Neonatal and Pubertal Development in Males of a Cooperatively Breeding Primate, the Cotton-Top Tamarin (Saguinus oedipus oedipus)

Abstract

In cooperatively breeding groups of mammals, reproduction is usually restricted to a small number of individuals within the social group. Sexual development of mammals can be affected by social environment, but we know little regarding effects of the cooperative-breeding system on males. Cotton-top tamarin (Saguinus oedipus oedipus) offspring typically do not reproduce in their natal group, even though they may be physically mature. We examined neonatal and pubertal development in captive male cotton-top tamarins as an example of reproductive development within a cooperative-breeding system and to compare cotton-top tamarins with the general primate model. Puberty was characterized using both hormonal and physical measures. Data were collected on urinary levels of LH, testosterone (T), dihydrotestosterone (DHT), cortisol, and the ratio of DHT to T; testicular development; body weight; and breeding age. We determined that 1) pubertal LH secretion began at Week 37, 2) a surge of T secretion followed at Weeks 41–44, and 3) an increase in the metabolism of T to DHT may have occurred at an average age of 48.6 wk. Most of the rapid weight gain was completed by Week 24, before hormonal increases and rapid scrotal growth. We concluded that rapid pubertal testicular growth in captive cotton-top males was completed by an average 76 wk, but that completion of the individual pubertal spurt can occur between 56 and 122 wk. In a cooperative-breeding system, the opportunity for successful reproduction is dictated by the social environment, but we found no evidence that male offspring were developmentally suppressed in their natal social groups. Our findings suggest that puberty in male New World callitrichid primates occurs more quickly than puberty in Old World primates, even though both have similar patterns of development.

INTRODUCTION

Cooperative breeding is a reproductive system wherein animals other than the genetic parents provide care for offspring. These alloparents, called “helpers at the nest” [1], are found in a wide array of taxa, including birds, fish, and mammals [2–4]. Among mammalian species, cooperative breeders include members of four families: rodents, carnivores, mole rats, and callitrichid primates (marmosets and tamarins). In many species, alloparenting is provided by the offspring from a previous birth. Such offspring helpers typically do not reproduce successfully until after leaving the natal group, even though they may be physically mature [2, 4].

In cooperative-breeding groups of mammals, reproduction is usually restricted to a mating pair or to a small number of individuals within the social group. A large number of studies have examined the mechanisms used to restrict breeding, even though most work on callitrichid monkeys has focused on the reproductive physiology and behavior of females (reviewed in [5]). Specifically, the presence of breeding females may prevent subordinate females from reproducing through hormonal suppression, behavioral mechanisms, or a combination of the two. For example, female cotton-top tamarin (Saguinus oedipus oedipus) offspring do not show ovarian cyclicity while housed in their natal groups, but they begin cycling and conceive rapidly once removed from their natal groups [6–9].

In comparison, we know relatively little regarding effects of the cooperative-breeding system on the reproductive biology or sexual development of male callitrichids. Even though mating pairs of cotton-top tamarins are reproductively dominant, no evidence has been found of a behaviorally mediated stress response or of altered physiology directly restricting breeding of adult male offspring [10]. That the social environment can have differential effects on the timing and extent of pubertal development has been well documented among mammalian species. In particular, subordinate social status can retard sexual maturation, whereas the presence of a mature male can accelerate development in females of many mammalian species ([11], reviewed in [12]). In males, the effects of social environment on reproductive development have been dramatically illustrated by noncooperatively breeding primate species: orangutans [13], rhesus macaques [12], and mandrills [14]. In each of these species, androgen levels and the expression of secondary sexual characteristics can be depressed in the presence of socially or reproductively dominant males for years at a time.

