Male weaponry in a fighting cricket

Abstract

Sexually selected male weaponry is widespread in nature. Despite being model systems for the study of male aggression in Western science and for cricket fights in Chinese culture, field crickets (Orthoptera, Gryllidae, Gryllinae) are not known to possess sexually dimorphic weaponry. In a wild population of the fall field cricket, Gryllus pennsylvanicus, we report sexual dimorphism in head size as well as the size of mouthparts, both of which are used when aggressive contests between males escalate to physical combat. Male G. pennsylvanicus have larger heads, maxillae and mandibles than females when controlling for pronotum length. We conducted two experiments to test the hypothesis that relatively larger weaponry conveys an advantage to males in aggressive contests. Pairs of males were selected for differences in head size and consequently were different in the size of maxillae and mandibles. In the first experiment, males were closely matched for body size (pronotum length), and in the second, they were matched for body mass. Males with proportionately larger weaponry won more fights and increasing differences in weaponry size between males increased the fighting success of the male with the larger weaponry. This was particularly true when contests escalated to grappling, the most intense level of aggression. However, neither contest duration nor intensity was related to weaponry size as predicted by models of contest settlement. These results are the first evidence that the size of the head capsule and mouthparts are under positive selection via male-male competition in field crickets, and validate 800-year-old Chinese traditional knowledge.

Figure 1. A male cricket showing the: a) dorsal view of the head capsule and pronotum, b) ventral view of the head capsule and pronotum (head has been tilted dorsally to expose ventral mouthparts), c) ventral view of the right and left maxillae, d) ventral view of the right and left mandibles, and e) left hind leg.

All drawings are at the same scale. Morphological measurements are shown as dotted lines and abbreviations are as follows: HW = head width, PL = pronotum length, MxS = maxillae span, PW = pronotum width, MxL = maxilla length, MdL = mandible length, and FL = femur length. The left maxilla, excluding the palp, is shaded grey in b) and labeled with abbreviations indicating the various parts in both b) and c): la = lacinia, g = galea, s = stipes, ca = cardo, and ms = maxadentes (after [81]). Note that the second maxadentes on the right maxilla shows an abnormality in being slightly shorter and blunted. Drawings by Janice J. Ting.

Figure 2. Scatterplot of log mean maxilla length versus log pronotum length showing the sexual dimorphism in mean maxilla length in a sample of 151 males (X) and 75 females (O).

Drawings are of a representative male (upper left) and female (lower right) showing the dorsal view as in Fig. 1a as well as the posterior view of the ventral surface of the head (left maxilla is shaded grey as in Fig. 1b). Arrows point to the individuals whose measurements are depicted. Drawings by Janice J. Ting.

Figure 3. Probability that the focal male would win a fight versus the difference in PC1 between the focal male and his rival for: a) experiment 1, and b) experiment 2.

Positive values of PC1 difference mean that the focal male had larger weaponry than his rival, whereas negative values indicate that the focal male had smaller weaponry than his rival. Observed values are represented by pluses (+) and the solid lines are plots of the logistic models.

Figure 4. Probability that the focal male would win a fight versus the difference in PC1 between the focal male and his rival for contests that: a) did not escalate to grappling, and b) escalated to grappling.

Positive values of PC1 difference mean that the focal male had larger weaponry than his rival, whereas negative values indicate that the focal male had smaller weaponry than his rival. Observed values are represented by pluses (+) and the solid lines are plots of the logistic models.