Article | Open | Published:

Costs of injury for scent signalling in a strepsirrhine primate

Scientific Reportsvolume 8, Article number: 9882 (2018) | Download Citation


Honesty is crucial in animal communication when signallers are conveying information about their condition. Condition dependence implies a cost to signal production; yet, evidence of such cost is scarce. We examined the effects of naturally occurring injury on the quality and salience of olfactory signals in ring-tailed lemurs (Lemur catta). Over a decade, we collected genital secretions from 23 (13 male, 10 female) adults across 34 unique injuries, owing primarily to intra-group fights. Using gas chromatography-mass spectrometry, we tested for differences in the chemical composition of secretions across pre-injury, injury and recovery, in animals that did and did not receive antibiotics. Lemur genital secretions were significantly dampened and altered during injury, with patterns of change varying by sex, season and antibiotics. Using behavioural bioassays (excluding odorants from antibiotic-treated animals), we showed that male ‘recipients’ discriminated injury status based on scent alone, directing more competitive counter marking towards odorants from injured vs. uninjured male ‘signallers.’ That injured animals could not maintain their normal signatures provides rare evidence of the energetic cost to signal production. That conspecifics detected olfactory-encoded ‘weakness’ suggests added behavioural costs: By influencing the likelihood of intra- or inter-sexual conflict, condition-dependent signals could have important implications for socio-reproductive behaviour.


For animal communication to effectively guide social behaviour, competition and mate choice, the signals used require a degree of honesty or ‘condition-dependence’1,2. Examples of condition dependency in signals and cues, including in vocalisations, visual ornaments, weaponry, behavioural displays and odours, are found throughout the animal kingdom1,3. Moreover, signal receivers use variations in signal composition and quality to assess both the stable and transient condition of conspecific signallers, altering their behavioural responses accordingly3,4,5. Because olfactory signals or cues are inextricably tied to an animal’s underlying physiology (perhaps even more so than are signals in other modalities), they are thought to provide a particularly reliable avenue for the advertisement and assessment of transient health, body condition or infection status6,7,8,9. Although evidence of condition dependence may imply a cost to scent production6,10, definitive evidence of such a cost (independent of pathogenic infection) is difficult to obtain without directly manipulating the physical condition of the signaller. Moreover, researchers rarely incorporate both chemical and behavioural methods in the same study. Using a strepsirrhine primate, the ring-tailed lemur (Lemur catta), we test for salient, condition-dependent variation in genital odorants associated with periods of wellness versus periods of naturally occurring injury. Specifically, we combined gas chromatography-mass spectrometry (GC-MS) with behavioural testing to examine (1) if lemur odorants vary chemically with injury and (2) if conspecifics are sensitive to such changes.

Indicator models of sexual selection predict that the expression of exaggerated signals should be condition-dependent, honestly conveying information about the signalling animal11,12,13. The expression of condition-dependent traits positively correlates with an individual’s acquired pool of resources13 and ability to withstand environmental challenges12, thus reflecting the degree to which an individual is impacted by poor nutrition, parasite load or physiological stress3,14. The maintenance of honesty in condition-dependent signals is a topic of current debate2,15,16 centring around two overarching, non-mutually exclusive principles: the ‘costly signaling hypothesis’ and the ‘index hypothesis’ (summarised by Weaver16).

Developed from Zahavi’s handicap principle17, the ‘costly signalling hypothesis’ posits that (1) signals are kept honest by production costs paid by the signaller and (2) low-condition individuals experience relatively greater costs than do individuals in good condition11,18. Recent critics2,15 argue that a handicap is not necessary to maintain signal honesty. Instead, signals may be kept honest, not by realised costs paid by honest signallers, but by the potential costs differentially paid by cheaters. For example, if the cost to signal production (however small) outweighs any benefit from investing in that cost, then ‘low-quality’ individuals will not likely cheat to produce a dishonest, ‘high-quality’ signal6,19. Researchers have variously expanded the definition of costly signalling to incorporate trade-offs associated with immunocompetence20, resource allocation13 and oxidative stress4.

In contrast to the handicap hypothesis, the ‘index hypothesis’2 does not require that honest signals be costly to produce. Instead, honesty is maintained by condition-dependent signals being mechanistically tied to a genetic or physiological pathway that is impossible to circumvent12,14,21. Regardless of the specific mechanism, evidence of condition-dependent, sexually selected traits derives overwhelmingly from studies of male visual5,16 and vocal22 signals. Nevertheless, the same principles may be applied to less easily measured traits, such as behavioural displays or olfactory signals23, including in females.

To communicate and coordinate sociality and reproduction, many vertebrates rely on complex chemical blends released from excretory products, saliva and scent glands7,24. Condition-dependent odorants and associated scent-marking behaviour are often sexually selected24, honestly conveying information on signaller traits, including sex, identity, age, reproductive state, dominance status and genetic quality7,24. Owing to their intimate ties to internal physiology, odorants are thought to be particularly sensitive to fluctuations in signaller health6,7. Thus, in addition to the life-history costs associated with signalling effort, experienced by any scent-marking species25, examples of condition-dependent odorant production and scent-marking activity have been found in lizards8,26, herpestids27 and laboratory rodents28,29,30. In the latter, researchers have shown that male mice inoculated with a pathogen or virus tend to invest less in scent-marking behaviour and produce depleted olfactory signals, and that female conspecifics prefer odorants from healthy males over those from infected males. Similar results are even obtained by challenging the immune system with non-replicating bacteria or lipopolysaccharides10,31,32. Such depleted investment in olfactory signals by immune-challenged animals implies an energetic cost to odorant production, creating a trade-off in investment between survival, reproduction and ornamentation10. To further our understanding of the condition dependence and potential costs of animal signals, it is crucial to consider other types of condition (beyond infection) that might affect energetic resources available for signal production, such as poor nutrition or injury.

Acute injury induces immediate, physiological responses from the mammalian immune and neuroendocrine systems33,34, both of which are known to affect the expression of sexually selected traits35. Nevertheless, experimental evidence of olfactory communication being influenced by injury is limited to the behavioural responses of bystanders to ‘alarm pheromones,’ which function as cues of predator-induced injury in conspecifics (e.g., flatworms36, mollusks37, crustaceans38, insects39 and fish40). Recently, Kimball and colleagues41 also showed that experimentally injured mice produce altered urinary cues that are salient to conspecifics; however, the authors were focused on developing an olfactory diagnostic tool for human brain injury and not on the implications of injury for rodent social communication. Despite the potentially serious physical and socio-reproductive consequences of injury (e.g., decreased competitive ability or social status, loss of access to potential mates), its effects on condition-dependent olfactory signals and on the behaviour of signal recipients remain to be described in natural systems.

In this study, we examined the effects of natural injuries on scent signatures and conspecific responses to such odorants in a socially complex mammal. Living in female-dominant, multi-male multi-female groups42, the ring-tailed lemur is an ideal model in which to examine the condition dependence of olfactory signals in both sexes. Ring-tailed lemurs arguably possess the most elaborate olfactory repertoire of any primate42,43,44: Both sexes possess scent glands (that are unique in the male), engage in conspicuous, multimodal scent-marking behaviour, and deposit chemically elaborate bouquets that contain information on the signaller’s sex, reproductive condition (e.g., breeding season, hormonal state), individual identity, neutral heterozygosity, diversity at the Major Histocompatibility Complex (MHC) and kinship43,45,46,47,48,49. Moreover, as a strepsirrhine with a functional vomeronasal organ50 and a derived increase in olfactory sensitivity51, ring-tailed lemurs of both sexes discriminate between different types of conspecific glandular secretions, modulating their behavioural responses seasonally and depending upon characteristics of the signaller45,48,49,52,53. Lastly, genital secretions in both sexes show salient, season-specific and potentially stress-induced decrements in chemical diversity (i.e., reflecting transient condition), that are most evident in individuals of low genetic quality (i.e., reflecting stable condition)46,47.

