A long postreproductive life span is a shared trait among genetically distinct killer whale populations

Abstract The extended female postreproductive life span found in humans and some toothed whales remains an evolutionary puzzle. Theory predicts demographic patterns resulting in increased female relatedness with age (kinship dynamics) can select for a prolonged postreproductive life span due to the combined costs of intergenerational reproductive conflict and benefits of late‐life helping. Here, we test this prediction using >40 years of longitudinal demographic data from the sympatric yet genetically distinct killer whale ecotypes: resident and Bigg's killer whales. The female relatedness with age is predicted to increase in both ecotypes, but with a less steep increase in Bigg's due to their different social structure. Here, we show that there is a significant postreproductive life span in both ecotypes with >30% of adult female years being lived as postreproductive, supporting the general prediction that an increase in local relatedness with age predisposes the evolution of a postreproductive life span. Differences in the magnitude of kinship dynamics however did not influence the timing or duration of the postreproductive life span with females in both ecotypes terminating reproduction before their mid‐40s followed by an expected postreproductive period of about 20 years. Our results highlight the important role of kinship dynamics in the evolution of a long postreproductive life span in long‐lived mammals, while further implying that the timing of menopause may be a robust trait that is persistent despite substantial variation in demographic patterns among populations.


| INTRODUCTION
The evolution of an extended female postreproductive life span is extremely rare in nature and is at present only known in five species of wild mammals (Croft et al., 2015;Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017;Ellis et al., 2018). Outside of the prolonged postreproductive life span seen in humans the only other species of mammals in which females have evolved early cessation of reproduction are toothed whales: short-finned pilot whales (Globicephala macrorhynchus), narwhals (Monodon monoceros), belugas (Delphinapterus leucas), and resident-ecot ype killer whales (Orcinus orca) (Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017;Johnstone & Cant, 2010). Some evidence suggests that also the false killer whales (Pseudorca crassidens) have a substantial postreproductive period (Photopoulou et al., 2017). In resident-ecotype killer whales, for example, adult females typically give birth to their last calf in their mid-30s to early 40s followed by a postreproductive life span that may span many decades (Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017). In the classical view of evolutionary theory, early termination of reproduction is not a beneficial trait (Hamilton, 1966;Williams, 1957) and understanding why and how the postreproductive life span has evolved remains a considerable challenge for evolutionary biology.
Adaptive explanations for the evolution of a long postreproductive life span have tended to focus on the inclusive fitness benefits of helping kin in late life (Kim et al., 2012;Nattrass et al., 2019;Sear & Mace, 2008). Females can gain inclusive fitness benefits in late life by ceasing reproduction and instead invest their energy in helping existing offspring survive and reproduce ("the mother hypothesis") (Williams, 1957). Further, through behaviors that help increase the survival of grandchildren, such as providing ecological knowledge (Brent et al., 2015) or provisioning (Hawkes et al., 1997), postreproductive females can increase their inclusive fitness ("the grandmother hypothesis") (Hawkes, 2004). In humans, grandmother benefits appear to be key for the evolution of a long postreproductive life span (Hawkes, 2004;Hawkes & Coxworth, 2013) and recent work in resident killer whales provides support for both the mother and grandmother hypothesis with the presence of both mothers and postreproductive grandmothers having a positive impact on the survival of their adult offspring and grandofffspring (Foster et al., 2012;Nattrass et al., 2019). However, the inclusive fitness benefits from helping are likely not on their own sufficient to explain the timing of menopause in both killer whales and humans (Hill & Hurtado, 1991) leading to the search for additional mechanisms that can contribute to the early termination of reproduction (McComb et al., 2001;Moss, 2001). Recent work has shown that kin-selected costs, as well as benefits, are important for the evolution of extended postreproductive life spans (Cant & Johnstone, 2008;Johnstone & Cant, 2010).
Demographic patterns with either female-biased dispersal and local mating or natal philopatry of both sexes and non-local mating give rise to age-specific changes in the relatedness of an individual to its group (kinship dynamics), in particular an increase in average female relatedness to other group members with age (Cant & Johnstone, 2008). In the case of resident killer whales, females are  (Ford & Ellis, 2006) 4.33 (1-16) (Center for Whale Research) Natal philopatry of males and females, nonlocal mating (Bigg et al., 1987) (but see Ford et al. (2018)) Northern residents Fish (mainly salmon) (Ford & Ellis, 2006) 6.5 (1-19) (Towers et al., 2020) Natal philopatry of males and some females, nonlocal mating (Barrett-Lennard & Ellis, 2001;Bigg et al., 1987) Bigg's West Coast Transients (WCT) (Sharpe et al., 2019) Marine mammals (mainly pinnipeds and small cetaceans) (Ford & Ellis, 1999) 3.9 (1-8) (Towers et al., 2019) Some dispersal of males and females from natal group, non-local mating inferred (Baird & Whitehead, 2000;Ford & Ellis, 1999) born into a social unit ("matriline") consisting of their mother, siblings, and other more distant relatives (Bigg et al., 1990). As they age, their own sons replace more distantly related males in the matriline, increasing their average local relatedness to the group over time . This ultimately leads to an asymmetry in selection for helping and harming with age, which means that older females that are more related on average to the group are under stronger selection to help, while younger females are under stronger selection to harm (Johnstone & Cant, 2010). Thus, in competition for reproduction older females experience a larger inclusive fitness cost compared with younger females .
