Feeding Behaviour of Seven Icefish Species (Channichthyidae) in the Ross Sea, Antarctica

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The document in question has a degree of similarity of 30% verified through Turnitin. Although it is understood that the authors have other works based on the same research, it is necessary that they review the similarities attached to this review.

It is not clear why they do an ANOVA and then a PCA. I do not consider the number of samples sufficient to do ANOVA and the normality of the data has not been tested to do so.

In order to perform an analysis of variance (ANOVA), the following requirements must be met:

 

Independence: the observed values are independent of each other. One way to ensure the independence of observations is by adding randomness to the sampling process.

Homoscedasticity: there must be homogeneity in the variances, that is, the variability of the residuals is constant.

Normality: the residuals must be normally distributed, or in other words, they must follow a normal distribution.

Continuity: the dependent variable must be continuous.

with less than 30 individuals these conditions are difficult to have.

I suggest use only non-parametrical statistics

Comments on the Quality of English Language

No comments

Author Response

1. Summary
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files
2. Questions for General Evaluation	Reviewer’s Evaluation	Response and Revisions
Does the introduction provide sufficient background and include all relevant references?	Can be improved	We have preferred to leave the Introduction as it is, and develop the manuscript with more details in the other sections
Are all the cited references relevant to the research?	Yes
Is the research design appropriate?	Yes
Are the methods adequately described?	Can be improved	A figure has been added to clarify how the measures used for the analyses had been taken. Furthermore, some more details added and some parts have been re-written.
Are the results clearly presented?	Must be improved	The major comments from the authors have been taken into account. In particular, we have improved some of the  figures to better highlight the results of the statistical analyses
Are the conclusions supported by the results?
3. Point-by-point response to Comments and Suggestions for Authors
Comments 1: The document in question has a degree of similarity of 30% verified through Turnitin. Although it is understood that the authors have other works based on the same research, it is necessary that they review the similarities attached to this review.
Response 1: Thank you for pointing this out. Some parts, especially in the Materials and Methods section, are very technical and it results often difficult to change words respect to previous articles (comprised our previous articles). Anyway, we have now tried to re-write some paragraphs to avoid as much as possible to re-use words from previous works.
Comments 2: It is not clear why they do an ANOVA and then a PCA. I do not consider the number of samples sufficient to do ANOVA and the normality of the data has not been tested to do so. In order to perform an analysis of variance (ANOVA), the following requirements must be met: Independence: the observed values are independent of each other. One way to ensure the independence of observations is by adding randomness to the sampling process. Homoscedasticity: there must be homogeneity in the variances, that is, the variability of the residuals is constant. Normality: the residuals must be normally distributed, or in other words, they must follow a normal distribution. Continuity: the dependent variable must be continuous. with less than 30 individuals these conditions are difficult to have. I suggest use only non-parametrical statistics
Response 2: Thank you for pointing this out. ANOVA and PCA have been applied for two different scopes: 1.      ANOVA and Welch’s ANOVA were used to investigate significant differences in MA and SI among species. The number of the specimens was low, but in many cases of morphometric investigation the small variance among specimens of the same species justifies the low number of samples. Normality and homoscedasticity were tested through Shapiro-Wilk and Levene tests. In our cases, all the assumptions for the use of ANOVA were satisfied for MA, while SI data resulted non-homoscedastic and Welch’s ANOVA was used. Now we have indicated it in Materials and Methods (Section 2.4 – Statistical analyses) and in the Results. 2.      PCA instead was performed to investigate which of the considered traits (not SI and MA, but single specific morphological traits) contribute to the differentiation among species.
4. Response to Comments on the Quality of English Language
Point 1: Minor editing
Response 1: The English was verified by the native speaking co-author.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting study of ecomorphology of seven species from the family Channichthyidae. This is a contribution to our knowledge of ecology of Antarctic fishes, which is limited and rare, due to access and logistical difficulties that represent to work in the antarctic ecosystem.The authors describe morphological traits associated with feeding behavior, related mainly with head and mouth structures. The study finds significant differences between species, relate to suction capabilities and the authors establish the ram-feeding as the main feeding mode among channichthids, in agreement with previous studies. 

I have some major comments and some suggestions for improvement, as well as a few minor comments. Once these comments are addressed, the paper would be of interest to the readership of Fishes and would make a fine contribution to the field and to the journal. The methodology need more details, such as the software used to perform different analyses, that should not be missed, mainly for reproducibility. Also, in some points I think it is not clear where the authors wants to highlight their results and guide the discussion. I think the author could extend the arguments and the discussion that leads to their main conclusions. 

 

Specific comments

 

Line 141. As I am not an expert on morphology, and this is a journal with a broader audience, I think you could improve the explanation of this section by adding a figure showing the different structures you measured to calculate SI and MA. Something similar to the reference you use (Carrol et al. 2004), but maybe with colors. This could help the readers, not experts in this field, and to the understanding of your results.