In addition, dominance and reproductive squabbles in lemurs are settled by the outcome of aggressive interactions54, such that intra- and inter-sexual aggression, resulting in naturally occuring injuries, can be relatively common. Along with increased scent-marking activity by both sexes, intra-male and intersexual aggression peak during the breeding season55,56,57,58,59; intra-female aggression also increases during the birthing and lactation seasons56,60,61,62,63. Physical aggression in ring-tailed lemurs is characterized by cuffs, lunges, chases and bites, occurring on the ground and in the trees42,56,64; subsequent injuries, whether from conspecifics or from falls, may be severe or even lethal65 (see Fig. 1b60). Injured animals, most often males, may pay both physical costs (e.g., physiological stress, injury) and social costs (e.g., loss of dominance status, reduced reproductive access, eviction from the group60,62,65,66,67,68). If olfactory signals are honest indicators of lemur condition that are costly to produce and maintain, then injured animals might be challenged to preserve normal scent signatures whilst their energetic resources are diverted towards recovery. Such costs in compromised animals could manifest as decreased chemical richness, decreased diversity and/or altered composition of odorants, relative to pre- or post-injury periods, and any of these chemical changes should be salient to conspecifics.

Figure 1
Figure 1

Representative gas chromatograms derived from the genital secretions of two male (af) and one female (gi) ring-tailed lemur (Lemur catta), obtained during pre-injury (left), injury (center) and recovery (right) phases, showing that scent signatures are significantly depressed and altered during injury. All of these injuries were of ‘moderate’ severity: Sample (b) was collected one day post-injury, following a fight with group members resulting in hair pulls, a 1.5-cm laceration under the left eye, and a deep laceration to the right hand that required sutures. Sample (e) was collected two days post-injury from an animal with a fractured 4th digit of the hind foot. Sample (h) was collected one day post-injury, following a fight with group members that resulted in puncture wounds to the right thigh and a 4.5-cm, shallow laceration to the right hand. Samples are scaled to the internal standard peak (hexachlorobenzene, rt 11.74 min; shown by arrows), except for (d) and (g), for which endogenous peaks were greater than that of the internal standard.


Injuries in relation to season, the animal’s sex, and wound severity

We noted over twice as many nonlethal injuries in ring-tailed lemurs during the breeding season (n = 23) than during the nonbreeding season (n = 11), as well as slightly more injuries in males (n = 20) than in females (n = 14), but these differences were not statistically significant (all chi-squared tests P > 0.10). Nevertheless, these patterns are consistent with the significant seasonal and sex differences in injury reported in a previous study of the same population, but reflecting an earlier 35-year span from 1971–200660. Injury severity, scored on a three-point scale (1 = ‘minor’, 2 = ‘moderate’; 3 = ‘severe’; see Materials and Methods), did not differ between seasons (mean severity scores, breeding season: 2.14; nonbreeding season: 2.07; Welch two sample t-test: t34 = 0.35, P = 0.73). In this female-dominant species, injuries sustained by females (mean severity score: 2.29) tended to be more severe than those sustained by males (mean score: 2.00; Welch two sample t-test: t46 = 1.46, P = 0.15). Both the time of year and the injured animal’s sex thus emerged as key variables affecting the likelihood of injury in lemurs.

Consequences of injury on the chemical complexity of lemur scent signals

The chemical complexity of genital gland secretions was significantly altered during periods when lemurs were injured (‘injury’), relative to periods either before (‘pre-injury’) or afterwards (‘recovery’), when they were uninjured (Fig. 1; Table 1). Whilst controlling for potential covariates, such as injury severity or veterinary-prescribed antibiotic treatment in either sex, or hormonal contraception in females (see Materials and Methods), the specific patterns of chemical change by injury status varied by sex and season.

Table 1 Summary of the relationships between the chemical complexity of lemur genital gland secretions, as measured by the richness and Shannon diversity indices, and various explanatory variables by sex.

Notably, injuries occurring in males during the breeding season were associated with a significant decline in chemical complexity, as measured both by richness (pairwise contrasts, pre-injury vs. injury: t36 = 4.70, P < 0.001; injury vs. recovery: t37 = 2.93, P = 0.035; Fig. 2a) and by Shannon diversity (pairwise contrasts, pre-injury vs. injury: t36 = 5.92, P < 0.001; injury vs. recovery: t37 = 3.23, P = 0.014; Fig. 2b). The injury-induced changes in chemical richness represented, on average, a 10.6% loss in the number of compounds present. There were no significant differences in chemical richness or diversity between the males’ two ‘uninjured’ phases (Richness, Shannon indices, pre-injury vs. recovery: all pairwise contrasts P > 0.60). During the nonbreeding season, however, males showed no such injury-associated decreases in chemical complexity (Richness, Shannon indices, all pairwise contrasts P > 0.90). Although not directly tested because of data skewness, similar seasonally dependent declines in Simpson diversity during injury were also apparent (Fig. 2c). Regardless of season, the chemical complexity of male scrotal signals did not change significantly according to injury severity or concurrent veterinary-prescribed antibiotic treatment (Table 1; see Materials and Methods).

Figure 2
Figure 2

Mean + standard error chemical complexity of genital secretions collected from male (ac) and female (df) ring-tailed lemurs during the breeding (black) and nonbreeding (white) seasons, across pre-injury, injury and recovery phases, showing significant effects of injury during the breeding season. Chemical complexity is measured using three diversity indices: Richness (a,d), Shannon (b,e) and Simpson (c,f). The numbers of samples are indicated in (a) and (d). Simpson data were too left-skewed to test robustly. Significant relationships are indicated as follows: *P < 0.05; ***P < 0.001.

The chemical complexity of female labial signals also tended to vary with injury status (i.e., pre-injury, injury, recovery), but not exclusively with injury (Fig. 2d-f; Table 1). Although on average, chemical richness in females did not differ according to injury status, unexpectedly, post-hoc tests showed no significant difference in richness between samples collected prior to or during injury (pairwise contrasts, pre-injury vs. injury: t25 = 0.95, P = 0.61), but richness then increased significantly during recovery (pairwise contrasts, injury vs. recovery: t26 = 2.58, P = 0.02; Fig. 2d). Shannon diversity was also significantly greater during recovery, relative to pre-injury (pairwise contrasts, pre-injury vs. recovery: t22 = 2.73, P = 0.017) and injury (pairwise contrasts, injury vs. recovery: t27 = 2.86, P = 0.012; Fig. 2e). There was also no significant change in Shannon diversity during the period of injury, relative to pre-injury (pairwise contrast, pre-injury vs. injury: t21 = 0.079, P = 0.99). We observed similar patterns for the Simpson index (Fig. 2f). Both richness and Shannon diversity varied with female hormonal contraception, but not with concurrent antibiotic treatment, season or injury severity (Table 1).

Consequences of injury on the chemical composition of lemur scent signals

The most common components of lemur genital secretions, as revealed by linear discriminate analyses (LDAs), varied according to whether the animals were uninjured, injured or injured and receiving concurrent antibiotic treatment. We retained principal components (PCs) with eigenvalues >1 separately for males and females, during the breeding and nonbreeding seasons (males, breeding season: n = 15 PCs, explaining 92.2% of variation across samples; males, nonbreeding season: n = 13 PCs, 97.6%; females, breeding season: n = 16 PCs, 95.8%; females, nonbreeding season: n = 10 PCs, 98.5%). In males, the LDAs for each subset of PCs correctly classified 73.8% and 90.0% of the samples collected during the breeding and nonbreeding seasons, respectively (breeding season: Wilks’ λ = 1.31, P = 0.19, Fig. 3a; nonbreeding season: Wilks’ λ = 0.94, P = 0.57; Fig. 3b). In females, the LDAs correctly classified 95.8% and 83.3% of samples collected during the breeding and nonbreeding seasons, respectively (breeding season: Wilks’ λ = 5.49, P < 0.001, Fig. 3c; nonbreeding season: Wilks’ λ not calculated due to small sample size, Fig. 3d).

Figure 3
Figure 3

Representation of the differences between the chemical composition of genital gland secretions in ring-tailed lemurs that were uninjured (open circles), injured (filled circles) and injured receiving concurrent antibiotic treatment (shaded triangles). Shown are separate linear discriminate analyses for males in the breeding (a) and nonbreeding (b) seasons, and for females in the breeding (c) and nonbreeding (d) seasons.

