Function predicts the allometry of contest-related traits, but not sexual or male dimorphism in the Amazonian tusked harvestman
Description
Abstract: The functional allometry hypothesis proposes that 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. Methods: We measured 9125 adult males and 100 adult females. We photographed all individuals in dorsal and lateral view using a digital camera (CANON T6i). Then, we used the software ImageJ (Schneider et al. 2012) to measure the following traits: (1) dorsal scute (i.e., carapace) length; (2) interocular distance; (3) femur length of all four right legs (hereafter ‘legs I to IV’) and right pedipalp; (4) length of the apophysis on the coxa of the right leg IV (hereafter ‘coxa IV apophysis’); and (5) tusk length. We measured both the right and left tusks and used the measure of the longest one because tusks can be damaged at the tips, and we wanted to ensure we were analyzing the original full size of the structure. We investigated male dimorphism in all traits we measured. To do so, we first inspected the distributions for bimodality using nonparametric kernel density curves. Second, we parameterized the distributions as mixtures of two skew-normal distributions or as a single skew-normal distribution with the package mixsmsn (Prates et al. 2013) for R version 3.6.1 (R Core Team 2019). We compared the fit of one distribution to the fit of a mixture of two distributions via their bias-corrected Akaike information criterion (AICc). Finally, if the best fit was a mixture of two distributions, we assigned individuals to either major or minor morph based on the probability of these males belonging to either morph. Some males, however, did not belong to either morph (based on an 80% probability threshold calculated by the mixture model). After assigning males to their morphs, we calculated the static allometry of each trait (regressed on dorsal scute length as a proxy for body size) for majors and minors using standardized major axis regressions fit with the R package lmodel2 (Legendre 2018). We compared the slopes between morphs using likelihood ratio tests implemented in the sma() function of the R package smatr (Warton et al. 2012). Because we are using female traits as controls, we also tested for bimodality in their dorsal scute length, pedipalp length, length of legs I to IV, and length of coxa IV apophysis. We also calculated the allometric slopes of these female traits using the same procedure described above and compared the slopes between males (or male morphs separately, if the trait was male-dimorphic) and females, also as described above.