Function predicts the allometry of contest-related traits, but not sexual or male dimorphism in the amazonian tusked harvestman

Animal contests involve threatening displays and physical coercion, which are respectively performed by threat devices used in mutual evaluation of size or strength, and weapons used for grasping, stabbing, striking, or dislodging a rival. According to the functional allometry hypothesis, directional selection consistently favors hyper-allometry in threat devices, whereas the allometry of weapons depends on the way they are used in contests. Here, we tested this hypothesis using the Amazonian tusked harvestman Phareicranaus manauara (Arachnida: Opiliones), a male-dimorphic species, as a study system. Behavioral observations allowed us to recognize four contest-related traits and three control traits, not used in contests. Two weapons used to grasp or prod the opponents from afar and one threat device were hyper-allometric, whereas one tactile signaling device (used to tap the opponent) and all control traits were either iso- or hypo-allometric. These findings support the hypothesis that function predicts the allometry of contest-related traits. However, function does not explain allometric differences in homologous traits between males and females (whose traits also were used as controls). We suggest that if a trait used in contests by males is used by both sexes in another context, natural selection and cross-sexual genetic correlations may constrain its developmental trajectory, preventing the evolution of sexual dimorphism in allometric slopes. Therefore, using female traits as controls for homologous contest-related male traits may not be appropriate. Finally, we show that function does not explain differences between male morphs in the allometric slopes of male-dimorphic traits. Thus, an important next step in allometric studies is to understand what factors affect the slopes of male-dimorphic traits.