The optimal allocation of reproduction and fertility across the lifespan constitutes a key life-history problem (Stearns, 1976; Roff, 1992; Healy et al., 2019). Theoretical frameworks that predict a trade-off between life-history traits therefore include reproductive senescence, i.e., a decline in reproductive performance, with age (Lemaître et al., 2015, Lemaître and Gaillard, 2017). While some authors have distinguished between aging (the process) and senescence (the outcome), we follow others (e.g., Monaghan et al., 2008; Lemaître et al., 2015) in not making this distinction and using only senescence below. In humans and a few other mammalian species, reproductive senescence terminates in menopause, i.e., the complete cessation of fertility before the average lifespan is reached (Walker and Herndon, 2008; Johnstone and Cant, 2010, 2019; Lieberman et al., 2021). Despite long-held skepticism (Medawar, 1952), there is now compelling evidence for the widespread existence of senescence in a variety of animal species (Monaghan et al., 2008; Nussey et al., 2013). Comparative studies based on long-term data from various wild vertebrate populations have confirmed the presence of early late life trade-offs, suggesting that resource limitation leads individuals to trade somatic maintenance later in life for high allocation to reproduction early in life (Lemaître et al., 2015), and have begun to identify environmental factors that modulate their timing and intensity (Gaillard and Lemaître, 2020). Studying intra- and interspecific variation in longevity and fertility, their complex relationship, as well as the factors modulating them, is therefore crucial for understanding variation in fitness, but our understanding of these processes remains limited, partly because data from very long-lived taxa remain scarce (Gaillard and Lemaître, 2020). Here, we present results of a comparative study exploring these patterns and sources of age-specific female reproduction in relatively long-lived basal primates to contribute to a redressal of this data gap.

As in the vast majority of mammalian species with relatively long lifespan (Lemaître et al., 2020), such as African elephants (Hayward et al., 2014), reproductive rates of female primates are expected to decline with age. Gestation and lactation are energetically demanding, and older female mammals should balance the successful rearing of current offspring against the probability to survive and to produce future offspring (Lemaître and Gaillard, 2017). Indeed, in wild, and thus unprovisioned, chimpanzees and yellow baboons, an age-related decline in female reproductive output was found, but only at relatively old maternal ages and at the same rate as the probability to survive declined (Thompson et al., 2007; Altmann et al., 2010). In some baboon populations, female reproductive activity even ceased before individuals reached their average lifespan (Packer et al., 1998). Further comparisons among six anthropoid (baboons, blue monkeys, capuchins, chimpanzees, gorillas, muriquis) and one lemur species (Verreaux’s sifakas) also suggested that reproductive senescence was present – at least in baboons and sifakas statistically significant effects could be demonstrated – but most individuals died before fertility cessation (Alberts et al., 2013). Another comparative study of 13 primate species showed that age-specific fertility declined with increasing age, also in captivity (Caro et al., 1995). Furthermore, in wild but provisioned rhesus macaques, rates of reproductive senescence were disproportionately higher than those of somatic decline (Lee et al., 2021). Finally, female muriquis, blue monkeys, baboons, gorillas and chimpanzees showed longer interbirth intervals (IBIs) in the oldest or youngest age classes, or both, and the oldest females also showed relatively fewer completed IBIs (Campos et al., 2022). Thus, under various conditions of food availability, reproductive senescence has been reported for various primate species. However, the majority of field studies used estimated or statistically modeled age classes and could not draw on known-aged individuals; one of the practical challenges of long-term field studies (Kappeler et al., 2012, 2017).

Whenever longitudinal data on reproductive rates are not available, cross-sectional data on IBI may provide relevant information on reproductive senescence because the probability of prolonged IBIs is predicted to increase with age. The current evidence for this effect is mixed, however. In Barbary macaques, for example, younger mothers wean their offspring indeed significantly earlier than older ones (Paul et al., 1993), indicating that the successful rearing of offspring is age-related, at least in some species (Atsalis and Margulis, 2008). In particular, IBIs were the longest for very young and very old females in this population (Paul et al., 1993). In chimpanzees, IBIs also increased significantly with maternal age, but the effect size was rather small and this effect was probably more due to individual maternal health, as IBI length in this species is closely related to somatic senescence (Thompson et al., 2007). In the comparative study of 13 captive primate species, the majority of correlations between IBIs and female age were negative, but only significant in humans, orangutans and chimpanzees (Caro et al., 1995). Female baboons seemed to experience shorter IBIs with advancing age in this study, but again this correlation was not significant. In wild capuchin, howler and spider monkeys, IBIs were not affected by age, but also not by parity, rank or body weight (Fedigan and Rose, 1995). Among several populations of captive common marmosets, Smucny et al. (2004) found a significant non-linear correlation between maternal age and IBI length. Thus, female primates appear to account for the increasing costs of reproduction with increasing age by adjusting IBIs in some cases.

