Contextual Complexity of Chemical Signals in Callitrichids

Abstract

In nearly 4 decades our research and that of others on chemical signaling in Callitrichid primates suggests a high degree of contextual complexity in both the use of signals and the response to these signals. We describe our research including observational field studies, behavioral bioassays (“playbacks”), functional imaging, and conditioning studies. Scent marking in both captive and the wild is used for more than just territorial marking. Social contextual effects are seen in responses by subordinate females responding with ovulatory inhibition only to chemical signals from familiar dominant reproductive females. Males detect ovulation through changes in scent marks. Males responded behaviorally and hormonally to chemical signals of novel ovulating females as a function of their reproductive status (fathers, males paired with a female but not fathers, and single males). Multiple brain areas are activated in males by female chemical signals of ovulation including areas relating to memory, evaluation, and motivation. Furthermore, males can be conditioned to respond sexually to a non-sexual odor demonstrating that learning plays an important role in response to chemical signals. Male androgen and estrone levels changed significantly in response to infant chemical signals as a function of whether the males were fathers or not, whether the odors were from their own versus other infants, as well as the infant’s stage of development. Chemical signals in Callitrichids are providing a rich source of understanding the context and function of the chemical sensory system and its stimulation of neural, behavioral and hormonal actions in the recipients.

Our research collaboration began at the start of 1983. Our goals were to understand the reproductive behavioral biology of Callitrichid primates, but studying chemical signals soon became an important part of our work. It was obvious that scent marking was an important behavior, and scent glands were visible in our female cotton-top tamarins (Saguinus oedipus). The true pioneer of Callitrichid chemical signals was Gisele Epple, and her research on tamarins had shown that chemical signals provided information on species, subspecies, reproductive state, sex and individual identification. We collaborated with and adopted many of her methods in our own studies (e.g. Ziegler et al., 1993). In a review chapter Epple (1986) described the presence of scent glands throughout primates, from prosimians to great apes, but Callitrichids were among the taxa with the most obvious chemical signaling.

Over the course of our research we have come to appreciate the complexity and importance of chemical signals and their responses in Callitrichid behavior. We recognize that the responses to these signals are not simple or innate but contextually dependent. Olfaction and taste cannot be easily separated and given that the vomeronasal gland (involved in processing chemical signals in many species) opens through the mouth, it is difficult, if not impossible, to separate the two senses. Thus, “chemical signaling” is a better term than either “olfactory” or “gustatory” and we will use that term here, but when obvious marking behavior is observed, we use the term “scent mark”.

Drea (2015) has provided a recent comprehensive review of chemical signals involved in reproduction in primates, which is one important aspect of chemical signaling. Ziegler (2013) has reviewed chemical and other signaling in marmosets and tamarins in the context of reproductive function. The present review focuses on a broader range of topics involving chemical signals. Marmosets and tamarins have been the subject of a variety of studies on chemical signals, and they are also the species we have studied and where we have the most expertise and what we focus on here.

We will briefly review the sources of chemical signals and their structure and then discuss the studies we and others have done to illustrate the contexts and functions of chemical signals in Callitrichids. We have used a variety of methods including observations, playbacks (behavioral bioassays), functional brain imaging and conditioning studies.

Sources of chemical signals and detection

Epple (1986) described the presence of anogenital, suprapubic, and sternal glands in each Callitrichid species for which data were available. In many species overt scent marking was equal between sexes, but in saddleback tamarins (Saguinus fuscicollis) and cotton top tamarins adult females had more developed glands than adult males and marked more frequently. The presence of multiple gland regions suggested that each might have evolved with separate secretions and distinct functions. In addition to specific glandular secretions there are many other potential sources of chemical signals including sweat, urine, feces, saliva and vaginal secretions, which can potentially be mixed with glandular secretions to create complex signals. Epple (1986) noted that urine is typically mixed with circumgenital secretions in each of the Callitrichid species she studied.

The vomeronasal organ is part of the secondary olfactory pathway and has been implicated in pheromone detection in other species. Anatomical work by Evans (2006) showed that Callitrichid primates have functioning accessory olfactory systems with vomeronasal organs that could lead one to conclude that these scent marks functioned as chemical signals affecting the internal physiological state of the donor.

