and the rhesus appear to be the best evidence for 62 the permanent behavioral changes which early hormone exposure can induce. Measures of groom, groom solicit and groom receive failed to substantiate a sex difference for these behavior patterns. Age affected the display of these patterns. In this regard the data agree with studies which have found that the groom- ing performed by female rhesus increases with age (Lindburg, 1973) and that the frequency of grooming by males increased as they grew older (Miller, Kling, and Dicks, 1972). The rearing and testing procedures followed in the present study may well have affected the measures of the grooming behavior in this study. Although the animals failed to show much grooming prior to the third social run, it has been reported that if infant rhesus are reared with their mothers in small groups of four or five females and an adult male, grooming may appear by 3 weeks of age (Hinde, Howell, and Spencer- Both, 1964). It is difficult to say a great deal about a possible sex difference in grooming behavior. Studies of adults have sug- gested that females are more avid groomers than males (Lindburg, 1973), but Hinde and Spencer-Booth (1967) reported no sex differences for the first 2-% years of life. A recent review (Sparks, 1967) brought together data suggesting that factors such as age, sex, claSs, dominance status, and kinship affect the patterns of grooming behavior shown by several primate species. Consequently broad generalizations will not be attempted from this study. 63 One of the major goals of this study was an examination of the activational action which gonadal hormones play in the modification of social and sexual behavior patterns of intact male rhesus. It is clear that the high levels of gonadal hor- mones associated with puberty in the male rhesus and present from about the age of 3 years on are unnecessary for the dra— matic decline in play behavior which occurred under our social testing conditions. It would appear that social development may be the critical factor moderating the frequency with which male rhesus exhibit patterns of play behavior. Other studies have also documented a dramatic prepuberal decline in play behavior by young males (Hansen, 1966; Hinde and Spencer-Booth, 1967; Ruppenthal, Harlow, Eisele, Harlow and Suomi, 1974). Certainly adult male monkeys do play and when placed in certain situations this play may occur at fairly high levels (Redican and Mitchell, 1974) and the absence of play initiators may have been what accounted for the steady decline in play in our groups. The lack of difference in the play behavior of castrates and intacts cannot be taken as evidence against the effect which gonadal hormones may have on play behavior. Gonadal hormones may interact with social deve10pment and ex- ert an influence on play behavior. It has been reported that four-year-old male rhesus play twice as much as eight-year-old male rhesus, and this could be a reflection of an effect of lengthy exposure to gonadal hormones (Kaufman, 1967). It 64 does not, however, appear that the decline over time in this behavior is causally related to the endocrine activity charac- teristic of puberty. The occurrence of erection in intact males was much more frequent than in castrates suggesting that even low androgen levels may affect this response. There are data from the human which suggest that from the time of birth to approx- imately three months of age, serum testosterone levels are rising (Forest, Cathiard, and Bertrand, 1973). From the time of three months of age, peripheral testosterone levels show a decline and remain fairly low until puberty. It appears from the data on erection that the gonadal hormone pattern for the rhesus may be similar to that of man. It is well documented that castration affects the occurrence of erection in the rhe- sus and that subsequent androgen therapy can restore this beha- vior pattern (Michael and Wilson, 1974). In view of this the decline in the occurrence of spontaneous erection during the period of the first social testing run could be suggested to reflect a similar decline in circulating testosterone from the age of about three and one-half months on. The occurrence of erection by castrates remains low from the time of social test run one on, and this appears consistent with a lack of gonadal hormone activity on the part of these animals. The frequency with which intact males exhibited this behavior in- creased on runs three, four, and five from a low level during run two. It should be remembered that erection is sometimes noticed during grooming as well as occasionally prior to a mount. 65 However, castrates did not differ from intact males for either of these behavioral categories, and thus it is difficult to ascribe differences in the occurrence of erection to a social factor or to social behavior differences such as frequencies of grooming or mounting by the intact males. Occurrence of erection in male rhesus monkeys may be an extremely sensitive index of hormonal state. Recent work on testosterone levels of group living rhesus has revealed a cor- relation between testosterone level and social rank (Rose, Holaday and Bernstein, 1971). Field reports on baboons and vervets have documented a high frequency of erection in adult males and have suggested a social communication function for this display (Wickler, 1969, 1972). Consequently social posi- tion may regulate, through peripheral androgen levels, the occurrence of erection. Yawning is a sexually dimorphic behavior pattern in adult rhesus monkeys. Yawning is reduced by castration but can be restored with androgen treatment (Phoenix, 1973; Phoenix, Slob and Goy, 1973; Phoenix, 1974) and females will yawn if given exogenous testosterone (Goy and Resko, 1972; Eaton, Goy and Phoenix, 1973). It is very reasonable, then, that the in— tact males were the only animals to exhibit this behavior at a high level and that the appearance of this behavior pattern was during a postpubertal social testing run. To the extent that yawning has been considered a dominance display for baboons (Hall and Devore, 1965) its occurrence may also be modulated through peripheral androgen levels in a 66 fashion similar to that previously suggested for erection. However, it is important to mention that social factors p35 'sg are not sufficient to stimulate frequent occurrence of this behavior in male rhesus;for castrates, although they occupied social positions superior to the females of their groups and equivalent to those occupied by the intact males in their groups failed to exhibit yawning frequencies equivalent to intact males postpubertally. Mounting behavior for both intact and castrated miles was very similar both in terms of frequency and development. Both groups exhibited mounting during all five of the social runs and both showed a gradual shift from immature mounts to single and double foot clasp mounts characteristic of adult mating behavior (Bielert and Goy, 1973). It doesn't appear that gonad- al hormone activation is critical to the occurrence and de- ve10pment of the mature mounting pattern in rhesus (Phoenix, 1973). Although mounting may not be radically affected by cas- tration, the occurrence of intromission definitely is (Michael, Wilson, and Plant, 1973). Intact and castrate animals in the present study displayed intromission, although with the social rearing condition under which they were followed, only about 30% of each group achieved intromissions. Three intact males (1636, 1558 and 1618) started snowing intromissions during social run number five. Male 1558 displayed only one intro- mission; male 1618 exhibited a total of nine, and male 1636 had a total of 129 intromissions during the 50-day test period. 67 Male 1636 was also observed to ejaculate a total of nine times. Starting in social run number four, castrates 4859, 4861 and 4874 began to exhibit intromission; castrate 4859 exhibited one, and castrates 4861, and 4874 each exhibited a total of three. Castrates 4859 and 4861 also exhibited the ejaculatory response pattern during this test run. During social run five, the same three castrates again displayed intromissions. Cas- trate 4874 who had exhibited a total of three intromissions during run fear, exhibited a total of eighteenintromissions during run five and this same castrate exhibited the ejacu— latory response pattern a total of four times. The castrates were never seen to ejaculate seminal fluid, nor were their fe- male partners ever observed with a plug of coagulated seminal fluids. The accessory sex glands of rhesus require androgen for their maintenance (Zuckerman and Parkes, 1938; Zuckerman and Sandys, 1939) and castrate rhesus fail to exhibit seminal emission (Phoenix and Jensen, 1973). The ejaculatory response which the castrates exhibited was indistinguishable from that of an intact male. The observation of this response in cas- trates is apparently the first to be reported for prepuberally castrated males. The occurrence of ejaculatory behavior can be taken as evidence for a certain degree of hormone indepen- dence in the sexual behavior of male primates. Apparently the potential for the diSplay of complete mating behavior is pres— ent in these nonhuman primates even in males which have only received prenatal androgen exposure. The critical nature of such exposure is confirmed by the observation of ejaculation 68 in female hermaphrodite rhesus under testosterone stimulation (Eaton, Goy and Phoenix, 1973). The fact that one of the three intact males in the present study intromissions displayed at a level far above that of any of the castrates could suggest a positive effect for gonadal hormones on the display of this measure of male sexual behavior. In this regard, it would appear that the role of androgen in rhesus monkey sexual behavior may be that of a behavioral potentiator. Its effects may be mediated through a variety of paths or systems, and interactions may well exist between them. As an example, the effects may be mediated through the maintenance of the penis or the nerves servicing it, for it has been shown that integrity of penile feedback is critical to the display of sexual behavior in the male rhesus (Herbert, 1973). The effects of androgen may also be expressed through the male's general demeanor and behavioral attitude. There is recent evidence suggesting that the interaction of male-female rhesus pairs is altered by castration of the male partner (Michael and Wilson, 1974). A third possibility is that the effects of testosterone may be mediated through visual or olfactory cues to the female. Data from two Species have suggested that the sex skin color of the male is androgen dependent (Zuckerman and Parkes, 1938, 1939; Vandenbergh, 1965; Dixson and Herbert, 1974). At the moment however, it is impossible, to determine how cues as skin color might affect the interaction between a rhesus pair. Although androgen is certainly of biologically critical importance to male pri— 69 mate reproduction, its importance for male sexual behavior patterns may be truly unique in regard to the degree of inde- pendence which may exist for the deve10pment and display of mating behavior. The observation of the ejaculatory response pattern in prepuberally castrated rhesus may have a human ana— log in the occurrence of the orgasm far in advance of the time of puberty when gonadal hormone concentrations show a dramatic rise (cited as general knowledge by Bermant and Davidson, 1974). A very interesting contrast existed for intact males and castrates in regard to the level of aggression exhibited in the first block of ten tests during social runs four and five. It has been reported for several age groups and species of primates that the period of time immediately following a group formation is characterized by high levels of aggression (Bernstein and Mason, 1963; Bernstein, 1964; Bernstein and Draper, 1964; Bernstein, 1969). The high level of aggression during block one of test runs four and five is consistent with these reports for at this time the animals of the social groups were being reunited after a period of single cage housing. A comparison of the levels of aggression during this period by the intact males and the castrates indicates that gonadal hormones may have had an influence on the frequency of aggres- sive interactions which occurred during this period, for these monkeys. Attempts at the manipulation of social position through hormonal means in