Behavioral strategies and hormonal profiles of dominant and subordinate common marmoset (Callithrix jacchus) females in wild monogamous groups.код для вставкиСкачать
American Journal of Primatology 67:37–50 (2005) RESEARCH ARTICLE Behavioral Strategies and Hormonal Profiles of Dominant and Subordinate Common Marmoset (Callithrix jacchus) Females in Wild Monogamous Groups MARIA BERNARDETE CORDEIRO SOUSA1,2n, ANA CLAUDIA SALES DA ROCHA ALBUQUERQUE3, FABIOLA DA SILVA ALBUQUERQUE1, ARRILTON ARAUJO1, MARIA EMILIA YAMAMOTO1, and MARIA DE FATIMA ARRUDA1 1 Curso de Pós-graduação em Psicobiologia, Departamento de Fisiologia, Universidade Federal do Rio Grande do Norte, Natal, Brazil 2 Curso de Pós-graduação em Cieˆncias da Saúde, Departamento de Fisiologia, Universidade Federal do Rio Grande do Norte, Natal, Brazil 3 Departamento de Cieˆncias Biológicas, Universidade Estadual do Rio Grande do Norte, Mossoró, Brazil New insights into the mating systems of common marmosets suggest that they are mainly monogamous, although polygyny and polyandry occasionally occur. Long-term monitoring of wild common marmosets has shown that some reports of polygynous groups (i.e., groups that contain more than one reproducing female) in fact indicate an unbalanced reproductive output associated with extragroup copulation. In this study we describe the behavioral and hormonal profiles of common marmoset (Callithrix jacchus) females living in three wild monogamous groups (Q, PBf, and T), varying from five to 11 individuals, at Nı́sia Floresta field station, RN, Brazil. The mating system of the groups was previously characterized in terms of affiliative, sexual, and mate-guarding behaviors. Behavioral data were collected once a week, and fecal samples were collected at least twice a week for 10–16 months, depending on the group. A preferential allogrooming relationship was recorded between dominant males and females. Under field conditions the reproductive inhibition of subordinate females appears to be more behavioral than hormonal, since subordinate females of the three groups ovulated and two conceived during the study. In these cases, the subordinate and dominant females reproduced 1 month apart, and infanticide (one case confirmed and one suspected) appeared to be part of the reproductive strategy of dominant females. Following the infanticide, ovarian inhibition (group T) or emigration and return to the natal group (group PBf) were observed. In the third group (Q) the subordinate female Contract grant sponsor: CNPq; Contract grant numbers: 464103/00-2; 52.1186/1997; 305216/20035; 470601/2003-1; 524409/96; Contract grant sponsor: CAPES; Contract grant number: 864/91-11. n Correspondence to: Maria B.C. Sousa, Universidade Federal do Rio Grande do Norte, Departamento de Fisiologia, Caixa Postal, 1511, 59078-970 Natal, RN, Brazil. E-mail: firstname.lastname@example.org Received 28 September 2004; revised 3 February 2005; revision accepted 3 February 2005 DOI 10.1002/ajp.20168 Published online in Wiley InterScience (www.interscience.wiley.com). r 2005 Wiley-Liss, Inc. 38 / Sousa et al. showed hormonal profiles compatible with pregnancy, but no infants were seen. These findings reflect the different alternatives that wild subordinate common marmoset females use to reproduce. Am. J. r 2005 Wiley-Liss, Inc. Primatol. 67:37–50, 2005. Key words: wild common marmosets; monogamy; sexual strategies; dominance; steroid hormones INTRODUCTION The various mating systems of neotropical primates (Callitrichidae family), such as monogamy, polygyny, and polyandry, have been described by numerous authors (e.g., C. jacchus [Digby & Ferrari, 1994; Roda & Pontes, 1998], C. humeralifer [Rylands, 1986], C. aurita [Coutinho & Correia, 1995], and L. rosalia [Dietz & Baker, 1993]). However, recent data regarding the reproductive output of dominant and subordinate females from nine wild groups, including two that were analyzed in the present study, have revealed two profiles: 1) both dominant and subordinate females reproduce and their offspring survive in almost the same proportion, and 2) although both females give birth a few days apart, only the offspring of the dominant survive [Arruda et al., 2005]. Common marmosets are one of the most widely studied neotropical primate species. They are considered to be singular cooperative breeders [Snowdon, 1996], with females competing to reproduce [Abbott, 1984]. The hormonal profiles of dominant and subordinate females living in captive groups were fully described in previous studies [e.g., Abbott et al., 1984, 1987, 1998; Alencar et al., 1995; Saltzman et al., 1997a]. These studies demonstrated that dominant females inhibit the ovarian cycling of subordinate females, probably by means of odor signaling [Abbott et al., 1993; Smith & Abbott, 1998]. Captive studies have also shown that reproductive inhibition of subordinate