Measurements of the chemical composition of scent marks have revealed complex chemical composition of scent marks. Belcher, Epple, Küderling & Smith (1988) analyzed the components of cotton-top tamarin scent marks using gas chromatography/mass spectrophotometry and found several different organic compounds within the mark. Smith, Tomlinson, Mlotkiewicz & Abbott (2001) found similar complexity in the scent marks of common marmosets (Callithrix jacchus). They found both different concentrations and different chemicals between individuals showing that chemical signals could be individually specific. Recognition of species, subspecies, sex and individuals based on chemical signals has been shown in several studies (e.g. Epple, 1986, Smith, 2006; Smith, Abbott, Tomlinson, & Mlotkiewicz, 1997). Thus, the information we have on scent mark composition in Callitrichids suggests that they are odor mixtures with the potential to code complex information.

Chemical signals and territory defense

In our first study of chemical signals (French & Snowdon, 1981) we observed that females used suprapubic marking in the context of territorial disputes, when challenged with a novel intruder, but they appeared to use anogenital marking in a sexual context, since male partners appeared to be attracted to these marks. Observations of scent marking in wild marmosets and tamarins have found that scent marking functions for more than territorial defense. Marmosets frequently anogenital scent mark around gouges in trees that provide for exudate feeding. Aggression has been observed at these feeding sites, suggesting competition for access. Two studies of Callithrix penicillata hypothesized that the marks denoted ownership of the tree gouges, which must be prepared hours in advance of any exudate feeding (Lacher, da Fonseca, Alves & Magalheas-Castro, 1981; Rylands, 1985). Only adults were observed to scent mark at exudate sources and mainly at new feeding sources. However, multiple groups fed at the same locations (albeit at different times, Lacher et al., 1981) countering a simple territory defense hypothesis. Other hypotheses are that these marks communicate reproductive status to other group members and may serve to maintain reproductive suppression in subordinates (Rylands, 1985, see below).

In wild common marmosets (Lazaro-Perea, Snowdon & Arruda, 1999), animals marked throughout their home ranges during travel and intergroup encounters with no increase in marking at territory boundaries. As with the other marmoset studies, scent marks were more common at exudate trees than on fruiting or sleeping trees. However, subordinate females were observed to mark more often than the reproductive female, possibly due to the greater involvement of subordinate females in border encounters with other groups (Lazaro-Perea, 2001). In studies of wild saddle-back and moustached (Saguinus mystax) tamarins Heymann (1998, 2001) observed much higher rates of suprapubic marking in female moustached tamarins than in males with females more often over marking male scent marks. In saddleback tamarins, there was no evidence that scent marks served to mark territories. Marks did not increase during intergroup encounters, did not lead to avoidance of encounters, and did not prevent intrusions or overlap in feeding resources (Lledo-Ferrer, Pelaez & Heymann, 2011). These authors suggested that marks may serve to exchange information about reproductive opportunities, similar to what was proposed for common marmosets by Lazaro-Perea (2001). Furthermore, sympatric saddleback tamarins and moustached tamarins differed in rate, intensity, degree of over marking and substrates used for marking despite the two species being overall very similar in use of their common habitat (Heymann, 2001). In golden lion tamarins (Leontopithecus rosalia) Miller, Laszlo & Dietz (2003) noted that males used scent marks to communicate dominance whereas females did not. Reproductive females did scent mark in the presence of other groups. Lion tamarins also marked location of food sources. Overall, field researchers have concluded that scent marking does not have a territorial function in Callitrichids. Roberts (2012) has argued that all mammalian scent marking has a territorial function and criticized the conclusions of Lledo-Ferrer et al. (2011). Those authors in turn have argued that there are likely to be multiple functions, including advertising status and marking resources (Lledo-Ferrer, Pelaez and Heymann, 2012). In summary, in contrast to our early observations and conclusions on captive tamarins, field observations showed that marking is not exclusively territorial in function, but occurs in multiple contexts. The many species differences present complications in terms of providing simple functional explanations for scent marking.

