Sustainable Agriculture: Nutritional Benefits of Wheat–Soybean and Maize–Sunflower Associations for Hibernation and Reproduction of Endangered Common Hamsters
Round 1
Reviewer 1 Report
Dear Authors,
you made a very sound study to experimentally check the effect of crop-associations on the performance (hibernation and reproduction) of hamsters.
I have some minor comments on the manuscript:
Line 72 –“…and PUFA content” – please explain abbreviation when it is first used…see Line 108
Line 126 “…both the main food and the supplement….” – what is the “main food” whats the “supplement”? You include new terms without clarification. please explain that the main food is a maize or wheat + the supplements sunflower or radish or soybean (see line 192 )
Line 123-129: Please make clear what type of material of the plants you are using. Seeds or leaves ? i understood you used seeds.
Line 129 “…Lumbricus terrestris..” please in italics
Line 376: “In the two later diet groups, “.. you mean “latter” I think.
Line 376-377: The whole sentence reads not very well. Please rephrase it to me more clear. I still don’t get the meaning. “hamsters’ Δbody mass during hibernation corresponded approximately to 25% of hamsters’ body mass at the onset of hibernation” –“ Δbody mass during hibernation ” – I thought the delta is the difference from before and after hibernation. –“During” -- I don’t get the sentence.
Your "story" i very well told and i think there is neither space nor the need to add more ecological factors.
As it is with maize. I interpret your results maybe a bit differently. You strongly emphasize to include more flowering stribes (with sunflower), less maize, more legumes.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
A paper worth publishing. Only minor changes should be made. In general, authors should look once more critically at the description of the methodology and the interpretation of the results. In the methodology, the choice of diets is not entirely clear, and in the discussion, statistically insignificant results are often interpreted as significant. This needs improvement.
Detailed comments are provided below.
Line 20-21. This result looks great, but N = 2. I think it's worth adding this information.
Line 28-29. Does this sentence apply to the whole Earth or only to Europe? Maybe don't look ahead until 2500, because the topic of your work doesn't justify it.
Line 62-63. Already in the introduction, it should be clearly stated that it is about the seeds of these crops.
Paragraph line 55-70. This paragraph needs to be improved. Decide whether it should still be a review of the state of knowledge or a presentation of paper goals or methodology. Give up the methodological details that you will describe in the next chapter.
Line 80. You should be more specific when describing why you selected soybeans, sunflower, and radish seeds for your research. At this point, it is not known whether the selection was made on the basis of oral consultations, questionnaires, what other factors were taken into account and, most importantly, how many types of seeds were considered. If your decision to choose soybeans, sunflower and radish was arbitrary, you should write about it openly. If the selection was based on specific criteria, their weights and the ranking of the options considered should be given.
Line 102-112. The sentences concerning impact of the hoards on hibernation quality and reproduction should be moved to the introduction.
Line 105. What's the point of writing about fat-storers here?
Line 168. Activity index. Why do you use such ranges instead of analyzing the specific result obtained for each individual? Such an approach may blur the differences between individuals fed with different types of food. Do you have more basic data on the number of breaths per 30 seconds to try a more detailed analysis?
Line 238-240. This sentence is not true.
Line 280. It sounds awkward "interaction between the diet and the type of food". Use the same terms consistently throughout the paper so that it is clear what interaction is being considered. The terms used in the methodology and results should be consistent.
Line 317-324. Your results fluctuate between 20 and 83%, but there are no statistically significant differences between these values. Consider whether the analysis you used is purposeful.
Figure 4A. I don't understand this graph. You monitored the number of females that initiated parturition. In Wrad group there were 5 females and only one initiated parturition, ie. 20% of females. What does whisker in Wrad bar mean then?
Figure 4B. mean litter size
Line 347. ‘’…were overall more adapted…’’ Each of the diets had some effect on the various parameters you studied. In order to identify the optimal diet, the authors could consider performing a multi-criteria analysis and creating a ranking of options. On this basis, they could indicate which diet is recommended for conservation activities. When preparing this ranking, the small sample size should be taken into account, for example, when analyzing the "litter size" factor. If there are only single observations for a given parameter, the random effect should be assumed to be of great importance.
Line 348-349. Is it already known from literature?
Line 371. ‘’…activity index and Δbody mass…’’. Repetition.
Line 385 ‘’… and still lost more mass, although non-significantly, than…’’. Can we interpret statistically insignificant dependencies in this way?
Line 389. Write uniformly throughout the paper using a lowercase letter.
Line 397 ‘’… a greater proportion of females gave birth…’’ and ‘’Females from the Mrad group were intermediate.’’ Above you claimed that the difference is statistically insignificant.
Line 412. Unfortunately, the research group was small and the random effect must also be taken into account.
Line 449. After reading this section for the first time, it seems necessary and interesting. On the other hand, however, the question arises whether the activities implemented since 2017 have an impact on the greater survival of the hamster or its reproductive success. Unfortunately, we do not find out about it. Therefore, I think authors should rethink this paragraph, shorten it and reformulate it to better suit the subject of this paper.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
A. Overall comments
This is an interesting study that contributes to an existing body of lab and field studies looking at mechanisms driving impacts of traditional agricultural practices on the common hamster and using mechanistic understanding to develop and test potential conservation solutions. The engagement of key stakeholders in the design of the study is a small (in amount of text needed to describe) but compelling aspect – a great model for sustainability-focused evaluations. The underlying study and manuscript are solid, and for the most part the manuscript is clear while being both thorough and concise. Except where noted below, I found the discussion to be very clear, interesting, and insightful.