 

Line 184. In figure 1 caption, change “from left to right…” to where (a) correspond to the photo acquisition setting, (b) the registered images and the reconstructed points and (c) the 3D mesh, textured and with landmarks, and the final annotated model. 

 

Line 193. I imagine that the 3D representations took a lot of time and work, but could you explain why you did not make models for all your 7 species studied here? Are P. macropterus ans C. hamatus representative of all feeding behaviors observed in channictids? 

 

Line 197. Add to the figure caption: full 3D animations are provided in the supplementary material. 

 

Lines 200-203. Please provide more information of how and where (with what software) you transform your data and performed normality and ANOVAs analyses. 

 

Line 205. Six of seven species? Also, where did you performed the PCA?

 

Line 230. You did not mention the Tukey’s test in the material and methods section, please provide more details. 

 

Line 243. I really believe that the readers could benefit with the incorporation of a figure indicating how did you measure LinSI, LoutMA, etc.

 

Line 269. CSAepax is the same that CSAepaxialis? Please be consistent. 

 

Line 291. It seems N.ionah and P. macropterus would not correspond to one cluster, instead the individuals of both species did not overlap, so I would tend to say the are two separate clusters? 

 

Line 310. Except for P. macropterus and N. ionah? As you describe bellow? Or this species also exhibit ram feeding behavior? I think it is not very clear since you discuss the suction power of both. 

 

Line 312. “…much lower than any other Antarctic and non-Antarctic 312 fish studied so far…” How much lower? Please provide more information, like values for other fishes from the order Notothenioidei and others.

 

Line 322. “Due to its more developed epaxial muscle and longer head, P. 322 macropterus is able to produce a slight suction” Which is the evidence that support this? Do you have any references, or this is based on your results? The way it is now is not clear. 

 

Line 333. How much higher? Is it significant? Why do you expect this difference in MA values to imply a higher degree of flexibility in the diet?

 

Line 336. You did not mention in the discussion that you considered P. macropterus and N. ionah as one cluster. Would you consider them as on group due to similarities in suction capabilities or as different strategies? 

 

352. Here you started to discuss about the third cluster? I think you can organize better the discussion to highlight your results, if you continue to talk about 3 (or less) clusters that you identified, the differences between them and what factors influence the clustering of different species. Also, I think it is not clear how your discussion benefits with the figure 5, since the relationship between C. wilsoni and C. hamatus is closer than the one you are trying to highlight, but this species exhibit different ecomorphology. 

 

Line 357. Is there any reason why you did not discuss phylogenetic relations among the other species? It seems you only take this information into account comparing the species in the cluster including C. wilsoni, C. dewetti and P. antarcticus.

 

Line 381. It seems to me that you attempt to discuss about phylogenetic relationships, but not providing sufficient references or linking your result of each species (or clusters) described in your results section. If you really want to go that way I will suggest improving your discussion in that aspect. How do you give a phylogenetic explanation when you have more related species (e.g. C. wilsoni and C. hamatus) forming two different cluster and grouped with other species? Maybe this is not what you want to highlight, but the way is written in the current manuscript it is not clear. You mentioned the hypothesis of Parker et al. 2022, but there is much information that you could discuss, such as their finding of repeated ecomorphological convergences among distantly related species of icefishes.

 

Table 2. Please change P. macropeterus by P. macropterus

 

Figure 3. Where did you construct the boxplots?

 

 

Figure 4. (a) Please change Dim 1 and Dim 2 for PC1 and PC2 respectively, (b) Please provide axis label (PC1 and PC2). Also for (b), you did not mention it in the materials and methods section. Please provide more information of the analyses performed and the software used. 