a group situation have by and large (Mirsky, Pi 70 1955; Green, Whalen, Rutley and Battie, 1972) been unsuccess- ful. This may partly be due to the fact that alliances and allegiances exist in monkey social groups and result in what has been described as a dependent social rank phenomenon (Kawai, 1965; Varley and Symmes, 1966). The existence of social inertia may be a critical factor resulting in failure of andITgen manipulation to alter the social structure of a group social inertia has been suggested where males remain dominant in a troop even after injury or disability (Southwick and Siddiqi, 1967). However, one study stands in marked contrast to the others (Joslyn, 1973). In this study, androgen injections to young female rhesus resulted in an up- ward shift of their positions in the dominance hierarchy of their group. An important factor contributing to the success of this project may have been the experimental design which was employed. Androgen injections were begun while the group was in a daily social testing run and these injections were continued following the end of the social testing, at which time the animals were singly caged. After several months, the group was reunited and the females who were continuing to receive androgen injections were observed to have shifted in their hierarchy positions. These shifts occurred at a time of social disorder and it may be the case that androgen's ef— fects on social position are most pronounced at such a time. An unstable social situation maximizes the probable occur- rence of aggression and andxxgen may well affect the intensity 71 of this particular behavior pattern, and through this means, exert an effect on an individual's dominance position. An important consideration in the androgen-aggressive behavior in- teraction is that of early experience and the effects which the early social environment may have on the prOpensity to display aggressive behavior.” Early social restriction has pronounced effects on measures of agonistic behavior for the monkey (Harlow and Harlow, 1965). Because there was reason to believe that the rearing condition of the males in this study may have had an effect on their display of certain behavior patterns such as mounting (Goy and Goldfoot, 1973; Coy, Wallen, and Goldfoot, 1974), a further examination of aggressive be- havior was carried out. Although the social experience of the groups of intacts and castrates compared in this study was very similar, groups 30 and 43 received more constant living experience than the others. Group 30 received several more months of social exposure than the two other intact groups; this group is one which has been described as "socialized" because of this fact (Goy and Goldfoot, 1973). Using addi- tional data collected in a fashion exactly like those already described in the methods section for intact groups nos. 28, 29, and 32, which also have been termed socialized (Goy, Wallen, and Goldfoot, 1974) an analysis of variance was carried out on the data for aggression from social test runs two, three, four and five. The males of groups 36 and 37 (the nonsocialized animals) were compared to the males of groups 28, 29, 30, and 32 (the socialized animals). The duration ,uo do (I) l‘ 72 And occurrence of these constant living periods for the animals of the additional three groups are as follows: group 28: six months between runs 2 and 3; group 19: five months between runs 2 and 3; group 32: five months between runs 1 and 2, two months between runs 2 and 3. The data are illustrated in Figure 2.10. The analyses of variance failed to reveal a statistically significant difference between the socialized and nonsocialized groups, although there was a statistically significant effect for both years and block, and, in addition, there were statistically significant interactions between social experience and block, year and block, and social experience, year, and block. A check of Figure 2.10 verifies the fact that the level of aggression which was exhibited in block 1 of runs four and five was different for these two groups. The socialized males failed to exhibit the high levels of aggression shown postpubertally by the nonsocialized males during these periods of social instability. Although many explanations could be offered, two seem very reasonable. One explanation involves the social interaction network which may exist in socialized groups. It could be that the allegiances and alliances which exist between the animals of such a group contribute a moderating effect on the agonistic behavior shown during periods of social instability. A second interpretation, somewhat related to the first, is that of the effect which behavioral maturation within a socialized environment may have on aggressive behavior. In consideration of this second inter- pretation, data from studies which have examined the interaction between testosterone levels and aggression in humans is offered. 73 Figure 2.10 Aggression shown by intact peer-reared and intact socialized peer—reared males during four social testing runs. MEAN FREQLENCY SCORED 30 MIN. PER TRIAL 3O 25 25 24 22 20 IS IS I4 74 AGGRESSION SOCIAL TESTIm RUN NO. 2 no.3 No.4 no.5 AGE IN MOS. I2'l5 24'27 36‘39 4B‘5I Y J "Hmnm PEER ‘F . “apnea MALES 4 I 1....me socusuzzo _ :83! mo mes .. r «r- I- d1- 1. '! l'i 1;: 411 L11; 345l2345|2345 IOTRIALBLOCKS 75 mania and Basu (1971) reported accmelatian betweena pathological measure— ment of aggression and testosterone level in adult males. This study suggested that a relation did exist between aggres- sion and testosterone level in their sample. Recent work (MeyeréBahlburg, Boon, Sharman and Edwards, 1974) has failed to replicate this correlation and has suggested that the Persky 33.5l, study may have been influenced by the blood sampling procedure. Work which has been carried out with a criminal pOpulation also failed to find a correlation between testosterone and aggressive behavior when a broad population examination was done, but did find a meaningful relation- ship between these two measures when a subsample was critically examined (Kreuz and Rose, 1972). In this paper a hypothesis was presented that within a population that is predisposed by virtue of social factors to develop.antisocial behavior, levels of testosterone may be an important additional factor in placing individuals at risk to commit more aggressive crimes in adolescence. The data from the groups of monkeys examined in this study seem in some regards consistent with the hypo- thesis proposed by Kreuz and Rose, for it appears that a re- lationship between intensity of aggressive behavior and Postpubertal testosterone levels may exist for those animals which for some reason exhibit a propensity for that behavior during periods of social instability such as at the start of a social run. Section 3: TESTOSTERONE TREATMENTS OF CASTRATE MALE RHESUS OF VARYING AGES IN GROUP AND PAIR ' TESTING SITUATIONS 3.1 INTRODUCTION Male rhesus monkeys are dependent upon gonadal hormones for the display of Species typical sexual behavior (Michael, 1972, Michael, Wilson and Plant, 1973, Phoenix, Slob and Goy, 1973, Michael and Wilson, 1974). Adult levels of androgen pro- duction normally begin about the age of three years as suggest— 'ed by the physiological changes associated with puberty begin. Puberty as measured by mating tests and testicular biopsies has been reported to occur in laboratory reared male rhesus between the ages of 2 years 11 months (van Wagenen and Simpson, 1954) to 3 years 6 months (Hartman, 1943). Reports on free- ranging male rhesus utilizing only testicular biOpsy suggest that puberty has taken place by 3 years 6 months of age (Conaway and Sade, 1965). Measurement of gonadal hormones in laboratory housed rhesus indicate that peripheral testosterone remain low until the age of about 2 years 10 months (Resko, 1967). Mature mating behavior seems tied to the events of puberty for reports on rhesus followed on Cayo Santiago have trailed to record frequent sexual behavior by males of an age less than 5 years. It is well known, however, that males one, two and three years old are capable of exhibiting 76 77 the mature mounting pattern, i.e. a double foot clasp mount (Goy and Goldfoot, 1974). In the free-ranging troop situ- ation it may be that young males though fully capable of mat- ing at the time of puberty, are prevented from doing so by older males (Koford, 1963). A recent laboratory study on pubertal rhesus has documented the fact that under pair test- ing Situations, in which conSpecific interference is eliminat- ed, males of pubertal age exhibit complete mating behavior (Michael and Wilson, 1973). Another explanation of the low level of sexual behavior on the part of pubertal males is that related to an incomplete deve10pment of secondary male sexual characteristics (Harlow, Joslyn, Senko and Dopp, 1966). Age affects the level at which certain components of the typical mating pattern are displayed with young males exhibiting more mounts prior to ejaculation (Michael and Wilson, 1973). Observations on complete mating behavior by pubertal laboratory primates tested in a pair situation have been reported by Rosenblum and Nadler (1971) for bonnet macaques. Pertinent data has also come from enclosure group studies. Hanby and Brown (1974) have followed the mating behavior Shown by young Japanese macaques studied in a large coral and report that prepuberal males exhibited the mature mating pat- ‘tern.with the exception of the intromission and ejaculation Components. It would appear that the deve10pment of the mount— jlug component of mature mating behavior for several species 0f. primates occurs at an age far in advance of that for phys- 78 iological puberty. In birds and rodents it has been possible by administration oerxogenous testosterone to elicit pre- puberal sexual behavior (Hamilton, 1938; Stone, 1940; Beach, 1942; Noble and Zitrin, 1942; Riss, Valenstein, Sinks, and Young 1955; Gerall, 1958, 1963). As of yet, no attempt has been.made to examine the induction of precocious sexual behavior in any primate species. However, a clinical condition, "pubertus praecox", is well documented for the human and its consequences are profound, "we find in the records that Salamis, the son of Euthmenes, grew to 4'6" in his third year, he walked slowly and had a bass voice, and was carried off by a sudden attack of paralysis when he turned three" (Pliny, Natural History). The psychological ramifications of this syndrome are little known, although Dr. John.Money and co-workers at The Johns HOpkins University have reported data on the psychosexual state of some two dozen cases of idio- pathic sexual precocity (Money and Alexander, 1969; Money and Walker, 1971). Their work has suggested that in boys psycho- sexual development was clearly not an automatic by-product of the precocious appearance of gonadal hormones. The body became sexually mature first and intellectual and psycholog- ical maturity followed. However, in comparison to similarly Euged controls the boys exhibited a tendency to high I.Q., eazfly'occurrence of the capacity for frank sexual-imagery 111 dreams and daydreams, and early establishment of the ca- Pa-city for erotic and sexual arousal in relation to visual imagery and visual perception as well as tactile sensation. 79 In consideration of human clinical data it is important to note that these boys were limited by social and parental pressures to a peer group environment and that the girls with which they had contact were at a fairly immature psychosexual stage themselves. It would appear that social factors 235 ES may have had an inhibitory effect on the sexual behavior of these boys and it is Open to speculation as to whether they might have, under conditions free of social constraint, exhib- ited some form or attempts at apprOpriate sexual behavior. Attempts have been made to alter the social hierarchy and social interactions in established primate groups through exogenous hormonal manipulation (Mirsky, 1955; Green, Whalen, Rutley, and Battie, 1972; Joslyn, 1973). By and large these attempts have been unsuccessful; however, one study did report success in the manipulation of dominance through exogenous testosterone treatment of genetic females tested in a group Situation (Joslyn, 1973), and under pair testing situ- ations testosterone has been shown effective in the modifi- cation of social status for apes (Clark and Birch, 1945, Birch and Clark, 1946). In light of the scarcity of information on the activa- tion testosterone might have on sexual behavior in young male primates as well as its possible effects on the social inter— «actions within a group, a series of manipulations was carried Chat in which castrate male rhesus of two different age classes Were observed while receiving testosterone pr0pionate injec- 17ions. The behavior of these animals was measured in a 80 social group testing situation and later these animals were tested under a pair testing situation with estrogen-primed stimulus females. 