Facial asymmetry tracks genetic diversity among Gorilla subspecies

Abstract

Mountain gorillas are particularly inbred compared to other gorillas and even the most inbred human populations. As mountain gorilla skeletal material accumulated during the 1970s, researchers noted their pronounced facial asymmetry and hypothesized that it reflects a population-wide chewing side preference. However, asymmetry has also been linked to environmental and genetic stress in experimental models. Here, we examine facial asymmetry in 114 crania from three Gorilla subspecies using 3D geometric morphometrics. We measure fluctuating asymmetry (FA), defined as random deviations from perfect symmetry, and population-specific patterns of directional asymmetry (DA). Mountain gorillas, with a current population size of about 1000 individuals, have the highest degree of facial FA (explaining 17% of total facial shape variation), followed by Grauer gorillas (9%) and western lowland gorillas (6%), despite the latter experiencing the greatest ecological and dietary variability. DA, while significant in all three taxa, explains relatively less shape variation than FA does. Facial asymmetry correlates neither with tooth wear asymmetry nor increases with age in a mountain gorilla subsample, undermining the hypothesis that facial asymmetry is driven by chewing side preference. An examination of temporal trends shows that stress-induced developmental instability has increased over the last 100 years in these endangered apes.

Keywords: asymmetry; geometric morphometrics; great apes; inbreeding; stress.

Figure 1.

An extreme example of facial asymmetry in a female Virunga mountain gorilla cranium (Tayna, individual GP.148), shown as a three-dimensional surface model with texture. This individual was not included in the sample because she was dentally immature at the time of death, but she exhibits an extreme version of the asymmetric pattern documented in this study. (Online version in colour.)

Figure 2.

The approximate ranges of the four Gorilla subspecies. The three subspecies analysed in this study are labelled (western lowland gorillas, Grauer gorillas and mountain gorillas), while cross river gorillas (G. g. diehli) are shown in red at the top left corner of the map. Major river boundaries (blue), elevation grades (grey) and country borders (white) are also shown. Modified from Tocheri et al. [15]. (Online version in colour.)

Figure 3.

Facial shape variation in gorillas. (a) Principal component analysis (PCA) of the symmetric component of facial shape variation. Three-dimensional surface models of the mean shape warped along the positive and negative PC1 and PC2 axes. (b) PCA of the asymmetric component, with three-dimensional models of the mean shape warped along the positive and negative axes. Two models show distance-based heatmaps of the intensity of asymmetry across the face with red indicating more asymmetry and blue indicating less asymmetry. In the PCA, specific individuals are circled and shown in panel (c), which depicts actual surface models of the most asymmetric individuals. Arrows denote directionality of asymmetry in the lower midface. Green and red points represent fixed and sliding semilandmarks used in this study, respectively. (Online version in colour.)

Figure 4.

(a) Regressions of unsided tooth wear asymmetry in lower first molars (LM1, n = 20) (r_s = –0.19, p = 0.422) and upper first molars (UM1, n = 21) (r_s = –0.43, p = 0.054) versus facial asymmetry scores in the Virunga mountain gorilla subsample. Shading shows 95% confidence intervals. (b) Density plots of tooth wear asymmetry of LM1 (n = 41) and facial asymmetry scores (n = 22) in Virunga mountain gorillas. At left, positive values indicate greater LM1 wear on the right side, while negative values indicate greater LM1 wear on the left side. At right, larger numbers denote greater facial asymmetry overall. (c) Facial asymmetry scores of individual western lowland, Grauer and mountain gorillas (n = 114). Facial asymmetry increases through time when controlling for sex and species differences in the magnitude of asymmetry (F_1,107 = 4.95, p = 0.028). Grey shading shows 95% confidence intervals. (Online version in colour.)