Human uniqueness illustrated by life history diversity among small-scale societies and chimpanzees

Background We compare life histories and selection forces among chimpanzees and human subsistence societies in order to identify the age-specific vital rates that best explain fitness variation, selection pressures and species divergence. Methods We employ Life Table Response Experiments that quantify vital rate contributions to population growth rate differences. Although widespread in ecology, these methods have not been applied to human populations or to look at species differences among humans and chimpanzees. We also estimate correlations between vital rate elasticities and life history traits to investigate differences in selection pressures and test predictions of life history theory. Results Chimpanzees’ earlier maturity and higher adult mortality drive species differences, whereas infant mortality and fertility variation drive differences among humans. Human fitness is decoupled from longevity by postreproductive survival, while chimpanzees forfeit higher potential lifetime fertility due to adult mortality attrition. Infant survival is often lower among humans, but lost fitness is recouped via short birth spacing and high peak fertility, thereby reducing selection on infant survival. Lastly, longevity and delayed maturity reduce selection on child survival, but among humans, recruitment selection is unexpectedly highest in longer-lived populations, which are also faster-growing due to high fertility. Conclusion Humans differ from chimpanzees more because of delayed maturity and adult mortality than child mortality or fertility rates. In both species, high child mortality reflects bet-hedging costs of quality/quantity tradeoffs borne by offspring, with high and variable child mortality likely regulating human population growth over evolutionary history. Among human subsistence societies, positive correlations between survival and natural fertility lead selection pressures in human subsistence societies to differ from modern populations undergoing demographic transition, due in part to positive correlations between longevity and natural fertility and negative correlations between recruitment elasticity and reproductive effort.


32
Humans and chimpanzees, whose recent common ancestor dates to 4-8 million years ago [1,2], 33 share behavioral adaptations and life history traits that distinguish them among other primates [3, hunter-gatherers are closer to chimpanzees than they are to today's lowest-mortality 37 industrialized populations [7]. Nevertheless, there is great variation among human and 38 chimpanzee life histories [8]. Here, we characterize human uniqueness by identifying the vital 39 rates that drive life history differences within-and between-species. We interpret population life We identify the vital rates that are most important in driving population growth and 71 decline using fixed LTREs [15], which decompose individual vital rate contributions to 72 observed differences in population growth rates. We compare these results with the more 73 widely-used elasticity analyses that prospectively estimate the potential fitness effects of vital 74 rates [15,19]. Differences between realized fitness contributions and the potential suggested by 75 elasticities may indicate constraints on life history evolution. 76 In this context, we evaluate three predictions based on fitness elasticities: (P1) juvenile

81
P1 relies on the high elasticities of infant and child survival, which are larger than 82 elasticities to adult survival or fertility, to predict that recruitment will be most important for 83 population fitness differences [20]. However, if stabilizing selection reduces variation in 84 important vital rates [22], fitness contributions of high-elasticity rates are likely to be small [23]. 85 More generally, when elasticities overestimate fitness effects this may reflect constraints on 86 stabilizing selection that would otherwise reduce variation in these important vital rates, whereas 87 underestimation implies that vital rate differences are more important than a priori predictions 88 from the force of selection.

89
P2 is the intuitive prediction that population growth should reflect both survival and 90 reproduction, since either will increase population growth, all else equal. However, longevity and 91 fertility may trade off [24][25][26]. For instance, greater life expectancy is associated with lower 92 fertility across modern industrial nations [27], driving a negative correlation between life 93 expectancy and population growth [28]. Therefore, the degree (and even the sign) of the 6 94 correlations of population growth with fertility vs. longevity are empirical questions that we 95 answer in the case of natural fertility subsistence populations and chimpanzees.