females occurs mainly when the females are not related [Saltzman et al., 1997a,b; Ziegler & Sousa, 2002]. However, although a wide range of information has been collected from captive groups, few data are available for wild groups. To better understand the dynamics of the social relationship between breeding males and dominant and subordinate common marmoset females in wild monogamous groups, we performed a long-term field study in northeastern Brazil, addressing three main questions: 1) what are the differences in social dynamics among dominant and subordinate females and breeding males, 2) are subordinate females hormonally inhibited in monogamous groups, and 3) what strategies are used by subordinate females to reproduce? MATERIALS AND METHODS Study Area This study was undertaken at the National Forest Station (FLONA) in Nı́sia Floresta county, 45 km from Natal, Brazil, which includes open areas and secondary forest [Santee & Arruda, 1994]. We studied three groups that occupied two different areas within the field station. One group occupied an experimental plantation area [Albuquerque et al., 2001], and the remaining two groups used previously described areas [Santee & Arruda, 1994]. Monogamy in Wild Common Marmosets / 39 Animals In this study we examined wild monogamous groups of common marmosets (Callithrix jacchus). The mating system of the groups was previously characterized with the use of long-term data from studies based on affiliative, sexual, and mateguarding behaviors, as well as female reproductive output. These groups were monitored from December 1997 to October 2000. In each group we followed the breeding male and two females (one dominant and one subordinate). Dominant status was determined based on the relationship with the breeding male (dominant females copulated and spent more time interacting with breeding males compared to subordinates). Identification of the breeding male was based on either long-term records or the results of a 2-month pilot study in which affiliative, sexual, and mateguarding behaviors were recorded. During the pilot study all individuals were habituated to the presence of an observer, captured, and marked with collars. Specific areas of the body were painted with picric acid for identification. In some animals, trichotomy of the tail was also used to help with individual identification. Behavioral and Hormonal Data Collection We observed each group from 5:00 a.m. to 2:00 p.m. using on-the-minute samples for 15 min each hour for the breeding male and the dominant and subordinate females. No data were recorded without a 40-min minimum interval between recordings. Ad libitum records were also obtained for qualitative data, such as intra- or extragroup copulations and aggressive displays. The analyzed data did not include information obtained during acute changes in the group’s routine, such as intergroup encounters. Observation sessions took place once a week, and the order of observations of focal individuals was randomly established before data collection began. The behavioral and hormonal sampling both lasted 10–16 months, depending on the group. Table I shows additional information about the groups. The behaviors recorded during the study included grooming, scent-marking, agonism (vocalization and chasing), and sexual activities (copulation and attempts). The behavioral definitions followed those used by Lazaro-Perea  and Lazaro-Perea et al. . TABLE I. Summary Information About Group Composition, Reproductive Performance of Females and Study Duration of the Three Monitored Groups Groups Q Total of individuals Monitored females/no. of parturition 6–5 D (Grécia)/2 Study duration September 1999 to October 2000 No. 1-minute records No. fecal samples by females 8,533 D=62 S (Glenda)a/0 PBf 5–9 D (Patricia)/1 S=66 December 1997 to April 1999 12,537 S (Paloma)/2 T 9–11 D (Tereza)/2 S (Tina)/1 a Daughter of dominant female. D, Dominant females; S, subordinate females. D=69 S=102 October 1999 to October 2000 9,266 D=68 S=49 40 / Sousa et al. Fecal samples were collected from adult females twice a week [Sousa et al., 2002], and assays were performed as described in Sousa and Ziegler , with solvolysis used to break up conjugates in the samples [Ziegler et al., 1996]. Enzyme conjugates and antibodies for both hormones were provided by the University of California–Davis. The intra- and interassay coefficients of variation for cortisol and progesterone were 11.5% and 23.6%, and 13.6% and 20.3%, respectively. Hormone values are provided as ng/g of feces. Data Analysis All data were examined for normality before the parametric tests were conducted, and the data were found to be normal. We used the software Statistics version 5.5 (Statsoft Inc., 1999). Student’s t-test and an analysis of variance (ANOVA) one-way post hoc Tukey test were used to compare the frequency of allogrooming and