Primates represent one end of the spectrum of mammalian pubertal development, characterized by a pronounced interval between birth and the onset of puberty [15]. Unlike puberty in small-bodied, rapidly maturing species such as rodents, puberty in primates is discretely separated from perinatal ontogeny [15]. The anthropoid primate model of male pubertal development is based largely on studies of sexually dimorphic catarrhine primates, particularly Old World monkeys, apes, and humans [15]. Callitrichid primates, however, are small-bodied (150–1000 g), monomorphic New World species with a markedly different social and reproductive system than the harems or multimale/multifemale breeding systems that typify Old World species. Two detailed studies of the common marmoset [16, 17] comprise the sum total of our knowledge regarding male reproductive development in cooperatively breeding callitrichid species. Male callitrichids, therefore, present an important model of sexual development for several reasons, including a mammalian example for the effects of a cooperative-breeding system on male sexual development and fertility, and a means of understanding the patterns of pubertal development in a small-bodied New World primate compared to the much better-studied Old World primate species. The goal of the present study was to describe the reproductive development of captive cotton-top tamarin males by examining physical growth and hormonal activity in the natal group. Specifically, we attempted to characterize male puberty in this species using hormonal and physical measures and to determine if adult males housed in their natal groups differed from breeding males in the expression of hormonal and physical sexual characteristics. By examining these questions, we also attempted to assess whether evidence exists for reproductive suppression of male cotton-top tamarins in their natal groups.

To address our specific aims, we collected data on four different parameters: 1) urinary levels of LH, testosterone (T), dihydrotestosterone (DHT), and cortisol; 2) scrotum size; 3) body weight; and 4) colony breeding age. Data were gathered through a variety of noninvasive techniques designed to be nondisruptive to complex social groups of tamarins with infants and young [18–23]. Testosterone is the urinary androgen most highly correlated with levels of circulating T in cotton-top tamarins [24]. Dihydrotestosterone is key to pubertal development of most male secondary sexual characteristics in primates [25], and it is believed to be critical to the development and functioning of sexual behaviors involved in successful copulation [26–31]. It is also the predominant intracellular androgen mediator, being metabolized irreversibly [24] from T in key peripheral reproductive tissues, such as the prostate, epididymis [32], and genital skin [25]. The ratio of DHT to T was calculated in an attempt to understand the utilization of plasma T before it was excreted in the urine. We might expect an increased conversion rate from T to DHT to occur as pubertal development in males progresses.

MATERIALS AND METHODS

Animal Care

Subjects were male offspring born in a captive colony at the University of Wisconsin-Madison. Subjects were reared by and, for the duration of the study, housed only with the natal group, which included the breeding pair (sire and dam) and siblings. Data were collected during a 4-yr period. Details of animal care and housing are described by Ginther et al. [10]. The research presented here was approved in March 1996 and February 1999 by the Institutional Animal Care and Use Committee of the University of Wisconsin, Madison, College of Letters and Science. Animals were housed and treated in accordance with NIH Guidelines for Animal Research, and research protocols were approved by the University of Wisconsin, College of Letters and Science Animal Care and Use Committee.

Hormone Data

A combination cross-sectional/longitudinal approach maximized use of data from available subjects over a 2-yr sampling period to create a developmental hormone profile representing ages from 9 to 124 wk (Week 1 = the first week postpartum that had ≥3 days remaining in that calendar week). Urine was collected from a total of 12 different subjects two times per week; each sample was assayed for LH (n = 12 subjects), T (n = 10 subjects), DHT (n = 10 subjects), and cortisol (n = 12 subjects). Urine collection and preparation before assay is described by Ziegler et al. [6, 7] and Ginther et al. [10]. Subjects and their social groups were well habituated to the procedure before the start of the study. The first morning urine void was collected directly into hand-held containers from inside the home cage of each subject. In humans and rhesus monkeys, initial pubertal increases in T secretion occur nocturnally and up to several months before a daytime pubertal increase [15]; in S. oedipus oedipus, the first morning urine void represents production of urine during the entire nocturnal period. The procedural details of enzyme immunoassays for urinary T, DHT, cortisol, and creatinine (Cr) are described by Ginther et al. [10]. The LH RIA detects alpha and beta subunits in urine [33]; procedural details have been reported by Ginther et al. [10]. Each sample was assayed for creatinine (mg/ml urine) to adjust for variations in water content between urine samples. The LH, T, DHT, and cortisol concentrations for each sample were divided by creatinine concentration of the same sample, and hormone levels were expressed per milligram of creatinine. Due to limited resources for T and DHT measurement, urine was pooled in 2-wk blocks before extraction and chromatography for some subjects.

