The cost-to-benefit ratio of group-living is thought to vary with group size: 24 individuals in "optimal" groups should have higher fitness than individuals in groups that 25 are too large or small. However, the relationship between group size and individual 26 fitness has been difficult to establish, a gap we address here in the gelada. We 27 demonstrate group size effects on the production of surviving offspring and on female 28 mortality rates, which are largely explained by group-size variation in infanticide risk and 29 foraging competition. We also identify a mechanism by which females may alter group 30 size: in large groups, females groomed with less than half of their group, increasing the 31 likelihood of fissions. Our findings provide insight into how and why group size shapes 32 fitness in long-lived species. 33 34 survival 36 37 38 39 3Main Text: 40 Variability in group size within species reflects a delicate balance of the costs and 41 benefits of group living. For example, in large groups the costs of group-living (e.g., 42 foraging competition: 1-3) may begin to outweigh the benefits (e.g., decreased 43 predation risk : 1, 4, 5), to the detriment of individual group members. Individuals in 44 "optimal" groups, by contrast, should have the highest lifetime reproductive success 45 (i.e., fitness: 6) compared to others in the population (7). Although the link between 46 group size and fitness has long been suspected, it has been challenging to actually 47 demonstrate with empirical data. This is, in part, because fitness in long-lived species is 48 difficult to measure. Instead, short-term measures of reproductive performance (e.g., 49 birth rate) are often used as proxies for fitness. But, perhaps more problematic is that 50 the accuracy of measures of fitness (which are more accurate over a longer period of 51 time) are inversely related to the accuracy of measures of group size (which are more 52 accurate over a shorter period of time). To circumvent this problem, we developed a 53 novel measure of reproductive performance that accounts for changes in group size. 54 55Here, we used a wild primate, geladas (Theropithecus gelada), as a test species to 56 identify whether we have evidence for an optimal group size for females and, if so, 57 whether we can identify the selective forces that shape it. Geladas are long-lived, 58 folivorous primates that live in social groups (hereafter, "units") that vary in size from 1 59 to 13 adult femalesvariation that is present even within the same population (8). This 60 variability may be due to ecological pressures (e.g., due to the seasonal environment: 9) 61 as well as social pressures. For instance, infanticide is the leading cause of infant 62 4 mortality: immigrant males kill over half of the dependent infants in a unit (10). 63 Moreover, larger units experience more takeovers than smaller units (11). Females in 64 larger units might therefore incur disproportionately higher costs than females in smaller 65 units.
67We analyzed 9...