Coevolution between harmful male genitalia and female resistance in seed beetles

Rönn et al. 10.1073/pnas.0701170104.

Supporting Information

Files in this Data Supplement:

SI Figure 3
SI Text
SI Table 2
SI Table 3
SI Figure 4




SI Figure 3

Fig. 3. The amount of harm to the female copulatory duct increases with the number of matings in seed beetles. The most obvious sign of this are melanized scars in the wall of the female bursa copulatrix of mated females, in the area where male spines reside during copulation. This figure shows the bursa copulatrix of three representative C. maculatus females, that were virgin (A), mated once only (B), and mated multiply (C). Arrows indicate location of scarred tissue.

SI Figure 4

Fig. 4. The phylogenetic reconstruction used in the comparative analyses reported represents a well resolved phylogeny based on Bayesian analyses of a total of 1,651 bp of sequence data from three different mitochondrial genes (1). Numbers by nodes represent Bayesian posterior probabilities.

1. Tuda M, Rönn J, Buranapanichpan S, Wasano N, Arnqvist G (2006) Mol Ecol 15:3541-3551.

Table 2. Trait values for all species

Table 3. Univariate and multiple conventional regressions of the effects of harmfulness of male genitalia and the amount of connective tissue in the copulatory tract of females across all species

SI Text

Species-Level Analyses

Raw data for all species are given in SI Table 2. Our analyses revealed a sizeable phylogenetic signal (see main text), and inferential models taking the phylogenetic relationship between the taxa into account are thus most reliable. Nevertheless, we also analyzed species-level data, ignoring shared ancestry of the taxa included, using conventional regressions for three reasons. This analysis allows assessment of (i) the extent to which our conclusions depend on the particular phylogenetic reconstruction used (1), (ii) whether the somewhat restricted number of taxa used represents an inferential problem (2), and (iii) whether collinearity between male persistence and female resistance renders estimation of their independent effects difficult (3).

The results of the species-level analysis were basically congruent with the PGLS models (see SI Table 3), showing that our main conclusions do not rely on a particular phylogenetic hypothesis.

Because the number of taxa included in our study was somewhat restricted, all models were assessed both with conventional tests of significance (based on the t distribution) and with more robust randomization tests [based on 1,000 random complete permutations of data (4)]. The harmfulness of male genitalia was significantly related to the amount of connective tissue in the female reproductive tract across species (b = 0.115, P_{t, dir} = 0.011, P_{rand, dir} = 0.015; see Fig. 2) and randomization tests of all regressions models closely matched conventional t tests (SI Table 3). We note that these randomization tests are also less sensitive to any inferential problems that may occur as a result of collinearity (4).

Although the low condition index of the multiple regression models in SI Table 3 (ci = 9.7) strongly suggests that collinearity is not a problem (3), we further assessed this by using principal component analysis. We first converted the relationship between harmfulness of male genitalia and connective tissue in females to a first (PC1) and second (PC2) principal component. Here, PC1 represents an axis describing the covariance between harmfulness of male genitalia and connective tissue in females, and PC2 is orthogonal to this axis and thus parameterizes the imbalance between male and female adaptations that is of interest here. We then performed multiple regressions of species-level data, using the uncorrelated PC1 and PC2 as independent variables. In complete congruence with our trait-based multiple regressions, these models showed that imbalance between male and female adaptations (i.e., the focal independent variable PC2) was significantly related to the cost of mating (b = -0.192, t = 3.017, P_dir = 0.024), but not to female lifetime offspring production (b = 11.59, t = 0.885, P_dir = 0.266) or scarring of the female reproductive tract (b = 0.583, t = 0.734, P_dir = 0.326). In sum, this computational exercise shows that collinearity did not cause inferential problems in our models.

1. Losos JB (1999) Anim Behav 58:1319-1324.

2. Garland T, Harvey PH, Ives AR (1992) Syst Biol 41:18-32.

3. Belsley DA, Kuh E, Welsch RE (1980) Regression Diagnostics: Identifying Influential Data and Sources of Collinearity (Wiley, New York).

4. Anderson MJ, Legendre P (1999) J Stat Comp Simul 62:271-303.

• Early Edition
• Archives
• Online Submission

• Feature Articles

  (Expanded research articles of exceptional breadth)

• Commentaries

  (Expert commentary on leading research)

• Letters

  (Online-only letters to the editor)

• Inaugural Articles

  (Articles by recently elected NAS members)

• PNAS Profiles

  (Biographical profiles of Academy members)

• Sustainability Science

  (Interactions of human and environmental systems)

• Collected Papers

  (Editorials, Perspectives, Reviews, Colloquia, etc.)

• Special Features

  (Solicited articles on exceptional research topics)

• Sackler Colloquia

  (Scientific reports held under NAS auspices)