Discussion
The link between mating system, mate choice and dispersal has rarely been studied in pair-living mammals. Here, we demonstrated that coppery titi monkeys are mostly genetic monogamous, as we did not find evidence for EPP in our study population. We also did not find evidence for relatedness- or heterozygosity-based mate choice. Despite the absence of evidence for active inbreeding avoidance via mate choice, pair mates in our study population had low average relatedness. This finding suggests that in our study population, natal dispersal generates sufficient level of genetic dissimilarity between females and males to render both active inbreeding avoidance and EPC unnecessary.
Coppery titis are only the second primate species and the seventh pair-living mammal with no evidence of EPP found in a study with an adequate sample size (the study on Bornean gibbon was based on just four infants from four family groups 37, Table 1). The absence of EPP in titis is not unexpected, as they are consistently pair-living, pair mates spend most of the day within a few meters from each other, sleep together at night and engage in frequent joint visual displays and duetting at the territorial borders 36,38–41. This high level of proximity and coordination should make mate guarding easy and effective enough to prevent EPC.
The opportunities for EPC are likely limited, too. The home ranges of our study groups have very little overlap (1.4% on average (0–4.7), unpublished data; Fig. 1), and to find extra-pair mates, individuals would need to intrude into the neighboring home ranges, risking aggression from the same-sex residents. Another way to obtain EPC could be mating with floaters, solitary non-territorial individuals ranging over a wide area after having dispersed from their natal groups. There is accumulating evidence for the importance of floaters in population dynamics of both birds and mammals 42,43. For example, in Azara’s owl monkeys who are very similar to titis in all aspects of their social system, mated individuals experience intense intra-sexual competition from floaters of both sexes 43,44. However, the evidence from Azara’s owl monkeys and many other bird and mammal species indicate that floaters do not copulate with the mated animals as often as might be intuitively expected, and EPP are attributed to the neighboring individuals in most cases (e.g., Barelli et al., 2013; Nimje et al., 2019; Petrie & Kempenaers, 1998; but see Cohas, Yoccoz, Da Silva, Goossens, & Allainé, 2006; Kenyon, Roos, Binh, & Chivers, 2011). In titis, only anecdotal reports of replacements by intruders exist 49,50, but given the difficulty of detecting floaters, it is possible that they are present in titi populations, too. However, given the high levels of proximity and coordination between pair mates, EPC with the floaters are probably not easier to obtain than EPC with the neighboring individuals. Furthermore, EPC, whether with floaters or neighboring animals, might be costly, with the risks including the higher probability of acquiring sexually transmitted diseases and, for females, the retaliatory withholding of parental care by males 51,52.
Opportunities for EPC are also affected by population density, with the higher densities making the encounters between individuals and, consequently, EPC less likely 53. The positive relationship between population density and EPP rates was demonstrated, e.g., in Eurasian beavers, Castor fiber, and in many bird species 8,45,54,55. Notably, the only published record of EPC in titis comes from a population of Plecturocebus ornatus (previously C. moloch) living in a disturbed habitat with exceptionally high density of 406 individuals/km2 56. For comparison, the population density at our study site was estimated at 34 individuals/km2 (unpublished data), and the average size of the home range of one group was 7.2 ha (Supplementary Table S1), bigger than the entire forest fragment of 6.9 ha inhabited by nine groups in the study of Mason (1966). The density of our study population was within the average range of values reported for undisturbed populations of titis 57–59; this relatively low density likely limited the opportunities for EPC. It should be mentioned, however, that for a population of Plecturocebus discolor from undisturbed habitat, a preliminary analysis reported three cases of EPP in a sample size of 16 offspring, although these data has not been published yet (Van Belle et al., 2016b, conference abstract). The density of this population (57 individuals/km2) was higher than that of our study population, the home ranges were on average smaller (5.0 ha) and the percentage of home range overlap was larger (4.8%) (Fernandez-Duque and Fiore, 2020), possibly accounting for the higher EPP rate.
Although in all cases of assigned paternities the social fathers were identified as genetic fathers for the group offspring (17 offspring born in 9 groups, up to 5 offspring generations per group), we cannot fully exclude the possibility of a low EPP rate in our study population. First, for one offspring (Group 10), paternity remained unassigned, as neither social father nor any other male from our sample was identified as the most likely father. While this case could be classified as neither extra- nor intra-pair paternity with confidence, it remains possible that this offspring was sired by an unsampled extra-pair male. In this case, the EPP rate in our study population would be 6%. Alternatively (if we assume that the social father is indeed not the genetic father of the offspring), this case could be the result of a male replacement in a group. Adult replacements are known to happen in titis, with the breeding positions vacated after the disappearance (presumable deaths) of adults being occupied by same-sex immigrants 33. Replacements can create groups that do not consist of biological parents and their offspring, leading to the apparent deviations from genetic monogamy even in the absence of EPC. As Group 10 was only habituated shortly before the genetic sample collection and no older offspring were present in it, we could not reconstruct its demographic history. Our data indicates that adult replacements do happen in our study population. The adult female of Group 4 was not identified as the genetic mother of the group’s juvenile offspring, while the adult male was indicated as the genetic father. When we started following this group, the juvenile was estimated to be 7–8 months old based on its body size and the fact that it walked independently (juvenile titis start to walk on their own most of the time at the age of ca. 4.5 months: Jantschke et al., 1995). Lactation in titis lasts ca. 6.5 months 62, and we did not see the female nursing. Therefore, we assume that the female replacement must have happened within ca. 2 months before we started following the groups, after the juvenile had been weaned.
Second, given the sample size of 17 offspring, maximum possible EPP level (assuming no EPP has been found and estimated with 95% confidence) will be 16.2%, calculated as y = 1–(1–x)17, where y is the probability of producing at least one extra-pair offspring (0.95); x is the frequency of EPP, and n is the sample size 63. This confidence limit is a product of the sample size and does imply that there is 16.2% EPP rate in our study population. To narrow down the confidence interval to at least 5% of EPP, we would need a sample size of 58 offspring, which is difficult to achieve in a reasonable period in a secretive arboreal primate with slow life history, living in pairs and giving singleton births only once a year.
Contrary to our predictions, we did not find evidence for relatedness- or heterozygosity-based mate choice in our study population. Interestingly, despite the absence of evidence for active inbreeding avoidance via mate choice, the pair mates in our study population were on average not related (mean
r = -0.033) and never shared the same mtDNA haplotype (Supplementary Table S1, Fig. 2). Only in one case the pair mates were second-degree kin with
r = 0.285. Low relatedness between mating partners in the absence of active inbreeding avoidance was demonstrated in many other populations of mammals and birds, e.g., grey wolves,
Canis lupus, arctic foxes,
Vulpes lagopus, great reed warblers,
Acrocephalus arundinaceus, and blue tits,
Parus caeruleus 24,64,65. In fact, active inbreeding avoidance via mate choice, although demonstrated in some birds and mammals (e.g., Hoffman et al., 2007; Leedale et al., 2020), was found in most pair-living species (García-Navas et al., 2009; Hansson et al., 2007; Schwensow et al., 2008; Sommer, 2005; reviewed in Jamieson et al., 2009). It has been suggested that in most situations, dispersal may be sufficient to avoid inbreeding
24. By disrupting close-kin associations, dispersal can make the probability of encountering close kin relatively low, rendering active inbreeding avoidance via mate choice unnecessary
66. In such cases, kin discrimination mechanisms might fail to evolve, and low inbreeding levels that will occasionally occur in such systems will be tolerated
66.