Reproductive inhibition and sexual maturation

Studies in tamarins and marmosets have found clear developmental patterns among females. Females living as non-reproductive animals in family groups rarely scent marked and they had poorly developed scent glands, but, upon removal from the family, scent glands developed rapidly and scent marking increased. The change in gland development and scent marking coincided with an increase in production of estrogen hormones and the onset of reproductive cycling (cotton-top tamarins, French, Abbott & Snowdon, 1984; French and Cleveland 1984; Heistermann, Kleis, Pröve & Wolters, 1989; saddle-back tamarins, Epple & Katz 1984; common marmosets, Abbott & Hearn, 1978). These studies, among others, suggest that female marmosets and tamarins, which are not the breeding female, are reproductively suppressed and, indeed, in tamarins typically only one female ovulates and reproduces within a family group. We extensively sampled over 30 subordinate females and not a single one ever ovulated within their natal group while living in intact families (Snowdon & Ziegler, 2007).

When a subordinate female is removed from a family group, she starts reproductive cycling (in as little as 8 days), suggesting that some cues from the family may be involved in maintaining suppression of ovulation. Marmosets and tamarins can discriminate between the odors of reproductively cycling and non-cycling females (common marmosets, Smith & Abbott, 1998; cotton-top tamarins, Washabaugh & Snowdon, 1998). One possible hypothesis is that chemical signals from the reproductive female may be involved in inhibiting ovulation in other females. To test this, one can do a “playback” study similar to those done in studies of vocal communication. Several researchers have collected samples of scent marks from the reproductive female and made daily transfer of the scents to the previously subordinate females after they had been removed from the family group. In common marmosets (Barrett, Abbott & George, 1990), saddleback tamarins (Epple & Katz, 1984) and cotton-top tamarins (Savage, Ziegler & Snowdon, 1988), the onset of ovulation was delayed significantly compared with control females who did not receive scent transfers.

Our initial thought was that there was a fixed pheromone production common to all reproductive females that had a uniform effect on subordinate females. We began to look for a specific chemical that might serve as the “suppression chemical” without success. Abbott, Saltzman, Schultz-Darken and Smith (1997) subsequently found that reproductive inhibition through scent transfer was effective only on those female common marmosets that had previously lived with the scent donating female. When presented with scents transferred from an unfamiliar, reproductive female, previously subordinate females ovulated as quickly as those who received no scent transfers (Figure 1). The implication of this finding is that reproductively subordinate females are not being influenced innately by a specific chemical, but rather that they have learned to recognize the unique chemical signature of the reproductive female with whom they had lived, and they are suppressed (or inhibited) only by that specific chemical signature.

Figure 1.

Days to ovulation in previously subordinate female cotton-top tamarins and common marmosets after removal from their natal group and pairing with a mate as a function of transfer of odors from a familiar or unfamiliar reproductive female. Data from Savage et al. (1988), Barrett et al. (1990), Abbott et al. (1997).

Male attraction to female odors

In our first study (French & Snowdon, 1981) we observed that anogenital marking appeared to be used to attract males. Marmosets and tamarins show no directly observable signs of ovulation- no sexual swelling, no changes in coloration of genital regions, no signs of menstruation. We observed what appeared to be cyclic patterns of males investigating females’ anogenital regions and scent marks, but once we were able to directly determine the day of ovulation through noninvasive measures of estrogen (Brand & Martin, 1983; French et al., 1984), we found that this increased interest in scent marks was not linked to ovulation but actually occurred early in pregnancy. However, subsequent work (Ziegler, Bridson, Snowdon & Eman, 1987) measuring both luteinizing hormone and estrogen found that urinary estrogen peaked after ovulation unlike prior to ovulation as seen in most other primates..

Nonetheless, adult captive cotton-top tamarins have a very high rate of conception. The vast majority (85%) of the ovulations that we recorded in our colony through hormonal measures resulted in pregnancy (Ziegler et al, 1987). There are two possible mechanisms. Copulations occur frequently throughout the ovulatory cycle as well as during pregnancy so one solution is simply to copulate daily, with no signals needed. Alternatively, there may be a chemical signal not observable by human researchers, which informs males when ovulation occurs. Ziegler et al. (1993) collected a scent mark on a glass stopper each day for 25 days from an ovulating female tamarin (normal cycle length is 21 days) and transferred that scent to the cage of another tamarin pair where the female was pregnant and, therefore, not ovulating. We took measures of frequency and latency to approach and contact the scent, sniffs or licks of the stopper, anogenital and suprapubic marking as well as male erections, sniffs and mounts of his mate. The observer was blind to the condition of the scent donor female, and 8 pairs were tested. Daily hormonal samples from the donor female were used to determine the periovulatory period (the day before, day of, and day after ovulation). Recipient females made more frequent contact with the scent platform on the donor’s periovulatory days but otherwise showed no difference in sniffing, licking or over marking the scent. Males were equally likely to explore the scent regardless of ovulatory state. However, males showed significantly more mounts of their own mate and rates of erection on days when the samples were from the periovulatory period than they did to marks of the same donor female when she was not ovulating. The results were clear that males can detect ovulation in females, based on changes in chemical signals. Qualitative aspects of the chemical signal appear to vary with the ovarian cycle.

