Forest Fragmentation and Developmental Stability of Wood Mice Apodemus sylvaticus: A Food-Mediated Effect?

Round 1

Reviewer 1 Report

This manuscript has aim to evaluate if fragmentation negatively affects the survival of young woody mice and if such effects are mediated by food availability in small woodlots. I have some doubts on the sampling and in the statistical analysis that could affect the quality of the manuscript.

It is essential to make sure that the individuals do not move between forest fragments. If this premise was not taken, the individuals from the non-supplemented fragment could move to the supplemented one causing confounded results. Also, sampling plots from regions 1-2-3, 4-5-6 and 7-8-9 seem to be close enough to not assume total independence of data. This way, a GLMM approach, considering this three regions as a random factor, could be more indicated for the analysis. The sample size of youngs are too small for robust analyses. It seems odd to me that there is a difference in fluctuating asymmetry in youngs but not in adults, it would mean that the fluctuating asymmetry disappears when animals grow? If so, would this asymmetry be really important for the development of the youngs? The hypothesis of diferential mortality raised by the authors to justify the absence of diference in adults is not easy to support with this low sampling size of youngs. It is more plausible to assume that the difference in assimetry found is an statistical artifact from the small sample size of youngs (3 in each site) that do not appear when sample size increases as in adults (28,29 and 18 indiviuals in each site).

Additionally, figure 1 could improve, it could show forest patches from a satellite image and numbers should be described in the figure legend, but as the field work was done in 1996 I think it would not be possible to retrieve an image that actually represents the site at the time. The authors should explain how food supply was disposed at the woodlots

Author Response

This manuscript has aim to evaluate if fragmentation negatively affects the survival of young woody mice and if such effects are mediated by food availability in small woodlots. I have some doubts on the sampling and in the statistical analysis that could affect the quality of the manuscript.

We have now included explanations on these queries that we did not include in the first version to avoid repetition with the paper in which all these details were developed (Díaz, M.; Alonso, C.L. Wood mouse Apodemus sylvaticus winter food supply: density, condition, breeding, and parasites. Ecology 2003, 84, 2680-2691. doi: 10.1890/02-0534; ref. #30). Care was taken in the wording of sentences to avoid self-plagiarism

It is essential to make sure that the individuals do not move between forest fragments. If this premise was not taken, the individuals from the non-supplemented fragment could move to the supplemented one causing confounded results.

We now quote briefly that wood mice in our study areas are known to forage in fallows close to permanent refuges inside fragments and forests during winter (lines 123-125). We have no direct evidence based on marked individuals, but work done in other systems with radio-tracked mice (also cited in this manuscript; Tattersall et al. 2001. Habitat use by wood mice (Apodemus sylvaticus) in a changeable arable landscape. J. Zool. 255, 487-494., Ref #51) shows how mice retreat to woodlots and forest borders after harvest due to increased predation risk outside. We also quoted this briefly in the discussion. We hope the new quotation will suffice to show that movement among study fragments during winter is at very unlikely. Besides, results of the Ecology paper indirectly prove this lack of movement on the basis of the significant responses to food addition by adult mice.

Also, sampling plots from regions 1-2-3, 4-5-6 and 7-8-9 seem to be close enough to not assume total independence of data. This way, a GLMM approach, considering this three regions as a random factor, could be more indicated for the analysis.

We did not make these analyses neither here nor in the Ecology paper (where we recognize the existence of three clear landscape blocks) since we do not expect block effects if mice populations within fragments and forests were effectively isolated. Besides, the proper random structure of the GLMM (three landscape blocks nested in the fixed factor control and food supplemented fragment, and two transects nested within the forest treatment) would have decreased power to a great deal, while we had to combine different FA measurements to increase such power.

The sample size of youngs are too small for robust analyses. It seems odd to me that there is a difference in fluctuating asymmetry in youngs but not in adults, it would mean that the fluctuating asymmetry disappears when animals grow? If so, would this asymmetry be really important for the development of the youngs? The hypothesis of diferential mortality raised by the authors to justify the absence of diference in adults is not easy to support with this low sampling size of youngs. It is more plausible to assume that the difference in assimetry found is an statistical artifact from the small sample size of youngs (3 in each site) that do not appear when sample size increases as in adults (28,29 and 18 indiviuals in each site).