To analyze appropriate sample sizes, data for each hormone were averaged across 4-wk blocks for each subject. Six subjects with the largest longitudinal sampling of androgen levels were used for analysis of the DHT:T ratio; for these subjects, data were available for analysis in 2-wk age blocks. For each of these six subjects, ng DHT/mg creatinine was divided by ng T/mg creatinine for every pooled or original urine sample. Profiles for all hormones and hormone ratios were analyzed by repeated-measures analysis (Statistical Analysis System [SAS]; SAS Institute, Inc., Cary, NC) using a mixed model with AR(1) covariance structure based on random subjects and fixed effect of time [34]. Difference of least-squared means was compared for each pair of age blocks. Significance was predetermined as P < 0.05.

Scrotal Width Data

A cross-sectional/longitudinal approach maximized use of data from available subjects (n = 28) to create a scrotal growth profile of ages ranging from 9 to 200 wk. Frontal scrotal width of unrestrained animals was measured systematically using a template-card technique according to the protocol described by Ginther et al. [21]. Once every 2 wk, subjects were lured onto the mesh walls of their home cages, and maximum frontal width of the scrotum was systematically measured using a template card of a series of rectangles of known width (1-mm intervals). From three to five replicates per subject per measurement session were averaged. The technique was validated by Ginther et al. [21]; card measures correlated highly with caliper measures of scrotal width and testicular dimensions on the same subject. All measurements for this study were made by the first author. Raw data were not reviewed, summarized, or tabulated until all data collection was completed.

For each of the 28 subjects, data were averaged into 2-wk age blocks. Each subject contributed data to 20.68 ± 3.06 (mean ± SEM) age blocks. Scrotal growth profiles were analyzed by repeated-measures analysis (mixed model, AR[1] covariance structure, random subjects, fixed effect of time) [34], followed by paired comparison of least-squared means for each pair of age blocks. Significance was predetermined as P < 0.05.

In a separate analysis, changes in testicular growth rate were modeled individually for 23 subjects as a linear relationship between age in days (Day 1 = day of birth) and scrotum size. For each individual, scrotal width of each measurement session was plotted according to age (in days), which revealed an initial phase of rapid scrotal growth. For a subset of nine individuals representing data over an extended period, this initial phase appeared to be followed by a period of slower growth. For each of these nine subjects, a segmented-regression technique [35] was used to fit two regression lines representing the two periods of differing growth rate, thus estimating the age of transition between the first and second growth periods.

Testicle Descent

All colony offspring were observed for testicle descent from the time they ceased riding natal group members and first approached the observer on the cage mesh. During scrotal width measurement sessions, the scrotum of each subject was visually inspected for descended testicles according to the criteria of Ginther et al. [21]: 1) a clear outline of both testes within the scrotum was visible; 2) no loose skin was at the bottom of the scrotum; and in immature subjects, also 3) the scrotal skin was observed moving over the testicles when the animal moved or shifted position; 4) the testes remained visible within the scrotum even after the subject locomoted.

Body Weight Data

A cross-sectional/longitudinal approach maximized use of data from available subjects to create a weight gain profile representing an age range from 1 to 122 wk. The body weights of 24 subjects were measured. Unrestrained animals were weighed at least once every 2 wk in their home cages after they were lured onto a digital platform scale; weights for infants being carried (e.g., <9 wk of age) were calculated by subtracting the weight of the carrier animal from the weight of the infant and carrier together. Weights were measured between 1330 and 1730 h and at least 30 min following the midday or afternoon feeding. Further details regarding the weighing technique are described by Achenbach and Snowdon [36].

For each subject, data were averaged across 2-wk age blocks. Each of the 24 offspring subjects contributed data to 19.00 ± 2.64 (mean ± SEM) age blocks. Weight profiles were analyzed by repeated-measures analysis (mixed model, AR[1] covariance structure, random subjects, fixed effect of time) [34], followed by paired comparison of least-squared means for each pair of age blocks. Significance was predetermined as P < 0.05.