Regarding overall chemical composition, male genital secretions varied with injury status in a season-specific manner (PERMANOVA main test, season*injury Pseudo-F2,35 = 2.17, P = 0.006). During the breeding season, the composition of scrotal secretions differed during injury compared with pre-injury (pairwise contrast, t13 = 1.59, P = 0.017); we did not find such a difference during the nonbreeding season (pairwise contrast, t5 = 1.25, P = 0.20). Scrotal secretions did not differ compositionally between injury and recovery phases, in either season (pairwise contrasts, all Ps > 0.30). Overall chemical composition varied significantly depending on the individual animal (Pseudo-F18,35 = 2.14, P < 0.001), but not with antibiotic treatment (Pseudo-F1,35 = 0.78, P = 0.63) or injury severity (Pseudo-F2,35 = 0.53, P = 0.93). Random forests69,70, a type of classification tree analysis that assigns samples to categories (in this study, injury status) based on predictor variables (chemical compounds), did not reliably predict injury status in males.

In females, the overall composition of genital secretions tended to vary with injury status, although not significantly (PERMANOVA main test, injury status Pseudo-F2,17 = 1.43, P = 0.081), and this effect was not dependent upon season (PERMANOVA main test, injury status*season Pseudo-F2,17 = 1.05, P = 0.040). Overall composition varied depending upon the individual animal (Pseudo-F10,17 = 1.75, P < 0.001), but not with antibiotic treatment (Pseudo-F1,17 = 0.89, P = 0.54) nor injury severity (Pseudo-F2,17 = 0.43, P = 0.95). Lastly, random forests, based on overall labial secretion composition, correctly classified 71% of samples from females according to injury status. The three compounds contributing most to classification accuracy were fatty acid esters (rt 36.96 min, mol. wt. 508; rt 37.06 min, mol. wt. 452; rt 41.22 min, mol. wt. unknown): The first decreased during injury (mean relative abundance when uninjured: 0.24%; mean relative abundance when injured: 0.14%), the second was undetected in samples from injured animals and the third tended to increase during injury.

Behavioural evidence of injury detection

Male ring-tailed lemurs (hereafter, the signal ‘recipients’) varied their investigation of and response to conspecific scent depending upon the injury status and sex of the animal from which the odorants derived (hereafter, the signal ‘donors’). When presented with two secretions from the same male donor, collected whilst the donor was ‘injured’ versus ‘uninjured’, the male recipients directed increased sniffing (z = 3.46, P < 0.001), decreased licking (z = 0.354, P = 0.043) and increased wrist-marking (z = 2.26, P = 0.024) to the scent of injured donors (Fig. 4). Whereas sniffing and licking are investigatory (potentially in response to volatile and nonvolatile components, respectively), wrist marking is a competitive form of counter marking71. We did not detect significant variation in other behavioural responses, nor did we observe any statistically significant differences in the behaviour of male recipients responding to the odorants from injured or uninjured female donors (all Ps > 0.10). Neither the time the odorant had been in storage nor the number of trials in which the recipient had participated showed any relation to behavioural responses (all Ps > 0.10).

Figure 4
Figure 4

Mean + standard error behavioural responses showing discrimination by male ring-tailed lemurs between matched odorants from male conspecifics in either ‘uninjured’ (white) or ‘injured’ (black) condition. Shown are data from n = 15 bioassays. Significant relationships are indicated as follows: *P < 0.05; ***P < 0.001.


Following long-term study of an aggressively female-dominant, group-living species – the ring-tailed lemur – we used an integrated analytical approach, to provide the first direct evidence of socially relevant changes in olfactory signals consequent to naturally occurring injury. When injured, ring-tailed lemurs of both sexes produced genital scent-gland secretions that were less complex than normal and altered in their chemical composition. Consistent with the proposition of energetic trade-offs affecting mechanistic pathways associated with signal production, these chemical deficits imply a cost to odorant production. Moreover, male signal recipients, relying on these odorants alone, could discriminate conspecific injury status and modulated their competitive behaviour accordingly, showing that honest olfactory advertisement of condition can have consequences on social behaviour.

Independent of any seasonal variations in injury severity, injury-induced alteration of male olfactory secretions was strongly evident during the breeding season and appeared to be associated with delayed recovery of the full suite of a signaller’s odours. By contrast, injury had only a weak, non-significant effect on a subset of male odorant components during the nonbreeding season. This differential effect may owe to seasonal patterns in the experience of physiological stress: Relative to the nonbreeding season, males in reproductive condition have raised concentrations of testosterone56,57,59 and corticosterone59,72, and show both increased scent-marking activity55 and heightened aggression56,57,60,64. During this intensely competitive time, male ring-tailed lemurs may be energetically challenged46,56,72 and thus less able to sustain the production of complex olfactory ornaments. Previously, we had observed males of low neutral heterozygosity being unable to sustain their normal signals during the breeding season46. Here, because our males were of average heterozygosity (see Supplementary Material), we instead suggest that injured males in reproductive condition had impaired ability to mitigate the physical costs of injury without drastically depleting their olfactory signatures.

In behavioural bioassays, male ring-tailed lemurs modulated their responses to the scents of male conspecifics depending upon their injury status, implying a function for condition assessment in same-sex competition. Examples of olfactory-based male assessment of competitor condition and fighting capacity are also found in lizards73,74, hamsters75 and laboratory mice76. When coupled with mechanisms for individual recognition45,52, sensitivity to changes in competitor condition helps avoid risks and fitness costs of unnecessary fighting by allowing males to (1) reliably assess competitor fighting ability, (2) assess their likelihood of winning an aggressive encounter with a potential competitor, and (3) selectively engage in aggressive interactions with animals of compromised or poor competitive ability5,77,78. Wrist marking and tail anointing are multimodal (i.e., combined visual, olfactory and, sometimes, auditory) displays of dominance in male ring-tailed lemurs55,79,80, whose dominance hierarchies are fluid. Status maintenance may thus require continual scent marking and assessment of competitor marks46,80. The higher rates of wrist marking we observed directed at odorants from injured animals (relative to odorants obtained when the same animals were healthy) is consistent with recipient males using counter-marking strategies to gain social dominance over competitively weak conspecifics76,81. Similarly, dominant resident males are more likely to engage in conspicuous tail-anointing and ‘stink-fighting’ behaviour, which functions as a potentially costly ‘badge of status’ relevant to both male and female recipients79. Alternately, injured animals might modulate their overt displays of aggression, including scent-marking behaviour, to avoid being attacked by dominant or otherwise healthy individuals75.

Unlike the situation in males, the genital secretions of female ring-tailed lemurs tend to increase in complexity from the nonbreeding season to the breeding season43,47,82, potentially suggesting some immunity of female signals to breeding season stressors. Nevertheless, as in males, females when injured produced genital secretions that differed chemically, albeit weakly, from those produced when they were uninjured, particularly during the breeding season. Although labial secretions also tended to be less chemically rich (but similarly diverse) during injury, lack of statistical significance could be attributed to a combination of both (1) the smaller number of samples available for females than for males and (2) considerable inter-individual variation in chemical diversity among females prior to injury. Alternatively, female olfactory signals may be differentially affected by injury: Females could have shown minimal decrements in chemical richness and complexity, but nevertheless experienced significant changes in the specific ratios of different compounds, such as the proportions of fatty acid esters. The latter have been shown to predict genetic quality47 and, now, injury status. Intriguingly, the complexity of female scent signatures tended to be greater during recovery, regardless of season, suggesting that, following an injury, females may express more chemically complex odorants than normal, perhaps to signal their return to vitality and to re-establish their dominance status within the group.