scent-marking behaviors performed by focal animals. For allogrooming, all interactions performed and received by the focal animals were analyzed. For both behaviors, hourly frequencies were grouped according to social status (dominant or subordinate). The criteria used to characterize ovarian functioning were as follows: cycling=increase of progesterone during the length of the luteal phase of ovarian cycle (by what increase, i.e., 50% or higher?); noncycling=progesterone levels continuously low (or ovarian cycles and pregnancies were determined as described in Ziegler and Sousa ). Spearman’s test was used to correlate the cortisol fecal concentrations in dominant and subordinate females. Significance was pre-established at Po0.05 for all tests. RESULTS Behavioral Profiles The pooled data (from three groups) of allogrooming interactions between the focal animals (dominant and subordinate females and breeding males) are shown in Fig. 1. As illustrated in Fig. 1A, the frequencies of grooming performed by dominant females on both subordinate females and males did not differ. On the other hand, the amount of grooming received by dominants from males was significantly higher than that received from subordinates (t-test(1,48)=3.877, Po0.001). No differences were recorded in the frequency of allogrooming performed by subordinate females on dominant females and breeding males (Fig. 1B). Subordinate females also received an equal amount of grooming from both dominant individuals. Males displayed the highest frequency of grooming behavior in comparison with the values recorded for females, regardless of whether they were dominants or subordinates (ANOVA(2,20)=2.326, Po0.05). As illustrated in Fig. 1C, the mean values of allogrooming performed by males were preferentially directed to the dominant females, and allogrooming received from the breeding male by dominant females was significantly higher than that received by subordinate females (t-test(1,52)=3.423, Po0.01). Scent-marking behavior did not show significant differences between dominant and subordinate females living in monogamous groups, although the mean frequencies for subordinates were higher (Fig. 2). It is interesting to point out that although n is small (n=2), the mean scent-mark frequencies for the subordinate females from group Q (Glenda) and group T (Tina) were higher when they were not cycling compared to those recorded during ovarian cycling (Glenda Monogamy in Wild Common Marmosets / 41 A BM SF Mean frequency 20 15 * 10 5 0 Performed Received B Mean frequency 20 DF BM 15 10 5 0 Performed Received C DF SF Mean frequency 20 15 * 10 5 0 Performed Received Fig. 1. Mean hourly frequency (+SEM) of grooming behavior performed on and received from (A) dominant females (DF), (B) subordinate females (SF), and (C) breeding males (BM) in monogamous wild groups of common marmosets. n ANOVA, Po0.05. 42 / Sousa et al. (noncycling)=5.4071.02; (cycling)=3.2071.30; Tina (noncycling)=11.4377.02; (cycling)=5.4473.26). Agonistic behaviors between dominant and subordinate females (i.e., vocalization toward the other individual, and chasing behavior) and sexual behaviors were recorded opportunistically (Q=6; PBf=4; T=5 episodes). Nevertheless, in all recorded instances of agonism between the two females, it was exclusively directed by the dominant toward the subordinate. We also recorded all occurrences of sexual behavior (Fig. 3). In groups Q and T, the dominant females were seen copulating and attempting to copulate with the breeding male (intragroup copulations). Subordinate females, on the other hand, were never seen copulating with the breeding male. However, they were 10 Mean frequency 8 6 4 2 0 Dominants Subordinates Fig. 2. Mean hourly frequency (+SEM) of scent-marking behavior by dominant (white bar) and subordinate (black bar) females living in monogamous groups of wild common marmosets. IGC or attempts BM 6 IGC or attempts SM Frequency EGC or attempts 4 2 0 D Group Q S D S Group PBf D S Group T Fig. 3. Total frequency of sexual behaviors (copulation and attempts to copulate) directed by males to dominant (D) and subordinate (S) females in three monogamous groups of common marmosets. IGC=intragroup copulation; EGC=extragroup copulation; BM=breeding male; SM=subordinate male. Monogamy in Wild Common Marmosets / 43 observed copulating with extragroup males (extragroup copulations) or, in the case of Q group, also with a nonbreeding male of the same group. In group PBf, both females were seen copulating and attempting to copulate with the breeding male, but only the subordinate was seen copulating with extragroup males. Hormonal Profiles of Dominant and Subordinate Females Progesterone. Group Q. Prior to this study, group Q had been hormonally monitored [Albuquerque et al., 2001] from August 1996 to March 1997, when another subordinate female (Graça) showed an ovulatory