Sexual dimorphism in feeding and trenching (when alone)
Male and female beetles fed on milkweed leaves and trenched latex canals at approximately equal rates. At the first census, 1 hour after initiating the experiment, 40% of beetles had fed and 26% had trenched the leaf. Females and males showed nearly identical probabilities of each behavior (n = 42, Fisher’s Exact test, p = 1 for each behavior). By 24 hours after initiating the trial, 93% had fed and 60% had trenched, but again there was no difference between the sexes (n = 42, Fisher’s Exact test, p = 1 for each behavior).
Size differences between the sexes did not appear to affect the probability of trenching and feeding, although males and females were size dimorphic (MANOVA including length, width, mandible width, and dry mass: exact F4,36 =5.06, p = 0.002). All measured size components were significantly different themselves: females weighed 35% more, were 10% longer, 5% wider, and had mandibles that were 11% wider than males (all Ps < 0.01). Using PCA, a single principal component explained 72% of the variation in our four size measures. Neither this PC nor mandible width was predictive of feeding or trenching in the initial census (n = 42, logistic regression, PC1: feeding: χ2 = 0.65, p = 0.42; trenching χ2 = 0.03, p = 0.86; mandible width: feeding χ2 = 1.38, p = 0.24; trenching χ2 = 0.46, p = 0.50) or after 24 hours (PC1: feeding χ2 = 0.18, p = 0.68; trenching χ2 = 0.84, p = 0.36; mandible width: feeding χ2 = 0, p = 0.97; trenching χ2 = 2.12, p = 0.15).
We used the setup of additional experiments (Experiment 3, Trial 2, and Experiment 4, described above) to address the same question of sexual dimorphism in trenching. In Experiment 3, Trial 2, after 24 hours, beetles made substantially more trench marks than in the previous trial (range 0–25, mean 8.6, compared to a mean of 1.2 trench marks in trial 1), and females made 65% more trenches than males (females 10.68 ± 1.70 males 6.48 ± 1.19; F1,48=4.08, p = 0.049). We did not measure beetle size in this trial. For Experiment 4, male and female beetles showed no difference in number of trench marks (F1,98=0.56, p = 0.455), although larger beetles made fewer trenches (size PC1 F1,98=7.03, p = 0.009), and sexes were again dimorphic in size (MANOVA exact F3,99=21.41, p < 0.001). Nonetheless, in trial 4 females did puncture 55% more veins after 24 hours compared to males (mean ± SE, females 4.21 ± 3.45, males 2.72 ± 0.44; F1.98=4.39, p = 0.039). Thus, although evidence for size-related dimorphism is abundant in this system, we only detected marginal differences in trenching between the sexes.
The role of sexual dimorphism and leaf damage in attracting conspecifics
In a natural population, our 135 treated stems were colonized by 68 female and 68 male T. tetrophthalmus across all census times. We found significantly higher average colonization rate on stems with a caged female beetle than on stems with a male beetle or on controls (ZI Poisson glm, Wald χ2 = 10.02, p = 0.007; pairwise comparisons, control-female p = 0.085, control-male p = 0.12, female-male p = 0.003) (Fig. 2). These pairwise comparisons were similar when we separately considered male vs. female colonizers, as the only significant difference was that females attracted more beetles than males.