For female-dominant species, such as ring-tailed lemurs, signals for health and vitality that are sensitive to variations in physical condition could be of critical importance for female reproductive fitness83,84, analogous to condition-dependent signalling in males83. There are several qualitative and mechanistic similarities between male and female intra-sexual competition, that are intensified by group-living85,86. Female ring-tailed lemurs use scent marking in resource defence87 and to assert dominance over same-sex competitors52,88, whilst also closely monitoring the odours of other females52. Given that, as in males, female lemurs produce recognizable scent signatures43, discrimination of injury status and competitive ability could minimise the number of potentially costly, aggressive interactions undertaken with vigorous, healthy individuals77,83. Such recognition might therefore entail fitness benefits, both for signal producers and recipients. The functions and fitness consequences of olfactory signals in female competition are poorly understood, and may be improved with empirical studies across a range of taxa85. Endler89 notes that we often lack critical information, both about the relative importance of signals to different recipients and about the type of information being advertised. We suggest that this gap in our understanding is particularly true with regard to female signals. We would predict that female lemurs might be especially attentive to changes in the condition of other females during late gestation and lactation, when female competition and the fitness costs of losing aggressive encounters intensifies49,56,60,61,62,63. Evidence of female assessment of injury status in same-sex competitors might be found in other female-dominant species, including meerkats (Suricata suricatta) and spotted hyaenas (Crocuta crocuta), or in species that compete aggressively for resources and mates83,84.

A major challenge for the study of condition-dependent signals is to ensure that the type and degree of experimental stressor is biologically or environmentally relevant to the system in question11,35. In our study, the signaller’s condition was naturally altered during periods of injury, and the physiological and energetic trade-offs associated with injury were amplified during the breeding season. The production of condition-dependent chemical signals may share metabolic pathways with critical cellular processes12, such as mitochondrial respiration14,90. Signal production pathways might also become limited by perturbations in insulin-like growth factors23,91, or by oxidative stress4,92, immune activation10,31,32 or inflammation93. For example, injury induces the production of protein complexes, termed ‘inflammasomes’, that are involved in the inflammatory response and tissue repair, and trigger pyroptosis, a form of cellular death94. Such processes could contribute to an ‘injury-specific’ odour, akin to disease-specific changes in body odour described in human patients95. Data on the concurrent physiological state of signallers are needed to describe the specific mechanisms underlying honest, semiochemical production.

Although researchers have previously argued that cost is not necessary to maintain honesty7,12,15, we suggest that our study provides strong evidence of trade-offs in resource allocation during injury. Moreover, a cost to lemur odorant production could function as a ‘revealing handicap.’ Changes in the production of potentially costly compounds, such as lipids and fatty acid esters, may be mediated by energetic trade-offs between immunological and physiological regulation of somatic repair following injury, and allocation of essential nutrients to chemical signal production46,47. Additionally, the genetic quality of individuals will likely affect both their condition and ability to buffer the effects of environmental stressors11,13. We echo previous calls for more empirical research in a variety of taxa and signalling systems23, with such research incorporating both observational and manipulative approaches18 to better tease apart the non-mutually exclusive influences of genotype, costs and condition-dependent signals.

Lastly, along with becoming altered during injury, lemur odorants were further perturbed by concurrent veterinary-prescribed antibiotic treatment, consistent with the putative removal of fermentative bacteria crucial for odour production96. In both males and females, antibiotic treatment was associated with concurrent changes in the composition of the most commonly occurring volatile chemical components, but not with changes in overall chemical complexity or composition. These results suggest that the most widespread compounds in lemur genital secretions may be modified and/or produced, at least in part, by bacteria and, consequently, are sensitive to broad-spectrum antimicrobials. Because the composition of commensal microbes shaping an individual’s scent signature might be altered by infection or host health6, elucidating the specific contribution of bacteria to lemur social odours requires an experimental approach in healthy animals.

By using a unique system involving natural alteration of the physical condition of the signalling animal, our study provides, to our knowledge, the first supporting evidence for a socially relevant olfactory indicator of naturally occurring injury. Given the social and physical costs of injury, particularly in an aggressively female-dominant species, lemurs of both sexes could benefit from being attentive to the health status of conspecifics and being selective about engaging in aggressive behaviour with specific individuals. Further research will elucidate the specific mechanisms by which physical injury can alter sexual signals, including potential and realised costs associated with the production of condition-dependent signals.

Materials and Methods

Subjects and housing

Our subjects were 27 adult ring-tailed lemurs (17 males, 10 females; mean ± standard error or s.e. age at the time of study: 7.61 ± 0.53 yrs, range: 1.8–25.5 yrs). Of these, 23 (13 males and 10 females) provided genital secretion samples, collected between 2007–2016 (Table 2), and nine males served as focal subjects in behavioural bioassays, conducted in 2016 (see sections on sample collection and behavioural bioassays, below). All of the subjects were captive-born and housed socially at the Duke Lemur Center (DLC; Durham, NC, USA)43,56. The animals’ social housing conditions allow for exposure to conspecific visual, auditory and olfactory cues, as well as for naturally occurring interactions, including those of aggression and their subsequent injuries (see injury section, below)48,56,60. Most subjects are semi-free ranging, with access both to forested outdoor enclosures (3–7 ha) and to temperature-controlled, indoor areas. A minority of subjects are housed indoors year-round. All of the animals are provided with a mixed diet of commercial primate chow, fruit, vegetables, fresh browse, and water56,60. Our research protocols (Protocol Registry Numbers A232–06–07, A171-09-06, A143-12-05 and A111-16-05) abided by the regulations of the United States Department of Agriculture and were approved by the Institutional Animal Care and Use Committee of Duke University. The DLC is fully accredited by the American Association for the Accreditation of Laboratory Animal Care; information on the DLC’s conservation, education and research mission is available at

Table 2 Number of unique injury events, involving three phases, that are represented by odorant samples collected from adult ring-tailed lemurs at the Duke Lemur Center, from 2007–2016.

Injury identification, occurrences, and classification

All DLC animals are monitored closely each day: If an aggressive interaction is directly observed or suspected, the individuals most likely to be involved are captured for closer examination and, if necessary, veterinary treatment. In the case of severe wounding or continued targeted aggression, an animal may be temporarily or permanently removed from its group. Veterinary records are added to a medical records database (Species360, Bloomington MN), detailing the nature and severity of any injuries, their cause (if known), any prescribed medications, follow-up care and treatment outcomes. We communicated with the veterinary staff about all injuries, as they presented, but retrospectively used the Species360 database to verify the condition for all of our subjects.

We report on 34 unique injury events (affecting 13 male individuals and 10 females, with some animals being injured multiple times during the study; Table 2). Most (28/34 or 82%) resulted from fights or probable fights with members of the animal’s own group, but altercations also occurred between members of neighbouring groups. Of the injuries resulting from fights, those sustained by males were most often inflicted by other males (n = 10 or 62.5%), but also by females (n = 2) or an unknown assailant (n = 4). The remaining injuries in males (n = 4) owed to unknown causes. Injuries in females were inflicted or likely inflicted by other females (n = 7) or unknown assailants (n = 5), or resulted from accidents (n = 2).

Some injuries (n = 5) changed in severity over the course of treatment (i.e., they improved as infections cleared or worsened with subsequent amputations or infections). We thus differentiated 39 injuries based on a three-point scale of severity: ‘Minor’ injuries (n = 8) included superficial scrapes, hair pulls, punctures or small (< 2 cm) lacerations requiring minimal veterinary intervention; ‘moderate’ injuries (n = 21) included lacerations (2–10 cm), digit fractures or dislocations; ‘severe’ injuries (n = 10) included fractures, amputations, and lacerations (> 2 cm) that damaged tendons or other deep tissues. Some injuries (minor: n = 2; moderate: n = 7; severe: n = 8) subsequently became infected and required antimicrobial treatment (most commonly involving amoxicillin or enrofloxacin, both of which are broad-spectrum antibiotics).

Individuals were uninjured during the periods before and after each injury (‘pre-injury’ and ‘recovery’ phases, respectively). The latter occurred minimally two weeks after the initial insult, as determined from veterinary records indicating that the subject was in good health (i.e., the wound had healed, there was no new evidence of trauma, no pathological signs of disease or illness, and normal behaviour had resumed).