3.2 METHODS FOR PART I: TESTOSTERONE STIMULATION AND GROUP TESTING. Egbjects and treatments. The animals used were all born in the laboratory of wild-caught mothers who were members of either the Wisconsin Regional Primate Research Center Breeding Colony or the Oregon Regional Primate Research Center Breeding Colony. A total of ten castrates comprise the exper— imental subject pool for these studies. Pertinent data on these animals are given in Table 3.1. Castrations were per- formed under general anesthesia by way of a subumbilical incis- ion. The operations were carried out in a fashion consistent with routine sterile surgery practice. Castrates received testosterone (1 mg/day/kg) in a corn oil (Mazola) solution. This dosage has been shown to be effective in the restoration of sexual behavior in the adult male rhesus castrate (Phoenix, Slob and Goy, 1973). Intact males were given corn oil injec- tions of a volume adjusted to that of .05 cc/kg/day as a control procedure for the injections which the castrates received. ‘Rearing procedures and social testing: All groups re- ceived similar social rearing. Oregon group 43's early rearing was identical to that of the younger groups for the first two years of life. All animals as infants were allowed 81 o N NNSNNN oats 335 SN N N NNSNNN 248m posed 3N N o NN-N-N NNNNNN 22 8828 N3 N N N NN-NN-NN 223 N35 SN 5883: N N NN-NN-NN odes N83 SN N N. NNNN NNNNNN 32 3233 min N N NNNNSN 32 83.: SN N N NZNN NNNNSN oNas 8838 NN-N N a N NTNNSN oNasoN N935 Na 5283: N N NZNN NTNNSN oNss boosts NNN N N NNNaN Edam 835 NNN o o SNNN 323 «used NNNN N N SNNN 223 N33 Nos. N. N SNNN SNNN 32. Seats N2: N4 N N ao-NN-N SNNN oan ethos :3. comes N N SANN SANN 22 8838 No: N N SNNN SNNN Bus Stand 9.: 295.489 engages 839530 8.8 Sam 833.8 N x8 N Nana N 98o 8338 SNNNNoa NaNuoN NsNuoN .385 mass 395mm”. mo man—ohm Noon Ransom .1». 35s.“. 82 e 4 NT: N24 oNss 3838 N3 N N NE... 32. posed :N N N NE... east 835 N; N o o NN.-NN-N £83 835 NTN 53832 N N NZNN o3. 83S NNN N N NNoNN NN-oN-N oNas Sansone EN estate... 325355 83838 38 SEE 833.8 N son a N23. a .595 ENNNooa SNNNNoa NsNoom NaNooN GASES N.N oNons 83 to live with their mothers until the age of 3 months at which time they were separated from them weaned and caged individ- ually. Social groups were originally formed of animals born roughly within the span of one month and once a peer group had been set up its membership remained constant. Daily one- half hour group tests were carried out during the first year of life. For a session of daily testing all the animals of a group were removed from their home cages and released into a small testing room containing a stainless steel ramp and plat- form. Individual animals were scored on a rotating basis by one observer for 5 minutes per animal. During the same test- ing period, a second observer kept a record on the occurrence of a number of sociosexual behaviors, noting the behavioral category and animals involved. This second observer scored the entire group for the full one-half hour test session. Following the 100—day test run, the animals remained caged individually for a period of time. All the anmmals of a group ‘were given opportunities for 24 hour continuous living experi- ence sometime following the end of their IOO-day test run. During this constant living period, the group was released into a large cage and there they remained until one month prior to their next social run at which time they were again caged individually. Following their first test run Wisconsin groups 1, 2 and 3 received 3 months of constant liv— ing experience; the older Oregon group had received 4 % months of such exposure following but not until the finish of its second social testing run. The animals of all three Wisconsin 84 groups were again tested for their social behavior as a group When the mean age of the animals in the group reached 12 months. The groups were tested 5 days a week for a total of 50 test days. Group 43 received social testingwhen the group members mean age reached 24, 36 and finally 48 months of age. In be— tween these test runs the animals of this group received periods of single cage housing as well as constant living experience. The 50—day test run when the younger groups were 12 months of age and when the older group was 48 months of age constituted what will be called the pretreatment baseline test period. All castrates were started on daily testosterone prOpionate injections at the end of this pretreatment period and the intact males were started on similar daily injections of corn oil. All groups were followed for a 50-day test run treatment period. Following this treatment period, all inp jections ceased, and testing was continued for an additional period of 50 test days. Data from this last 50-day testing period was only available for Wisconsin groups 1 and 2. For the two other groups testing was discontinued sometime during what would have been the post-treatment test period; for group Wisconsin 3 this followed thedeath of female 8-75 and castrate 8-70 from shigellosis and Oregon 43 was discontinued due to severe aggression.between the two females of the group which left female 4857 physically debilitated. . Social hierarchy positions were determined by use of a constructed fear grimace matrix. A matrix is constructed by plotting an animal against those group members he fear grimaces 85 to with alpha being the animal of a group which never exhibits fear grimaces to any group member. Definitions. There were eighteen separate behavior cate- gories which were scored for their occurrence during the thir- ty minute social testing period. In addition, a total of twenty- one behavioral categories were scored for their occurrence during the five minute period in which each animal received focal animal scoring treatment. For purposes of this dis- cussion, only the behaviors to be discussed will be defined. Those behaviors scored for one-half hour included: Disoriented mount: An erect stance in which the hands are placed on the partner's back or head and apprOpriate hindquarters orientation is lacking with thrusting being directed either to the partner's side or head. Deficient mount: A stance in which full hind leg exten- sion is never achieved and which thus is carried out in a crouched posture. Thrusting movements must accompany the mount. No foot-claspgmount: A hind-quarter oriented stance be- hind another animal in which the hands are placed on the part- ner's hips, perineum, or back. No clasping by the hind feet occurs. Thrusting movements must accompany the mount. Single or double foot clasp mount: An erect stance associated with coitus in the monkey, in which the hands are placed on the partner's hips, perineum, or back. One or both feet of the actor clasp one or both legs of the partner. Pel- vic thrusting may or may not accompany these mounts. 86 Intromission: Penile insertion in thevagina, recognized by deep and regular pelvic thrusting. Ejaculatory response pattern: A recognizable behavioral state characterized by full-body tenseness, immobility, and slight quivering of the flanks. Present: A posture in which the hind quarters are raised and oriented towards another animal, while the front legs are generally rigid or slightly flexed. The behavior is not scored if more than one animal could have been the receiver of the social movement. Erection: Engorgement and subsequent turgidity of the penis sufficient to permit extension of the glans beyond the prepuce. This turgidity may or may not be accompanied by penis flips, Erection was scored when obvious such as dur- ing grooming bouts, after or preceding occasional mounting attempts as well as under situations of seemingly unproved circumstances. Yawn: Hyperextension of the mandible and retraction of the lips, exposing all of the teeth. The expression closely resembles that of the stereotypic of human beings. Aggression: A vigorous biting of a portion of the body of another animal. Head and body shaking may accompany this contact; bleeding and laceration sometimes result from this interaction. 87 Fear Grimace: A stereotyped submissive facial expression in which the lips are fully retracted to expose the teeth. Those behaviors scored for only 5 minutes per test period include: Groom: A stereotypical sorting or brushing aside of the fur of another animal. This behavior pattern may involve one or two hands and is often accompanied by lip-smacking. Groom receive: A quiet and relaxed response to social grooming. The groomer may exhibit changes of posture or limb movement as if to accommodate the grooming partner. Groom solicit: A sustained body presentation limited in this study to one of two postures. 1) a prone display of the entire or nearly entire ventral body surface; 2) a seated posture in which the legs are slightly Spread and the head is tilted back in a fashion to extend the neck. A reference to Figures 3.1 and 3.2 will make this behavioral category some- what clearer. Autogenital manipulation: A manipulation of the genitals with the front or hind feet in the case of males. Genital manipulation received: The acceptance of genital examination or manipulation by another animal. This manipula- tion may be manual or oral in nature. Weights and measurements: The animals of Wisconsin groups 1, 2, and 3 were periodically weighed. On a weighing day all 88 Figure 3.1 Castrate S-55 exhibiting the prone body display described as groom solicit while under testosterone propionate therapy Figure 3.2 Castrate S—55 exhibiting the seated body presen- tation described as groom solicit while under testosterone propionate therapy. mm —‘ _.— _. ’4 limp 90 of the animals of a Specific group were removed from their home cages by means of a transport cage which was then placed on the scale for weight determination. During the treatment period the weights were used to adjust the testosterone pro- pionate injections for thecastrates and the oil injections for the intacts. This same procedure was followed for the animals of Oregon 43 during their treatment period. Measurements were also taken on the penis size of the young castrates and intacts. These measurements were done with calipers as well as a flexible plastic measuring tape. 3.3 RESULTS FOR PART I Physiological Measures: By the second week after the start of injections the castrates weighed significantly more than their pre—treatment weight (Figure 3.2) (Mann-Whitney U test; u = 4, p4<.026). The testosterone treatments resulted in a pronounced reddening of the sex skin of the young castrates. This effect has been reported previously for other castrate rhesus (Zuckerman and Parkes, 1938; Vandenbergh, 1965). During the period of treat- ment the penes oftme castrates reSponded with rapid grOWth, (Table 3.2)) By day 40 from the start of testosterone prOpionate treatment the penes of all castrates had shown some response to the stimulation. The blanks in Table 3.2 reflect penile development stages of marked immaturity. In such stages the glans is often completely contained by a prepuce which is tightly adherent in the coronal area. Similar juvenile penile 91 , Figure 3.3 Weight gains by young male rhesus monkeys prior to, during, and following an experimental treatment period. WWI“) 4.00 P 300!- 2.00 - l A 1 L L A A .. ’0’ A hr, -.P 92 +——mscwous———-I o. 5 IO I5 WEEKS FR“ THE BEGIMNG CF TREATKNT 93 Table 3.2 Penis measurements for young male rhesus prior to and during an experimental treatment period. 