The authors investigated whether disassortative mating and overdominance are potential mechanisms for maintaining body color variation in Chelymorpha alternans populations. As a result, the authors found little evidence for disassortative mating and strong evidence for overdominance. The experiments were well designed and statistical analyses generally seem sound. I only have a few issues to raise as listed below.

1. In the present study, the authors did not find strong evidence for disassortative mating. However, if the difference in the survival of offspring from the pairs with different color pattern and the pairs with the same color pattern also exists in natural populations, preference for disassortative mating should be selected and evolved in nature. I recommend to add some discussion on why disassortative mating is not evolved in nature even with a large fitness advantage observed in this study.

2. The authors performed many statistical tests in this study and p-values of some of them need to be adjusted to control family-wise error rate. Specifically, p-values of the statistical tests for the results presented in Figure 1, Figure 3, and Table S3 need to be adjusted.

3. The authors explains that the high death rate of the larvae from the pairs with the same color pattern may be due to the secondary metabolites from host plants. If this it true and Chelymorpha alternans utilize multiple host plant species, larval survival needs to be examined on multiple host plant species to evaluate their fitness. Although additional experiments would be difficult, this point needs to be discussed.

Minor issue

Fig 2: Please add explanations on the error bars in Fig. 2B and 2C. Labels of x and y axes are distorted and difficult to read.

The authors responded to the queries well but there remain several unclear points. My responses to the responses of the authors are as follows.

1. “L398-406: A more detailed discussion has been added considering lack of disassortative mating in light of strong advantages to offspring survival from disassortative pairings:”

The authors added a paragraph in the discussion section to give a convincing discussion on the point that I made in the previous round of review.

1. “These p-values have been adjusted using a Holm’s Bonferroni correction.”

The authors claimed that they performed p-value adjustment using Holm’s Bonferroni correction, there is no description of the p-value adjustment in the main text and in figure legends. The authors need to describe how they adjusted the p-values explicitly.

1. “Figure 2: The error bars were removed for the proportional data of Figures 2B and 2C. The x and y axis have also been adjusted for clarity.”

My original request was to add explanations on the error bars in Fig. 2B and 2C. If the experiments were replicated, error bars need to be shown. I want the authors to explain why they removed the error bars.