96
P3 arises as a consequence of the slow life history of primates [29]. When infant 97 mortality is low, more survive to maturity, thereby reducing selection on recruitment. A longer 98 lifespan also permits replacement of dead offspring with new births, while low fertility raises the 99 average age of a population. Because all of these effects make newborn survival less important to 100 population fitness, elasticity to child survival is predicted to correlate negatively with life 101 expectancy but positively with fertility (P3, [21]). We extend this logic to predict that 102 recruitment elasticity should also correlate positively with the pace of fertility and thus 103 negatively with mean age at first birth (AFB), mean age of childbearing (MAC) and inter-birth 104 intervals (IBI) since smaller values increase fertility, but positively with age at last birth (ALB).

106
Demographic Data 107 We examine fertility and mortality rates published for ten contemporary, non-industrial 108 small-scale societies with natural fertility and minimal to no access to modern medicine during 109 the period corresponding to the demographic data (S1, S2 Tables; S1 File contains ethnographic but are used to reflect "best-case" scenarios for chimpanzees: low mortality in the protected and 119 provisioned Gambia population and high fertility in the captive breeding program at Taronga 120 Zoo. Because fertility estimates for Ngogo chimpanzees are not published, we estimate 121 contributions applying fertility estimated at nearby Kanyawara. Also, because the Taronga Zoo 122 mortality data includes few chimpanzee deaths we use mortality data averaged across three zoo 123 populations [30]. 124 We employ parametric models of mortality and non-parametric models of fertility to 125 obtain smoothed annual rates (see S1 File for details is updated by applying the population projection matrix A to the population age structure n (n = 8 143 {n i }) and stable asymptotic population growth is described by the dominant eigenvalue λ 144 (n(t+1) = A n(t) = λ n(t)). From the matrix A we calculate vital rate sensitivities (s ij ) reflecting 145 the force of selection on a vital rate as well as elasticities (e ij ) scaling the proportional effect on 146 population growth (e ij = (∂ λ / ∂ a ij ) = s ij (λ / a ij ) [15,19]. Because elasticities conveniently sum 147 to unity (1 = Σ i,j e ij ), we can add elasticities across vital rates across age x to estimate the total 148 elasticity to survival (E s = Σ x e x+1,x ) or to fertility (E f = Σ x e 1x ; 1 = E s + E f ), or sum across 149 specific ages (e.g., before or after reproductive maturity at age α) to distinguish the elasticity to   that sum to the total difference in population growth rates (Δλ = Σ i,j C ij ), we also examine 163 combined effects (C ij* = |C ij | / Σ i,j |C ij |), which are analogous to elasticities in that they sum to 164 unity (1 = Σ i,j C ij* ) and which reflect the relative effect of fitness contributions (e.g., survival 165 across childhood vs. adulthood) (C c = Σ x<α C x+1,x* ; C a = Σ x≥α C x+1,x* ; C s = Σ x C x+1,x* ; C f = 166 Σ x C 1x* ; 1 = C s + C f ). For comparison, fertility is binned into early, prime and late fertility 9 167 effects at the ages when completed fertility is 0-25%, 25-75% and 75-100% of TFR in the 168 hunter-gather reference (ages 0 to 22, 23 to 35, and 36 to 50, respectively). To aid interpretation, 169 we calculate standard demographic rates: mortality hazard (μ x ), survivorship (l x ), life expectancy 170 at birth (e 0 ), total fertility rate (TFR), mean age at first birth (AFB), mean age of childbearing 171 (MAC), mean age at last birth (ALB) and mean inter-birth intervals (IBI) (S2 Table; S1 File 172 contains calculations). 173 We also evaluate three predictions of population biology and life history theory (P1-P3). , we expect child survival effects to be smaller than elasticities predict. 177 We calculate a scalar ratio (Z ij ) that reflects the actual realized contributions of vital rates, 178 relative to the potential suggested by elasticities (Z ij = C ij* / e ij ). Because both vital rate effects 179 (C ij* ) and elasticities (e ij ) sum to unity, we can estimate Z across all of childhood (Z c = C c / E c ) or 180 across adulthood (Z a = C a / E a ) and for lifetime fertility (Z f = C f / E f ).    226 † Missing population data complemented with comparable rates (for Ngogo using fertility from 227 nearby Kanyawara, for Taronga using mortality averaged over three zoo populations)  to 2) in our sample is among the managed Gambia chimpanzees, and the lowest mortality  Table).  gathereres (p = 0.008). These chimpanzee IBIs are also shorter than the 5. birth spacing. Closer examination shows population differences in the tempo of fertility (Table   257 1, S3 Table). 1, S2 Table). Fertility elasticities may climb rapidly with age (e.g., Herero, Yanomamo and Taï 264 chimpanzees) or slowly (e.g., Ache, Hadza and Tsimane) depending on the pace of fertility, but 265 decrease at approximately the same rate as survival elasticities due to mortality attrition 266 affecting both simultaneously (Fig 1C,D).

