​​​​

This paper by Oddou-Muratorio et al. analyses the different selection components of phenology. In their conceptual framework, the authors explore fecundity and viability selection on phenological traits and their relationship with sexual selection by assortative mating, using two plots dominated by Fagus sylvatica (European beech) with 300 m of difference in altitude. Authors regressed fecundity vs. the phenological mismatch on the timing of bud burst (TBB) as a proxy of the sexual selection. The regression between fecundity and TBB was used as the proxy of fecundity selection. And the regression of TBB vs. seedlings growth (greenhouse experiment) was used as the proxy of viability selection. 

I find that this manuscript presents an interesting collection of field and greenhouse data. The authors combine genetic analyses from seeds and seedlings collected in the field for estimating paternity and parentage with seedling growing data from a greenhouse with a drought treatment. The paper is well written and the statistical analyses seem appropriate (although I do not understand why phenological mismatch has negative values in fig. 5 are negative, see my comments below). My main concern is that results seem to be too specific of the study plots, female or male fertility, age of individuals and altitude, being difficult to generalize. Also, the combination of data needs more rationale behind. For instance, the spatial design of the study includes an elevational gradient, although with only two plots separated by only 300 m in altitude. This is not so much regarding the distribution range of beech in the area (from 750 to 1700 m a.s.l). But above all, it is not clear how the altitudinal gradient is going to affect different selection types on the phenology. I am missing a more formal hypothesis regarding this. Similarly, the inclusion of a drought treatment for the seedling experiment in the greenhouse needs more justification and link with the main objective of the manuscript (i.e., detecting selection gradients). I have the impression that the manuscript collates data originally designed for other independent studies and the internal logic of the design is difficult to discern.

In the discussion section, the authors explain assortative mating at the lowest elevation because phenology was more spread out. But this could be also being an effect of higher beech density. How can these two factors be separated?

In short, I believe that the manuscript is a very valuable piece of science, although I consider that it should clarify a bit more the rationale for the combination of methods. Also, it has numerous tables that I believe that could be summarized or included in the supplementary material. A more organized description of the results according to the original framework (three selection types, three proxies of it) would clarify the message for the reader.

Minor edits:

L22. Include the species name here.

L23. Here and elsewhere, the authors use the term “low altitude” for referring to plot N1, which is 1,020 m a.s.l. I consider that this gives a fake impression of low altitude to the reader, whereas it corresponds to a relatively high altitude compared to other systems. I suggest to use another term throughout the text such as “mid altitude” or something equivalent.

L49. Define microevolution.

L84. I suggest including here a reference that all the hypotheses are for “temperate ecosystems”.

L93. I believe that this sentence is incomplete. Please, rewrite it.

L147-150. These two sentences read very repetitive, try to improve the writing of them.

Figure 1. Try to improve the aesthetics of this figure. For instance, I cannot understand why you use a flag for assortative mating, the same for the circle with the cross. The right sided arrows are supposed to indicate delays in timing, but this is not intuitive.

L163. Remove hyphen for Mont Ventoux.

L166. I suggest changing the name of the plots for more informative ones. Plots N1 and N4 seem to me as names used for field work, but they are not intuitive for readers not working at Mont Ventoux (the majority). For instance, why they are not included plots N2 and N3 in the analyses? I recommend to change these names to “mid-altitude” and “high-altitude” plots. In fact, the authors have done so in figure 5, entitling the sites as “low altitude” and “high altitude”. However, I consider that 1000 m a.s.l. (N1) is not exactly low altitude, so a better term should be used. In addition, when reading results, it is difficult to remember which plot was high and which one was low with the current names. I believe that is better to include a name for the plots that gives information on the elevation gradient instead of the field design.

L190. As the number of visiting dates was different per plot, should not the phenological score sum (PSS) be divided by the total number of visits?

L237. Reference for the MEMM software is missing.

L331. This is not a clear reference, try to include the entire URL to the supplementary file instead.

L351. Italics for lmer and lme4. Citation of the package needs to be included.

L356. Instead of the mean, it is more informative to include the median or even the peak for phenological information on days (because it is a discrete variable).

L375. Put space always after and before =

Table 2, heading. The name of the function used should not be included in the results, but in the description of methods. Also, include the reference of the package where this function comes from. 

Figure 5. Why do you have negative value for the absolute mismatch (while the formula indicates the absolute value)?

L454. Put the exact reference of the appendix (including URL) here.

Fig. 8s. If you find that this figure is critical, include it in the main text