Integration, Modularity, and Homoplasy in the Forelimbs of Apes

Abstract Covariation constrains and biases the evolution of morphological traits, leading to similar phenotypes appearing repeatedly in certain clades. Here, I test whether this phenomenon can explain the evolution of locomotor adaptations in the forelimbs of apes, which share a suite of traits thought to have evolved in parallel despite the locomotor diversity of extant apes. Three hypotheses have been proposed: (1) apes have a shared, derived pattern of covariance that biases their evolution along evolutionary lines of least resistance; (2) apes have greater autonomy and increased modularity that improves their ability to evolve into novel areas of morphospace; and (3) the large number of potentially homoplastic traits in apes are really a smaller, more parsimonious number of trait complexes. I find that apes have higher evolvability and respondability but lower autonomy and flexibility than do monkeys. This is true both along hypothetical evolutionary trajectories and when averaged across random selection vectors. Evolvability along trajectories is lower than average evolvability in apes, indicating that these trajectories are not evolutionary lines of least resistance. Apes are generally best represented by models of modularity that include more modules than models that best represent monkeys, but several models of modularity receive comparable support across all taxa. Overall, all three hypotheses receive some support but the predictions of none are fully met. Instead, these results may indicate a partial breakdown and realignment of the P matrix in apes in response to directional selection, demonstrating that the relationship between integration, selection, and evolution is complex and nonlinear.