Fitness Contributions
268 All ten small-scale societies and two wild chimpanzee populations are growing, but two 269 chimpanzee groups are declining slowly and one is collapsing (Fig 2; Table 1). However, due to 270 wide variation among our small sample, population growth differences are not statistically 271 significant (r = log λ, p > 0.1; S1 Table). Compared to the hunter-gatherer reference, lower population growth of non-foragers is mainly due to higher survival at all ages, but offset by 293 prime and late-age fertility, which are lower than the hunter-gatherer reference (Fig 3).   Table).

314
Because so few chimpanzees survive to advanced ages, large differences in the potential 315 for late-life fertility contribute little to population growth. Among humans, high survival drives 316 population growth among non-foragers; among hunter-gatherers, lower early fertility effects are 317 offset by higher prime-and late-age fertility (Fig 3A; S5, S7 Figs). Aborigines was due to high survival offsetting low fertility at all ages, whereas the Ache grew 335 faster due to high prime and late fertility. Very rapid growth was due to survival and early 336 fertility among the Yanomamo and due to survival and late fertility among the Tsimane.  were larger among non-foragers (C s < C f ; p = 0.008; S4 Table).  Table). there is no correlation between E 0 and TFR within or across species (p > 0.1; Table 2). Across  was an important driver of population-and species-level differences (33% of all effects across 394 human populations and 37% across chimpanzees), but adult survival was also an important 395 driver (14% of all effects across humans and 27% among chimpanzees; Table 1). However, As predicted (P2), population growth rates were positively correlated with longevity 436 across our two-species sample, but they were not correlated with fertility. Hadza life history is close to the hunter-gatherer composite reference, suggesting perhaps that 500 they may best represent the "typical" contemporary hunter-gatherer population. While the !Kung survived contact, our sample may over-represent growing populations especially since these 510 short-term data may have captured transient growth periods in population cycles with rapid 511 declines and prolonged recovery [14].

513
Since divergence from chimpanzee-like ancestors, human survival has increased so much that 514 even pre-industrial human and chimpanzee mortality profiles barely overlap. While species 515 differences in adult mortality have been widely recognized [17], we report additional species 516 differences and similarities: hunter-gatherers have similar, and sometimes higher, infant 517 mortality than chimpanzees, while fertility is much more variable across human societies and 518 overlaps the range of chimpanzees, especially across prime childbearing years. However, due to 519 high mortality attrition, the force of selection on chimpanzee fertility is much lower than for 520 humans and more strongly favors younger mothers.

521
Our findings suggest that the trajectory forward from the life history of our most recent 522 common ancestor with the chimpanzee was likely not a monotonic decline in mortality and that 523 high and variable infant mortality likely played a large role in regulating population growth over 524 evolutionary time. We also find that fertility differences have substantial effects on population 525 growth despite low elasticities, and that older individuals may contribute more to population-526 level fitness differences than younger individuals with higher reproductive values. The diverse 527 environments humans inhabit are partly responsible for observed variation in reproductive 528 success across populations, but quality-quantity tradeoffs between fertility and juvenile survival,