Low sample sizes would produce low power to detect differences, not false significant effects, whereas large samples sizes increase confidence on lack of significant effects, not the opposite as quoted by the referee. This point is a basic assumption of inferential statistics, as developed in all statistical textbooks (such as the classical Sokal and Rohlf 1981. Biometry, 3rd edition. Freeman, New York, Ref#45) cited in our manuscript. Hence, our results cannot be attributed to samples size effects, and conclusions cannot be considered as biased by samples sizes. We did not quote this explicitly in the text because it would be redundant with such basic assumption. Nevertheless, we now add in the Abstract a brief sentence emphasizing that significant differences among young mice were found in spite of low sample sizes, thus indicating that effect sizes were large enough to be detected even with these small samples.

Additionally, figure 1 could improve, it could show forest patches from a satellite image and numbers should be described in the figure legend, but as the field work was done in 1996 I think it would not be possible to retrieve an image that actually represents the site at the time. The authors should explain how food supply was disposed at the woodlots

This map is provided in Fig. 1 of the ref #30 (Díaz, M.; Alonso, C.L. Wood mouse Apodemus sylvaticus winter food supply: density, condition, breeding, and parasites. Ecology 2003, 84, 2680-2691. doi: 10.1890/02-0534), whose copyright belongs to the Ecological Society of America. This fact is quoted in the figure legend. If editors think that the point raised by the referee is important, they would ask the Ecological Society of America the necessary permission to reproduce the figure. Please find it inserted below

Author Response File: Author Response.pdf

Reviewer 2 Report

The presented manuscript is devoted to an important scientific problem - effects of forest fragmentation on food availability and development of Apodemus sylvaticus. The evaluation of the influence was carried out according to the results of the study of the fluctuating asymmetry of the lower jaws. However, the chosen method for estimating the fluctuating asymmetry seems to me to be unsuccessful.

138-139 - “Asymmetry was calculated as the absolute difference between the values of each distance between the right and the left lower mandibles”.

-- In relation to this sentence, I think that authors allowed a methodological error, which has influence on the further interpretation. Namely, authors didn't describe theoretical base of the usage of "absolute differences." There are modern proven methods for estimating the fluctuating asymmetry (Klingenberg Ch. 2015. Analyzing Fluctuating Asymmetry with Geometric Morphometrics: Concepts, Methods, and Applications. Symmetry 2015, 7(2), 843-934; https://doi.org/10.3390/sym7020843)

In adition I didn't understand a samples count that used for analysis. Therefore, I ask authors to clarify the follow aspects: (1) Reference to a modern study, where authors used absolute values in view of assymetry -- my opinion propose usage two var-covar matrices, each for left and right side, with final correlation, e.g, via Mantel test; (2) Clear information for specimen number.

Also, it is desirable to write the purpose of the study more clearly at the end of the Introduction.

Comments for author File: Comments.pdf

Author Response

The presented manuscript is devoted to an important scientific problem - effects of forest fragmentation on food availability and development of Apodemus sylvaticus. The evaluation of the influence was carried out according to the results of the study of the fluctuating asymmetry of the lower jaws. However, the chosen method for estimating the fluctuating asymmetry seems to me to be unsuccessful.

138-139 - “Asymmetry was calculated as the absolute difference between the values of each distance between the right and the left lower mandibles”. -- In relation to this sentence, I think that authors allowed a methodological error, which has influence on the further interpretation. Namely, authors didn't describe theoretical base of the usage of "absolute differences." There are modern proven methods for estimating the fluctuating asymmetry (Klingenberg Ch. 2015. Analyzing Fluctuating Asymmetry with Geometric Morphometrics: Concepts, Methods, and Applications. Symmetry 2015, 7(2), 843-934; https://doi.org/10.3390/sym7020843) In adition I didn't understand a samples count that used for analysis. Therefore, I ask authors to clarify the follow aspects: (1) Reference to a modern study, where authors used absolute values in view of assymetry -- my opinion propose usage two var-covar matrices, each for left and right side, with final correlation, e.g, via Mantel test; (2) Clear information for specimen number.