Odorant sample collection

All odorant sampling of lemur genital (i.e., male scrotal and female labial) secretions occurred in triplicate (i.e., three swabs were taken), following previously published procedures43 (see Supplementary Material). We later divided the replicate samples for use in GC-MS analyses and in behavioural bioassays (Table 2; for details on the analytical and bioassay procedures, see below). In early years of the study, we routinely (i.e., monthly) collected odorants from all healthy ring-tailed lemurs at the DLC throughout the breeding (November–February in the northern hemisphere56) or nonbreeding seasons (March–October). These samples provided the pool from which we selected all of the ‘pre-injury’ and many of the ‘recovery’ samples. Additionally, we opportunistically collected ‘injury’ odorants when an animal was brought in for veterinary care, typically on the day the insult occurred or shortly afterwards, whilst the animal was still showing active signs of injury (mean ± s.e.: 6.25 ± 0.90 days post-insult, range: 0–26 days). In later years of the study, sample collection was more sporadic, specifically targeting the collection of injury or recovery samples. The mean (± s.e.) time span between collection of pre-injury vs. injury samples was 1.36 ± 0.18 years (range: 10 days – 6 years) and between collection of injury vs. recovery samples was 1.42 ± 0.27 years (range: 14 days – 6 years). The maximum time span across all sampling relating to a single injury was eight years.

Obtaining complete sequences of samples for a given injury (representing pre-injury, injury, and recovery) was not always feasible owing to logistical challenges, including matching the samples by season or female reproductive state (see section on statistical analyses, below). We obtained 21 complete sequences; nevertheless, for each injury, we minimally obtained one set of samples during an uninjured phase, either before or after the injury. In 9 of the 34 unique injuries described, we also collected additional (n = 1–3) triplicate sets of odorant samples when the veterinary staff reassessed injury severity during follow-up examinations. Such resampling allowed us to track changes in chemical composition, either as the severity of the injury changed or as the animal began receiving antibiotics.

Gas chromatography-mass spectrometry

We used our previously published GC-MS procedures43,46 (see Supplementary Material) to describe the chemical composition of lemur odorants collected during pre-injury, injury, and recovery phases (Table 2). Because it is not possible to control for the absolute amount of secretion collected or analysed, we present data on relative abundances. We have previously shown that individual-specific lemur scent signatures are stable across years and with storage time48,52, which we also verified here (see Supplementary Fig. S1).

Behavioural bioassays

We performed 30 behavioural bioassays52,71 within a five-day period in late October 2016, at the onset of the breeding season for ring-tailed lemurs in the northern hemisphere56. We used male recipients because they respond reliably to scents from both males and females52 and, during the breeding season, respond most strongly52 and scent mark most intensively55. We presented nine healthy male ‘recipients’ (each receiving 1–5 bioassays) with a choice between two odorants collected from a given, conspecific ‘donor,’ one collected whilst the donor was uninjured and one whilst injured (Table 2; see Supplementary Material). Nine males and six females served as odorant donors (in 1–4 bioassays each). Five of our recipient males also acted as odorant donors. We presented recipients only with the odorants from unfamiliar donors (defined as never having belonged to the recipient’s group and whose odorants had not been encountered by the recipient in the previous two years). We controlled for seasonal variation in odorant composition by matching the paired odorants according to the timing of sample collection (i.e., where possible, we used samples collected within the same calendar months, even if collected across years). We also did not use odorants from individuals concurrently treated with antibiotics or from females that were either pregnant, lactating or hormonally contracepted.

Statistical analyses

We first investigated the chemical complexity of odorants deriving from uninjured versus injured lemurs by calculating the following three indices for each odorant type: Richness, Shannon and Simpson46,97. Richness is simply the number of peaks retained for statistical analyses, whereas the Shannon and Simpson indices apply weight to peaks based on their relative abundance: The Shannon index is most sensitive to those peaks of intermediate abundance, whereas the Simpson index is most sensitive to the most abundant peaks97. Shannon and Simpson indices were calculated using the vegan package (version 2.4–498) in the program R99 and R Studio (version 1.0.136100). For these analyses, we controlled for seasonal variation and individual-specificity in chemical profiles by matching ‘pre-injury,’ ‘injury,’ and ‘recovery’ odorants collected sequentially from the same animal within either the breeding or nonbreeding seasons. We further controlled for female reproductive status by (1) matching odorants according to female contraceptive treatment (contracepted, not contracepted) and (2) not using odorants from pregnant or lactating individuals. When possible, we matched odorants collected within the same season (n = 19 injuries); other injuries were seasonally matched with odorants collected across years (n = 15).

To test for variations in chemical complexity between injured and uninjured animals, we analysed each chemical diversity measure separately in a series of general linear mixed-effects models (GLMMs) in the lme4 package in R (version 1.1–14101). Because males and females have different genital secretions (i.e., scrotal vs. labial43) and show different seasonal patterns in chemical complexity46,82, we analysed the data for each sex separately. Although group composition and size varied over the course of the study, this variance was not likely to impact our results, as we have previously found no significant effects of housing condition on lemur chemical profiles48,82. We therefore excluded aspects of group composition from the analyses. Similarly, we also excluded animal age and dominance rank, as previous studies in adult ring-tailed lemurs have not shown these variables to significantly affect odorant composition43 (see Supplementary Fig. S1), or their salience to conspecifics52. In each GLMM, we included as fixed effects the animal’s injury status (pre-injury, injury, recovery), season (breeding, nonbreeding), and their interaction, along with injury severity (minor, moderate, severe), and female contraceptive treatment (contracepted, not contracepted). Some animals were prescribed antibiotics as part of their veterinary treatment: Because the ‘fermentation hypothesis’ posits that beneficial microbes inhabiting animal scent glands contribute to host social odours96,102, we additionally controlled for concurrent antibiotic treatment (antibiotics, no antibiotics). A random effect, ‘identity’, was also included, which incorporated both the animal’s identity and its specific injury, thus controlling for cases of multiple injuries for some individuals (e.g., 19 injuries derive from eight individuals). The significance of all fixed effects was assessed using t-statistics and degrees of freedom (Satterthwaite approximation) estimated in the lmerTest package (version 2.0–33103) in R. To confirm the robustness of our models, we verified the normality of residuals using Q-Q plots and Shapiro-Wilk tests. We tested the significance of post hoc contrasts using Tukey-adjusted P-values (multcomp package version 1.4–8104). We used Gaussian distributions and applied transformations to Shannon index data for males and females (^2 and ^5 respectively, to improve skewness). Simpson data were too skewed to test robustly, but for comparative purposes we present the mean ± s.e. for all three indices.

To address the possibility that two scent samples could be equally complex, but different in composition, we next examined changes in lemur odorant composition associated with being uninjured, injured, and injured whilst receiving concurrent antibiotic treatment. As in our analyses of chemical diversity, we divided the chemical data on relative abundances into four subsets (one for each sex in each season) and conducted separate multivariate statistical analyses for each subset. First, using linear discriminate analysis (LDA), we tested for changes in composition based on the most common chemical components. To reduce the dimensionality of the data, we also calculated principal components (PCs), using the relative abundance of peaks occurring in minimally 80% of each data subset. We combined samples collected ‘pre-injury’ and during ‘recovery’ into a single ‘uninjured’ category, thereby reducing the number of classification categories. We used PCs with eigenvalues >1 as variables in LDAs, classifying samples according to the injury status of the donor animal (i.e., ‘uninjured,’ ‘injured’, ‘injured + ABX’). We calculated PCs and LDAs using JMP (version 13).

We next tested for differences in the overall composition of chromatograms (retaining peaks that comprised >0.05% of the total chromatogram area and occurring in >5 samples), using a permutational MANOVA (‘PERMANOVA’) in Primer-E (version 7.0.13105) with the PERMANOVA+ add-on106. We first applied a square root transformation to reduce the influence of the most abundant peaks, then calculated a Bray-Curtis dissimilarity matrix. PERMANOVAs were performed using type III sums of squares, a reduced fit model procedure and 9999 permutations. As in the analyses of odorant diversity, we included as fixed factors the interaction between animal injury status and season, plus antibiotic treatment, contraceptive treatment (females only) and injury severity, along with identity as a random factor.