How do females respond to other ovulating females? Washabaugh and Snowdon (1998) presented scents from the reproductive female in a group along with scents from novel, non-reproductive females and novel, reproductive females. These were collected at times other than the periovulatory period. We presented the scents in counter-balanced order on separate days and recorded behavioral measures of male and female response. Females showed increased olfactory investigation of and spent more time with novel scents, compared with their own. Furthermore, females displayed increased sexual solicitation toward their own mates in the presence of a scent from a novel, reproductive female but not with the scent of a novel, non-reproductive female. Males showed similar increases in olfactory investigation of and time spent with scents from novel reproductive females, but, unlike females, showed no changes in sexual behavior as a function of scent stimulus. In a similar study with common marmosets, Smith and Abbott (1998) found that both sexes showed more investigation of novel, reproductive versus novel, non-reproductive female scents but only with the periovulatory scents of reproductive females. Taken together these three studies indicate that marmosets and tamarins can detect the reproductive status and ovulatory condition of females, and they respond differently to familiar versus unfamiliar females.

Our next question was whether males react differently to novel female scents as a function of the male’s reproductive status. Ziegler, Schultz-Darken, Scott, Snowdon, and Ferris (2005) presented male common marmosets with scents from novel ovulating females or vehicle for 10 minutes and then obtained a blood sample for hormonal monitoring 30 minutes after first presentation of scents. Both the rate of sniffing and duration of erections were significantly greater in response to the novel female scent than to the vehicle control. Non-significant increases in rates of licking, touching and scent marking were observed as well as a non-significant decrease in latency to first sniff. Serum testosterone levels increased significantly within 30 minutes of initial presentation of the ovulatory odors, but we observed no change in cortisol levels. In a second experiment, males were presented with scents from novel ovulating females or a vehicle control. Serum levels were measured 30 minutes after initial exposure. Serum testosterone levels increased in some, but not all, males. Males who were fathers showed no difference in testosterone levels from control levels whereas both paired males (without infants) and singleton males showed significant increases in serum testosterone in response to ovulatory scents compared with vehicle (Figure 2). Thus, how males responded to novel females was conditional on their reproductive status. Paternal males appeared to show little interest in cues from novel, reproductive females. We suggest that male marmosets have a natural behavioral and neuroendocrine response to sexually relevant cues such as ovulatory scent marks from novel females. However, under stable family conditions, there may be an inhibitory or a dampening process that prevents males from exhibiting a full response.

Figure 2.

Changes in testosterone levels within 30 minutes of presentation of an odor from a novel, ovulating female as a function of male social status. There was no difference in levels for parental males, but significant increases in testosterone for paired and single males (P’s < 0.05) (adapted from Ziegler et al., 2005).

We then became interested in how these chemical signals are processed in the brain. As was noted earlier, Evans (2006) has shown that there are functioning vomeronasal organs and accessory olfactory pathways in many Callitrichid species. These have been hypothesized to be important in the processing of chemical signals. In addition, lesioning studies in common marmosets showed that the anterior hypothalamus and medial preoptic areas were important in regulating male sexual arousal (Lloyd & Dixson 1988). Male pre-copulatory and copulatory behavior was significantly impaired in lesioned males. Females continued to sexually solicit lesioned males despite their lack of responsiveness (Dixson & Lloyd, 1989). Given that non-paternal males respond behaviorally and hormonally to scents from ovulating females, it was logical to hypothesize that chemical signal information is conveyed through the accessory olfactory system to target the anterior hypothalamus and medial preoptic areas to induce sexual arousal.