The combination of such inclusive costs of intergenerational reproductive conflict and inclusive fitness benefits of helping kin are hypothesized to be key predictors for the evolution of a long postreproductive life span in mammals (Cant & Johnstone, 2008;Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017;Johnstone & Cant, 2010).
Investigating kinship dynamics and age-specific life history changes requires long-term social and demographic data that captures most of the life span of animals. These data are therefore rare in long-lived mammals. The long-term data collected on different populations of killer whales in the coastal waters of the USA and Canada now extends over more than four decades, providing a unique opportunity to examine the link between kinship dynamics and life history evolution in a long-lived marine mammal. In addition to the support for the mother and grandmother hypotheses in resident killer whales (Foster et al., 2012;Nattrass et al., 2019), there is strong support for the reproductive conflict hypothesis with offspring of older females that are born into conflict with offspring of a younger female having a 1.7 times higher mortality risk . However, it is still unknown whether these traits are shared between different populations of killer whales. Killer whales are among the most widely dispersed mammals on the planet and are found in all oceans (Baird et al., 1999;Ford, 2009). Lineages that differ morphologically (Baird & Stacey, 2011) and behaviourally (Riesch et al., 2012) and are genetically isolated (Morin et al., 2010) (Table 1; as well as a third, rarely encountered offshore ecotype not considered here) (Morin et al., 2010;Olesiuk et al., 2005;Parsons et al., 2013). Both populations of residents are specialist fish-eaters with salmon making up almost all of their prey (Bigg et al., 1990;Olesiuk et al., 1990), whereas Bigg's killer whales are specialized in hunting marine mammals (Ford & Ellis, 1999). This differentiation in diet is reflected in the social behavior of the ecotypes with resident killer whales typically being observed traveling in larger social groups consisting of several maternal groups, compared with Bigg's killer whales (Towers et al., 2019(Towers et al., , 2020. The mean group size of cohesive maternal groups however is similar for the two ecotypes (Table 1; Towers et al., 2020). In resident killer whales, there is almost no dispersal of males and limited dispersal of females from the maternal group. In contrast, there is a dispersal of both sexes from the maternal group of Bigg's killer whales (Baird & Whitehead, 2000;Ford & Ellis, 1999), which may be related to maintaining optimal group foraging size for predating on marine mammals (Baird & Dill, 1996;Krause & Ruxton, 2002).
The difference in demography observed between the resident and Bigg's ecotypes is predicted to generate different patterns of kinship dynamics (age-specific changes in relatedness) which can be illustrated using the theoretical model presented by Johnstone and Cant (2010). Previous theoretical work on kinship dynamics in resident killer whales predicted an increase in local relatedness with female age (Johnstone & Cant, 2010) which has been confirmed using individual-based demographic and social data from resident killer whales . Here, we use this established modeling framework to predict the patterns of age-related changes in kinship for female Bigg's killer whales allowing us to directly compare the predicted kinship dynamics between the resident and Bigg's ecotype. Using this approach, it is predicted that female local relatedness will increase with age for both killer whale ecotypes, albeit with a weaker relationship under the demographic conditions of Bigg's killer whales (Figure 1; Johnstone & Cant, 2010). This increase in female local relatedness is opposite to the pattern observed in most mammals (male dispersal and local mating), and it predicts that there will be selection for a postreproductive female life span in both killer whale ecotypes. We hypothesize however that the difference in the strength of the kinship dynamics will lead to a lower potential for inclusive fitness benefits from helping kin in late life and inclusive fitness costs of reproductive conflict with younger females for Bigg's females . Given the predicted differences in kinship dynamics, and assuming the costs and benefits of reproduction with age are equal and that these do not change with changing dispersal, we predict that selection for a postreproductive life span will be weaker in Bigg's killer whales compared with resident killer whales (Johnstone & Cant, 2010) and we further hypothesize that this will be reflected by an older age at last reproduction, and a shorter postreproductive life span in female Bigg's killer whales.
Here, we test for the presence of a postreproductive period in Bigg's killer whales and compare female postreproductive life span between the sympatric resident and Bigg's killer whale ecotypes.
Using over 40 years of individual-based demographic data, we model survival trajectories using a Bayesian hierarchical framework (Colchero & Clark, 2012) and calculate the postreproductive repre-