Age of First Breeding

All breeding data (n = 41 sires) from the colony database were compiled spanning the time period from 1977 to 1997. The age of each sire in the colony’s 20-yr history [20] was computed at the time of first identified conception (calculated from the date of first full-term live or nonlive birth minus an average gestation of 184 days; [6, 7]). No data from wild-caught males were used. Colony sires were not paired with mates according to a specific age protocol.

RESULTS

Hormone Profiles

Urinary LH was lower during Weeks 9–12 than at any other point in the developmental profile (Fig. 1) and was excreted in increasing amounts through Week 16. The average level of LH remained between 3 and 4.5 ng/mg Cr from Weeks 17 through 124, except for two significantly elevated 8-wk periods beginning at Week 37 and at Week 77. Urinary levels of both T and DHT were significantly elevated during Weeks 9–12 postpartum. After Week 12, average DHT levels decreased from 0.67 ng/mg Cr and remained between 0.09 and 0.25 ng/mg Cr through Week 120. Urinary T decreased from 6.2 ng/mg Cr and remained between 1.3 and 4.1 ng/mg Cr, except for a significant peak at Weeks 41–44.

FIG. 1.

Average developmental levels of LH, T, and DHT for captive male cotton-top tamarins in a series of 4-wk age blocks (mean ± SEM). The sample size for each age block is given on the upper axis. Upper panel) The LH age block 9–12 is significantly lower than all other age blocks (*). Blocks with the letters a and b are significantly different from one another, and blocks with the letters c and d are different. Middle panel) Blocks with the letters a and b are significantly different, and blocks the letters c and d are different. Lower panel) The DHT age block 9–12 is significantly higher than all other age blocks (*).

Five of the six subjects used to construct a longitudinal curve for the DHT:T ratio showed a dramatic increase in the androgen metabolite ratio. In a post-hoc analysis beginning with Week 23, the jump point of the ratio for each male was defined as the first rise in ratio value greater than or equal to two standard deviations above the mean of the three previous age blocks [37] and that exceeded all previous ratio values. The ratio profiles for the 5 males were standardized according to the time of the jump point; the average age at the jump point was 48.6 ± 5.42 wk (mean ± SEM) (Fig. 2). These standardized profiles for the DHT: T ratio were analyzed by repeated-measures analysis (mixed model, AR[1] covariance structure, random subjects, fixed effect of time) [34], and the difference of least-squared means was compared for each pair of standardized blocks. For approximately 5 mo before the jump point, the DHT:T ratio was significantly lower than that of the first 2-wk block at the jump point and of most of the blocks thereafter (Fig. 2). In addition, the average DHT:T ratio was significantly lower before (0.03 ± 0.01, mean ± SD) than after (0.09 ± 0.04) the jump point (P < 0.05, t = 2.608, df = 4, one-tailed paired t-test).

FIG. 2.

Developmental changes for the averaged ratio of DHT to T during sexual maturation. Two-week age blocks (mean ± SEM) were standardized across subjects according to the defined jump point. For each subject showing a jump in DHT:T ratio (n = 5 of 6 subjects), the first age block at the jump point is defined as period 1; the block before the jump point is period −1. The different letters (a and b) represent significant differences. Sample sizes are shown for every other block and are representative of the entire profile.

Cortisol was excreted in decreasing amounts from Week 9 through Week 16 postpartum (Fig. 3) and remained basal [10, 22] thereafter, except for a significant peak at approximately 1 yr of age.

FIG. 3.

Average developmental levels of cortisol during sexual maturation in a series of 4-wk age blocks (mean ± SEM). The sample size for each age block is given on the upper axis. Age block 9–12 is significantly higher than all other blocks (a). The age blocks labeled b are significantly higher than all remaining blocks but do not differ from those labeled b1. Blocks labeled b1 are significantly higher than most unlabeled blocks.

Scrotal Width Profile

The average developmental profile for scrotal growth is shown in Figure 4. The profile was divided into phases according to patterns of statistical significance resulting from the comparison of means for each pair of age blocks [34]. None of the pairs of age blocks between Weeks 11 and 25 (phase s1) were different from one another, with the exception of Weeks 19–20 and adjacent blocks. Between Weeks 25 and 75 (phase s2), rapid growth occurred; every block in this phase was significantly different from every other block in the entire profile, except for a few adjacent pairs. After Week 75 (phase s3), the profile began to asymptote, and after Week 82, only three of the adjacent pairs of blocks were significantly different. After Week 123 (phase s4, not shown in Fig. 4), the average scrotum size was identical to that in phase s3, and few of the pairs of age blocks differed from one another for all pairs across the entire phase.