We participated in the first study to use non-invasive functional magnetic resonance imaging (fMRI) methods in awake, conscious common marmosets. Based on the lesioning work of Dixson and Lloyd, we hypothesized that ovulatory scents would lead to arousal in the anterior hypothalamus and medial preoptic areas of male marmosets. Functional magnetic resonance imaging measures changes in blood oxygenation levels. Neuronal activity is thought to increase metabolic rate leading to changes in blood flow and volume to the active area. Deoxygenated hemoglobin is paramagnetic and leads to reduced magnetic resonance signal. Oxygenated blood has little magnetic susceptibility so that areas with enhanced blood flow will have a stronger magnetic resonance signal. (See Ferris & Snowdon, 2005, for more on the theory behind fMRI and methods of testing nonhuman primates.)

We presented male marmosets (none of them fathers) in a magnetic resonance spectrometer with three conditions (a neutral control; scent from an ovariectomized female; scent from an ovulating female) lasting 7 minutes each with a 10-minute wash-out period between each scent presentation to clear the olfactory receptors. We predicted that there would be significant activation in the anterior hypothalamus and medial preoptic areas in response to the ovulatory scent compared with the control, and this was what we observed (Ferris, Snowdon, et al., 2001). Significantly more voxels (3 dimensional pixels) were activated in response to ovulatory scents compared to ovariectomized controls.

However, unlike lesioning or single unit recording techniques fMRI can be used for exploratory research as well as hypothesis testing. The whole brain can be imaged during stimulus presentations and one can examine a variety of other areas that might be influenced by the different types of scents. When we examined multiple areas, we found significant differences in response to scents from ovulating and ovariectomized females not only in sexual arousal areas but also in the prefrontal cortex involved in executive functioning, the temporal cortex involved in auditory perception and integration, the hippocampus involved in memory and spatial orientation, the insular cortex involved in subjective emotional experience in humans, the cingulate cortex active in learning and memory, the putamen influential in regulating movement and learning, and the cerebellum that integrates a variety of sensory inputs and controls voluntary motor actions (Ferris, Snowdon et al., 2004, Figure 3). Thus, the scents from ovulating females are not simply involved in regulating male sexual arousal but are influencing a variety of brain areas including many that are involved in sensory integration, learning, memory, attention and movement. The scents of an ovulating female marmoset are not just inducing sexual arousal but stimulate a variety of brain areas involved in cognition, decision making and behavioral regulation. These indicate a significant degree of cognitive appraisal and decision making on the part of the male as he interprets female scents.

Figure 3.

Mean number of voxels activated in some brain areas involved in sensory integration, learning, memory, attention and movement in response to odors from ovulating and ovariectomized females. Within subject comparisons significant at p < 0.05. (Adapted from Ferris et al., 2004).

The most powerful way to demonstrate that the behavioral and hormonal responses to chemical signals are not simple reflexes is to explicitly condition animals to respond when a novel, previously irrelevant odor is used as a conditioning stimulus. Although sexual conditioning has been demonstrated in a wide variety of taxa ranging from fish, through birds, rodents and humans there have been no previous studies with nonhuman primates. If we could condition males to respond sexually to a completely arbitrary stimulus, this would illustrate the importance of learning in mediating sexual responsiveness to scents.

Snowdon, Tannenbaum, Schultz-Darken, Ziegler and Ferris (2011) presented nonpaternal male marmosets with conditioning trials involving lemon scents (that they had not experienced before). In preconditioning trials males were presented with 6 20-minute sessions each with a wooden disk scented with lemon oil or with a no scent control disk. In the conditioning trials males received 12 trials with a lemon scent disk initially presented to the male. The scent stimulus was removed. The male was then given access to a cage containing a box with a novel periovulatory female who was released from the box shortly after the male was placed in the cage. The marmosets were allowed to copulate to ejaculation or to interact with each other for up to 15 minutes. These trials were counter-balanced with control trials, where males were presented with the unscented disk and placed in a cage with a novel periovulatory female. In the control condition the female was not released from the box and so no copulation occurred. After a 5 day delay, the males were tested again as in the pre-conditioning phase. Each male received 6 sessions with a lemon scented disk and 6 sessions with an unscented disk. However, as in the pre-conditioning trials, no females were in the box during these trials.