We are aware that there are alternative methods to measure and analyse fluctuating asymmetry. We in fact cite the recent review by Graham et al. (2010. Fluctuating asymmetry: methods, theory, and applications. Symmetry 2010, 2, 466-540, ref #49.) as well as the classical books and papers by Møller and Swaddle (1997. Asymmetry, developmental stability, and evolution. Oxford University Press, Oxford, ref #31), Palmer and Strobeck (1986. Fluctuating asymmetry: measurement, analysis, patterns. Annu. Rev. Ecol. Syst. 1986, 17, 391–421; ref #33) and Swaddle et al. (1994. The analysis of fluctuating asymmetry. Anim. Behav. 1994, 48, 986-989. Ref#40). We have chosen this one for consistencty with both our own work with fluctuating asymmetry on other organism (e.g. Díaz et al. 2004. Herbivore effects on developmental instability and fecundity of holm oaks. Oecologia 2004, 139, 224-234, ref #41) and with recent papers on fluctuating asymmetry in wild mice (Shadrina and Vol’pert 2016. Fluctuating asymmetry of craniological features of small mammals as a reflection of heterogeneity of natural populations. Symmetry 2016, 8, 142, ref #34). We now quote this explicitly in lines 237-238, also citing the alternative method based on geometric morphometrics. Concernig simple sizes, we gave this information in the first paragraph of the results section, but now we include them again in the legend of Table 1, following also suggestions of referee 1.

Also, it is desirable to write the purpose of the study more clearly at the end of the Introduction.

Done, thanks. We have moved most of the former last paragraph of the introduction to the section ‘Fluctuating asymmetry of lower mandibles’ in MatMet, as suggested by referee 1, and added the sentence ‘To evaluate if food limitation in young mice could affect their survival prospects, we compared the developmental stability of individuals born in supplemented small woodlots with respect to that of controls and large forest fragments, and also the stability of young and adult mice. Given the previously observed food limitation observed in adults wintering in small fragments [30], we expected a) higher developmental instability of mice born in small woodlots, b) that food supplementation would decrease the levels of FA of animals to similar values of those found in individuals inhabiting forest habitats. Finally, we expected that differences in FA should be stronger in young individuals than in adults, as the latter could have compensated higher initial levels of FA during development or incurred in differential mortality [31]. (lines 97-106)

 

Author Response File: Author Response.pdf

Reviewer 3 Report

I was surprised when I found that the methodological approach was based on snap trapping, since, to my knowledge, this method is rarely or no longer used in research today. However, data collection took place way back in 1996 and at that time this data collection approach was still a quite common practice.

All in all, one has to certify that the authors have done a good job here. However, there is one central issue: the biological effect detected is very small. Against the background of the extremely small sample size of juvenile individuals (n=3 per study area), the result on which the entire work is based can be questioned. Nevertheless, the authors manage to adequately interpret this effect – a higher degree of fluctuating asymmetry in juvenile animals on the control area. One has to acknowledge that the paper offers little reason for criticism beyond that, as it is consistent in itself and the authors avoid overinterpretation and adequately classify the result in the context of scientific literature. However, the key question here is whether the detected biological effect is strong enough to justify publication. This is certainly a borderline case, but I think the work deserves to be published. At the same time, it would be desirable if the authors could emphasize this limitation of interpretability even more clearly, especially in the abstract.

I have a few further minor comments:

Title: I am not convinced that the authors studied the effect of forest fragmentation on the winter food availability, as insinuated by the title, suggest to delete the term “winter food availability” from the title.

Introduction: Lines 50-54: please check, whether there is some redundancy here – please reword.

Figure 1: Please insert a North arrow.

Methods: Some general remarks about the study design:

1)      Snap traps often destroy the skull, which can make it difficult or impossible to take measurements like CBL – could the authors indicate the ratio of individuals where this happened.

2)      To my understanding, the analysis here is solely based on the measurements taken from the mandibles. It would be interesting to see, whether FA is correlated with other measurements, like body mass or body condition score...

Line 203: please spell out “FA” when first mention it in a chapter.

Line 224: “more exposed” than what? Probably the authors refer to continuous forest but should say it here.