Lastly, we used random forests69, each comprising 1000 classification trees, to identify the compounds most affected by being injured (versus uninjured) and, thus, potentially costly to produce. For these analyses, we again pooled the ‘pre-injury’ and ‘recovery’ phases to minimize the number of categories, and used the randomForest package107 in R. We generated random forests separately for males and females, and excluded from our analyses the samples obtained from animals treated with antibiotics. We report on compounds that contributed most to model classification accuracy (i.e., we minimized ‘out-of-bag’ error rate70). We compared model accuracy with that of a ‘default’ model, which simply assigns all samples to the most numerous class.

To test for differences in male behavioural responses to odorants derived from injured versus uninjured donors, we used GLMMs and the glmmADMB package (version in R. Due to the limited number of bioassays we could perform, we could not robustly test for differences in male responses depending upon both the sex and reproductive condition of the donor animal. Therefore, we restricted our analyses to bioassays involving odorants collected from females during the nonbreeding season (n = 11 bioassays) and from males during the breeding season (n = 15). We ran separate GLMMs for each of these two donor categories, using each behavioural response as the dependent variable. We scored behaviour (sniff, lick, wrist mark) as counts and/or as bouts of varying duration (seconds per response). Tail marking, biting and ‘threat yawns’ were not observed sufficiently often to analyse robustly. We used Poisson, negative binomial or gaussian distributions as appropriate. In each GLMM, we included the sample’s relative storage time (i.e., whether it was the ‘older’ or ‘newer’ of the pair) and trial number (i.e., the total cumulative number of bioassays in which the recipient had participated) as fixed factors, and donor identity nested within recipient as a random factor. We used a stepwise GLMM selection procedure, sequentially dropping variables with the greatest P-value from the GLMM, until only significant factors remained. We then added each excluded factor back into the final model to confirm statistical non-significance109. In all of these statistical analyses, we set significance at α < 0.05.

Data accessibility

GC-MS odorant sample data and bioassay. ‘CSV’ files are available in the Dryad repository, at

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Bradbury, J. W. & Vehrencamp, S. L. Principles of animal communication. Second edn (Sinauer Associates, Inc., 2011).

  2. 2.

    Biernaskie, J. M., Grafen, A. & Perry, J. C. The evolution of index signals to avoid the cost of dishonesty. Proc R Soc B Biol Sci 281, 20140876 (2014).

  3. 3.

    Andersson, M. Sexual selection. (Princeton University Press, 1994).

  4. 4.

    Garratt, M. & Brooks, R. C. Oxidative stress and condition-dependent sexual signals: more than just seeing red. Proc R Soc B Biol Sci 279, 20120568 (2012).

  5. 5.

    Emlen, D. J. The evolution of animal weapons. Annu Rev Ecol Evol Syst 39, 387–413 (2008).

  6. 6.

    Penn, D. & Potts, W. Chemical signals and parasite-mediated sexual selection. Trends Ecol Evol 13, 391–396 (1998).

  7. 7.

    Wyatt, T. D. Pheromones and animal behaviour: chemical signals and signatures. (Cambridge University Press, 2014).

  8. 8.

    Martín, J. & López, P. Condition-dependent chemosignals in reproductive behavior of lizards. Horm Behav 68, 14–24 (2015).

  9. 9.

    Poirotte, C. et al. Mandrills use olfaction to socially avoid parasitized conspecifics. Sci Adv 3, e1601721 (2017).

  10. 10.

    Zala, S. M., Potts, W. K. & Penn, D. J. Scent-marking displays provide honest signals of health and infection. Behav Ecol 15, 338–344 (2004).

  11. 11.

    Cotton, S., Fowler, K. & Pomiankowski, A. Do sexual ornaments demonstrate heightened condition-dependent expression as predicted by the handicap hypothesis? Proc R Soc B Biol Sci 271, 771–783 (2004).

  12. 12.

    Hill, G. E. Condition‐dependent traits as signals of the functionality of vital cellular processes. Ecol Lett 14, 625–634 (2011).

  13. 13.

    Morehouse, N. I. Condition-dependent ornaments, life histories, and the evolving architecture of resource-use. Integr Comp Biol 54, 591–600 (2014).

  14. 14.

    Hill, G. E. Cellular respiration: the nexus of stress, condition, and ornamentation. Integr Comp Biol 54, 645–657 (2014).

  15. 15.

    Számadó, S. The cost of honesty and the fallacy of the handicap principle. Anim Behav 81, 3–10 (2011).

  16. 16.

    Weaver, R. J., Koch, R. E. & Hill, G. E. What maintains signal honesty in animal colour displays used in mate choice? Philos Trans R Soc Lond B Biol Sci 372, 20160343 (2017).

  17. 17.

    Zahavi, A. Mate selection: a selection for a handicap. J Theor Biol 53, 205–214 (1975).

  18. 18.

    Kotiaho, J. S. Costs of sexual traits: a mismatch between theoretical considerations and empirical evidence. Biol Rev 76, 365–376 (2001).

  19. 19.

    Higham, J. P. How does honest costly signaling work? Behav Ecol 25, 8–11 (2013).

  20. 20.

    Folstad, I. & Karter, A. J. Parasites, bright males, and the immunocompetence handicap. Am Nat 139, 603–622 (1992).

  21. 21.

    Hamilton, W. D. & Zuk, M. Heritable true fitness and bright birds: a role for parasites? Science 218, 384–387 (1982).

  22. 22.

    Casagrande, S., Pinxten, R. & Eens, M. Honest signaling and oxidative stress: the special case of avian acoustic communication. Front Ecol Evol 4, 52 (2016).

  23. 23.

    Warren, I. A., Gotoh, H., Dworkin, I. M., Emlen, D. J. & Lavine, L. C. A general mechanism for conditional expression of exaggerated sexually‐selected traits. Bioessays 35, 889–899 (2013).

  24. 24.

    Johansson, B. G. & Jones, T. M. The role of chemical communication in mate choice. Biol Rev 82, 265–289 (2007).

  25. 25.

    Gosling, L. M., Roberts, S. C., Thornton, E. A. & Andrew, M. J. Life history costs of olfactory status signalling in mice. Behav Ecol Sociobiol 48, 328–332 (2000).

  26. 26.

    García-Roa, R., Sáiz, J., Gómara, B., López, P. & Martín, J. Dietary constraints can preclude the expression of an honest chemical sexual signal. Sci Rep 7, 6073 (2017).

  27. 27.

    Mitchell, J., Cant, M. A., Vitikainen, E. I. & Nichols, H. J. Smelling fit: scent marking exposes parasitic infection status in the banded mongoose. Curr Zool 63, 237–247 (2017).

  28. 28.

    Penn, D., Schneider, G., White, K., Slev, P. & Potts, W. Influenza infection neutralizes the attractiveness of male odour to female mice (Mus musculus). Ethology 104, 685–694 (1998).

  29. 29.

    Yamazaki, K. et al. Presence of mouse mammary tumor virus specifically alters the body odor of mice. Proc Natl Acad Sci USA 99, 5612–5615 (2002).

  30. 30.

    Kavaliers, M., Choleris, E. & Pfaff, D. W. Genes, odours and the recognition of parasitized individuals by rodents. Trends Parasitol 21, 423–429 (2005).

  31. 31.

    Lopes, P. C. & König, B. Choosing a healthy mate: sexually attractive traits as reliable indicators of current disease status in house mice. Anim Behav 111, 119–126 (2016).

  32. 32.

    López, P., Gabirot, M. & Martín, J. Immune activation affects chemical sexual ornaments of male Iberian wall lizards. Naturwissenschaften 96, 65–69 (2009).

  33. 33.

    Desborough, J. The stress response to trauma and surgery. Br J Anaesth 85, 109–117 (2000).

  34. 34.

    Molina, P. E. Neurobiology of the stress response: Contribution of the sympathetic nervous system to the neuroimmune axis in traumatic injury. Shock 24, 3–10 (2005).

  35. 35.

