The adaptive value of facial expressions has been debated in evolutionary biology ever since Darwin’s seminal work. Amongst mammals, primates, including humans, exhibit the most intricate facial displays. Although previous work has focused on the role of sociality in the evolution of primate facial expressions, this relationship has not been verified in a wide sample of species. Here, we examine the relationship between allomaternal care (paternal or alloparental) and the morphology of three orofacial brainstem nuclei (facial; trigeminal motor; hypoglossal) across primates to test the hypothesis that allomaternal care explains variation in the complexity of facial expressions, proxied by relative facial nucleus size and neuropil fraction. The latter represents the proportion of synaptically dense tissue and may, therefore, correlate with dexterity. We find that alloparental care frequency predicts relative neuropil fraction of the facial nucleus, even after controlling for social system organization, while allomaternal care is not associated with the trigeminal motor or hypoglossal nuclei. Overall, this work suggests that alloparenting requires increased facial dexterity to facilitate nonverbal communication between infants and their non-parent caregivers and/or between caregivers. Accordingly, alloparenting and complex facial expressions are likely to have co-evolved in primates. 

We collected: 1) volumes for the medulla and three brainstem nuclei (trigeminal motor: Vmo; facial: VII; hypoglossal: XII); and 2) the neuropil fraction for the same brainstem nuclei for 46 nonhuman primate species from Sherwood et al. (Sherwood et al., 2005). The neuropil fraction represents the proportion of an area that is not occupied by the projection profiles of all stained elements of tissue containing neuronal cell bodies and glial/endothelial nuclei. The data in Sherwood et al.  (Sherwood et al., 2005) were originally reported as Grey Level Index (GLI), which is the inverse of neuropil fraction. We converted GLI values by subtracting them from 100. Volumes were from the left side only, so values were doubled. 

We collected continuous measures of paternal and alloparental care from Heldstab et al. (Heldstab et al., 2019), which supplemented the databases of Isler & van Shaik (Isler & van Schaik, 2012) and Heldstab et al. (Heldstab et al., 2017).

We also collected categorical data on activity periods (i.e., the period(s) of the day during which an animal is most active; categories = diurnal (daytime), nocturnal (nighttime), cathemeral (daytime and nighttime)) and social system organization (i.e., the composition and grouping patterns of individuals; categories = group-living, pair-living, solitary; this categorization scheme represents the three fundamental types of social organization (Kappeler, 2019)) from published literature sources (DeCasien & Higham, 2019).