If the detrimental age-effects cannot be fully compensated by reduced reproductive rates, infant survival may also be compromised. Even if females continue to reproduce at regular intervals throughout their lifetime, advanced maternal age may result in increased rates of miscarriages or negatively affect offspring survival because of reduced birth weights, milk production or other complications (Atsalis and Margulis, 2008; Lemaître and Gaillard, 2017; Comizzoli and Ottinger, 2021). In fact, infant survival in female olive baboons rapidly declined with age, and mothers beyond 21 years of age were more likely to experience a miscarriage (Packer et al., 1998). Yet, although the maternal age effect on offspring survival in semi-free ranging Barbary macaques was statistically not significant, infants of older mothers were more likely to survive than those of younger females (Paul et al., 1993). Thus, the efforts by senescence mothers to reproduce successfully may also be compromised by reduced infant survival.

The social system of a species may also have evolutionary consequences for age-dependent female reproductive strategies. Humans and beluga whales, killer whales, narwhales and short-finned pilot whales have evolved an extreme form of reproductive senescence (i.e., menopause), where the post-reproductive phase can last at least as long as the reproductive phase (Ward et al., 2009; Ellis et al., 2018; Johnstone and Cant, 2019). In these species, menopause appears to be the outcome of kin selection in social systems characterized by appreciable or effective female dispersal by reducing the costs of intergenerational reproductive conflict (Johnstone and Cant, 2019). Reproductive effort can also be compromised by competition with other females independent of age, however (Beehner and Lu, 2013). For example, in Hanuman langurs conception probability was lower and IBIs were longer in groups containing more females (Sommer and Rajpurohit, 1989; Sommer et al., 1992). Thus, the evolution of key life-history traits appears to be also shaped by features of the social system, including patterns of female kinship, mate quality and reproductive competition.

Other factors can modulate reproductive performance on more immediate levels and time scales. In the wild, animals are exposed to natural fluctuations in food availability, rainfall, temperature and parasites, all of which can have instantaneous consequences for the current or next reproductive cycle, and thus, for reproductive rates. These relationships can be investigated by either cross-sectional comparisons across multiple populations or groups, longitudinal contrasts across multiple seasons or years for the same individuals, or by comparing wild and captive populations of the same species or genus. In the present study, we will adopt the latter two approaches, focusing on two lemur genera in the wild and in captivity.

Long-term studies of wild populations are often compromised by practical challenges related to the longevity of the study subjects, which may exceed the academic lifespans of the researchers, and other practical difficulties (see above). In contrast, unmanaged captive populations represent a convenient opportunity to investigate the dynamics of maximal female reproductive rates under conditions of food abundance, veterinary care and reduced parasite exposure. Such effects might be far more significant and easier to detect in species with short lifespans and high reproductive rates (Tidière et al., 2016). Nonetheless, compared to their wild relatives, semi-captive ring-tailed lemurs did not seem to experience reproductive senescence (Lemaître and Gaillard, 2017). However, in two species of captive ruffed lemurs, litter size as well as infant survival follow an inverted U-shaped pattern, with a decreasing point after mothers reached the age of 9 and 11.5 years, respectively (Tidière et al., 2018). In Hanuman langur females, members of free-ranging populations exhibited a later onset of about 10 months of first conception than they do in captivity (Sommer et al., 1992). Similarly, the timing of first reproduction in wild female common marmosets was delayed by several months compared to their captive counterparts (Tardif et al., 2008). Comparisons of reproductive effort between wild and captive populations can therefore provide information about the magnitude of phenotypic plasticity in these traits, offering a benchmark for maximal reproductive effort to which data from populations with various environmental constraints can be compared.

Non-human primates share fundamental features of their life histories with humans (van Schaik and Isler, 2012), so that studying a basal group of primates can contribute valuable comparative information on the evolution of reproductive senescence in general and (human) menopause in particular. Despite their relatively long lifespans, slow development and reproductive rates, long-term field studies of primates have generated valuable demographic data for comparative studies of fundamental questions in life-history evolution (Bronikowski et al., 2011; Morris et al., 2011; Colchero et al., 2016, 2021; Campos et al., 2017, 2022). Lemurs, the endemic primates of Madagascar, have been underrepresented in these comparative studies because of the limited number of long-term study sites (Jolly, 2012; Kappeler and Fichtel, 2012a; Sussman et al., 2012; Wright et al., 2012; Kappeler et al., 2017). They are nonetheless interesting for studies of reproductive senescence because they tend to live longer than most anthropoid primates of similar body mass, with a remarkable maximum longevity of over 30 years for mammals of < 10 kg, at least in captivity (Zehr et al., 2014). Lemurs also exhibit impressive interspecific variation in age at first reproduction, litter size and reproductive rates (Kappeler, 1995, 1996; Ross, 1998; Tecot et al., 2013), but patterns and drivers of intraspecific variation have not yet been studied in great detail.