94 Table 3.2 Csstrsts my 40 my 75 Intsct Day 40 thy 7S Ant-11 Pn- post post Animal Pro post post W 1W _II_ TP 1? Mar Ibssurmt _l|_ oil 011 5-24 A 1.5 1.5 1.5 5-59 A 1 S 1.5 1 S I 1.0 1.9 2.5 B --- --- --- C 0.6 1.0 1.0 C --- 0.5 0.5 8-51 A 1.5 1.5 1.5 5-47 A 1.5 1.5 1.5 5 0.7 2.5 3.0 B --- --- --- C 0.5 1.0 1.2 C 0.5 0.5 0.6 5-45 A --- 1.5 1.5 5-57 A 1.5 1.5 1.5 B --- 0.5 2.5 B --- --- --- C --- 0.7 0.7 C 0.5 0.3 0.3 5-55 A 1 5 1.5 1.6 5-71 A 1.5 1.5 1.5 B --- ~-- 2.5 B 0.1 0.1 0.1 C 0.2 1.0 1.0 C 0.5 0.5 0.5 5-70 A --- 1.5 1.5 5-77 A 1.5 1.5 1.5 B --- 2.0 2.4 B --- --- --- C --- 1.0 1.5 C 0.2 0.2 0.2 5-70 A --- 1.0 1.5 l --- 2.4 5.0 C --- 0.5 1.0 A. Circa-form of pains shaft behind glans in cm. 3. Distance of glans frat pupae in on. C. Distant of opening of urethrs flu: and of glam in on. 95 development conditions have been reported for young talapoin (Dixson and Herbert, 1974). Behavioral Measures: Prior to the start of daily injections there was only one behavior category for which intact males and castrates differed, that behavior category was the brief contact & rough and tumble play category Table 3.3. Other work has suggested that castrates do not differ from intact males on this play behavior category and the difference in this study may have been related to the Specific animals involved and the age of these animals at the time of the study (Bielert, unpublished data). Brief contact & rough and tumble play have been described as sex dimorphic and in this regard a statistical comparison between the castrates and the females in their social groups was carried out. The females exhibited brief contact and rough and tumble play at a frequency of 3.6 i 1.3 and differed from the castrate at a significant level (t test, df = 11 p‘<.005). During the period of treatment the young castrates were shown to differ from intact males on a total of six behavioral measures. For these six behaviors castrates also Showed statistically significant differences from their pretreatment testing period. Only one behavior category that of brief contact and rough and tumble play differed in a Similar fashion for the intact control males. It should be added here that females exhibited these play behaviors at a frequency of 8.9 i 2.1 during the treatment period, but castrates remained statis- tically different from them (t test, df = 11 p‘<.025). Table 3.3 96 Behaviors of young male rhesus prior to and during an eXperimental treatment period. All behaviors tested in the upper half of the table were scored for a one half hour period each test session; the behaviors listed in the lower half of the table were scored for a five minute period and test session. Means are for fifty day test periods. 97 Table 3.3 Before Injections Durg‘ g Injections Behavior: 6 if d .i Disoriented 8 5.5 1 3.9 2.1 1 1.2 3.2 1 1.1 2.3 1 1.4 deficient mounts: No foot claSp 2.0 1 1.0 0.3 + 0.2 1.8 + 1.7 0.2 + 0.1 mounts: — T " Single or double 0.2 1 0.1 0.8 1 0.7 0 4 + 0.2 0 6 + 0.4 foot clasp mounts: presents: 4.0 1 1.7 6.9 1 3.1 6.2 1 3.1 8.1 1 3.9 Erections: 0.8 10.4 1.5 10.7 1.2 10.5 25.319.31' Yawns: 0.6 10.5 1.21o.7 0.4 10.4 36.9 18.91“ Aggression: 0.7 1 0.3 0.3 1 0.2 1.1 1 0.3 0.7 1 0.3 . 4 ' 1- a Brief contact a 21.4 1 1.9 12.9 1 2.5 32.6 1 1.7 19.7 1 3.81 rough G tumble play: Groom: 0.0 1 0.0 0.3 1 0.2 0.4 1 0.1 0.8 1 0.4 Groom solicits: 0.1 1 0.0 0.5 1 0.4 0.1 1 0.0 5.7 1 2.51" Groom receive: 0.0 1 0.0 0.2 1 0.1 0.2 1 0.1 1.4 1 0.51“ Autogenital 0.5 1 0.5 1.0 1 1.0 0.0 1 0.0 0.2 10.1 manipulation: Genital manipula- 0.9 10.2 1.8 10.7 1.310.6 5.9 11.31“ tion received: Mean 1 S. E. - 4 Significant difference (using the 1 test, p< .05) between castrates and intacts. ISignificant difference (using the _t_ test for matched pairs, p < .05) before after oil injections to intact males. ISignificanr difference (using the _t_ test for matched pairs, p< .05) before and after '1'? injections to castrates. 98 No increases in aggression or for any measure of mount- ing were seen during the treatment period. However, erection and genital manipulation received did increase during the period of testosterone administration. One of the more conspicuous changes was groom-solicit. Frequently a castrate would solicit and then receive grooming and or genital manipu- lation from another animal. Sometimes an erection would occur during this period of social interaction, often however, an erection was evident at the time of the solicitation. Figures 3.1 and 3.2 show 8-55 exhibiting this solicitation posture in its two forms. Erections and yawning often occurred during the grooming periods but both of these behaviors also occurred in situations seemingly independent of any social stimulation. The castrates themselves devoted very little attention to their own genitalia and no increases in the incidence of autogenital manipulation were obtained nor were any incidents of mastur- bation observed during testosterone treatment. For the older group of castrates six behavior categories showed changes in frequency during the treatment period, (Table 3.4). As had been the case for the younger castrates there were no increases in mounting or in aggression. In fact aggres- sion decreased significantly in frequency during the treatment period. Two behaviors, erection and autogenital manipulation, showed statistically significant increases during the testos- terone injection period. The measure of genital manipulation received was not affected by the testosterone treatment. Table 3.h 99 Behaviors of adult neonatally castrated male rhesus prior to and during a testosterone pro- pionate therapy period. All behaviors listed in the upper half of the table were scored for a one half hour period each test session; the behaviors listed in the lower half of the table were scored for a five minute period each test session. Means are for fifty day test periods. 100 Table 3.4 Behavior: Before TP DuringTP Single or double foot clasp mounts: 6 4:2 1 4'711’0 I: Presents: 9.3 1 1.9 4.2 1 1.4 Erections: 3.9 1 1.9 33.3 1 9.1* Yawns: 0.510.3 58.1113.8* * Aggression: 3.3 1 0.7 0.7 1 0.3 Brief contact 6 rough E: tumble play: 3'1 i 1'9 O 1 :- 0 l a Groom 4.5 1 1.3 3.7 1 1.6 Groom solicits: 1.1 1 0.8 0.7 1 0.2 Groom received: 6.4 1 0.6 8.7 1 3.1 Autogenital * 'pulation: 2.1 1 1.2 S 3 1 2 1 Genital manipula- 0.6 + 0.2 1 4 + 0 6 tion received: Mean 1 S. E. * Significant difference (using 1 test for matched pairs, p < .05) before and duringTP administration. LAJ C) s": 101 Autogenital manipulations by these animals were occasionally of a masturbatory nature and ejaculations by way of masturba- tion occurred 16 times for 4874 and once for 4859. Figure 3.4 is of 4874 during a period of autogenital manipulation. Castrate 4863 exhibited two ejaculations with female 4868 during the period of testosterone therapy; however, prior to testosterone therapy he had exhibited the ejaculatory reaponse pattern four times with this same partner. During the period of testosterone administration the castrates did expel visible quantities of seminal fluid. No groom solicit increases were noticed for these older castrates; and there were decreases in grooming by these animals. Table 3.5 presents a pretreatment test run behavior com- parison for the young animals and the older castrates. Fre- quencies for certain sex behavior categories differed markedly between these two age classes of males. In terms of mounting behavior the older castrates were showing almost exclusively the mature mount pattern. The younger males occasionally showed mature mounts, i.e. single or double foot claSp mounts, but the ratio of mature mounts to immature mounts was only about .45 on the average. The older castrates were also showing a higher incidence of erection. Aggression was also higher for the older males. Play by the older castrates was infrequent but grooming occurred at a fairly high frequency. These dif— ferences in social interaction appear strictly deve10pmentally related and gonadal hormone independent. 102 Figure 3.4 Castrate male h874 showing autogenital manipu- lation while under testosterone prOpionate therapy. Table 3.5 104 Behaviors for young male, young castrate male, and adult neonatally castrated male rhesus during a fifty day test period. All behaviors listed in the upper half of the table were scored for a one-half hour period each test session; the behav- iors listed in the lower half of the table were scored for a five minute period each test session. 105 Table 3.5 Behavior: Young} and 3 01d 1‘ Disoriented 6 deficient mounts: 3'7 :- 1‘9 0'5 -+- 0'5 No foot claSp ts: 1.1 1 0.5 0.0 1 0.0 Single or double * foot clasp mounts: 0.5 1 0.4 6.4 1 2.1 Presents: 5.6 1 1.8 9.3 1 1.9 Erections: 1.2 1 0.4 3.9 1 1.9" Yawns: 0.9 1 0.4 0.5 1 0.3 a Aggression: 0.5 1 0.2 3.3 1 0.7 Brief contact a 1 * rough 5: tumble play: 12.9 1 2.5 3.1 1 1.9 Groom: 0.2 10.1 4.511.3" Groom solicits: 0.5 1 0.3 1.1 1 0.8 Groom received: 0.1 1 0.1 6.4 1 0.6. Autogenital manipulation: 0.8 1 0.5 2.1 1 1.2 Genital manipulation 1.4 1 0.4 0.6 1 0.2 received: Mean 1 S. E. 1Mean here is for young castrates only, since young castrates differed significantly from young intact males . 1: Significant difference (using 1 test, p < .05) between old and young males and old castrates. 106 In only one group, Wisconsin 2, was there a change of social hierarchy during the period of testosterone treatment. Table 3.6 summarizes data on dominance shifts. The occurrence of this social order shift allowed a consideration of the fash— ion in which social position exerts an effect on the display of certain behavior patterns which appeared testosterone stim— ulated. In Wisconsin 2 the two castrate males S-45 and S-SS rose in the dominance hierarchy from positions 4 and 6 to po— sitions 1 and 3, respectively. This occurred on day 39 of social testing during the period of treatment. Table 3.6 presents the frequency data from the six behaviors for which testosterone treatment values differed from those of the pre— test period. Erection and yawn were the two behavior patterns which exhibited the most marked effects. It should be mention- ed in this regard that the values for the ten-day period pre- ceding the dominance shift were very typical of those shown during other ten-day testing blocks as reference to table 3.6 substantiates. The data for the four castrates that were followed through a 50—day post-treatment test run are presented in Table 3.7. Brief contact & rough and tumbleplay showed no difference across the comparisons for the pre-treatment, treatment, and post-treatment period. This was also the case for the measure cf groom solicit, but the pattern of changes for these two behaviors was dissimilar. In the case of play there was an increase in the level from the pre-treatment 107 Table 3.6 Behavior scores for two castrate rhesus prior to and following a dominance shift during testosterone treatment. 108 Table 3.6 Castrate S-4S Castrate S-SS Behavior: A. B C .A B C Erection 2.3 3.0 21.0 45.0 59.0 131.0 Yawn 54.5 88.0 105.0 4.5 6.0 40.0 Brief Contact 8 Rough & Tumble 32.3 26.0 17.0 10.5 9.0 6.0 Play Groom Solicit 5.3 3.0 5.0 17.0 15.0 21.0 Groom Received 1.5 1.0 0.0 3.3 3.0 0.0 Genital “mm" 8.5 7.0 4.0 8.5 10.0 9.0 lation Received A.= the mean frequency for ten-day test blocks prior to dominance shift B = the absolute frequency fbr the ten-day test block prior to dominance shift C = the absolute frequency fbr the ten-day test block fbllowing dominance shift 109 Table 3.7 Behavior scores for young castrate male rhesus prior to, during, and after testosterone prOpio- nate treatment. Means are for fifty day test periOds e 110 Table 3.7 Anim1 Pre Post Behavior No. R ___lg__ a. —, d. Erection: s—24 0.4 14.0 3.6 Pre H vs. 8* s-31 5.2 43.4 13.4 I: vs. posta . s-45 0.0 6.0 1.6 5'55 2a 9.2;}. M SE 1.9 31.4 8.9 Yawn: S-24 0.0 22.8 1.6 Pre B- vs. I? 5-51 4.0 42.4 4.6 H vs. Post 8* 5-45 0.0 64.6 7.8 5'55 M 2A5. L3. 3? 1.0 35.3 4.2 Brief Contact 5-24 12.8 25.2 27.4 Pre B vs. 8 n.s. 3 Rough and s-31 13.8 26.2 26.8 I: vs. Post a n.s. Table Play: 8-45 20.0 29.2 29.4 5'55 M .949. fl 31' 15.8 22.5 23.1 Groom s-24 0.0 1.0 1.2 Pre B vs. a n.s. Solicit: S-31 0.0 4.6 4.6 B vs. Post R n.s. s-4s 0.6 5.2 0.4 5'55 2.4. 11;! .L-E i 0.7 7.1 1.9 Groom s-24 0.0 1.0 1.0 Pre B vs. 1» Receive: S-31 0.0 3.2 2.4 B vs. Post R n.s. s-45 0.8 1.2 0.0 5'55 °_-_2. 2.19. M SE 0.3 2.0 0.9 Genital s-24 2.0 3.6 1.4 Pre Rvs. R" Bhnipulation S-31 1.0 8.4 2.8 H vs. Post R‘” Received: 5-45 4.8 7.6 0.0 5'55 33.4. ' 9.:9. 222. i 2.5 7.1 1.8 “Significant difference (wing 1 test for matched pairs, p < .05). "Significant difference (using 1 test for matched pairs, p< .025)- ***S'1gnificant difference (Using 1 test for matched pairs, p < .001)- 111 period during the time of testosterone injections and the level of play continued at this level during the post—treat~ ment period. In contrast to play the groom solicit measure ’ showed an increase during the treatment period followed by a decline during the period of testosterone withdrawal. 