FIG. 4.

Average developmental increases in scrotal width (upper panel) and body weight (lower panel) in a series of 2-wk age blocks (mean ± SEM). Sample sizes are shown for every other age block and are representative of the entire profile. The times of key changes in hormone levels are indicated above the x-axes. Upper Panel) In phase s1, none of the pairs of blocks differed significantly (with one exception). In phase s2, every block was different from every other block in the entire profile, except for a few adjacent pairs. In phase s3, the values of adjacent pairs of blocks did not differ, except for three adjacent pairs. Lower Panel) In phase w1, each block was significantly different from every other block. During phase w2, the values of adjacent pairs and triplets of blocks did not differ, except for one or two blocks. In phase w3, none of the weights differed among any of the age blocks.

Construction of a developmental profile by averaged age blocks resulted in the smooth asymptotic curve shown in Figure 4; this curve depicts the average developmental stage at each time point. However, individual profiles constructed longitudinally for each individual subject revealed that each subject underwent a more abrupt change in growth rate that was not evident in the averaged population curve. According to the fits of individual regression lines, the average scrotal growth rate during the first period was 2.18 ± 0.13 mm per 100 days (n = 22 subjects) (Table 1); during the second period, this rate had slowed to 0.28 ± 0.07 mm per 100 days (n = 10 subjects) (Table 1). The estimated age at transition from the first to the second growth period was 74.9 ± 2.75 wk from birth (mean ± SEM, n = 9 subjects) (Table 1). The average age of change according to these individual regression analyses corresponded roughly to the time of change from phase s2 to phase s3 in the population curve for scrotal growth (Fig. 4).

TABLE 1.

Scrotal growth (mm/100 days) as determined by segmented regression.^a

ID	Family	Period 1 (mean ± SEM)	Period 2 (mean ± SEM)	Age of change in wk (days)
Alfa	1	2.21 (0.13)	0.17 (0.02)	77 (539)
Balo	3		0.07 (0.02)
Bari	2	1.72 (0.41)
Chec	2	1.31 (0.41)
Esch	3	2.94 (0.13)	0.17 (0.002)	70 (490)
Falc	5	2.83 (0.30)
Grov	4	2.30 (0.07)
Hers	6	2.25 (0.08)	0.19 (0.11)	76 (535)
Indi	1	1.59 (0.08)	0.15 (0.02)	83 (584)
Juni	1	1.51 (0.04)	0.41 (0.09)	74 (516)
Kafk	2	2.29 (0.18)
Lark	1	1.99 (0.08)	0.17 (0.20)	80 (557)
Leni	2	1.82 (0.11)
Mati	3	3.53 (0.18)
Mone	3	1.83 (0.07)	0.33 (0.03)	75 (523)
Nash	3	2.78 (0.27)
Tali	3	2.40 (0.18)
Titi	3	2.18 (0.59)
Urso	2	1.32 (0.17)
Wate	4	1.99 (0.18)
Wolf	5	2.95 (0.09)	0.83 (0.06)	56 (393)
Yelt	2	1.40 (0.16)
Zhiv	2	2.74 (0.19)	0.30 (0.07)	83 (578)
Mean ± SEM		2.18 ± 0.13	0.28 ± 0.07	74.9 ± 2.8 (523.9 ± 19.1)

^aGrowth rates (mm/100 days) for the rapid-growth period (period 1) and the slower-growth period (period 2) are shown for each individual in separate columns. Age of change from the rapid to the slower period was estimated as the age at transition between the two regression lines, as calculated for each subject.