B. Questions about analyses
While the analyses conducted made sense and mostly seemed very solid, there is insufficient information provided about some aspects, particularly the model-selection analysis. The authors conduct several modeling analyses using AIC to select a top model, then reporting significant test results for terms in that top model. Although the full models and final selected model is stated for each analysis, more clarity is needed about the set of candidate models (all the models in each model set) for each analysis.
For example, what is meant by "all" models (line 197) and are the rules determining "all" the same for all the model-selection analyses? For example, for the Activity Index analysis, I assume "all" means" would include an intercept-only model, 15 models with only main effects, and something like 30 models with one or more 2-way interactions but not violating the marginality restriction. Did models vary only in fixed effects (i.e. all have the same random effects structures)? What is the "relative weight of each variable" referred to on lines 246-247; the methods didn't mention summed AICc weights or anything of that sort. What is the justification for ignoring model selection uncertainty and just using a single top model for each analysis?
The authors can satisfactorily address my comment by a) clarifying in their methods text what the model set was for each analysis (clarify what's encompassed in "all") and other details, and b) putting the relevant AIC model tables (models, log-likelihood, number of parameters, and dAICc scores) in supplemental material.
Was the goodness-of-fit of the Poisson model examined, particularly whether there was evidence of overdispersion (lines 218 forward)? If not, it should be. That's a more critical concern for Poisson modeling than normality of residuals is for the LM/LMM analyses. Unaddressed overdispersion can greatly inflate the probability of Type 1 error. Unaddressed underdispersion, more likely with some litter-size data sets, has the opposite effect.
C. Interpretation of hibernation "strategies" in discussion (lines 347-356)
This is an interesting interpretation / synthesis of the results. My concern is that it is difficult to judge how compelling this synthesis when looking back at the results. One constraint is simply how the results are presented. Presenting one response per subplot in Figures 1 and 4 is essential for reporting the individual pieces of the results, but it does not lead to easy visualization of bivariate/multivariate patterns – i.e., doesn't help me examine whether this interpretation of "two strategies" is a convincing one. Adding additional figures synthesizing patterns across response variables to supplemental material could be helpful.
A second issue is that the statistical results don't seem to indicate as clean a divide as this text indicates. Certainly 1a seems to show 2 general groups based on activity (although see comment D9 below), but M-Sunf animals, which had high activity, were in the "lower" food-intake group (based on group difference codes in Fig. 1b, contrary to what is said in line 352). Based on group difference codes in 1c, M-Sunf had low loss of body mass but this net loss was not statistically different from W-rad animals, making "greatest activity and lowest d-body mass" statement on line 357 of mixed accuracy. (See also comment D10 below.) Overall, I think the "two strategy" discussion needs to be make clear why a couple of the results details are consistent with the idea of two strategies.
D. Minor comments
- Line 34: Add a qualifier indicating the general range in which it is a "flagship species" – "of European farmlands"? "Eurasian farmlands"?
- Line 35: Change "it" to "this area".
- Line 51: "Hibernators" reads as if you mean "hibernating individuals of common hamster". Tweak wording – "In hibernating small mammal species, hibernators with the best body condition…." Or something of that sort.
- 81: If the intended audience is international, clarify through phrasing what a LIFE+ project is.
- Line 105-106: Wording to work in the group names is a little awkward. "Some species store internal fat reserves ("fat storers"), whereas others such the common hamster hoard … ("food harders")."
- I am unable to view portions of Table 1 to the right of the Humidity column. I felt comfortable that I could fill in the details sufficiently to proceed with my review despite this.
- Figures, especially 1 and 3, would benefit from changes in letter size, point size and weight, etc. Even zoom to 150% on a desktop monitor, Fig. 3 symbols are either difficult to see (asterisks) or differences in shapes are unable to be discriminated.
- The interpretation of Fig. 3 makes sense based looking at the figure and caption, other than that the issues noted in the previous comment hinder seeing fine details. However, I found the methods text of 2.6 (line 158 forward) to be of little stand-alone value: looking at the figure allowed me to make sense of the text, rather than preparing me for the figure. I do not have any suggestions for addressing this issue so maybe there is no problem here. With the figure in view, it's not a confusing concept and doesn't warrant an example figure.
- For the activity index results, the reader is not really given a mental picture of how much activity varied among diet treatments. There aren't dramatic visual differences among overall mean index values (Fig 1A; I recognize that this is just a simple visual summary not directly tied to the actual modeling analysis). Simple summaries in supplemental material such as the average or across-individual proportion of observations in each activity category (1, 2, 3), rather than just the means could help the reader better visualize how general activity varied. And/or, adding a few key descriptive details to the results in a sentence or two could be helpful.
- In Fig. 1C, the "statistically different" codes ("a", "b") seem randomly distributed with respect to the actual mean mass losses: e.g. WRad and MSunf have the highest and second lowest means, respectively, yet share an "a". Based on Table S2, it doesn't appear that mean mass at the onset of hibernation differed among treatments.
- Talking about using "lower" to indicate "less negative / not as far below zero" for the negative delta-mass results is a confusing word choice (e.g. line 276).
- Figure 4: Why not provide letters indicating differences among groups for 4b and 4c comparable to Figure 1? I understand that 4b and 4c are split only for visual presentation.
- Section 4.4 was a fascinating section but the transition at the start ["Once such knowledge is developed…"] was not a smooth one. Lines 450-454 came across as "too many words and too general of a lead-in" rather than tightly worded like the rest of the discussion.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