Line 227: “worse” than what?

Author Response

Referee 3

I was surprised when I found that the methodological approach was based on snap trapping, since, to my knowledge, this method is rarely or no longer used in research today. However, data collection took place way back in 1996 and at that time this data collection approach was still a quite common practice.

All in all, one has to certify that the authors have done a good job here. However, there is one central issue: the biological effect detected is very small. Against the background of the extremely small sample size of juvenile individuals (n=3 per study area), the result on which the entire work is based can be questioned. Nevertheless, the authors manage to adequately interpret this effect – a higher degree of fluctuating asymmetry in juvenile animals on the control area. One has to acknowledge that the paper offers little reason for criticism beyond that, as it is consistent in itself and the authors avoid overinterpretation and adequately classify the result in the context of scientific literature. However, the key question here is whether the detected biological effect is strong enough to justify publication. This is certainly a borderline case, but I think the work deserves to be published. At the same time, it would be desirable if the authors could emphasize this limitation of interpretability even more clearly, especially in the abstract.

Thanks for the criticism. We are aware that samples size was small, but we are reluctant to increase it by killing more individuals. Our measurements are destructive, and were taken in the context of another study were killing was necessary to obtain physiological and parasitological measurements of the individual trapped. We here just try to extract as much information as possible from this old material. We now develop all these ideas in the abstract and here and there in the introduction and discussion.

 

I have a few further minor comments:

Title: I am not convinced that the authors studied the effect of forest fragmentation on the winter food availability, as insinuated by the title, suggest to delete the term “winter food availability” from the title.

Thanks for the remark. We have rephrased the title to ‘Forest fragmentation and developmental stability of wood mice Apodemus sylvaticus: a food-mediated effect?

 

Introduction: Lines 50-54: please check, whether there is some redundancy here – please reword.

Done, thanks. The sentence now reads ‘Habitat loss and fragmentation are the global change drivers with the strongest negative impacts on biodiversity [1-3]. Most research effort on these drivers have been focused on habitat specialists, as it is accepted that negative effects of fragmentation should be strongest for the species more dependent on the fragmented habitat [1, 4-5]. (lines 55-58)

 

Figure 1: Please insert a North arrow.

Done, thanks

 

Methods: Some general remarks about the study design:

1)      Snap traps often destroy the skull, which can make it difficult or impossible to take measurements like CBL – could the authors indicate the ratio of individuals where this happened.

We now include this information at the beginning of Results (lines 266-267). 20% skulls were too damaged by snap traps to be used for estimating FA.

2)      To my understanding, the analysis here is solely based on the measurements taken from the mandibles. It would be interesting to see, whether FA is correlated with other measurements, like body mass or body condition score...

We did not include these correlations because a) we did extensively analyse effects of fragmentation and food addition on adult body condition in the study published in Ecology from which the mandibles analysed here came from and b) there are no clear mechanistic relationships between measurements of body condition and survival prospects, whereas there is a mechanism, developmental selection, that relates directly developmental instability and survival. We thus think that including these analyses in the MS will not add any relevant information to the study.

Line 203: please spell out “FA” when first mention it in a chapter.

Done; now line 307.

 

Line 224: “more exposed” than what? Probably the authors refer to continuous forest but should say it here.

Yes; now we quote this explicitly in lines 338-339

 

Line 227: “worse” than what?

The sentence has been deleted

 

Author Response File: Author Response.pdf

Reviewer 4 Report

Review of Diversity-2116327

Effects of forest fragmentation on the winter food availability and developmental stability of wood mice Apodemus sylvaticus

Dear Authors,

I read with interest your study on FA in wood mice in forest fragments in Spain. Despite small sample sizes for young mice, and FA being uncertain as a metric, I believe that your work adds an important element in our understanding on how forest fragments function as potential sink habitats for forest-dwelling species. I am supportive of this work being published. Thank you for doing this work. I also commend you for sharing results from a study done over 25 years ago.

Your study is well designed and executed, and the manusript is largely well written. As such, I have few substantive comments. That said, the writing could be polished, and I have provided below some suggested minor edits.

I hope you find my comments helpful in revising your manuscript.