cycle during the pregnancy of the breeding female (Grécia). Hormonal data from this group were also collected from April 1998 to May 1999. Another subordinate female (Gertrudes) gave birth in the natal group, but the offspring was a victim of infanticide by the dominant. In the current study, dominant and subordinate females from group Q were monitored from 4 September 1999 to 8 October 2000. As shown in Fig. 4A, the progesterone profiles of the dominant (Grécia) and subordinate (Glenda) females fluctuated, and during this period we observed two births from the dominant female (12 March and 15 August 2000). An increase in progesterone levels for the subordinate female was recorded during this period, which was compatible with ovarian cyclicity. Although fecal progesterone remained high for 5 months (from 8 December 1999 to 3 May 2000, as shown in Fig. 4A, which corresponds to the duration of a full pregnancy), no infants were seen in the group. As previously presented (Fig. 3), sexual interactions between the subordinate female and an extragroup male and another male within her group were recorded. The subordinate female showed a new increase in progesterone 2 months later. From that time on, fecal collection became discontinuous and most of the Q-group members disappeared, including the dominant female, breeding male, and subordinate female. Immediately before the dissolution of the group, this subordinate female gave birth to two live infants, which, along with their mother, were not seen afterward. A positive correlation between the cortisol levels of the dominant and subordinate females (r=0.54, Po0.01) was found during gestation #1 of the dominant, which was coincident with the maintained increase in progesterone of the subordinate. Group PBf. This group was formed by fission from group PB after the reproductive female died in April 1997 (see Lazaro-Perea  for a full description). Between April and November 2000, two females bred, but one of the infants of the subordinate (Paloma) was killed by the dominant female (Patricia), who gave birth 1 month later. The data presented here correspond to the monitoring carried out from December 1997 to March 1999. These are the same two females that were previously studied [Lazaro-Perea, 2000]. Figure 4B shows fecal progesterone levels in the dominant and subordinate females of the PBf group. During this time, the subordinate female reproduced twice, whereas the dominant female reproduced once. Six days after parturition the subordinate female’s infants were killed by the dominant female, which gave birth 1 month after the subordinate. Both females remained cycling in the following period until the subordinate emigrated 4 months later (August 1997). The dominant female died in September 1997, and only 2–3 days afterward the subordinate female returned to the group and remained as the breeding female. She gave birth again in February 1999. 44 / Sousa et al. Q 2000 D S 1500 1000 500 10/4/2000 9/4/2000 8/4/2000 7/4/2000 6/4/2000 5/4/2000 4/4/2000 3/4/2000 2/4/2000 1/4/2000 12/4/1999 11/4/1999 10/4/1999 0 9/4/1999 Fecal progesterone (ng/g) A Date B PBf * 1500 1000 + 500 36246 36218 36195 36173 36148 36117 36099 36074 36042 36021 35991 35965 35941 35916 35895 35874 35847 35817 0 35782 Fecal progesterone (ng/g) 2000 Date C * 1500 1000 500 36825 36787 36763 36739 36720 36697 36677 36658 36645 36627 36601 36575 36559 36539 36512 36497 36481 36468 0 36449 Fecal Progesterone (ng/g) T 2000 Date Fig. 4. Fecal progesterone profile of dominant and subordinate females in groups Q, PBf, and T during monitoring. The dotted vertical line indicates parturition of the subordinate female, and the full line indicates that of the dominant females. The asterisk (n) indicates the occurrence of confirmed (group PBf) or suspected (group T) infanticides. The arrow corresponds to the date when a subordinate emigrated from her natal group, and + corresponds to the death of the dominant female. Group T. Group T was monitored for a total of 12 months, from 16 October 1999 to 16 October 2000. Similarly to observations made in group PBf, the subordinate female gave birth twice, and her first birth (8–9 February) occurred only 1 month before the parturition of the dominant (9 March), as Monogamy in Wild Common Marmosets / 45 illustrated in Fig. 4C. The offspring of the subordinate were likely killed, and although no newborn infanticide was witnessed, aggressive episodes from the dominant male and female directed toward the subordinate female suggest the occurrence of infanticide. Another finding that suggests infanticide was a cortisol increase in both females around that time, as was noted in group PBf when infanticide occurred. The subordinate female of group T remained in the group without any apparent sign of pregnancy until April 2001 (when she was last seen in the field station area, a month before the