All pairs copulated to ejaculation on every lemon trial. In the pre-conditioning phase there was a mean of 2.0 approaches and interactions with the box that would contain the female in conditioning trials on both lemon and no scent trials. However, following the post-conditioning trials there was a significant difference between lemon and no scent trials, with a mean of 1.8 box directed acts in the no scent trials, but a mean of 17.5 box-directed acts on lemon trails. The strongest evidence for sexual conditioning was that erections were observed on 79% of the post-conditioning lemon trials whereas there were no observed erections on any pre-conditioning trials or to the no scent presentations following conditioning.

There are several implications to these results. Most studies of sexual conditioning in other species have tested animals either a few hours or at most one day after conditioning. In this study there was a five-day break between conditioning and testing, and testing was carried out over a 4 day period with no evidence of any reduction in response. Thus, sexual conditioning was long lasting. The fact that male marmosets can be conditioned to arbitrary cues associated with sexual activity with a female suggests that males may naturally learn about the specific chemical signals of a specific female. Marmosets form pair bonds, and males are essential to assist with infant care in these cooperatively breeding monkeys. Conditioning to the cues of one’s mate coupled with the long duration of this conditioning may be important in maintaining the relationship throughout pregnancy so that the male is available for assisting in infant care after birth. Finally, this study has demonstrated that males can be conditioned to arbitrary cues. Thus there may be nothing specific to the scents produced by females which are innately sexually arousing. These results further suggest that scents of ovulating females are individual specific and that marmoset males may learn specific characteristics of odors of females providing a possible basis for mate identification.

Male response to infant cues

In a study of male responsiveness to infant separation calls, fathers responded equally well to cries of both their own and other infants, crossing a bridge from one enclosure to another to gain access to the infant (Zahed, Prudom, Snowdon & Ziegler, 2008). Either the infants did not have distinctive calls or fathers did not discriminate between their own and other infants and are responsive to all infants. An indiscriminant response to all infants may not be adaptive if a male invests much effort in infant care. However, it is possible that males might respond to infant scents more than to infant cries. Prudom et al. (2008) presented common marmoset fathers and non-fathers with scents of infants and exposed fathers to the scents for 10 minutes. A serum sample was taken 20 minutes after scent presentation. Fathers, but not non-fathers, showed a significant reduction in testosterone after exposure to infant scents compared to vehicle control. Thus, paternal status influences hormonal responses to infant scents (as it did to odors from novel ovulating females). However, this study did not answer the question of whether fathers can identify their own infants based on scent cues.

Subsequently Ziegler, Peterson, Sosa and Barnard (2011) tested common marmoset fathers with scents of their own infants, scents of other infants, and a vehicle control both at an age when infants were still dependent on continuous care and three to four months later when the infants were fully independent of paternal care. As in previous studies the males were exposed to odors for 10 minutes with a serum collection at 20 minutes after initial exposure. Fathers showed a significant decrease in serum androgens after exposure to scents from their own infants but not to scents from unfamiliar infants. Furthermore, this decrease occurred only when the infants were in need of continuous care and not when they were independent. An opposite response was observed with serum estrogen levels. Estrogen increased in fathers exposed to scents of their own infants but not to scents of unfamiliar infants, and no changes were observed when infants were independent.

These collective results suggest that male endocrine responses to infant scent cues are conditional upon the male actually being a father; upon the scents being from the male’s own infant and not another infant, and that the scents are collected from infants when they are dependent. This is a complex, contextually-dependent response that appears to require sophisticated sensory processing.

Discussion

As we have worked on chemical signals over nearly four decades our understanding of chemical signals has moved from the conceptualization of pheromones as species-specific chemical compounds that have a direct and universal effect on all recipients within the species, to a much more nuanced vision of chemical communication as being greatly influenced by a variety of social and environmental contexts indicating cognitive complexity in the response to these signals.

Studies of chemical signals in Callitrichid primates have demonstrated that they are composed of multiple chemical components and that scents from different individuals have different distributions of chemicals allowing for individual recognition. Multiple scent glands suggest that different glands may have different functions. Since glandular secretions can be mixed with many other compounds such as urine, they provide multiple sources of signal complexity. However, in contrast to results from earlier experimental studies on captive animals, field observations of scent marking in Callitrichid monkeys have shown that marking occurs in a variety of contexts and may be used differently in different species. Chemical signal “playback” studies have shown that responses are based on who is producing the signal- a familiar or an unfamiliar animal, an ovulatory or anovulatory female, one’s own infant or a novel infant. The degree of dependence of an infant is also a factor that influences responses. Responses to chemical signals are also based on the social status of recipients. Paternal males failed to respond with increased testosterone to the odors of novel ovulatory females whereas nonpaternal males showed considerable interest. Paired males were sexually aroused by novel female odors but increased sociosexual behavior with their own mates as a response. The social environment of the recipient can have a profound influence on how chemical signals are interpreted.

