Sexually divergent selection, allometric constraints, and the evolution of sexual dimorphism in cichlids from Lake Tanganyika

The evolution of sexual dimorphism is widely acknowledged as manifestation of sex-specific genetic architecture. Although empirical studies suggested that sexual dimorphism evolves as a joint consequence of constraints arising from the genetic architecture and sexually divergent selection, it remains unclear whether and how these established microevolutionary processes scale up to the macroevolutionary patterns of sexual dimorphism among taxa. Here, we studied how sexual selection and parental care drive sexual dimorphism in cichlid fishes from Lake Tanganyika. We found that male-male competition, female choice, and maternal mouthbrooding are associated with sexual dimorphism in body length, body color, and head length, respectively, despite strong allometric relationships between body length and head length. Within-species (static) allometry of head length on body length evolved as sex-specific responses to mouthbrooding where females evolved higher intercepts while males evolved steeper slopes. Thus, selection to increase mouth size in mouthbrooders may have broken down and reorganized the pattern of allometric constraints that are inherently strong and concordant between sexes. Furthermore, sex-specific responses to mouthbrooding left a remarkably clear signature on the macroevolutionary pattern, resulting in a decoupling of co-evolution in parameters of static allometries observed exclusively within maternal mouthbrooders. Our study provides multiple lines of evidence that are consistent with the idea that macroevolutionary patterns of sexual dimorphism in Lake Tanganyika cichlids result from sexually divergent selection. Our approach illustrates that an examination of within-population phenotypic variance in the phylogenetic comparative framework may facilitate nuanced understandings of how macroevolutionary patterns are generated by underlying microevolutionary processes.

Methods
In total, 1289 samples from 39 species were used in this study, out of which 1056 samples from 33 species were photographed. All photos were resized to a uniform scale. Measurements of standard length and head length, as specified in the Supplementary Data, were then obtained either from the photographs, where available, or directly from the fish samples.