Dunbar addresses three critical questions about grouping: what is a large vs a small group; what glues individuals together in groups; and what are the cognitive consequences of grouping. The vexed problem of “what is a group” is disregarded here since he limits his analysis to primates where groups can generally be defined and bounded, but addressing that question is still fundamentally important. As a non-primate, you don’t need glue (grooming) if you school or herd; you also don’t need complex recognition and inhibition mechanisms. However, fish (Kohda et al., Reference Kohda, Hotta, Takeyama, Awata, Tanaka, Asai and Jordan2019), birds (Emery et al., Reference Emery, Seed, Von Bayern and Clayton2007), and possibly even gregarious wildebeest (Saltz, Palmer, & Beaudrot, Reference Saltz, Palmer and Beaudrot2023) do have decision capacities in the absence of “stable” groups. Excluding olfactory and vocal communication as glue mechanisms also excludes many non-primate but social species. So the question of groupingness (e.g. Moffett, Reference Moffett2024) remains to be debated and determined.

Perhaps what makes primates distinctive is the lack of much interesting variation in their groupings, since they follow clear fractal constraints. Dunbar notes the chaotic social and grouping dynamics of some primates; specifically squirrel monkeys, talapoins and mandrills. Could these species provide more insight into what makes groups evolutionarily interesting than the “boring” stable baboons and macaques, with their predictable rank and kin-determined cliques (Kappeler, Reference Kappeler2019)? And if unpredictability underlies complex cognitive responses (Pellis, Pellis, & Himmler, Reference Pellis, Pellis and Himmler2014), then dynamic groupings should be particularly informative.

Thus, in order to understand the evolution of grouping, it might benefit us to focus on those very species that have dynamic structures; to try to make sense of the chaos in flexible groupings as evolved responses. What Dunbar calls the social infertility trap, the impact of social stress on female reproductive physiology, may be counterbalanced by shared infant care resulting in improved infant survival (Lee et al., Reference Lee, Fishlock, Webber and Moss2016). The push away from gregariousness due to stress and competition is countered by the pull of greater reproductive success, beyond that of minimizing predation risk as proposed by Dunbar. Interestingly, primates use male infant care more than that from females, while shared female care may be more prevalent among carnivores, rodents and ungulates (Isler & van Schaik, Reference Isler and van Schaik2012). Is this difference a function of cognitive capacity or of the energy costs of infant production, which are themselves linked since bigger brained babies cost mothers more in terms of both energy and time. A broader interspecific analysis may provide more answers to this question.

Species that have fission-fusion dynamics, where association and interaction decisions are made on a shifting temporal basis (Aureli et al., Reference Aureli, Schaffner, Boesch, Bearder, Call, Chapman and Schaik2008), within a stable unit (e.g. chimpanzees, hyenas, lions) are amenable to fractal analysis; species with hypervariable and unpredictable grouping dynamics (dolphins, elephants and mandrills) may not be. Despite this constraint, social network analysis has provided clear insights into the complex structuring of sociality for some of these species (Sueur & Maire, Reference Sueur and Maire2014; Weiss et al., Reference Weiss, Franks, Croft and Whitehead2019), and these studies have emphasized the importance of feedback between maintaining relationships and physiological needs, affecting network stability. In addition, Strier, Lee, and Ives (Reference Strier, Lee and Ives2014) argued that inbreeding avoidance expressed through dispersal regimes underlies thresholds for fission-fusion dynamics compared with permanent group fissions, with fission-fusion more likely under female-biased dispersal with male philopatry and permanent splits more likely under male-biased dispersal with female philopatry. Limiting analyses of the drivers of association to competition and fertility suppression may obscure other important patterns determining grouping.

Another issue with the fractal analysis relates to the finding that the highest variance occurs at optimum group sizes; an instantaneous observation of a species’ group size may or may not represent that species’ characteristic group size or the historical demographic or social processes that resulted in that size in a population (Strier, Reference Strier2017). Long-lived species such as primates are likely to experience environmental and demographic dynamics over a lifetime that influence their group size and composition and the observed behavioural variation at any specific time, as well as producing differing responses to similar stimuli (Strier, Reference Strier2017, p 8). These raise questions both about data quality from short versus long-term studies and more critically about shifting baselines – what was a historical group size, optimum or not, may never be known given anthropogenic alterations of primate habitats. Nor do the size metrics used here take into account the propensity for multi-species grouping seen in a number of primate and other species. However, ultimately the modelling of single species and multi-species networks should provide us with greater insights into what group sizes can or, under optimal conditions, should be, and Dunbar’s analysis provides us with a way to explore these questions.

Dunbar’s findings also have important implications for our understanding of primate population dynamics by providing a general framework against which we can assess how environmental processes impose group size deflections from a “norm”. Deflections can include both long term climate change (e.g. aridification and tree loss for the Amboseli baboons: Altmann, Hausfater, & Altmann, Reference Altmann, Hausfater and Altmann1985), immediate environmental catastrophes such hurricanes affecting Macaca populations (Testard et al., Reference Testard, Larson, Watowich, Kaplinsky, Bernau, Faulder and Brent2021), disease epidemics experienced by a variety of primate populations (Rouquet et al., Reference Rouquet, Froment, Bermejo, Kilbourn, Karesh, Reed and Leroy2005; Holzmann et al., Reference Holzmann, Agostini, Areta, Ferreyra, Beldomenico and Di Bitetti2010; Strier, Reference Strier2021), or abrupt regime shifts in mortality as seen in Brachyteles (Strier & Ives, Reference Strier and Ives2024). More generally, the findings imply, in terms of population resilience, that populations whose group sizes become synchronised by natural catastrophes at the low end of their distribution will be much more at risk of extinction than those with a wider range of group sizes simply because their fertility is so depressed. This is a testable hypothesis when the relative impact of fertility and mortality on group (and population) size can be assessed. Taking this association into account in our studies of primate populations could be of vital importance in assessing their ability for persistence in a changing world.