3.4 METHODS FOR PART II: TESTOSTERONE STIMULATION AND PAIR TESTING Sggject and treatments: The same group of males used in Part I of this series of experiments was used in Part II. In addition, a total of eight females were used in Part 2. Five of these females were used as partners for the young males and castrates. All of these five were from the social groups Wisconsin 1, 2 and 3. These females were given subcutaneous silastic tubing implants of crystalline estradiol. The pro- cedure employed for the making of these implants and their placement was similar to that which has been described (Karsch, Dierschke, Weick, Yamaji, Hotchkiss, and Knobil, 1973). In the case of these young females the exposed estrogen packed silastic tubing was a length of approximately 1.5 cm. No testing was carried out with the females until two weeks follow- ing estradiol implantation. For the older castrates a total of three females were employed as stimulus animals. One of the females, 4868, was familiar to the castrates, a second female, 4875, had been similarly reared and was similarly aged to 4868 but was an unfamiliar partner to the males of this 112 study; the third femlae, 1559, had been wild-caught and a mem- ber of the Wisconsin Regional Primate Research Center Breeding Colony for 2 years but was estimated to be of similar age to the other females. These females received implants of an estrogen packed silastic tubing with an exposure length of approximately 3 cm. They also received a two week post—oper— ative period prior to the beginning of behavioral testing. When the pretesting of all the males had finished they were put on a daily testosterone injection regime of 1 mg/kg/day. The young intact males received daily corn oil injections of .05 cc/kg/day. The animals received at least one month of such injections prior to retesting with the females. Testigg: At a mean age of 2.1 years, Which was at least 195 days following previous androgen treatment, the young castrate and intact males were given pretreatment sex tests with the familiar and unfamiliar estrogen=implanted stimulus females. At a mean age of 5.1 years, which was 86 days follow- ing previous androgen treatment, the castrates of group 43 were given pretreatment sex tests with their estrogen implant- ed stimulus female partners. All pair tests were of twenty minutes duration and were carried out in a balanced fashion. No male received more than one test a day.~ The tests were conducted in a large bare test room which contained a stain— leSs steel ramp and platform. A test began with the intro- duction of the female into the test room which already held the male. Two observers scored the pair's interactions util— 113 izing a scoring inventory which included a large variety of both male and female behaviors. Defnintions: In addition to scoring no foot clasp mounts, single foot clasp mounts, double foot clasp mounts, intromis- sions, ejaculations, yawns, erections and female presents, the following behavioral measures were scored. Approach: A directed walk at a moderate to fast rate towards the partner. Approaches may or may not be followed by a contact of the other animal. Contact: A light touching of the partner's perineum, back, hindquarters, or hips with one or both hands. Purselip: A facial expression in which the ears are drawn back as well as the brows and scalp while the chin is thrust forward and the lips puckered and rhythmically smacked. Threat: Any of the following behaviors or combinations of these behaviors which contain obvious and direct orientation towards the partner.' Contact threat: Brief bouts of gentle to moder- ate biting, shaking or slapping. Distinguished from aggression by degree of intensity. Bgdy lungs: A sudden body thrust with an intense and directed stare. May be followed by an aggres- sive interaction or cage shaking. Charge: A directed pursuit of another animal which may be accompanied by woofing vocalization and fol- lowed by body contact. 114 3.5 RESULTS FOR PART II Table 3.8 presents the behavioral data for the young castrates and males on their pretreatment and treatment test— ings. Only one measure, that of female present, differed sig- nificantly between the groups during the period of pretreat— ment testing. By the end of the first series of testosterone treatment tests the castrates differed from the intact males on the measures of purselip, and yawn. The measures of approach, contact, mount, intromission, and erection had higher means for the castrates, but the differences failed to reach a level of statistical significance. The castrates differed from the intacts at the time of the second testing during treatment, for measures of approach, contact, mount, erection, purselip, and yawn. By the first series of androgen treatment sex tests for the castrates the measures of approach and contact differed significantly from the pretreatment testing when matched pair 1_tests were carried out (p <'.005 for approach and p'<'.O5 for contact). Two castrates S-31 and S-78 showed intromissions during the first and second treatment testings. In addition, during both the first and second treatment test series, cas- trate S-78 ejaculated. Examination of the female partner on these occassions revealed evidence of seminal emission. The results for the older castrates were very individualis— tic. Three castrates 4859, 4875, and 4861 all showed approaches, Table 3.8 115 Sex behavior scores for young male and young castrate male rhesus prior to and during eXperi— mental treatment. The scores represent means for a total of five fifteen minute pair tests with young estrogen primed stimulus females. 116 I .moame 335 can mmpmuummo c0953 :8. v a .98» a one manna 88.8wa ucmuwmafim c: I moans 83R 98 mopmaummu 5853 So. v m .93» u wanna 8:08me 9:35.355 .3 .mode pumpfi can mounfiummo Eczema 3o. v a than p 05 manna codenamed 3803253 c K .m.fi «« «£fi «as «a v.m o.H m.o N.o m.m o.o v.m o.o N.N H. m.o 0.0 n.m 5.0 v.¢ w. o.m o.H 09.9530 can: mm amok. 9:953: Om.“ .mJH o.m u.o N.o o.o m.H N.o m.N o.o o.N H. w.o o.o ¢.N N.a N.m m.H v.v m.H cumuummu can: Z no.2. peasant. m.m macaw m . N a . o n.H o.o m.a m.o 0.0 0.0 m. N. N. 0.0 m.o o.H o.H H.n m.H m.H oumnpmmu cams” 3953505 ucomoum came—om umoafi. :33on can» 3.39:5 scammaouufi 9:392. eunucou Beaded. "Swanson 117 contacts, and mounts of their female partners during their pretreatment testing; in addition 4874 displayed an intro- mission. Following treatment, all four castrates displayed higher frequencies of approach, contact, and mount, and 4874 showed an increase in intromissions. Erections, yawns, and purselip were also diSplayed at higher frequencies during the first series of treatment test. In this regard no yawns or purselips, were seen during the pretest period and only one male (4874) had any erections. No consistent trend was appar— ent for measures of threat and aggression. The second round of pair tests was very similar to the first for all measures. No ejaculations were seen by any of the castrates during the period of androgen treatment, however, masturbation in- cidents occurred for two castrates, 4859 and 4874. 3.6 DISCUSSION In contrast to several previous studies (Mirsky, 1955; Green $1.11,, 1972) clear behavioral differences were obtained between castrates and intacts while under testosterone treat- ment and in the group testing situation. The behaviors of erection, yawn, groom solicit, groom receive, and genital manipulation all increased in frequency during the period of testosterone stimulation. The castrates-also exhibited be— havioral changes under the pair testing condition and the be- haviors which increased in frequency included approach, contact, mount, purselip, yawn and erection. 118 The concern of this study was with behavior and not dom- inance interactions within the group and because of this a contrast to those studies which were directed towards an examination of dominance hierarchy changes may be inap- prepriate for the data reported on social behavior in the pre- vious studies were limited. In only one of the four social groups examined in this study was there a shift, using fear grimace matrix, in dominance order during the period of tes- tosterone treatment. A similar occurrence was reported by Joslyn (1973) in which androgen injections were given to young females which were reared in a fashion similar to the castrates of this study and for which a dominance shift was reported during the treatment period. In the present study for two of the three groups no dominance changes occurred, but the dominance positions of the castrates were high to begin with in these groups. In a third group, Wisconsin 3, the castrates occupied low dominance positions. Possibly a shift might have occurred in Wisconsin 3 as a result of the testosterone treat- ments. Castrate S—78 did attempt to assert himself with male S-77 but these attempts were rebuffed by S-77 acting in unison with female S-75. It would appear that dependent rank is an important factor which may well affect the outcome of studies which have attempted to manipulate dominance through hormonal manipulation (Kawai, 1965; Varley and Symmes, 1966). In this regard, attention might be drawn to male S-47. Prior to the start of the treatment period, S-47 occupied the fifth 119 social rank position in his group. His preferred play part- ner was castrate S—45. During the treatment period a dominance shift occurred in this group and male S—47 moved to the second rank position. This male had received only oil during the entire treatment and his shift is probably best explained by a dependent rank process. It would appear that within a group setting social factors may well interact in a fashion to offset any influence gonadal hormones may play in the deter— minationcf a social hierarchy position. Although some studies (Rose, Holaday, and Bernstein, 1971; Rose, Gordon, and Bernstein, 1972) have suggested a relation- ship between testosterone and aggression for primates, the causality of the relation appears questionable (Eaton and Resko, 1974). No increases in aggression were seen in any of the groups of this study. In fact a further analysis was carried out on two motor components frequently associated with rhesus agonistic behavior, the "woof threat; and the "ear flip". And again no increases were found for these measures when the pretreatment period was compared to the treatment per- iod. One study has been carried out in which increases in aggression were reported to accompany androgen administration (Trimble and Herbert, 1968). This study employed adult ovar- iectomized female rhesus and similar hcreases have not been reported in studies which have utilized adult castrate male rhesus (Phoenix, Slob, and Goy, 1973; Michael and Wilson, 1974). It is at least possible that the physiological variables 120 affecting the diSplay of aggressive behavior for the two sexes are different. Consistent with this interpretation are sev- eral studies which have dealt with central nervous system control of aggression. Kling (1974) reported that female rhesus monkeys, when given amygdaloid nucleus lesions increased their levelsx of aggression while males which received the same lesion failed to show such increases. Work on cats (Inselman— Temkin and Flynn, 1973) utilizing central nervous system stimu— lation for the elicitation of attack has suggested that the neural substrates critical to this behavior in males and females reSpond differently to hormone priming. It would appear possible to suggest that aggressive behavior is mediated through different neural networks in the two sexes and that the effects which testosterone have on this behavior are closely related to those neural and physiological substrates critical to its expression. Work carried out on genetic female rhesus monkeys which were exposed to testosterone 11121312 (Goy and Phoenix, 1971) indicates that testosterone may well have an influence on the level of aggression exhibited by adult rhe- sus monkeys, but its effects on aggression are achieved by early hormone exposure and one intimately related to situational circumstances as well as individual behavioral prOpensities (Bielert, unpublished data). The social and individual behavior categories which were affected by testosterone treatments in the group testing sit- uation were different for the two age groups of castrates. 