TO IMPROVE
-The authors could add predictions and make the scope of the study larger (generalizing outside of Drosophila). Predictions will help the reader follow where the paper is going.
-If Dollo’s law was presented in the introduction, it could become the basis for a prediction, which would make the “return” to the normal trait even more surprising.
-Explain what are the implications for natural populations and explain what is the difference between loosing a trait versus changes of other nature in a trait.
-explain why we can exclude the effects of de novo mutations. Why are we sure that evolution uses SGV? Two points make it possible to say that it is cryptic genetic variation which allowed the adaptation observed in the various experiments: the speed of appearance of the phenotypic responses and their observation in several parallel lines. For those readers who are more accustomed to hearing about compensatory evolution by spontaneous mutations in microbial species, it may be important to emphasize this point further as it is the basis of the ms.
-In methods, the authors state about the BASE population: “This general population of flies should be segregating much of the natural variation in the FVW population.” It is unclear whether the introgression method is effective in getting rid of the majority of the alleles of the strain initially carrying the mutation. If it is well known that 8-10 backcross are sufficient to isolate a locus in Drosophila, it should be mentioned for those who know nothing about this model species and who read the paper. If this is not the case, then it is a limitation to discuss. It is important when comparing phenotypes of NS strains to those of BASE (Figure 4). If many other loci are associated with vc1 in BASE, some may interact inconsistently with alleles in the natural population. For example, when greater larval competitiveness is observed after natural selection, this higher phenotype could be caused by the removal of incompatible alleles associated with vc over generations, and not the selection of cryptic genetic variation.