Using noninvasive functional imaging technology, we have shown that multiple brain areas are activated by odors from ovulating females, incorporating not only brain areas known to be involved in sexual arousal, but also many other brain areas that are involved in sensory integration, learning, memory, cognitive evaluation and control. Thus, ovulatory scents have effects well beyond sexual arousal. Finally, in the first study of sexual conditioning in a nonhuman primate, we have shown that male marmosets can be sexually conditioned to arbitrary scents. This conditioning can be quite long lasting compared with sexual conditioning studies on other taxa. This is the strongest evidence that responses to chemical signals in Callitrichids are not simple innate, reflexive reactions. As a result our current understanding of chemical signals in Callitrichids is more complex and nuanced than it was when we first started this research 40 years ago.

Although mammalian signals have multiple chemical components, it is still possible that one or a few compounds are responsible for influencing behavior. A recent study on ring-tailed lemurs (Lemur catta) by Shirasu and colleagues (2020) found three main components in male scents that were effective in attracting attention of females. The three components acted synergistically to produce a stronger response than any single component. Thus, complex signals do not necessarily imply contextually complex responses. Contextual features need to be examined in both signal production and responses to signals.

We have described our trajectory as we learned more about chemical signals in Callitrichids, but what are the implications for understanding Callitrichid reproductive biology, our original starting point? Callitrichids are cooperative infant caregivers with both parents as well as other group members being critical for successful infant rearing. The adult mates also exhibit pair-bonding (in the captive populations we have studied) (definitions from Huck, Di Fiore & Fernandez-Duque, 2020).

This has several implications. First, mate choice decisions and intrasexual competition should be equally important to both sexes. Tamarin females develop scent glands when reproductively mature, and the secretions from these glands lead to male attraction and can serve to inhibit ovulation in subordinate females. Tamarin females respond to scents from novel reproductive females by increased sexual solicitation of their mates. The development of permanent female secondary sex characteristics at puberty is rare among nonhuman primates, but suggests the use of scents for both intrasexual competition and mate attraction. Sexually selected traits are often assumed to be more common in males than in females, but in biparental and cooperative breeding species, both sexes should show sexually selected traits.

Second, maintenance of a monogamous relationship requires identification of specific individuals. For arboreal animals ranging in densely forested habitats visual signals may be more obscure than vocal or chemical signals. The use of vocal and chemical signals for identification is likely. The presence of chemical cues from a novel female can be a potential threat to the relationships and, hence, increased sexual solicitation by females and increased erections and mountings of the female by the male may be a way to strengthen a relationship that is under potential threat.

Third, caring for infants is costly, especially when females give birth to twins that weigh up to 20% of her body weight at birth. A male, who will expend considerable energy in caring for infants, could be expected to identify his own infants and to allocate care only to those infants. Hence, recognition of infants is important. Marmoset males did not appear to discriminate between their own and other infants based on vocal cues, but these males did distinguish between their own and other infants using the more proximate cues of chemical signals.

Fourth, territory marking and defense would seem to be an important mechanism for maintaining a stable monogamous relationship and this has been shown in captive experiments introducing strangers to a pair. The failure to find clear evidence of territorial marking in the wild is somewhat puzzling. However, there are frequent encounters between groups (e.g. Lazaro-Perea, 2001) with sexual interactions between members of adjacent groups occurring often, especially by subordinate group members. Thus, territory defense may not be as critical in the wild. Instead the use of scent marks at tree gouging sites for exudate feeding may be a more important resource to defend than spatial boundaries.

Chemical signals play an important role in many aspects that are important for successful reproduction in cooperative breeding monkeys. Our research on chemical signals led us to the development of new methodologies such as functional brain imaging and sexual conditioning to scents that can be extended to other primate species. Our research has given us a better appreciation of the mechanisms involved in Callitrichid reproductive biology.