Testicle Descent

First observations of male offspring with fully descended testes were made during the following age blocks (no. observations, mean age ± SEM): Weeks 5–8 (n = 1, 8 wk); Weeks 9–12 (n = 3, 11.33 ± 0.33 wk); Weeks 13–16 (n = 4, 14.25 ± 0.48 wk); Weeks 17–20 (n = 2, 19.50 ± 0.50 wk); Weeks 21––24 (n = 3, 22.33 ± 0.67 wk); and Weeks 25–28 (n = 2, 25.00 ± 0.00 wk). For all but one subject, the testes had descended fully into the scrotum by the time the subject was first observed; therefore, mean ages do not represent the actual time of descent.

Only 1 subject approached the observer early enough to observe the testes in the process of descending into the scrotum. At Week 9, the testes in this subject were positioned at the border between the genital and abdominal skin but shifted position frequently with locomotion. At Week 13, the bottom edge of the testes were consistently below the level of the prepuce. At Week 18, both testes were consistently centered at the level of the prepuce, and at Week 22, both testes were consistently positioned at the bottom of the scrotum.

Body Weight Profile

The developmental profile for weight gain is shown in Figure 4. The profile was divided into phases according to patterns of statistical significance resulting from comparison of means for each pair of age blocks [34]. Body weight gain occurred steadily and sharply until Week 25 (phase w1); every block in this phase was significantly different from every other block in the profile. Between Weeks 26 and 75 (phase w2), the profile began to asymptote; pairs or triplets of adjacent blocks did not differ. After Week 77 (phase w3), none (except for 2 of the blocks) differed from any other in that phase. In summary, even though the timing of the developmental phase changes was identical between body weight and scrotal width profiles, the phases were offset. Rapid gain in body weight preceded the phase of rapid scrotal growth.

Age of First Breeding

Mean sire age (n = 41) at the time of first identified conception was 37.82 ± 2.07 mo (mean ± SEM). The youngest colony sire to impregnate his mate, did so at 21.59 mo, three additional males under 25 mo, and an additional two males between 25 and 26 mo. It is possible that some dams in our colony miscarried their first pregnancy; therefore, data only correspond with fetuses carried to term. Some males had earlier opportunities to breed than others and had inconsistent social or environmental contingencies that may have affected successful reproduction and complicated cross-species comparisons.

DISCUSSION

We examined reproductive development in captive male cotton-top tamarins as an example of development within a cooperative-breeding system and to compare cotton-top tamarins with the general anthropoid primate model. The urinary LH, T, and DHT:T profiles indicated that the onset of puberty occurred between 37 and 48 wk of age. We determined that 1) pubertal LH secretion began at Week 37, 2) a surge of T excretion followed at Weeks 41–44, and 3) a subsequent increase in the ratio of DHT to T occurred at an average age of 48.6 wk. Based on measurements of scrotal growth, we concluded that, in an average population, rapid pubertal development is completed by 76 wk, but that individuals vary greatly and that completion of the pubertal spurt can occur between 56 and 122 wk. On average, offspring achieved the scrotum size of breeding males (18–20 mm [21]) during phase s3 (~Week 85), even though size continued to increase through Week 122. The youngest animal in our colony to have sired offspring did so at approximately 2 yr of age. In a cooperative-breeding system, the opportunity for successful reproduction is dictated by restrictions of the social environment. Despite these restrictions, and corresponding with previous observations [10], we found no evidence that male offspring were developmentally suppressed in their natal social groups. Our findings suggest that the process of puberty in male New World callitrichid primates is more rapid than in Old World primates, even though they have similar patterns of development.

Hormonal Development

The neonatal resurgence of LH production seen in humans, macaques, and other catarrhine primates (reviewed in [38, 39]) was not evident in our data, but it may have preceded the sampling period. The significant 2-mo increase in LH excretion at Weeks 37–40 may represent the onset of puberty. In Old World monkeys, apes, and humans, increased plasma LH is likely coupled with the transition into puberty [15]. Pubertal activation of the gonads in primates is caused by increased gonadotropin secretion [15], and in cotton-top tamarins, onset of this gonadotropin increase (Week 37) preceded the increase in T at Weeks 41–44 and the average age of increase in DHT:T ratio (48.6 wk). The second elevation in LH, at Week 77, corresponded roughly with the developmental change from phase 2 to phase 3 (Week 75) for both the scrotal width and body weight profiles. The LH rose rapidly for only 2 mo across the averaged population. In humans, the pubertal increase in circulating LH is a rapid event [15].