entire group disappeared). Cortisol The mean values for cortisol excreted in feces were similar for dominant (129.12722.09 ng/g) and subordinate (116.97720.02 ng/g) females. The mean values for cortisol when subordinate females were cycling were higher than when they were not cycling (193.117476.91 ng/g, and 47.69790.78 ng/g, respectively). Also the values for fecal cortisol were higher during gestation #2 (30.92757.32 ng/g) compared to gestation #1 (298.607381.38 ng/g) for the subordinate female of PBf group, but not for the two successive pregnancies of the dominant females in the Q and T groups (173.467349.11 ng/g, and 122.357345.27 ng/g, respectively). During the two episodes of infanticide perpetrated by dominant females toward subordinate offspring in groups T and PBf, acute increases in cortisol were recorded in both dominant and subordinate females, as shown for group PBf in Fig. 5. For subordinate females, the increase was about twice that observed for dominants. Fecal cortisol (ng/g) 2000 * + 1500 1000 500 3/13/1999 2/10/1999 1/16/1999 12/19/1998 11/13/1998 10/24/1998 9/19/1998 8/19/1998 7/22/1998 6/19/1998 5/22/1998 4/24/1998 3/31/1998 3/1/1998 1/25/1998 12/18/1997 0 Date Fig. 5. Concentration of fecal cortisol excreted by dominant (-’-) and subordinate (–J–) females of group PBf during the monitoring. The dotted line indicates parturition (n=2) of the subordinate female, and the full line shows the parturition of the dominant. The arrow corresponds to the date when a subordinate emigrated from her natal group, + corresponds to the death of the dominant female, and n corresponds to the infanticide episode. 46 / Sousa et al. DISCUSSION The findings of our study show that a preferential relationship between breeding males and females occurs in wild monogamous common marmoset groups, even when more than one female gives birth. We found that allogrooming reciprocity within the dominant pair was significantly higher than that observed between breeding males and subordinate females, and between dominant and subordinate females. Also, males were the main groomers in all three groups, as recorded for six groups (group Q in a different period, and five other groups) in the same field area [Lazaro-Perea et al., 2004]. Although we found high mean values for scent-marking behavior performed by subordinates, they were not significantly different from those of dominant females. Our results are similar to those recorded for five groups of common marmosets living in the same field station, including group Q (prior to our study) [Lazaro-Perea et al., 1999]. This study found no strict correlation between scentmarking and reproductive dominance or territoriality. Higher scent-marking during the noncycling periods for subordinates may indicate avoidance of agonism from dominant females, since scent-marking during the cycling period may be seen as a challenge to the dominant’s status. Complete suppression of scentmarking by subordinates was not observed, given the importance of this behavior for attracting potential mates from neighboring groups, as suggested by LazaroPerea et al. . Other evidence from the literature points out the flexibility of this behavior in common marmosets, which frequently varies according to social context [Epple, 1970; Lazaro-Perea et al., 1999], light/dark cycle [Nogueira et al., 2001], and reproductive condition [Evans & Poole, 1983, Nogueira et al., 2001], among other factors. The mean dominant and subordinate female cortisol fecal levels were similar; however, we were unable to analyze basal levels because the females all experienced different reproductive conditions during the study, including noncycling, cycling, pregnancy, and postpartum phases. Consequently, since cortisol levels depend on the physiological condition of the female, these data could not be properly interpreted. However, cortisol levels increased dramatically during infanticide, and this may safely be regarded as an indication of extreme stress in both dominant and subordinate females (but mainly in the subordinates whose offspring were killed). Aggression was mostly low in the wild groups we monitored. Nevertheless, on the few occasions that we witnessed it (infanticide episodes included), it was directed from the dominant toward the subordinate female. This is in line with a previous study in which aggression between a mother and daughter was mild and infrequent [Saltzman et al., 2004]. The restricted cortisol peaks we recorded for wild females are likely related to specific conditions. This suggests that subordinate females are able to thwart potential aggression by the dominant female using mechanisms that have not yet been identified. In general, the progesterone fecal profiles during pregnancy showed a regular, repeated fluctuation pattern similar to that of plasma. However, the expected maintained progesterone increase, from