    Buchanan, K. L. Stress and the evolution of condition-dependent signals. Trends Ecol Evol 15, 156–160 (2000).

  36. 36.

    Wisenden, B. D. & Millard, M. C. Aquatic flatworms use chemical cues from injured conspecifics to assess predation risk and to associate risk with novel cues. Anim Behav 62, 761–766 (2001).

  37. 37.

    Smee, D. L. & Weissburg, M. J. Hard clams (Mercenaria mercenaria) evaluate predation risk using chemical signals from predators and injured conspecifics. J Chem Ecol 32, 605–619 (2006).

  38. 38.

    Ferner, M. C., Smee, D. L. & Chang, Y. P. Cannibalistic crabs respond to the scent of injured conspecifics: danger or dinner? Mar Ecol Prog Ser 300, 193–200 (2005).

  39. 39.

    Wisenden, B. D., Chivers, D. P. & Smith, R. J. F. Learned recognition of predation risk by Enallagma damselfly larvae (Odonata, Zygoptera) on the basis of chemical cues. J Chem Ecol 23, 137–151 (1997).

  40. 40.

    Chivers, D. P. & Smith, R. J. F. Chemical alarm signalling in aquatic predator-prey systems: a review and prospectus. Ecoscience 5, 338–352 (1998).

  41. 41.

    Kimball, B. A. et al. Brain injury alters volatile metabolome. Chem Senses 41, 407–414 (2016).

  42. 42.

    Jolly, A. Lemur behavior. (University of Chicago Press, 1966).

  43. 43.

    Scordato, E. S., Dubay, G. & Drea, C. M. Chemical composition of scent marks in the ringtailed lemur (Lemur catta): glandular differences, seasonal variation, and individual signatures. Chem Senses 32, 493–504 (2007).

  44. 44.

    Schilling, A. A study of marking behavior in Lemur catta in Prosimian Biology (eds Martin, R. D., Doyle, G. A. & Walker, A. C.) 347–362 (University of Pittsburgh Press, 1974).

  45. 45.

    Palagi, E. & Dapporto, L. Beyond odor discrimination: demonstrating individual recognition by scent in Lemur catta. Chem Senses 31, 437–443 (2006).

  46. 46.

    Charpentier, M. J. E., Boulet, M. & Drea, C. M. Smelling right: the scent of male lemurs advertises genetic quality and relatedness. Mol Ecol 17, 3225–3233 (2008).

  47. 47.

    Boulet, M., Crawford, J. C., Charpentier, M. J. E. & Drea, C. M. Honest olfactory ornamentation in a female-dominant primate. J Evolution Biol 23, 1558–1563 (2010).

  48. 48.

    Crawford, J. C., Boulet, M. & Drea, C. M. Smelling wrong: hormonal contraception in lemurs alters critical female odour cues. Proc R Soc B Biol Sci 278, 122–130 (2011).

  49. 49.

    Grogan, K. E., Harris, R. L., Boulet, M. & Drea, C. M. Genetic variation at the MHC influences both olfactory signals and scent discrimination in ring-tailed lemurs. bioRxiv, (2018).

  50. 50.

    Smith, T. D. et al. The vomeronasal organ of Lemur catta. Am J Primatol 77, 229–238 (2015).

  51. 51.

    Heritage, S. Modeling olfactory bulb evolution through primate phylogeny. PLoS ONE 9, e113904 (2014).

  52. 52.

    Scordato, E. S. & Drea, C. M. Scents and sensibility: information content of olfactory signals in the ringtailed lemur. Lemur catta. Anim Behav 73, 301–314 (2007).

  53. 53.

    Charpentier, M. J. E., Crawford, J. C., Boulet, M. & Drea, C. M. Message ‘scent’: lemurs detect the genetic relatedness and quality of conspecifics via olfactory cues. Anim Behav 80, 101–108 (2010).

  54. 54.

    Kappeler, P. M. Female dominance in Lemur catta: more than just female feeding priority? Folia Primatol 55, 92–95 (1990).

  55. 55.

    Drea, C. M. & Scordato, E. S. Olfactory communication in the ringtailed lemur (Lemur catta): form and function of multimodal signals in Chemical Signals in Vertebrates 11 (eds Hurst, J. L., Beynon, R. J., Roberts, S C, & Wyatt, T. D.) 91–102 (Springer, 2008).

  56. 56.

    Drea, C. M. Sex and seasonal differences in aggression and steroid secretion in Lemur catta: are socially dominant females hormonally ‘masculinized’? Horm Behav 51, 555–567 (2007).

  57. 57.

    Gould, L. & Ziegler, T. E. Variation in fecal testosterone levels, inter-male aggression, dominance rank and age during mating and post-mating periods in wild adult male ring-tailed lemurs (Lemur catta). Am J Primatol 69, 1325–1339 (2007).

  58. 58.

    Sauther, M. Reproductive behavior of free-ranging Lemur catta at Beza Mahafaly Special Reserve, Madagascar. Am J Phys Anthropol 84, 463–477 (1991).

  59. 59.

    Cavigelli, S. A. & Pereira, M. E. Mating season aggression and fecal testosterone levels in male ring-tailed lemurs (Lemur catta). Horm Behav 37, 246–255 (2000).

  60. 60.

    Charpentier, M. J. & Drea, C. M. Victims of infanticide and conspecific bite wounding in a female-dominant primate: a long-term study. PLoS ONE 8, e82830 (2013).

  61. 61.

    Sauther, M. The dynamics of feeding competition in wild populations of ringtailed lemurs (Lemur catta). In Lemur social systems and their ecological basis (eds Kappeler, P. M. & Ganzhorn, J. U.) 135–152 (Plenum, 1993).

  62. 62.

    Cavigelli, S. A., Dubovick, T., Levash, W., Jolly, A. & Pitts, A. Female dominance status and fecal corticoids in a cooperative breeder with low reproductive skew: ring-tailed lemurs (Lemur catta). Horm Behav 43, 166–179 (2003).

  63. 63.

    Pride, R. E. Foraging success, agonism, and predator alarms: behavioral predictors of cortisol in Lemur catta. Int J Primatol 26, 295–319 (2005).

  64. 64.

    Parga, J. A. Male mate choice in Lemur catta. Int J Primatol 27, 107–131 (2006).

  65. 65.

    Kittler, K. & Dietzel, S. Female infanticide and female-directed lethal targeted aggression in a group of ring-tailed lemurs (Lemur catta). Primate. Biology 3, 41–46 (2016).

  66. 66.

    Gould, L., Sussman, R. & Sauther, M. L. Demographic and life‐history patterns in a population of ring‐tailed lemurs (Lemur catta) at Beza Mahafaly Reserve, Madagascar: a 15‐year perspective. Am J Phys Anthropol 120, 182–194 (2003).

  67. 67.

    Vick, L. G. & Pereira, M. E. Episodic targeting aggression and the histories of Lemur social groups. Behav Ecol Sociobiol 25, 3–12 (1989).

  68. 68.

    Sauther, M. L., Sussman, R. & Cuozzo, F. Dental and general health in a population of wild ring‐tailed lemurs: a life history approach. Am J Phys Anthropol 117, 122–132 (2002).

  69. 69.

    Breiman, L. Random forests. Mach Learn 45, 5–32 (2001).

  70. 70.

    Cutler, D. R., Edwards, T. C., Beard, K. H., Cutler, A. & Hess, K. T. Random forests for classification in ecology. Ecology 88, 2783–2792 (2007).

  71. 71.

    Greene, L. K. et al. Mix it and fix it: functions of composite olfactory signals in ring-tailed lemurs. Royal Soc Open Sci 3, 160076 (2016).

  72. 72.

    Starling, A. P., Charpentier, M. J., Fitzpatrick, C., Scordato, E. S. & Drea, C. M. Seasonality, sociality, and reproduction: long-term stressors of ring-tailed lemurs (Lemur catta). Horm Behav 57, 76–85 (2010).

  73. 73.

    López, P. & Martín, J. Chemical rival recognition decreases aggression levels in male Iberian wall lizards. Podarcis hispanica. Behav Ecol Sociobiol 51, 461–465 (2002).

