Dear Prof. El-Kassaby,
thank you for your author response. Please fulfill the following points in your revised manuscript:

1. Reviewer comment: P14107L23-34: Here you jump from paleoecolgical evidence to observations from the 1950... totally different things, I believe...
Your Reply: "The core idea is that the environment plays an important role in seed size development. The paleontology data give evidence that climate change may account for seed size variation, followed by studies of how the impact of environment on seed size variation gradually brings our attention. We attempt to emphasize the importance of environment on seed size variation from the perspective of paleontology and previous studies."
Editor comment: Please make sure in the revised text that your clearly refer to processes that apply to the paleoontological scale and to the ecological time scale so that the reader knows which process belongs to which time scale.
2. Reviewer comment: P14110L12: I am really not convinced that you need so many climatic variables. they will be highly correlsted anyway, right?
Your reply:
ClimateWNA can easily generate many climate variables and each of them has its own specific meaning although many of them might be correlated. We wanted to take the full advantages of this program and let PLS to make selection among the climate variables. PLS is specifically suited for this purpose.
Editor comment: It is important that you justify and explain in great detail in your revised manuscript what each climate variable contributes to the overall analysis. It is very convenient that the ClimateWNA software can generate many climate variables, but it needs to be of advantage for the scientific analysis and generate more insight.
3. Reviewer comment: P14110L13: You should really explain what ClimateWNA does.
Your reply:
ClimateWNA has been widely used with over 200 citations according to Google Scholar over the past three years. The program downscales gridded PRISM climate data to scale-free point locations using partial derivative functions with respect to elevation. It also uses a delta method to downscale GCM climate data. Detail description of the methods can be found in Wang et al. 2012.
Editor comment: Please make sure that your revised manuscript sufficiently explains the important functionality of the ClimateWNA software. Even though it has been widely used, it is important for the reader to know the essential functionalities to understand and interpret the presented results without consulting Wang et al. 2012. The Wang et al. 2012 publication may be consulted for additional information but not to understand the software in principle.
4. Reviewer comment: P14110L20: Since you are using only one GCM, which is always risky and actually not very robust in terms of uncertainty, you should at least explain where in the range of CMIP5 GCMs CCSM4 is located.
Your reply: CCSM4 is near the average in both temperature and precipitation increases for BC, the study area.
Editor comment: Please make sure this is mentioned in the text and the quantitative difference to the average CMIP5 temperature and precipitation increase is provided.
5. Reviewer comment: P14119L20ff: Are these numbers of climate effects across different RCPs? then better to provide the range over the RCPs rather than the mean...
Your reply: I don’t know which line you denote. I guess it should be some line between L16 and L28. Through Fig. S8, you can see the mean and variance for each population. It therefore makes sense if the mean is used for general comparison. Otherwise, I should provide three ranges as per three RCP scenarios.
Editor comment: Please clarify this in the text, when you refer to the mean for each RCP. But it is important that you also provide the range for each RCP. This is important to understand the influence of the range across climate projections where each climate scenario represents a different realization.

I think the above mentioned points will help you to improve the clarity of the manuscript.
Best wishes,
K.Thonicke.

Dear Prof. El-Kassaby,
thank you for your revised manuscript and addressing the reviewer's concerns. It can now be published in Biogeosciences. Thank you particularly for your patience during the unusually long review process, it was for the first time nearly impossible to find reviewers that would also submit their reports.

Best wishes,
K. Thonicke.