Urinary levels of both T and DHT in our subjects were significantly higher during Weeks 9–12 than during the next 5 mo. This finding is suggestive of a high level of neonatal androgen production. A similar pattern has been described for neonatal plasma T concentrations in common marmosets. Plasma T was elevated postpartum but began to decline, reaching a nadir at approximately 100 days after birth, and then began to rise again during pubertal development [16, 17, 40]. In humans and other primate species, plasma T is elevated within the first few months postpartum and is suspected of having critical effects on sexual maturation of the gonads, endocrine system, and behavior [38–40]. The source of elevated neonatal T may be due to the persistence of fetal Leydig cells (FLC), a morphologically distinct form found in primates and other mammals [41, 42]. In rats, FLC are present and steroidogenic through 90 days postpartum and may activate the hypothalamo-hypophyseal-testicular axis for pubertal development [43].

Neither the T nor DHT profile fit the anticipated sigmoidal form that characterizes pubertal plasma levels in the common marmoset [16, 17] and most other anthropoid primate species studied [15]. Except for high postnatal levels in both hormones and a significant increase in T during Weeks 41–44, both androgens in our data remained relatively stable and with no measurable changes through 2.5 yr of age. This result may be, in part, an artifact of examining excreted metabolic products. Long-term urinary metabolite profiles are complicated by the possible differential utilization of plasma hormones within the body as an animal matures and target tissues undergo changes in hormone-receptor density and function. However, by examining the ratio of excreted DHT to T over the developmental profile, we hoped to identify potential changes in the utilization of T during the course of sexual development. A striking jump in the DHT:T ratio at approximately 49 wk of age was seen in five of six subjects. The DHT:T profile for these five subjects suggests that a rapid switch in conversion from T to intracellular DHT occurs in developing males before 1 yr of age. After the jump point, proportionately more DHT than T was excreted in the urine. In species such as the tamarin, in which hormones must be measured indirectly through urine, the DHT:T ratio might provide a critical marker for the onset of maturing endocrine physiology during puberty.

After parturition, the transformation into the adult form of the adrenal cortex requires a period of months in humans and macaques [44, 45]. In newborn callitrichid infants, the fetal adrenal cortex (FAC) is prominent and secretes high levels of cortisol [46], which fall dramatically over the first 6 mo. During Weeks 9–12 in our tamarin subjects, cortisol was produced at a rate well above basal levels [22, 23] and was significantly higher than at any other point during the first 2.5 yr. Production by the FAC may explain the extremely high cortisol levels seen in our subjects during the third postpartum month. A second possibility is the stress of weaning, a social process that begins in captive cotton-tops at approximately 4 wk postpartum and that may continue for several weeks [47]. Psychosocial stress has been implicated as a source of increased corticosteroid production in a number of primate species [48].

A significant increase in cortisol levels was observed during Weeks 53–56, even though these levels were still within the normal range seen in mature males [10, 23]. At this age, offspring in regularly reproducing family groups are typically involved in alloparental activities for the first time, caring for the second cohort of siblings. Initial alloparenting experiences in breeding groups of tamarins are often traumatic and result in some social disruption as the year-old sibling learns carrying skills (unpublished observations).

Genital and Physical Development

Body growth in the anthropoid primate model differs from that in other mammals by a lengthened duration of preadolescent growth and a bimodally shaped growth curve. The primate growth curve is comprised of a neonatal growth peak, an extended phase of slower juvenile growth, and a second acceleration of growth at adolescence [15, 49]. In other nonprimate mammals, puberty quickly follows weaning, and two separate phases of rapid growth are replaced by a general deceleration of growth [49]. The common marmoset shows a slower phase of juvenile growth before the onset of pubertal hormonal development, despite its small body size (300 g) [15]. For male cotton-top tamarins, the growth curves for both testes and weight showed only one phase of accelerated growth. However, rapid growth of testes occurred later than rapid growth in weight, so a phase of nondistinct scrotal growth was evident (phase s1). This suggests that, for sexual physical characteristics, tamarins adhere to the general primate model. In contrast, no slow phase of weight gain was evident before the rapid rate of weight gain that began a few weeks after birth.