| METHODS
Here, we use postreproductive representation (PrR) as a measure for postreproductive life span, which is measured as the proportion of adult years being lived as postreproductive (Levitis & Lackey, 2011), and in the context of this study, it is the proportion of female years which are lived postreproductively. While a decline in fertility with age is a general trait among mammals (Packer et al., 1998), the long postreproductive periods, often spanning more than a decade, observed in humans and some toothed whales is a rare trait. Reports of postreproductive life span in other species often reflect individual variation in senescence, rather than a general trait at the population level are calculated for populations living under artificial conditions often with reduced mortality risk (Cohen, 2004;Croft et al., 2015;Levitis et al., 2013;Packer et al., 1992). A significant advantage of the PrR measure is that it is a population-level measure that is directly comparable between species or populations with different lengths of life span as it is the number of female years lived postreproductively out of all years lived by females in a given population, while PrR also allows for a test of whether the postreproductive life span is significantly larger than what is expected by chance (Levitis & Lackey, 2011).  , 1976, 1987). Sex was determined based on the pigmentation of the ventral side of genital or mammary slits, the presence of neonates, or the size and shape of the dorsal fin of adults for Bigg's killer whales (Table 2). We therefore limited our analysis to a subset of the most commonly documented individuals in this population (Towers et al., 2019). We used the photo-identification data from all populations in the format of a capture-mark-recapture sighting matrix, that is, absence/presence for each individual during each year. The high overall annual recapture probabilities provide reliable long-term data on these populations (Bigg et al., 1990;Towers et al., 2019Towers et al., , 2020.

| Data type and collection
In all three datasets, year of birth was only included for animals that were born after the start of the given study, otherwise year of birth was zero, indicating birth year as unknown (Table 2).
In  (Towers et al., 2019). If year of birth and/or death were unknown for a given individual, these values were assigned a zero and would be estimated by the model (see Table 2 for number of individuals with known birth and death year).

| Estimating age-specific survival
As killer whales are a long-lived species, a common feature of the observation data from all three populations is that some individuals were born before the studies started (i.e., left truncated) and some were still alive at the end of the study (i.e., right-censored).
This gives rise to uncertainties regarding some birth and death years in the datasets. Further, some individuals have gaps in their sighting history of more than 1 year likely due to their preference for waters beyond the core study areas. This introduces uncertainty in the recapture probability and year of death of these individuals. To account for the uncertainties and missing observations in the data, we used a Bayesian hierarchical framework to estimate the agespecific survival and mortality for all three populations (Colchero & Clark, 2012). This framework estimates the unknowns and uncertainties as latent variables (i.e., variables to be estimated) and combines this with flexible parametric mortality functions to predict the age-specific survival of the three killer whale populations . Although the data on Southern residents are nearcomplete, we have used the same approach on all datasets to ensure the results are directly comparable. Given that males and females likely have different mortality trajectories with males often having a higher mortality rate compared with females (Lemaître et al., 2020; Olesiuk et al., 1990), the sex of individuals was included as a covariate in the analyses.
We fitted 10 different mortality and survival models to each of the three datasets using the BaSTA package  in R version 3.6.2 (R Core Team, 2019). The different functions tested either describes a constant mortality that is independent of age (Exponential), an exponential increase in mortality with age (Gompertz) (Gompertz, 1825), an increase or decrease in mortality as a power function of age (Weibull) (Pinder et al., 1978), or an initial exponential increase in mortality that plateaus after a given age (Logistic) (Vaupel et al., 1979). Adding a shape term to the mortality function allows the mortality to have an initial decline from birth (bathtub shape) (Siler, 1979) or an added constant mortality rate that is independent of age (Makeham) (Makeham, 1867