121 In the case of the three groups of young castrates, the most noticeable changes were for those behaviors concerned with grooming. The appearance of thebody display described as a groom solicitation was an unexpected outcome in this study. A number of observations on several species (Sparks, 1969) have documented the findings that grooming sessions are fre— quently associated with sexual interaction and may be influ— enced by endocrine factors (Michael and Herbert, 1963). Data from the present study suggests that for males the groom soli- citation may have a degree of testosterone dependence. In this regard it may be that groom solicitation postures are age related and that the older castrates in our study employed solicitation postures other than the two which were scored under the taxomony employed in this study. In contrast to an interpretation of age dependency, it could be suggested that the behavior response seen under this testing situation by the young castrates was idiosyncratic and related to the specific animals and their rearing background. There is at least suggestive evidence that such is not the case. Kling (1968) has reported that when a group of young 2 year old males were given testosterone treatment and pair tested with other young males, "the most consistent effect of testosterone . . . was the increase in grooming behavior", though it is impossible to tell from the report if this increase was related to an increase in groom solicitation. Further evidence for a role of testosterone allogrooming comes from Trimble and Herbert (1968); they reported that when adult 122 ovariectomized female rhesus were given androgen "there was an overall increase in the mean number of grooming invita- tions offered by the female while they were receiving the three doses of testosterone compared with the series in which they were not treated". In addition, Michael and Wilson (1974) have reported that female grooming of a constant male partner decreased following castration of the male. It would appear that the effect of testosterone on groom solicitation and the display of grooming by rhesus is clear but is probably partly determined by the individual animal involved and his age. The finding that the young castrates exhibited yawning at higher frequencies under testosterone treatment may be inter— preted to mean that the neural substates critical to this response were fully responsive to androgen stimulation. It has been suggested that yawning is a testosterone stimulated behavior pattern for both sexes (Goy and Resko, 1972) and further data consistent with this position has been offered by Eaton and associates (Eaton et al., 1973). It is clear that this behavior pattern normally makes its appearance post- puberally in intact males (Bielert, unpublished data). Data following the dominance shift in group Wisconsin 2 make it clear that social factors play a role in the diaplay of yawn- ing. Though yawning is rarely shown by females it can be displayed at high levels under social conditions where the dominance relationship between a heterosexual pair is dominated by the female (Bielert, unpublished data). 123 The occurrence of erections differed significantly be- tween the two groups of castrates prior to testosterone therapy. In both groups erection frequency was significantly increased by androgen treatment. One explanation for the pre— treatment difference might be that of social behavior profile differences. As has been previously mentioned the older group spent more time in grooming sessions. Frequently during these sessions erections were noted and it may have been through the increased grooming activity that the erection scores for the two ages differed. Though the occurrence of erection was clearly testosterone related it appears that this behavior pattern is also influenced by social factors. Following the dominance shift in Group 2 both castrates showed increases in the frequency of this behavior pattern. The testosterone dose for these two castrates remained constant throughout the period of treatment and thus appears unrelated to the frequency shifts. It has been suggested that erection has a social sig- nal value (Wickler, 1969, 1972). In this regard, it may well be that social position affects peripheral testosterone levels and through this mechanism the occurrence of erection but it also appears that social position 233.23.15 also capable of affecting the display of this particular behavior pattern. One of the most dramatic differences between the two groups was the sudden appearance of autogenital manipulation in the group of older castrates during the period of testos- terone treatment and its absence during the treatment period of the younger groups. To the extent that this behavior may 124 be termed sexual it would appear that the younger castrates were insensitive to testosterone's affects on its display. The fact that the erections occurred for both groups under androgen treatment may be taken as evidence for reSponsive— ness of the peripheral and neural tissues. If the absence of increased autogenital manipulation for the young castrates cannot be attributed to physiological insensitivity it may be attributable to behavioral immaturity. It may be that one year old laboratory-reared male rhesus are still sexually immature in regard to their autoerotic behavior patterns and that social experience‘pgg.§g is a critical factor in the effectiveness of testosterone in altering this particular sex behavior pat- tern. The failure of mounting behavior to increase during testosterone stimulation during the period of social testing is difficult to explain since mounting behavior was occurring in all of the groups though its display did differ for the two age groups with the older castrates exhibiting almost ex- clusively the mature mount pattern which characterizes adult rhesus sexual behavior (Bielert and Goy, 1973). It may be that social interactions were an inhibitory factor in this regard for dominance does appear to affect the individual sexual behavior pattern components displayed by both males (Perachio, Alexander, Marr, 1973) and females (Herbert, 1967; Goldfoot, 1971). It was with this consideration in mind that the pair tests were conducted. Under the pair testing situation all 125 testosterone treated males did show increases in mounts dur- ing the period of treatment. These increases were particular- ly noteworthy in the case of the young castrates. An im- portant consideration in terms of the apparent discrepancy between the social and pair testing results for mounting is the fact that it was only under the pair test situation that the young males and castrates had exposure to estrogen—primed females. This fact alone, however, was not enough to stimulate sexual behavior in the young intact males but it may have been a factor in the failure of the young castrates to respond during the treatment period which accompanied the social testing. The increases in approach, contact, and mount by the young castrates are a clear demonstration of the sexual behavior potentiality which these young animals have. The occurrence of intromission and ejaculation by these young males during the period of treatment is the first report of such behaviors in rhesus of this young age and could possibly be described as precocious. An interesting sidelight in terms of the castrate's response to testosterone was the increase in frequency of the facial expression "Purselip". Chevalier- Skolnikoff (1973) has suggested that the lip smack expression known to be connected with sexual interactions of rhesus pairs (Lindburg, 1971) may be influenced by the female's hormonal state. The data from this study suggests that in addition it is also influenced by the individual male's hormonal state. During a period of testosterone treatment at the termination of the pair testing the young castrates were tested with adult 126 estrogen primed females. The young castrates showed great interest in the females and purselipped to these partners, in addition they also exhibited the stiff-legged walk which is scored as strut. This behavior pattern has also been de— scribed for adult rhesus pair interactions (Lindburg, 1971) and the observations from this study suggest that some element of the female partner influences this particular display. Eaton 11‘11. (1973) have also reported an increase in this social behavior pattern during androgen treatment of female pseudohermaphrodites. One of the most startling observations to come from this study was the occurrence of the ejaculatory response pattern in prepubertal animals and also in older castrates not receiv- ing androgen. These observations for the rhesus monkey are consistent with human. In the case of the human, orgasm can occur for the male far in advance of the onset of phys- iological puberty (cited as general knowledge by Bermant and Davidson, 1974). These observations suggest that social and hormonal variables connected with puberty Operate on an already complete neural substrate, and that prepuberal sexual behavior may possibly provide positive reinforcement to the individual. Such reinforcement may be very important in spe- cies such as the rhesus in which there is a lengthy develop- ment of sexual behavior prior to the onset of puberty. SECTION 4 SEXUAL BEHAVIOR OF ADULT LABORATORY-REARED INTACT MALE RHESUS MONKEYS: COMPARISON WITH WILD-CAUGHT ADULT INTACT MALES AND LABORATORY- REARED PREPUBERALLY CASTRATED ADULT MALES 4.1 Introduction The rhesus monkey (Macaca mulatta) is a comparatively long-lived animal, and shares with many other primates includ- ing man a period of extended social deve10pment preceding sex- ual maturity. The sociality of primates and their well—organ— ized and stable interaction networks are features which are firmly rooted in phylogeny (Gartlan, 1973). When young rhe- sus are prevented from interacting with peers or reared with- out maternal care, serious and long-term effects on individual adjustment and social behavior may be observed (Harlow and Harlow, 1969). When the social testing has been carried out in adulthood, a time when infantile gonadal state should no longer be a consideration for the behavior of either the male or his stimulus partner, the sexual behavior of these restric— tion reared animals was revealed to be deficient (Harlow, 1965; Harlow, Joslyn, Senko, and D0pp, 1966). It would appear from the work which has been carried out that the social and sexual behavior patterns of male primates are sensitive to ear- ly social and environmental restriction. This proposition appears valid for both catarrhine monkeys (Meier, 1965; Missakian, 127 128 1969) and apes (Rogers and Davenport, 1969; Riesen, 1971). At least in the case of the rhesus these deficits don't appear to be moderated even after lengthy periods of rehabil- itative social exposure (Missakian, 1972). Though early reports were difficult to interpret (Thorek, 1924; Zuckerman and Parkes, 1938; 1939) it is now clear that the sexual behavior of the adult male primate is adversely affected by gonadal insufficiency. Bilateral gonadectomy has been shown to significantly affect the occurrence of the ejacu- latory pattern as well as the occurrence of intromission, mounting, spontaneous erection, and yawning (Michael, 1972; Michael, Wilson and Plant, 1973; Phoenix, Slob, and Goy, 1973; Michael and Wilson, 1974). Behavior can be restored to cas- trate male primates through exogenous androgen (Michael 11_g11, 1973; Phoenix 1973; Phoenix g1_§1., 1973; Michael and Wilson, 1974; Phoenix, 1974). Consistent with the effects of castra- tion are those obtained following treatment with the anti- androgen, cyproterone acetate. Cyproterone acetate treatment lowers peripheral androgen levels and in addition causes a deterioration in sexual behavior (Michael, Plant, and Wilson, 1973). Recently papers have attempted to assess the social and environmental factors which may have a determining influence on the development and exhibition of mature sexual behavior for primates under laboratory conditions (Goy and Goldfoot, 1974; Goy, Wallen and Goldfoot, 1974). However no attempt has yet been made to assess the extent to which social restric- 129 tion and gonadal state may interact to affect male primate sexual behavior. With this in mind an attempt will be made to describe and characterize the sexual behavior of adult laboratory reared male rhesus and place this behavior in per- spective through comparisons with wild.caught adult males. In addition, attention will be given to the sexual behavior displayed by adult neonatally castrated laboratory reared male rhesus. Further data will be presented for these cas- trates from a series of pair tests conducted while testosterone propionate therapy was being given. 