Representatives of two lemur families (Indriidae and Lemuridae) that evolved diurnal activity, group living and single births independently (Kappeler, 1998; Kappeler and Pozzi, 2019), thus resembling anthropoid primates in key life-history features, have been subjected to long-term field studies. The available published information about lemur reproductive senescence can be summarized as follows (for wild mouse lemurs see Rina Evasoa et al., 2018; for captive lemurs: Zehr et al., 2014). First, one study of Milne Edward’s sifakas (Propithecus edwardsi) revealed that, although no overall effect of age on fertility could be found, offspring survival was significantly decreased in years with below-average rainfall during the lactation period, but the age classification in this study was based on estimates of tooth wear rather than known ages (Wright et al., 2008). Second, in wild Verreaux’s sifakas (P. verreauxi) at Beza Mahafaly, which were also assigned to one of 5 age classes based on tooth wear analyses, females began reproducing surviving infants with 5 years of age, followed by a sharp increase in reproductive rates over subsequent years and a steady phase of about a decade before a decline in average fertility set in around the age of 18 (Richard et al., 2002). Nonetheless, many females in their early 20 s continued to reproduce regularly, and the oldest mother was estimated to be 28 years old. Controlling for body mass, these sifakas had a greater age at last reproduction than 14 other primate species for which these data were available at the time. Female reproductive performance in this species was also related to unpredictable climate because a severe drought decreased birth rates by almost 20%, and only few offspring born that year survived (Richard et al., 1991, 2000). At Kirindy Forest, the proportion of female Verreaux’s sifakas reproducing varied widely across years, and part of this variation was explained by variation in the number of co-resident females (Kappeler and Fichtel, 2012a). Finally, wild female ring-tailed lemurs, which were also grouped into age-classes, reproduced irrespective of age, but birth rates declined in groups with larger numbers of females and in years with a drought (Gould et al., 2003). In another population, old female ring-tailed lemurs reproduced at slightly, but not significantly, reduced rates, presumably because most of them had already died or disappeared before a clear pattern could have been detected (Ichino et al., 2015).

Additional data from long-term studies of age-specific fertility in the wild are obviously needed to further investigate reproductive senescence and its underlying mechanisms in lemurs. Here, we present such data on sympatric wild Verreaux’s sifakas and redfronted lemurs (Eulemur rufifrons), offering an opportunity for contrasting the life histories of representatives of independently evolved lineages to identical (fluctuations in) environmental conditions. Both species live in small multi-male, multi-female groups and feed on a broad range of fruits and leaves (Koch et al., 2017). Verreaux’s sifakas are slightly larger than redfronted lemurs (2.9 vs. 2.2 kg), their groups are of similar size (6.1 vs. 5.4 individuals), female reproduction begins later (5 vs. 3 years), but both species typically produce singletons and can live well over 20 years (Richard et al., 2000, 2002; Kappeler and Fichtel, 2012a,2016; Zehr et al., 2014). Despite female philopatry in both species, female Verreaux’s sifakas occasionally emigrate into other groups (Kappeler and Fichtel, 2012a), and female redfronted lemurs very rarely succeed in immigrating into other groups after being evicted from their natal one (Kappeler and Fichtel, 2012b). In order to assess the effects of ecological challenges on female reproduction, we compare our results with data from closely related captive Propithecus and Eulemur species, for which the Duke Lemur Center (DLC) (Durham, NC, United States) has accumulated an impressive data set under uniform captive conditions (Zehr et al., 2014).

The main aims of our study are therefore to evaluate female reproductive output and IBIs in two sympatric wild lemur populations as a function of age, the number of adult females (to account for female competition) as well as rainfall as a proxy for variation in food availability. To further investigate the relationship between ecology and life-history dynamics, we compare data on wild and captive populations of the same two genera to account for annual environmental variation and its consequences for reproductive senescence in the unpredictable environment of Madagascar. In line with the relevant evolutionary theory, we expected reproductive senescence (i) to occur in these relatively long-lived small mammals, and (ii) to be less pronounced in captive animals because they are buffered from natural hazards.