I look forward to soon seeing your work published.

Sincerely,

Thomas Jung

Adjunct Professor, University of Alberta

Substantive Comments:

1. While the results for young mice are the most interesting for this study, they are based on a very small sample size – 10 animals. This is likely a result of trapping in early spring, before most first litters; hence, mostly overwintering adults were most common. The limitation of small sample sizes needs to be more explicitly, and deeply, discussed in the Discussion. I would expect to see a recommendation for further study with larger sample sizes (perhaps from museum specimens). Maybe captured later in the year than March.

2. Do you have the actual permit numbers to report (Line 113)? I suspect not given how long ago this study occurred, and that is okay. But, if you do have the permit number than please provide it. At minimum, please state who issued the permits to snap trap mice for your study. Thanks.

3. Predictions presented at the end of the Methods (Lines 165-170) need to be explicitly identified at the end of the Introduction. Please move these up to the Introduction.

 

Detailed and Editorial Comments:

Line 3: I suggest putting the scientific name in brackets – i.e., “(Apodemus sylvaticus)”. Same for lines 14, 41, and 197.

Line 4: Authors’ names should be written out in full, following the format used for most other papers in Diversity.

Line 16: Rephrase “recruitment expectations”. Unclear what is meant. Awkward.

Line 22: Replace “its survival prospects” with “survival”

Line 25: Please rephrase “that can unbalance its role towards effective seed dispersal” for clarity and concision. Awkward and unclear.

Line 39: Replace “key important preys of higher trophic levels” with “important prey for predators”

Line 44: Replace “mice” with “mouse”

Line 47: Replace “population dynamics” with “abundance”

Line 51: Replace “thanks” with “due”

Line 53: Replace “thanks to” with “given”

Line 61: Replace “reproduction, that is,” with “reproduction. That is,” Break into two sentences.

Line 63: Replace “work we aim” with “study we aimed”

Line 65: Delete  “To this end,”

Line 68: Please rephrase “due to the inability of an organism to execute its developmental program”. Awkward and unclear.

Line 71: Replace “manipulative experiment” with “experimental approach”

Line 80: Nowhere does it clearly say that the work was done in Spain! Please ensure that it says so on line 82 and for the figure captions.

Line 86: This was done a long time ago. But that is okay.

Line 89: Replace “supply excess” with “provide supplemental”

Line 92: Replace “select” with “selected”

Line 109: Replace the first sentence with this one “We aimed to capture all individuals in the woodlots during four nights on 18–21 March 1996.”

Line 110: I am a bit confused if this work occurred in 1998 (Line 86) or 1996 (Line 110).

Line 112: Delete “Snap traps kill the trapped individuals at once reducing suffering.” Unnecessary.

Line 115: Insert “apart” after “15 m”

Lines 118 and 119: Delete the times in brackets. Unnecessary detail.

Line 137: Please delete the name here and instead include this person in the Acknowledgements stating their contribution.

Line 150: Please replace “(Apodemus sylvaticus)” with “wood mice (Apodemus sylvaticus)”. Consistently stick with the common name in the text.

Line 173: Replace “were able to capture” with “captured”

Line 185: Table 1 column headings need formatted to fit better.

Line 214: Replace “mice” with “mouse”

Line 216: This last sentence is quite awkward and unclear. Please reword for clarity and grammar.

Line 225: Replace “source monopolization” with “monopolization of resources”

Line 227: Replace “worse” with “poorer”

Line 250: Use the common name here please, for consistency.

Line 256-258: Absolutely. I agree this is a plausiable explanation.

Line 276: Replace “manipulative” with “experimental”

Line 280: Delete “of”

Line 294: Please include me a reviewer – Thomas S. Jung

Line 300: References are numbered in duplicate and this needs corrected.

 

Author Response

Referee 1 (edited MS)    We have moved most of the former last paragraph of the introduction to the section ‘Fluctuating asymmetry of lower mandibles’ in MatMet.  We think that the position of the point defining M4 towards the angular process is clear enough (point of maximum concavity).  Finally, we have included data on sample sizes in the legend of table 1 despite this information is clearly exposed in the first paragraph of the results section and in the legend of Figure 1.