4.2 Methods Subjects. The animals selected for this paper's analysis were all adult male rhesus of at least four years of age at time of testing. It has been reported that males of this age exhibit high peripheral levels of testosterone (Resko, 1967). A total of nineteen laboratory reared males were tested. Five of these nineteen were castrated. The castrations were performed on the day of birth in the case of two animals and the remaining three were castrated at approximately three months of age.. All castrations were performed under general anesthesia by way of a subumbilical incision. The Opera- tions were carried out in a fashion consistent with routine sterile surgery practice. A total of ten wild-caught adult male rhesus monkeys made up the wild-caught test group. Eight were proven breeders and had been members of the Wisconsin 130 Regional Primate Research Center Breeding Colony for at least a year. The females used were fully adult wild-caught rhesus which had been bilaterally ovariectomized. The females were brought into a receptive condition through estradiol benzoate injections of 10 ug/day or silastic tubing implants of crystal— line estradiol. The implants were placed subcutaneously in the subscapular region of the back. The surgical placement of the implants was done under anesthesia. The implants were similar to those which have been previously described (Karach, Dierschke, Weick, Yamaji, Hotchkiss, and Knobil, 1973). The silastic tubing had only a 3 cm crystalline estradiol exposure length packed. All animals were caged individually and maintained on commercial monkey chow supplemented with vitamins and periodic fruit and vegetables. Rearigg procedures. All of the laboratory-reared males tested received similar social experience during their first four years of life. The males were reared under what has been described as a peer—group procedure (Phoenix, Goy, and Young, 1967; Goy and Goldfoot, 1974; Goy, Wallen and Goldfoot, 1974). These groups varied in total membership from 4—6 individuals. All groups had at least two female members and varied in terms of whether the remaining members were intact males, neonatal castrates, female pseudohermaphrodites or a combination of several of these treatment classes. All infants were allowed to live with their mothers until the age of approximately 3 months, at which time they were separated from their mothers, 131 weaned, and caged individually. Peer—groups were made up of similarly aged animals and once a peer-group was set up as such, its membership remained constant. Daily one-half hour group tests were carried out 5-days a week for a total of 100 test days during thefirst year of life. For a daily testing, all the animals of a group were removed from their home cages and released into a small testing room which contained a stain- less steel ramp and platform. They were allowed to interact for one-half hour after which they were returned to their home cages. Following the 100-day test run, animals remained caged individually. The animals of a peer-group were again tested for social behavior when the mean age of the animals in a group reached 12, 24, 36 and 48 months. The groups were tested 5-days a week for 50 test days during these later social test runs. The only deviation from the previous test run was that the animals were tested in a slightly larger test room during these runs. Animal care and maintenance procedures were relatively constant and no unique manipulations were carried out on any of the peer-groups from which the animals with which this study is concerned were from. At various times and for varying periods of time peer=groups received 24 hour continuous living experience; during these periods the entire peer group was released into a large indoor—outdoor cage facility. Periodically animals of a specific peer-group were given round-robin testing during which two members of a peer- group would be released into a testing cage and allowed to 132 interact for a ten-minute period. These exposures were systematic and occurred during the animals' second, third and fourth years of life. Data collection. The data analyzed in this paper were collected over a period of time at both the Wisconsin Regional Primate Research Center and the Oregon Regional Primate Research Center. As a result of this not all animals saw the same female partners, but the partners were always strange to them and were always sexually experienced and fully adult. The males were tested on different occasions and saw a total of -from 3 to 6 females. The data for a particular male represents the mean for all the tests which he received. All tests were of 15 minutes in duration. The testing was conducted in either a testing cage or a small testing room. At the start of a test the female was introduced by way of a guillotine door into the testing area which already contained the male. Scoring of the pair's interactions was done by experienced observers utilizing a standardized scoring taxonomy. No attempt was made to hide the observers, although no efforts were made on their part to interact with the pair. Definitions. A total of five behaviors were scored whenever they occurred, no matter which animal exhibited them. 133 1. Fear_grimace: A facial expression in which the lips and checks are retracted so that the teeth, sometimes tightly clenched, and gums are exposed and deep furrows appear on the checks. The ears are usually flattened back and there may be a tenseness of body musculature. 2. Yawn: A hyperextension of the mandible and retrac- tion of the lips which exposes the teeth. This expression closely resembles that of the yawn displayed by human beings. 3. Groom: A stereotyped sorting and or brushing aside of the fur, very often accompanied by lip smacking. This behavior pattern may involve one or two hands and was only scored when directed towards another animal. 4. Aggression: ,A vigorous biting accompanied by head shaking as if to tear or rip. 5. Threat: Any of the following behaviors or combi- nation of these behaviors which contain obvious and directed orientation towards another animal. a) Contact threat: Brief bouts of gentle or moderate biting; shaking or slapping. Distinghished from aggression by degree of intensity. b) Body lungs: A sudden body thrust with, an intense and directed stare. This may be followed by genuine aggression or possibly cage shaking. c) Charge: A directed pursuit of another animal which may be accompanied by a woofing vocalization and followed by body contact. 134 A total of six behaviors were scored for only the male partner. 1) Contact: The male lightly touches the perineum, back, hind quarters, or hip (s) of the female with one or both hands. 2) M2211: An erect stance associated with coitus in the monkey, in which the hands are placed on partner's hips, peri— neum or back. One or both feet of the male clasp one or both legs of the partner. Pelvic thrusting may occur. The rhesus monkey usually exhibits a series of mounts which may or may not include intromissions preceding an ejaculation. 3) Intromission: Penile insertion into the vagina recog— nized by deep and regular pelvic thrusting. The rhesus monkey characteristically exhibits several intromissions preceding an ejaculation. 4) Ejaculation: Recognized during copulation by a series of deep pelvic thrusts followed immediately by full- body tenseness, immobility and slight quivering of the flanks. 5) Purselip: A facial expression in which the lips are slightly extended and rhythmically opened and closed. This expression is usually accompanied by a neck thrust and often associated with a strut like locomotion. During this strut the male walks in a stiff legged fashion with his tail crooked and curled over his back. 6) Erection: An engorgement of the penis sufficient to permit extension of the glans beyond the prepuce. This tur- gidity may or may not be accompanied by penis flips. Erection 135 was scored when obvious such as during grooming bouts, often preceding occassional mounting attempts, as well as under sit- uations of seemingly unprovoked circumstances. The behaviors scored for the female alone included the following: 1) Presenting: A stereotyped posture of receptivity characterized by quadrupedal immobility orientation of the perineum towards the partner, tail deflection, and modest extension of fore and hind limbs. Presenting was classified during testing as follows: a. Presenting to contact: Presenting posture assumed as partner makes a contact. b. Presenting spontaneously: Presenting without physical partner interaction. 2) Proximity: A behavioral measure of interindividual distance. A female was scored for proximity when she was seated within one foot and a half of her seated partner. Body orientation was not a part of the criterion for this behavior. 4. 3 RESULTS Comparisons between the two intact groups revealed marked differences for several measures of male sexual behavior. Peer-reared intact males contacted, mounted, intromitted and ejaculated at statistically significant lower levels than wild- caught males. Yawning frequency also differed between these 136 groups with the wild-caught males exhibiting this behavior at a significantly higher frequency than the intact peer-reared males. There were no differences for the measures of purselip or spontaneous erection, although the wild-caught males did purselip more and exhibit erections less frequently than the intact peer-reared males. Grooming by the males was rare and no difference between the groups was obtained for this measure. The agonistic behavior measures of threat and aggression also failed to differ significantly between the two intact groups (Table 4.1; Table 4.2). Behavior by the female partners was exhibited at signifi- cantly different levels for several patterns. Spontaneous presents occurred at a higher level during the tests with the intact peer-reared males. In marked contrast to the spontan- eous present measure presents to contact occurred at a much higher frequency when the female was with a wild-caught male, although in this regard it should be noted that the wild- caught males were showing more contacting. Females also showed proximity more frequently to the wild-caught males. The agon- istic behavior of the females appeared greatly affedted by the type of partner she was with. No female aggression was ever directed to a wild-caught male. Females also directed fewer threats to wild—caught males than they did to the intact peer-reared males (Table 4.1; Table 4.3). Comparisons between the two peer-reared groups also re- vealed significant differences related to the gonadal state of the male being tested. Castrate males exhibited fewer contacts 137 Table 4.1 Sex behavior scores for tests of castrate peer- reared, intact peer—reared and intact wild- caught adult male rhesus monkeys. 138 Table 4.1 Castrate Intact Intact Behavioral Measures Peer-reared Peer-reared Wild-Caught N - 5 N - 14 .\' - 1o Omtact 0. __ 1 1.7 1 0.5 4.6 1 0.6 bunt 0. 1 l 1.6 1 0.4 3.7 1 0.5 Intromission 0.6 1 0.3 1.7 1 0.3 Ejaculation 0.1 10.1 0.4 1 0.1 Male Erection 0.3 1 0.3 1.5 1 0.4 0.7 1 0.3 Purse Lip 0.9 1 0.5 3.1 1 1.2 6.5 1 1.7 Yuan 1.7 1 1.4 3.1 1 1.0 9.6 1 2.4 Aggression: 0.6 1 0.2 0.1 10.0 0.1 1 0.0 Threat 2.2 1 0.8 0.8 10.1 0.5 1 0.2 Groo- 0.3 1 0.2 0.9 1 0.5 0.1 1 0.0 59m 1.810.3 3.0 10.5 1.1 10.2 Cmtact Present 0.1 1 0.1 1.3 1 0.4 3.2 1 0.5 Peale Proximity 0.2 1 0.1 0.6 1 0.2 3.9 1 0.6 You 2.5 1 0.6 0.1 10.1 0.4 1 0.1 Aggression 0.9 1 0.4 0.1 10.0 Threat 1.l 1 0.4 0.6 1 0.2 0.1 1 0.0 M 1 S. E. for a 15 minute pair test. 139 Table 4.2 Results of the statistical analyses of male sexual behavior during pair tests for castrate peerareared, intact peerbrearcd, and intact wild- caught adult male rhesus monkeys. Behavioral Measure: Contact : Wild -caught Intact peer Mount : Wild-caught Intact peer Intranission: Wild- caught Intact peer Ejaculation: Wild- caught Intact peer Erection: Purse lip: Yawn: Wild- caught Intact peer Aggression: Wild-caught Intact peer Threat: Wild- caught Intact peer Male groom: VS. 5 '5 V5. V3. 5 5 V5. 