  74. 74.

    Martín, J. & López, P. Scent may signal fighting ability in male Iberian rock lizards. Biol Lett 3, 125–127 (2007).

  75. 75.

    Petrulis, A., Weidner, M. & Johnston, R. E. Recognition of competitors by male golden hamsters. Physiol Behav 81, 629–638 (2004).

  76. 76.

    Hurst, J. L. & Beynon, R. J. Scent wars: the chemobiology of competitive signalling in mice. Bioessays 26, 1288–1298 (2004).

  77. 77.

    Alberts, A. C. Constraints on the design of chemical communication systems in terrestrial vertebrates. Am Nat 139, S62–S89 (1992).

  78. 78.

    Searcy, W. A. & Nowicki, S. The evolution of animal communication: reliability and deception in signalling systems. (Princeton University Press, 2005).

  79. 79.

    Walker-Bolton, A. D. & Parga, J. A. “Stink flirting” in ring-tailed lemurs (Lemur catta): male olfactory displays to females as honest, costly signals. Am J Primatol 79, e22724 (2017).

  80. 80.

    Kappeler, P. M. Social status and scent-marking behaviour in Lemur catta. Anim Behav 40, 774–776 (1990).

  81. 81.

    Rich, T. J. & Hurst, J. L. The competing countermarks hypothesis: reliable assessment of competitive ability by potential mates. Anim Behav 58, 1027–1037 (1999).

  82. 82.

    Boulet, M., Charpentier, M. J. & Drea, C. M. Decoding an olfactory mechanism of kin recognition and inbreeding avoidance in a primate. BMC Evol Biol 9, 281 (2009).

  83. 83.

    Stockley, P., Bottell, L. & Hurst, J. L. Wake up and smell the conflict: odour signals in female competition. Philos Trans R Soc Lond B Biol Sci 368, 20130082 (2013).

  84. 84.

    Stockley, P. & Bro-Jørgensen, J. Female competition and its evolutionary consequences in mammals. Biol Rev 86, 341–366 (2011).

  85. 85.

    Stockley, P. & Campbell, A. Female competition and aggression: interdisciplinary perspectives. Philos Trans R Soc Lond B Biol Sci 368, 20130073 (2013).

  86. 86.

    Rosvall, K. A. Intrasexual competition in females: evidence for sexual selection? Behav Ecol 22, 1131–1140 (2011).

  87. 87.

    Mertl-Millhollen, A. S. Scent marking as resource defense by female Lemur catta. Am J Primatol 68, 605–621 (2006).

  88. 88.

    Palagi, E., Telara, S. & Tarli, S. M. B. Reproductive strategies in Lemur catta: balance among sending, receiving, and countermarking scent signals. Int J Primatol 25, 1019–1031 (2004).

  89. 89.

    Endler, J. A. Some general comments on the evolution and design of animal communication systems. Philos Trans R Soc Lond B Biol Sci 340, 215–225 (1993).

  90. 90.

    Koch, R. E., Josefson, C. C. & Hill, G. E. Mitochondrial function, ornamentation, and immunocompetence. Biol Rev 92, 1459–1474 (2017).

  91. 91.

    Kuo, T.-H. et al. Insulin signaling mediates sexual attractiveness in Drosophila. PLoS Genetics 8, e1002684 (2012).

  92. 92.

    von Schantz, T., Bensch, S., Grahn, M., Hasselquist, D. & Wittzell, H. Good genes, oxidative stress and condition–dependent sexual signals. Proc R Soc B Biol Sci 266, 1–12 (1999).

  93. 93.

    Arakawa, H., Cruz, S. & Deak, T. From models to mechanisms: odorant communication as a key determinant of social behavior in rodents during illness-associated states. Neurosci Biobehav Rev 35, 1916–1928 (2011).

  94. 94.

    Strowig, T., Henao-Mejia, J., Elinav, E. & Flavell, R. Inflammasomes in health and disease. Nature 481, 278–286 (2012).

  95. 95.

    Shirasu, M. & Touhara, K. The scent of disease: volatile organic compounds of the human body related to disease and disorder. J Biochem 150, 257–266 (2011).

  96. 96.

    Archie, E. A. & Theis, K. R. Animal behaviour meets microbial ecology. Anim Behav 82, 425–436 (2011).

  97. 97.

    McCune, B., Grace, J. B. & Urban, D. L. Analysis of ecological communities. (MjM Software Design, 2002).

  98. 98.

    Vegan: Community ecology package. (R package version 2.4-4,, 2017).

  99. 99.

    R Core Team. R: a language and environment for statistical computing. (R Foundation for Statistical Computing, 2016).

  100. 100.

    RStudio: integrated development environment for R (Boston, MA). URL: (2016).

  101. 101.

    Bates, D., Maechler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. J Stat Softw 67, 1–48 (2015).

  102. 102.

    Albone, E. S. Mammalian semiochemistry: the investigation of chemical signals between mammals. (Wiley, 1984).

  103. 103.

    Kuznetsova, A., Brockhoff, P. & Christensen, R. lmerTest Package: tests in linear mixed effects models. J Stat Softw 82, 1–26 (2017).

  104. 104.

    Hothorn, T., Bretz, F. & Westfall, P. Simultaneous inference in general parametric models. Biom J 50, 346–363 (2008).

  105. 105.

    Clarke, K. R. & Gorley, R. N. PRIMER v7: User Manual/Tutorial. (PRIMER-E, 2015).

  106. 106.

    Anderson, M. J., Gorley, R. N. & Clarke, K. R. PERMANOVA+ for PRIMER: guide to software and statistical methods. (Primer-E Ltd, 2008).

  107. 107.

    Liaw, A. & Wiener, M. Classification and regression by randomForest. R News 2/3, 18–22 (2002).

  108. 108.

    Fournier, D. et al. AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optim Method Softw 27, 233–249 (2012).

  109. 109.

    Zuur, A. F., Ieno, E. N., Walker, N. J., Saveliev, A. A. & Smith, G. M. Mixed effects models and extensions in ecology with R. Vol. xxii (Springer, 2009).

Download references


We gratefully acknowledge the DLC staff for animal and veterinary care, health records and sample collection, particularly E. Ehmke, D. Brewer, B. Schopler, C. Williams and S. Zehr. We thank T. Clarke for assistance with conducting bioassays and A. Stonehill for scoring the videos. G. Dubay kindly provided access to the GC-MS instrument and training in its use. We thank D. Nuttall for guidance on chemical quantitation, N. Grebe for helpful discussions on statistical analyses, and L. Greene for sample collection and guidance on sample preparation and statistical analyses. This research was supported by the National Science Foundation (BCS-0409367, BCS-1749465, IOS-0719003 to C.M.D.), Duke University Research funding (to C.M.D.) and the Duke Lemur Center Director’s Fund (to R.L.H.). This is DLC publication #1402.

Author information


  1. Department of Evolutionary Anthropology, Duke University, Durham, NC, USA

    • Rachel L. Harris
    •  & Christine M. Drea
  2. Department of Biology, Bishop’s University, Sherbrooke, QC, Canada

    • Marylène Boulet
  3. Department of Anthropology, Pennsylvania State University, State College, PA, USA

    • Kathleen E. Grogan
  4. Department of Biology, Duke University, Durham, NC, USA

    • Christine M. Drea


  1. Search for Rachel L. Harris in:

  2. Search for Marylène Boulet in:

  3. Search for Kathleen E. Grogan in:

  4. Search for Christine M. Drea in:


M.B. and C.M.D. conceived of the idea for the study. All authors collected odorant samples used for bioassays and chemical analyses. R.L.H. produced the olfactory chemistry data with input from M.B. and conducted the bioassays with input from K.E.G. R.L.H. analysed the data and wrote the first draft of the manuscript, with critical edits from C.M.D., K.E.G. and M.B. All authors have approved the final manuscript for publishing.

Competing Interests

The authors declare no competing interests.

Corresponding author

Correspondence to Christine M. Drea.

Electronic supplementary material

About this article

Publication history






By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.