The data presented in this study come from two sources. First, we collected demographic and life-history data from sympatric populations of Verreaux’s sifakas and redfronted lemurs at Kirindy Forest, western Madagascar, a dry deciduous forest exposed to pronounced climatic seasonality (Kappeler and Fichtel, 2012a). Beginning in 1995 (sifakas) and 1996 (redfronted lemurs), members of several adjacent groups of both species, including immigrants and newborns, have been regularly captured, measured and individually marked with Passive Integrated Transponder (PIT) tags and unique nylon collars. One individual per group has been equipped with a radio collar to facilitate multiple censuses per week during which each group’s composition, including the number of co-resident adult females, has been recorded. Birth dates of newborn infants are therefore known with a precision of a few days; in most cases we know the exact birthday. Ages of individuals at the beginning of the study could be determined with an accuracy of a year because both species are strict annual breeders and juveniles of different ages clearly differ in body mass from each other; ages of adults during first captures were estimated based on patterns of tooth wear. We therefore created two data sets for the analyses of age-dependent reproductive rates of these two species: dataset A contains only individuals with exactly known ages; dataset B also includes females that were adults at the beginning of the study and whose ages have been estimated. Finally, because climatic data were not collected regularly during the early years of the study, we used published rainfall data from the CHIRPS (n.d.) data base to assess interannual variation on precipitation.

Second, in February 2022 we extracted life-history data from the published records of the Duke Lemur Center (DLC; Zehr et al., 2014), which houses captive colonies of sister species of the two species studied at Kirindy. Because P. verreauxi and E. rufifrons are not kept at the DLC, we extracted reproductive data (i.e., age at first reproduction, IBIs, age at all subsequent births) for Coquerel’s sifaka (P. coquereli) and red lemur (E. rufus) females of known ages. Both species are similar to P. verreauxi and E. rufifrons in body size, social system and also inhabit the dry deciduous forests of western Madagascar in the wild (Mayor et al., 2004; Razafindramanana et al., 2020).

Previous research indicated that lemurs do not experience any form of reproductive cessation and thus, are principally capable of reproducing until death (Wright et al., 2008), whereas other studies revealed reproductive senescence late in life (Alberts et al., 2013). The present analyses revealed strong evidence that females in three out of four populations experience reproductive senescence, thereby offering a basis for an exploration of the sources and dynamics of age-specific female reproduction in independently evolved lineages under different environmental conditions. In line with our first prediction, we did find evidence for reproductive senescence in three populations of these relatively long-lived small mammals. In wild redfronted lemurs, captive red lemurs and captive Coquerel’s sifakas, the probability of giving birth decreased with mother’s age, but it was unaffected by maternal age in wild Verreaux’s sifakas. While the pattern observed in the first three of our study populations corresponds to the pattern also observed in about 70% of mammalian species to date (Lemaître et al., 2020), the absence of evidence for reproductive senescence in Verreaux’s sifaka is intriguing, also because it only partly corroborates results of previous studies with data from another population of this species, which reported reproductive senescence (Alberts et al., 2013) but no change in IBIs with age (Campos et al., 2022). Thus, populations may vary in patterns of reproductive senescence and/or different components of reproductive senescence may vary independently.

This outcome across populations rejects our second prediction, stating that reproductive senescence might be less pronounced in captive animals because they are buffered from external adversities. With respect to the relative importance of age in modulating reproductive rates, we found that neither the number of co-resident females, nor annual rainfall did predict variation in the probability of giving birth in wild Verreaux’s sifakas either. In wild redfronted lemurs, in contrast, mothers were more likely to give birth in years with more rainfall. Finally, birth rates did not co-vary with any proxy indicating potential selective appearance or disappearance in the population. i.e., longevity or age at first or last reproduction (van de Pol and Verhulst, 2006; Nussey et al., 2008). Only in captive Coquerel’s sifakas longevity co-varied by trend with birth rates, suggesting that longer-lived females might have been more efficient at giving birth throughout their life.

We investigated effects of age, female competition and variation in food availability on female reproductive performance and IBIs in two sympatric wild populations of true lemurs and sifaka and compared these data with captive populations of the same two genera to identify consequences for reproductive senescence. We found some negative effects of age on reproductive performance in three out of four populations. Wild and captive populations did not differ in these life-history traits in the predicted manner, but data from the captive populations were presumably compromised by husbandry decisions. In the wild, only some females live long beyond their last birth, offering no evidence for fertility completion in these basal primates, but reproductive senescence appears to set in shortly before the end of their natural lifespan. Our study contributes additional data on age-related life-history adaptations in monotocous primates and, hence, to a more comprehensive appraisal of reproductive senescence.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found 10.6084/m9.figshare.20036558.

The animal study was reviewed and approved by the Caff/Core Ministry of the Environment, Madagascar.

PK and CF collected the data and conceived research. LPr and CF analyzed data. LPe and PK drafted the manuscript. All authors contributed to writing the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.