Scrotal growth and weight gain in our subjects appeared to follow the same general pattern of neonatal and pubertal development, even though the analogous phases of growth (Fig. 4) were temporally offset. The overall shapes of the profiles were similar, but rapid weight gain (phase w1) preceded rapid scrotal growth (phase s2) and the points of significant increases in LH, T, and DHT:T ratio. These results are compatible with metabolic theories regarding the onset of puberty. For both male and female cercopithecene monkeys, humans, and other mammals, attaining critical body weight or growth is apparently necessary for the onset of sexual maturation. Changes in metabolic signals or the availability of metabolic fuels associated with decreases in somatic growth may facilitate the initiation of pubertal hypothalamic function [50–52]. Interestingly, the change from phase 2 to phase 3 for weight and scrotum size (Week 75) immediately preceded the second increase in LH excretion.

By examining individual scrotal growth profiles, it was apparent that individual animals underwent abrupt changes in testicular growth rate that were not evident in the averaged profile. Thus, phase s3 showed an asymptotic reduction in scrotal growth rate, which is a byproduct of summarizing across the abrupt developmental changes of many individuals. Similarly, phase w2 indicated a slowing of weight gain when averaged across the population as a whole. By examining individual plots of weight data, we identified what seemed to be two periods of first rapid, and then slower, growth, which is similar to the pattern shown by segmented regression of individual scrotal growth profiles. However, not enough longitudinal data were available to span a transition age and to perform statistical tests. Future studies should target the age range within phase w2 for longitudinal studies that might identify abrupt individual changes in weight gain. Individual differences in discrete points such as these in physical development could be a powerful tool for future studies to examine how developmental physiology and anatomy correlate with social, environmental, or genetic variables.

Assessment of Male Puberty in a Cooperatively Breeding New World Callitrichid Primate

Hormonal measures were the most useful to identify the onset of pubertal development, but physical measures provided some information about its completion. Pubertal onset in male cotton-tops (Week 37) corresponded with increases in plasma T in captive common marmosets at 36 wk [16] and 29 wk [17]. Marmosets reach adult body weight at 71–79 wk [16], corresponding with the time of change from phase w2 to phase w3 in our males. Rapid testicular growth in common marmosets occurs between 36 and 93–100 wk [16], similar to the time of rapid scrotal growth in our data. Common marmosets also exhibited considerable individual variation in gonadal growth rate and in both the start and end of the rapid growth phase [16].

Adult male cotton-top offspring older than 2 yr and living in their natal groups exhibit the scrotum size and levels of reproductive hormones of breeding males [10, 21]. The present study supports these previous findings. By 2 yr of age, levels of urinary hormones and scrotum size of male offspring fell well within the range given by Ziegler et al. [53] and Ginther et al. [10, 21] for males breeding successfully outside the natal group. Thus, male cotton-tops are apparently capable of achieving reproductive maturity within the cooperatively breeding social group. However, the time required for completion of puberty may differ according to the presence or absence of a dominant, breeding male. Different physiological aspects of fertility, such as ejaculate quality, were not measured in this study but may also play a role in achieving male fertility. Similarly, the stage of the breeding female’s reproductive cycle should be investigated in future studies with respect to levels of male reproductive hormones and a possible role in fertility.

In summary, male cotton-tops appear to mature fully and do not experience the degree of reproductive suppression seen in females of the species, an effect that is paralleled in other cooperatively breeding callitrichid, carnivore, and bird species [54–57]. The time course of puberty in male tamarins is similar to that in the other callitrichid primates studied thus far. Endocrine and physical patterns of development paralleled those of the general primate model of male neonatal and pubertal development based on Old World monkeys, apes, and humans. Even though the hormonal profiles did not show the sigmoidal form that typifies the general model, the key developmental events in maturation of the pituitary-gonadal axis were represented through excreted steroids. The bimodal growth curve that typifies maturation in catarrhine primates paralleled scrotal growth in our males; however, changes in body weight were similar to the growth curve expected for small-bodied, rapidly reproducing mammals.