Content of review 1, reviewed on June 26, 2020

The questions the authors propose around stressful environments and the release of cryptic genetic variation is an interesting but a difficult one. The field of plasticity and canalization is complicated and showing/proving you have indeed released cryptic genetic variation is difficult.
The idea of a release of phenotypic plasticity is a difficult one. I have not really thought about how the release of cryptic genetic variation via stressful environments will alter phenotypic plasticity per se but more so how it will impact phenotypic variation. I think the tricky thing here is you almost need a control environment to show what your baseline plasticity looks like in comparison to more stressful environments and then you need to show that the increase in plasticity from more stressful environments is in that direction. I don’t agree with the authors (page 11) that an increase in phenotypic variance, in general, represents a release of phenotypic plasticity. To show that cryptic genetic variation is influencing plasticity I think the increase in variation needs to be somewhat directional after all plasticity should be adaptive. Perhaps the difficulty here is that what you are measuring is both the trait and its plastic response, so how do you explicitly untangle plasticity from just cryptic genetic variation influencing the variation of the trait? I am not sure you can, at least not in the current design. I don’t think this makes your study any less interesting but I would consider not placing your results in a plasticity framework but rather a release of cryptic genetic variation framework.
I find it difficult to assess how plasticity has changed across the environments or the degree of stress each environment represents. Are some environments more stressful, would you expect some environments to release cryptic genetic variation and others not too? The release of cryptic genetic variation is unlikely to be continuous but instead threshold like i.e zone of canalization, how would a threshold model of the release of cryptic genetic variation influence your results. Because of the complexities surrounding predictions of plasticity and cryptic genetic variation I found the analysis difficult to interpret. Shifting the focus from cryptic genetic variation in plasticity to cryptic genetic variation might alleviate some of this.
I also wondered about the developmental stage at which fish were exposed to different environments and how this might impact the release of cryptic genetic variation. Canalization is often thought about in terms of environmental and genetic canalization and I wonder to what extent you will breakdown genetic canalization when you do not expose individuals entirely during development. Are you “releasing” environmental canalization and is this the same as releasing cryptic genetic variation. It has been a while since I have read the canalization literature but to prove that it is indeed a release of cryptic genetic variation you would need to follow this variation through to the next generation or see if you can select upon this variation.
Different traits will have different levels of canalization and hence will and will not accumulate cryptic genetic variation. Have you shown that these traits are indeed canalized across a certain environmental gradient?
Finally, I wonder about broad genome metrics and how these relate to traits. I guess you would expect if stressful environments do release cryptic genetic variation you should see an increase in heterozygosity with stress but is part of the problem that individuals were not exposed all through development? or that only some environments truly represent stress and your signal is being swamped by this.

Materials and methods
It is not clear in the materials and methods how many populations and how they were translocated across the 29 sites. This experiment took place over 4 years, were the translocations replicated across years? Did the same populations get moved into the same sites? How many individuals were replicated and re-caught from each site? It is clear that the sites were propagated at a certain density but how does this translate into actual fish numbers? The experimental design is unclear. And how stressful these environments actually are is also unclear.
Could you breakdown your plasticity into populations and look at the reaction norms across populations?
Do your figures show the coefficient of variation for centroid size and standard deviation for warps?

Source

    © 2020 the Reviewer.

Content of review 2, reviewed on October 18, 2020

The authors have done a good job at addressing the reviewers concerns and I have no further comments to add.

Source

    © 2020 the Reviewer.

References

    C., Y. M., J., F. D. 2021. Evaluating the correlation between genome-wide diversity and the release of plastic phenotypic variation in experimental translocations to novel natural environments. Journal of Evolutionary Biology.