| Testing the fit of the model
We used the deviance information criterion (DIC) to evaluate the model fit and predictive power of the different models, a measure analogous to Akaike's information criterion (AIC) (Spiegelhalter et al., 2002) (but see Spiegelhalter et al. (2014)). The importance of including the categorical covariate of sex was investigated using the Kullback-Leibler discrepancy, which is informed by the overlap of the posterior distribution of parameter estimates (Kullback & Leibler, 1951;Larson et al., 2016;McCulloch, 1989). For data including only individuals of known sex, the Gompertz model with a bathtub b Adult sex ratio is calculated as the fraction of all adult individuals that are male.
shape described the mortality and survival trajectory well for all three populations, as the second best fit for Southern and Northern residents and the best fit for Bigg's (Table 3). Mathematically, the Gompertz bathtub model consists of three elements.
where a 0 and a 1 define the initial decline in mortality from birth, c defines the mortality through the adult stage and b 0 and b 1 define the exponential increase in mortality at the senescent stage. This model therefore offers great flexibility in the age of onset of aging as well as changes in the rate of aging throughout the life span (Colchero & Clark, 2012;Siler, 1979) and we use this model for quantifying the postreproductive life span for all three populations to best ensure direct comparison between the three populations.
Given the lack of data on older whales and the related uncertainty around ages at death, we included weakly informative prior distributions that allowed the models to explore the parameter space while reflecting a plausible range of resulting values for survival. These prior distributions were informed using known life history traits for the populations (Olesiuk et al., 2005). For the final model (Gompertz with bathtub shape), we ran the model with the following prior means

| Permutations of individuals with unknown sex
The sex of a substantial portion of individuals was not determined in the Northern resident and Bigg's populations (Table 2). These individuals were all under 15 years. In mammals, mortality is often higher in early life, and by excluding juvenile individuals from the analyses, we are likely underestimating such early-life mortality.
Instead of including them as a third category of "unknown sex", which would-in essence-calculate the mortality trajectory of an artificial short-lived sex, we used a permutation approach to assign a sex to these individuals randomly. This way we are able to include the early-life mortality into the full age-specific mortality trajectory.
We implemented the same Bayesian hierarchical model that was the best overall fit for individuals of known sex, the Gompertz bathtub model, which was confirmed by a trial run on a set of random datasets as the best fit for the full data. Although only 21 individuals were of unknown sex in the Southern residents, we also ran the same number of permutations on this population for comparability.
We ran 1,000 permutations, where sex was randomly assigned to individuals of unknown sex for each permutation. Both populations have a female-biased adult sex ratio (Table 2), and we assumed a sex ratio of 1:1 at birth. We used the permuted output to calculate the

| Calculating postreproductive representation
To calculate PrR, two data series were obtained, the l x and m x series.
The first series, l x , is the probability of survival to a given age x, obtained from the survival model output (Figure 2). The second, m x , is calculated as the number of offspring born to females of a given age divided by the number of females that were alive at that age. We included all females with a known or estimated year of birth in our calculation of m x . Maternities were assigned from observations of mother-calf interactions following the first observation of the calf (Ford & Ellis, 1999;Towers et al., 2019). Only calves born within the study period were used to estimate age-specific fecundity. Females of a given age were classified each year as either (a) having given birth, (b) not having given birth, or (c) having an unknown birth status.
The latter classification was used when a female was not observed in a given year and the presence or absence of a new calf could not be confirmed. We only included females of a given age with a known birth status (1 or 2) to derive m x . The proportion of females that had a known status of birth was 78% in Northern residents, 70% in Bigg's killer whales, and 93% in Southern residents. The proportion with unknown birth status did not systematically vary by age in any of the three populations. Mathematically, PrR is based on l X and e x , where e x is the life expectancy at age x. A multiplication of these terms gives T X (the total individual years lived after age x). PrR is then calculated from T x at age B and age M, which are the ages, where 5% and 95% of female fecundity has been realized (inferred from m x ). Thus, the formula for calculating PrR is: The input was a lifetable consisting of l x (from the basta model output as this allows for the inclusion of both right-censored and left-truncated data, as well as individuals with missing data) and m x (calculated from the observational data). To test the statistical significance of the PrR value, it was tested against the null hypothesis that survivorship and fecundity declines with the same rate, which would lead to PrR = 0. We simulated 9,999 populations of 1,000 individuals, where this null hypothesis was true and compared each of these null populations with each permutation of the observed population. .

| RESULTS
Overall, 198 Southern resident, 568 Northern resident, and 510 Bigg's killer whales were observed in these populations over the >40 year period of observations (Table 1) and included in the analysis. The adult sex ratio for all datasets was female-biased (Table 2).
There was a substantial number of individuals, where the sex had not been determined with 202 in Northern residents and 209 in Bigg's killer whales (Table 2).

| Age-specific survival
The model (Gompertz with bathtub shape) reveals a clear age-and sex-specific pattern of survival for all three populations (Figure 2) that have similar average and 90% life span (Table 4). We assessed the fit of the model by plotting the estimated survival probability together with the observed survival probability of individuals of known age ( Figure S1).