the 12th week until shortly before parturition, was not well characterized on all occasions (for instance, during Tina’s pregnancy). In these cases, we used the pregnancy duration to estimate the beginning of the gestation period [Hearn, 1983]. Sexual activity (copulations and copulation attempts) by the dominant females in all three groups was observed exclusively with the breeding male. Subordinate females from all three groups were seen copulating with extragroup Monogamy in Wild Common Marmosets / 47 males. We recorded five births by dominant females and three by subordinate females in groups Q, PBf, and T, but only the infants born to the dominant females survived. Breeding vacancies are rare and unpredictable for female common marmosets. Therefore, nonbreeding females have to decide among many alternatives, such as delaying reproduction and remaining in their natal group, contesting the dominant female and attempting to breed in her natal group (which may open the way for inheriting the breeding position) with either an extragroup male or the breeding male (if he is not her father), filling in a breeding vacancy in a neighboring group, or migrating and trying to find a breeding vacancy [Arruda et al., in press]. All of these alternatives imply costs, and their potential for success is varied. Subordinate common marmoset females may or may not escape reproductive suppression. Data from captive groups show that up to 50% of subordinate females escape this suppression in family groups [Abbott, 1984; Saltzman et al., 1997a]. Recent hormonal data from wild groups suggest there is usually more than one ovulating female in every group [Albuquerque et al., 2001, 2004]. Alencar et al.  suggested that the ability to escape reproductive suppression is related to the type of dominant relationship that exists between dominant and subordinate females. Our results suggest that at least one subordinate female in each wild group is able to escape reproductive suppression. Although Emlen  considered that this may demonstrate increased tolerance by dominant individuals, allowing for limited breeding by subordinates, Clutton-Brock  offered a critique of this model and suggested that subordinates breed simply because dominants are unable to prevent it. Saltzman et al.  found that the replacement of the breeding male by an unrelated male in captive family groups may favor reproduction by subordinate females (daughters), depending on their relationship with their mothers, because only nonsubmissive, cycling subordinates were able to breed in the presence of a new, nonrelated male. Both our behavioral and hormonal data are in agreement with the conclusions drawn by Clutton-Brock  and Saltzman et al. . The allogrooming and aggression between females, and the differential relationship of grooming and sexual behavior between the breeding male and the two females suggest that there is a hierarchy between the females, and also that the bond between the male and the dominant female is stronger than that with the subordinate female. Nevertheless, subordinate females in all groups bred (in group Q, the subordinate female gave birth to two infants a few months after hormonal monitoring had ceased). This suggests that the dominant female was not able to prevent subordinates from either cycling or breeding. Subordinate females face another constraint, which is finding a suitable mate. In some groups, the breeding male may be related to subordinates (as the father or a brother) [Araújo, 1996]. Extragroup males could fit the role of unrelated mate, which would explain the extragroup copulations observed in the current study and by Arruda et al. [in press] in these and other groups from the same population. In fact, extragroup males are the likely source of social cues that induce the activation of the reproductive function in subordinate females, as suggested by Saltzman et al. [1997b, 2004]. The idea of extragroup males as potential or, occasionally, actual mates would also explain the higher levels of scent-marking by subordinate females that have also been reported for the same population [Lazaro-Perea et al., 1999]. However, group PBf does not fit this model. In contrast to the other two groups, the subordinate female received more grooming from the breeding male 48 / Sousa et al. compared to the dominant female, and was seen copulating with the breeding male, as well as with extragroup males. This group was formed after another group (PB) dissolved when the breeding female died. Group PB split [LazaroPerea, 2000] into two groups: all females in one group (PBf), and all males in the other (PBm). When we initiated this study, the PBf group was restructuring and both females were likely kin and competing for the breeding position. Paloma, the subordinate female, gave birth twice, and on both occasions her offspring were killed by