140 Table 4.2 intact peer . castrate . intact peer . castrate . intact peer castrate intact peer castrate . intact peer . castrate intact peer . castrate . intact peer castrate 13.57 10.35 5.83 6.84 2.22 2.85 5.73 9.78 5.88 1.16 Level of Sigiificance p < .001 p < .01* p < .01“ p< .005 p < .01* p < .01* p <.01 p< .01* p < .05" p < .005 p < .01" n.s. n.s. n.s. p < .01 p < .01* p < .001 n.s. p < .01" p < .01 n.s. p < .01* *Post hoc comparisons. (Newman-Kuels test) 141 Table h.3 Results of the statistical analyses of female sexual behavior during pair tests with castrate peer-reared, intact peer-reared, and intact wild-caught adult male rhesus monkeys. 142 Table 4.3 Behavioral Measure: _F__ Level of Si gificance Spontaneous Present: 5.49 p < .025 Wild-caught vs. intact peer p < .01* Intact peer vs. castrate p < .01* Contact Present: 8.08 p < .005 Wild-caught vs. intact peer p < .01* Intact peer vs. castrate p < .01* Proximity: 25.93 p < .001 Wild-caught vs. intact peer p < .01* Intact peer vs. castrate n.s. Yawn: 33.47 p < .001 Wild- caught vs. intact peer n.s. Intact peer vs. castrate p < .01" Aggression: 8.57 p < .005 Wild— caught vs. intact peer n.s. Intact peer vs. castrate p < .01* Threat: 4.38 p < .025 Wild-caught vs. intact peer p < .01* Intact peer vs. castrate p < .01* *Post hoc comparisons . (Neman-Kuels test) 143 and mounts than their intact counterparts. No castrate male was ever seen to exhibit an intromission or the ejaculatory response as defined for this study (Table 4.1; Table 4.2). One of the castrates, however did masturbate to a point of exhibition of the ejaculatory response. No seminal emission was seen but the flank quivering and the general body tense- ness was indistinguishable from that which has been described for the rhesus during ejaculation (Bielert and Goy, 1973). Interesting in this regard was the great similarity between the autogenital stimulation pattern and that which could normal- ly be expected to be associated with coitus. The rhesus ex- hibits a series of intromissions with pelvic thrusting prior to ejaculation and only ejaculates on a single intromission. The masturbatory pattern shown by this castrate consisted of a series of autogenital stimulation bouts finally culminating in the ejaculatory response pattern. -The great regularity between these stimulation bouts was similar to what is typical of normal heterosexual sexual interaction. The masturbatory behavior of the castrate in this study was similar to that of other intact males in regard to the pattern of stimulation exhibited. It is tempting to speculate that for Species in which the multiple intromission pattern is typical the mastur- batory behavior will follow a similar stimulation sequence, and that for species in which only one intromission precedes ejaculation a single autogenital stimulation session will frequently result in ejaculation and be the typical pattern shown during masturbation. 144 Castrates also exhibited fewer yawns than the intact peer reared males. The castrates failed to differ significantly from the other groups on the measures of purselip and erection, but they did exhibit the lowest mean values for both of these behavior patterns as reference to Table 4.1 clearly indicates. In terms of the measures of agonistic behavior followed in this study, the castrates exceeded all other groups. They differed significantly from the intact peer—reared males for measures of both threat and aggression (Table 4.2). Female interactions with castrate peer-reared males set them apart from the other groups in several respects. Cas- trates received significantly fewer spontaneous presents than the intact peer-reared males, and their tests also had fewer presents to contact. Females yawned much more when in the presence of a castrate male, and in addition they also exhibit- ed higher frequencies of threat and aggression to their cas- trate partners. In at least one regard the castrates were treated similarly to the intact peer-reared males and this was for the measure of female prozimity. Females did not move into proximity of the castrates with any great frequency (Table 4.1; Table 4.3). Though the differences across the groups were clear cut, a check of the data revealed a high number of nonresponders. Since one of the concerns of this analysis was that of the way in which sexual behavior had been affected, further data anal- ysis was carried out for the three groups involved. Figure 145 4.1 presents a graphic illustration of the way in which the gonadal state and rearing condition affected the measures of male sexual behavior examined in this study. For all measures examined the percentage of responders from the castrate group was the lowest. In addition it was only for the measures of contact and mount that the intact peer-reared males exhibited a response level of at least 50%. As one makes comparisons between the behavioral cate- gories presented, it is clear that the progression from contact to mount to intromission and finally ejaculation corresponds to a steady decline for all groups in terms of percentage of animals responding. This decline is the most rapid for the castrate group and least for the wild-caught male group. This progression is probably related to the mating pattern characteristic of the rhesus. This pattern is characterized by a series of mounts, some of which are accompanied by intro- mission and thrusting, following a series of intromissions, ejaculation is finally achieved. It was decided that for the two behavior patterns for which both the intact peer-reared males and the wild-caught males exhibited at least a level of 50% responders, comparisons would be made. Limiting analysis to responders only the intact peer-reared males differed from the wild-caught males in terms of frequency for measures of both contact and mount (contact, unpaired 17test = 2.67, df = 18, p (_.O1; mount unpaired 1¢test, t = 2.75, df = 16, p < .01 (Winer, 1962). 146 Figure 11.1 Percentages of castrate peer-reared. intact peer- reared, and intact wild-caught males exhibiting various measures of male sexual behavior. 147 - assume PEER mince 74-5 [:1 m7 PEER mo 74-14 E NTACT no CAUGHT mo T | 1!!! HHH I J i ‘HllIIIIHIHIlHllllHHlillHlilil ‘1 H 1 $WMLS mm NATLEASTGE TEST o5888883888 HHIHI!IF!1HlWIHIM1INllII!HilHHI1IHIlllllHHHllliiliHii !HllllIHIHIHIHIIIHIHHlNHHilllllllHlHllllll‘lllH 8.1HHHHIHIHHIHIHHHUIl so 59 80 so 59 noun numussuou BEND/ORAL WEOORES TESTS 53 Ed 8 8 W l ! I W, 'H HIIHIHI i Hlllllil l I 1 HHHIIHH II t l ! E 2 WI! 1 148 The agonistic interactions between the pair were also examined across the three groups. Figure 4.2 presents a graphic representation of the data. All castrate males dir— ected threats and aggression to their female partners. All intact peer-reared males also threatened their female partners although not all of the males aggressed the same animals. The wild-caught males had the lowest percentage of responders for theSe measures. Females responded agonistically to the castrate males and at a somewhat lower level to the intact peer-reared males. An examination of figure 4.2 makes it clear that no matter what the sex or gonadal status of the ani- mal aggression as defined for this study was exhibited by a lower percentage of the animals than was the case for the measure of threat. It was the disprOportionate frequency of agonistic behav- ior that led to a consideration of dominance interaction be- tween the pair and its possible affect on thebehaviors exam- ined. Dominance criteria were set up as follows: 1) an animal was considered to be the dominant animal of a test if its partner directed a fear grimace to him or her. 2) an animal was considered to be the dominant animal of a test if its threats or aggression were totally unreciprocated. The entire group of 149 tests was examined and using the two criteria in order listed, all tests were classified as one in Which the male was dominant, one in which the female was dominant, 149 Figure 4.2 Percentages of males and females showing agon- istic behavior during their sex tests. 150 CLASSES 0F MALES CASTRATE ER NTET PEER NTET WILD REAIED HIS REARED NIH CWT NOD - MEAT [:1 AGGRESSION TESTS 30303030 59595959 60606060 MALE LEW—J IMALE LEM IMALE LEMMA SEXOFACTOR 151 or as one in which no decision could be made using the cri— teria as set up. Figure 4.3 gives a graphic representation of the data. For both intact groups, the males were more fre- quently given the dominant animal status. However, females were dominant more frequently than males when their partners were castrate peer-reared males. The percentage of tests for which a dominance decision could not be made was very similar for both groups of intact males. In contrast to this the percentage of undeterminable tests for the group of cas- trates exceeded the percentage for which a dominance decision could be reached. When all the pair tests were considered, males were dominant to females more frequently than they were subordinate to them. The frequencies with which certain behaviors occurred seemed strongly influenced by the dominance relationship of the pair. Figure 4.4 presents data on the effect which part- ner dominance had with respect to the occurrence of certain behavior patterns. The data are from a total of 83 of the 149 tests for which a dominance decision was possible. Males were dominant in a total of 67 of the tests.’ This is equivalent to 45% of all tests given or 81% of tests for which a dominance decision with the criteria of this study could be made. Fe- males were dominant to their partners in 16 tests. The tests occurred across all three groups. The females were the dom- inant pair member in 11% of all the tests and 19% of the tests for which a dominance decision could be reached. The 152 Figure 0.3 Dominance relationship between a pair for cas- trate peer—reared, intact peer-reared, and intact wild-caught male rhesus monkeys. monomrm 06883883882 153 «9'- E i 1 11 I1 HIHHIIIHHEHH. 11111111111111111 154 Figure h.4 Dominance relationship between a pair and its effect on male sexual behavior measures. 155 883:: 1444‘ A L J 1 888888815 1 PERCENMEOFTESTSNWHICH memos OCCWREDIFLEASTGJCE - (I A [II I I 1IEHHHEIHHHHHHIHHHHHL E TESTS wm: FEMALE mm 74-78 [:1 TESTS wrm MALE comm u-e'r 156 occurrence of certain male behavior patterns seemed markedly affected by the factor of dominance. Males contacted, mounted, intromitted and ejaculated no a lower percentage of tests when they were tested with a female partner who was dominant. A statistical analysis was carried out on the data using a Yates analysis (Ferguson, 1959) and statistical significance was ob- tained for measures of male contact (X2 = 5.5, df = 1; p <:.025) and male mount (X2 = 7.12, df = 1; p <:.O1). Males also had a lower percentage of tests with occurrence of erection if their partner was dominant to them. The purselip facial expression was exhibited on a higher percentage of tests if the male was dominant and this was a statistically valid difference (x2 = 7.83, df = 1; p < .01). Female behavior measures were also influenced by the dominance interaction of the pair. Figure 4.5 presents a graphical display of the relation between dominance and female behavior measures. No female ever came into proximity of a male if she was the dominant partner. Females did, however, spontaneously present to males even whenthey were the domi- nant partner, but they did so at a level lower than that of the situation in which the male was dominant. Females were observed to mount males. These occurrences were rare and never occurred with a wild-caught male. Female mountings of males did occur on a higher percentage of tests if the female was the dominant member of the pair but there was no statistically significant difference. 157 Figure 4.5 Dominance relationship between a pair and its effects on female behavior measures. WGTESTSNW TEMMONLEASTM 00538888838 lllJllllLl 158 I l T PROXMTY m MOUNT FEMALE m PATTERNS _moomwnu-ne BMW 74-87 159 Figure 4.6 presents the data for two specific behavioral categories; those of groom and yawn. Grooming was not a fre- quent behavior shown by pairs under the testing conditions of this study. Females exhibited grooming on a higher percentage of tests than the males and exhibited this behavior regardless of their dominance status. Males never exhibited grooming on tests in which the female was the dominant partner. Yawn- ing was exhibited by both males and females but occurred in a higher percentage of the tests for males regardless of the What. mans-vhga-‘mm I dominance relationship they held. Female yawning occurred on a statistically significantly higher frequency when the female was the dominant member of the pair (X2 = 9.7, df = 1, p