| Postreproductive representation
All three populations had a postreproductive representation that was significantly greater than zero (p < 0.05), with estimated PrR values above 0.3 for all three populations (Figure 4). Female fecundity had an initial increase at ~12 years, and 95% of fecundity had occurred at age 37-41 (Table 5; Figure 3). For all three populations, median expected female life span as postreproductive was >20 years (Table 5) and the probability of surviving until age M was median >40%.

| DISCUSSION
We From the survival model output, the survival patterns are similar for all three killer whale populations, with females expected to live both substantially longer than males and to ages of >50 years ( Figure 2). Interestingly, the model predicts that male Bigg's killer whales may have a longer expected life span compared with resident males (Figure 2), a pattern that could be influenced by the differences in their environments, both social (Ellis, Franks, Nattrass, Cant, Weiss, et al., 2017;Lemaître et al., 2020) and ecological (i.e., prey availability (Ford et al., 2010;Shields et al., 2018)). However, this could also be an effect of male dispersal in Bigg's, resulting in more uncertainty for the model around the ages at death of males. In females, the survival trajectory of this study is generally supported by previous estimates of life span in killer whales (Olesiuk et al., 1990(Olesiuk et al., , 2005. We used a Bayesian hierarchical approach to estimate the age-specific survival, with the benefit of being able to include individuals with unknown age of birth and death. Through permutations, we were able to include individuals of unknown sex, which likely pro-  Bigg's 12 a 46 ± 2 b 37 a 32 ± 2 b a Fecundity (including age B and age M) is calculated from the number of births to females of a given age out of the number of alive females of that age. b Median expected life span (median ± SD) based on age-specific survival from permutated models. . This metric indicates that Bigg's females are expected to live longer than females from both resident populations with a difference of ~10 years (Figure 2).   Especially in humans, offspring can have a long period after weaning, where they rely on older individuals for help providing food (Johnstone & Cant, 2010). Our study demonstrates that females of both resident and Bigg's killer whales can expect to live more than 20 years after they cease reproduction, allowing for a substantial period with the potential for helping kin. In resident killer whales, mothers impact the survival of their offspring well into their adult years (Foster et al., 2012), and grandmothers are important repositories for ecological knowledge for relatives (Brent et al., 2015) increasing the survival of their grandoffspring (Nattrass et al., 2019).
Although dispersal of both males and females is more pronounced within the Bigg's ecotype, the postreproductive females are never observed on their own, but always with either their son(s) or daughter(s) (Towers et al., 2019). Further, the smaller matrilines may consist of up to three generations of maternally related kin and are regularly observed associating with individuals outside of the maternal group (Ford & Ellis, 1999;Towers et al., 2019). These patterns of association provide opportunities for social interactions that can lead to inclusive fitness benefits for older females (Brent et al., 2015;Towers et al., 2018). whales (Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017), which could be a result of difference in ecology between the two species. Moreover, there are several examples of primate species with female-biased dispersal and local mating (e.g., chimpanzee, bonobos, and gorillas (Eriksson et al., 2006;Nishida et al., 2003;Stokes et al., 2003)), and thus an increase in female local relatedness with age similar to humans, where a prolonged postreproductive life span has not evolved (Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017).
It is likely that we may find similarly prolonged postreproductive life spans in other killer whale populations or other mammal species as we gather more individual-based data on long-lived social mammals.
Some evidence already suggests that false killer whales (Pseudorca crassidens) (Photopoulou et al., 2017) and Asian elephants (Elephas maximus) have a long period as postreproductive, although for Asian elephants it is likely a social rather than physiological trait (Chapman et al., 2019).  (Ellis, Franks, Nattrass, Cant, Bradley, et al., 2017)). Nonadaptive mechanisms associated with physiological constraints on the reproductive life span have been proposed as a driver for the evolution of menopause in mammals (Croft et al., 2015;Tully & Lambert, 2011;Wood et al., 2000). However, other long-lived mammals, such as African elephants and baleen whales, breed for their entire life span (Mizroch, 1981;Moss, 2001) and accumulating evidence supports that adaptive benefits related to the postreproductive period for females in humans and killer whales are important drivers for this life history trait (Foster et al., 2012;Lahdenperä et al., 2004;Nattrass et al., 2019). It is possible that menopause cannot easily be reversed once evolved which may help explain the universality in the timing of menopause in humans, and likely also killer whales, despite the evident differences among societies, such as patterns of dispersal (Snopkowski et al., 2014).

| CONCLUSION
In conclusion, when taken together with previous work, our findings support the hypothesis that kinship dynamics play a key role in the evolution of a prolonged postreproductive life span in killer whales.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section.