Patricia, the dominant female [Arruda et al., 2005]. The infanticide witnessed in the PBf group shows that escaping reproductive suppression and finding a suitable mate are only the first steps toward successful breeding. Offspring survival depends on the type of relationship the subordinate mother has with the dominant female, and to what extent the latter will tolerate a second breeder in the group. Polygyny has been reported in wild C. jacchus [Arruda et al., 2005; Barreto, 1996; Digby & Ferrari, 1994], but if its costs are so high that infants’ survival is highly unlikely, other alternatives must be sought. This was the case for these three females. Paloma, the PBf subordinate female, tried unsucessfully to gain a breeding position. After two failed attempts to breed, she left the group. Shortly afterward the breeding female, Patricia, died, and Paloma returned to the PBf group. There she reproduced successfully with the same breeding male. The other two subordinate females left their groups after they lost their infants, and we have no further information regarding them. This outcome suggests that the main reason for migration is competition with the breeding female, even if this competition is not reflected in agonism levels. The likely costs of remaining in the group without breeding opportunities are higher than the possible costs of migrating. Data from other subordinate females from the same population suggest that migrating females may eventually find a breeding position in groups that have lost the breeding female, or groups that split and form two new groups [Arruda et al., 2005]. In conclusion, in response to the three questions posed in this study, our results show that the social dynamics of dominant and subordinate females were different in monogamous groups because the dominant females and breeding males spent more time on reciprocal grooming, and in at least two of the groups maintained exclusive sexual interactions, suggesting a preferential relationship between them in comparison with breeding males and subordinate females. We also found that under field conditions, the subordinate females’ reproductive inhibition appeared to be more behavioral than hormonal, since all monitored subordinate females showed ovarian functioning over the course of the study. In two groups (PBf and T) the subordinate females were able to reproduce, although their infants did not survive. We identified three different reproductive strategies used by subordinate female common marmosets living in wild monogamous groups: 1) staying in the natal group as nonreproductive members for a period of time and waiting for a more favorable opportunity to reproduce, 2) attempting to reproduce as a second female and emigrating unsuccessfully, and 3) eventually returning to the natal social group to occupy the breeding-female vacancy. ACKNOWLEDGMENTS We thank Herbert M. Santos, Maria Carla Lopes do Nascimento, and Patricia Guilhermino Ledo for their help with the fecal collection and hormone assays. We also thank the Brazilian Institute for the Environment and Renewable Resources (IBAMA) for permission to conduct this research at Flona, Nisia Floresta, RN. We Monogamy in Wild Common Marmosets / 49 are grateful to two anonymous referees for suggestions that improved the manuscript. This work was supported by grants from CNPq to A. Araujo (464103/ 00-2), M.B.C. Sousa (52.1186/1997, 305216/2003-5, and 470601/2003-1), and M.E. Yamamoto (524409/96), and from CAPES to A.C.S.R. Albuquerque and A. Araujo (864/91-11). REFERENCES Abbott DH. 1984. Behavioral and physiological suppression of fertility in subordinate marmoset monkeys. Am J Primatol 6:169–186. Abbott DH. 1987. Behaviourally mediated suppression of reproduction in female primates. J Zool Lond 231:455–470. Abbott DH, Barret J, George LM. 1993. Comparative aspects of the social suppression reproduction in female marmosets and tamarins. In: Rylands AB, editor. Marmosets and tamarins: systematics, behaviour, and ecology. Oxford: Oxford University. p 235–256. Abbott DH, Saltzman W, Schultz-Darken NJ, Tannenbaum PL. 1998. Adaptations to subordinate status in female marmoset monkeys. Comp Biochem Physiol C 119: 261–274. Albuquerque ACSR, Sousa MBC, Santos HM, Ziegler TE. 2001. Behavioral and hormonal analysis of social relationship between oldest (reproductive and non-reproductive) females in a wild monogamous group of common marmoset, Callithrix jacchus. Int J Primatol 22:631–645. Albuquerque ACSR, Nascimento MCL, Santos HM, Sousa MBC. 2004. Função ovariana e adrenocortical de fêmeas adultas de sagüi (Callithrix jacchus) em relação ao comportamento de emigração. In: Mendes S, Chairello A, editors. A primatologia no Brasil. Vol. VIII. Santa Teresa: Sociedade Brasileira de Primatologia/Museu Melo Leitão. p 301–316. Alencar AI, Yamamoto ME, Oliveira MS, Lopes FA, Sousa MBC, Silva NG. 1995. Behavior and progesterone levels in Callithrix jacchus females. Braz J Med Biol Res 28:591–595. Araújo A. 1996. Influence des facteurs écologiques, comportementaux et démografhiques sur la dispersion de Callithrix jacchus. Ph.D. dissertation, Université Paris-Nord (Paris XIII), Paris, France. 234 p. Arruda MF, Araújo A, Sousa MBC, Albuquerque FS, Albuquerque ACSR, Yamamoto ME. 2005. Two-breeding females within free-living groups may not always indicate polygyny: alternative subordinate female strategies in common marmosets (Callithrix jacchus). Folia Primatol 76:10–20. Clutton-Brock TH. 1998. Reproductive skew, concession and limited control. Trends Ecol Evol 13:288–292. Coutinho PEG, Correia HKM. 1995. Polygyny in a free-ranging group of buffy-tufted-ear marmosets, Callithrix aurita. Folia Primatol 65:25–29. Dietz JM, Baker AJ. 1993. Polygyny and female reproductive success in golden lion tamarins, Leontopithecus rosalia. Anim Behav 46:1067–1078. Digby LJ, Ferrari SF. 1994. Multiple breeding females in free-ranging groups of Callithrix jacchus. Int J Primatol 15: 389–397. Emlen ST. 1991. Evolution of cooperative breeding in birds and mammals. In: Krebs JR, Davies NB, editors. Oxford: Blackwell. p 301–337. Epple G. 1970. Quantitative studies on scent marking in the marmoset (Callithrix jacchus). Folia Primatol 13:48–62. Evans S, Poole TB. 1983. Pair-bond formation and breeding success in the common marmoset (Callithrix jacchus jacchus). Int J Primatol 4:83–97. Hearn JP. 1983. The common marmoset (Callithrix jacchus). In: Hearn JP, editor. Reproduction in New World primates. Lancaster: MTP Press. p 181–215. Lazaro-Perea C, Snowdon CT, Arruda MF. 1999. Scent marking behavior in wild groups of common marmosets (Callithrix jacchus). Behav Ecol Sociobiol 46: 313–324. Lazaro-Perea C. 2000. Behavioral and demographic changes following the loss of the breeding female in cooperatively breeding marmosets. Behav Ecol Sociobiol 48: 137–146. Lazaro-Perea C. 2001. Intergroup interactions in wild common marmosets, Callithrix jacchus: territorial defense and assessment of neighbours. Anim Behav 62:11–21. Lazaro-Perea C, Snowdon CT, Arruda MF. 2004. Grooming as a reward? Social function of grooming between females in cooperatively breeding marmosets. Anim Behav 67:627–636. Nogueira SL, Sousa MBC, Medeiros CFN, Costa MPO. 2001. Diurnal variation in 50 / Sousa et al. scent marking behavior in captive male and female common marmoset, Callithrix jacchus. Biol Rhythm Res 32:169–177. Roda AS, Pontes ARM. 1998. Polygyny and infanticide in common marmosets in a fragment of the Atlantic forest of Brazil. Folia Primatol 69:371–376. Rylands AB. 1986. Infant-carrying in a wild marmoset groupo, Callithrix humeralifer: evidence for a polyandrous mating systems. In: Mello MT, editor. A primatologia no Brasil 2. Rio de Janeiro: Sociedade Brasileira de Primatologia. p 131–144. Saltzman W, Schultz-Darken N, Abbott DH. 1997a. Behavioral and social correlates of escape from suppression of ovulation in female common marmosets housed with natal family. Am J Primatol 41:1–21. Saltzman W, Schultz-Darken N, Abbott DH. 1997b. Familial influences on ovulation function in common marmosets (Callithrix jacchus). Am J Primatol 41: 159–177. Saltzman W, Schultz-Darken NJ, Wegner FH, Wittwer DJ, Abbott DH. 1998. Suppression of cortisol levels in subordinate female marmosets: reproductive and social contributions. Horm Behav 33:58–74. Saltzman W, Pick RR, Salper OJ, Leidl KJ, Abbott DH. 2004. Onset of plural cooperative breeding in common marmoset families following replacement of the breeding male. Anim Behav 68:59–73. Santee D, Arruda MF. 1994. The Nı́sia Floresta common marmoset research station. Neotrop Primates 2:8–11. Smith TE, Abbott DH. 1998. Behavioral discrimination between circumgenital odor from peri-ovulatory dominant and anovulatory female common marmoset (Callithrix jacchus). Am J Primatol 46:265–284. Snowdon CT. 1996. Infant care in cooperatively breeding species. In: Rosenblatt JA, Snowdon CT, editors. Parental care: evolution, mechanisms and adaptative significance. San Diego: Academic Press. p 643–689. Sousa MBC, Ziegler TE. 1998. Diurnal variation on the excretion patterns of steroids in common marmoset (Callithrix jacchus) females. Am J Primatol 46:105–117. Sousa MBC, Albuquerque ACSR, Nascimento MCL. 2002. Fecal collection in wild common marmosets, Callithrix jacchus. Neotrop Primates 10:21–24. Ziegler TE, Scheffler G, Wittwer DJ, SchultzDarken NJ, Snowdon CT, Abbott DH. 1996. Metabolism of reproductive steroids during the ovarian cycle in two species of callithrichids, Saguinus oedipus and Callithrix jacchus, and estimation of the ovulatory period from fecal steroids. Biol Reprod 54:91–99. Ziegler TE, Sousa MBC. 2002. Parent–daughter relationships and social controls on fertility in female common marmosets, Callithrix jacchus. Horm Behav 42:356–367.