Author Response

1. Summary
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files
2. Questions for General Evaluation	Reviewer’s Evaluation	Response and Revisions
Does the introduction provide sufficient background and include all relevant references?	Yes
Are all the cited references relevant to the research?	Yes
Is the research design appropriate?	Yes
Are the methods adequately described?	Must be improved	A figure has been added to clarify how the measures used for the analyses had been taken. Furthermore, some more details added and some parts have been re-written.
Are the results clearly presented?	Can be improved	The major comments from the authors have been taken into account. In particular, we have improved some of the  figures to better highlight the results of the statistical analyses
Are the conclusions supported by the results?	Can be improved	The Discussion has now been reorganized and improved following the major suuggestions
3. Point-by-point response to Comments and Suggestions for Authors This is an interesting study of ecomorphology of seven species from the family Channichthyidae. This is a contribution to our knowledge of ecology of Antarctic fishes, which is limited and rare, due to access and logistical difficulties that represent to work in the antarctic ecosystem. The authors describe morphological traits associated with feeding behavior, related mainly with head and mouth structures. The study finds significant differences between species, relate to suction capabilities and the authors establish the ram-feeding as the main feeding mode among channichthids, in agreement with previous studies.  I have some major comments and some suggestions for improvement, as well as a few minor comments. Once these comments are addressed, the paper would be of interest to the readership of Fishes and would make a fine contribution to the field and to the journal. The methodology need more details, such as the software used to perform different analyses, that should not be missed, mainly for reproducibility. Also, in some points I think it is not clear where the authors wants to highlight their results and guide the discussion. I think the author could extend the arguments and the discussion that leads to their main conclusions.    Thank you for your comments and suggestions. As suggested, more details were given to the Methods. The software used for statistical analysis and a figure (Figure 1) showing the ecomorphological measures taken into account have been added. Also the Discussion was reorganised and expanded based on our results.
Comments 1: Line 141. As I am not an expert on morphology, and this is a journal with a broader audience, I think you could improve the explanation of this section by adding a figure showing the different structures you measured to calculate SI and MA. Something similar to the reference you use (Carrol et al. 2004), but maybe with colors. This could help the readers, not experts in this field, and to the understanding of your results.
Response 1: As suggested, we added a figure (Figure 1) showing the ecomorphological measures taken into account.
Comments 2: Line 184. In figure 1 caption, change “from left to right…” to where (a) correspond to the photo acquisition setting, (b) the registered images and the reconstructed points and (c) the 3D mesh, textured and with landmarks, and the final annotated model.
Response 2: As suggested, it was replaced.
Comments 3: Line 193. I imagine that the 3D representations took a lot of time and work, but could you explain why you did not make models for all your 7 species studied here? Are P. macropterus and C. hamatus representative of all feeding behaviors observed in channictids?
Response 3: Indeed modeling the 3D representation for animation and modeling the two different animations was time consuming; therefore we decided to give priority to the rendering of the two fishes with opposite behavior, the P. macropterus and the C. hamatus.
Comments 4: Line 197. Add to the figure caption: full 3D animations are provided in the supplementary material.
Response 4: As suggested it was added.

Comments 5: Lines 200-203. Please provide more information of how and where (with what software) you transform your data and performed normality and ANOVAs analyses. 

 

Response 5: More details were added in section 2.4. – Statistical analyses

 

Comments 6: Line 205. Six of seven species? Also, where did you performed the PCA?

Response 6: It has been corrected. As it is now indicated, all the statistical analyses were performed using R software.

 

Comments 7: Line 230. You did not mention the Tukey’s test in the material and methods section, please provide more details. 

 Response 7: We have now added it in the Material and Methods section.

 

Comments 8: Line 243. I really believe that the readers could benefit with the incorporation of a figure indicating how did you measure LinSI, LoutMA, etc.

 

Response 8: A figure was added (Figure 1) showing the ecomorphological measures taken into account.

 

Comments 9: Line 269. CSAepax is the same that CSAepaxialis? Please be consistent. 

 

Response 9: It has been corrected.

 

Comments 10: Line 291. It seems N.ionah and P. macropterus would not correspond to one cluster, instead the individuals of both species did not overlap, so I would tend to say the are two separate clusters? 

 

Response 10: We hope that the dendrogram that we have now added could explain the three clusters that resulted. N. ionah and P. macropterus are in the rightmost group, identified by green circle.

 

Comments 11: Line 310. Except for P. macropterus and N. ionah? As you describe bellow? Or this species also exhibit ram feeding behavior? I think it is not very clear since you discuss the suction power of both. 

 

Response 11: We have now specified how the sentence referred to a first general observation, while a more in-depth analysis, as the one we carried out, allows us to more detailed conclusions. We can state that icefishes are overall ram feeders, but the different species present different peculiarities that result in small differences in feeding modes (e.g. the light suction capability indicated for P. macropterus and N. ionah)

 

Comments 12: Line 312. “…much lower than any other Antarctic and non-Antarctic fish studied so far…” How much lower? Please provide more information, like values for other fishes from the order Notothenioidei and others.

 

 Response 12: SI and MA ranges available in the literature were added. The sentence has been modified: ". The shape and size of the icefish heads and mouths also explain the very low values of their biomechanical indices, much lower than any other Antarctic and non-Antarctic fish studied so far, which show ranges of SI values between 0.02 and 0.40, and MA values between 0.12 and 0.68”.

 

Comments 13: Line 322. “Due to its more developed epaxial muscle and longer head, P. macropterus is able to produce a slight suction” Which is the evidence that support this? Do you have any references, or this is based on your results? The way it is now is not clear. 

 

Response 13: It can be inferred from the value of SI. In the text it was clarified and the sentence was changed “Our results suggest that P. macropterus, due to a more developed epaxialis muscle than the other icefish investigated, is able to produce a slight suction.”

 

Comments 14: Line 333. How much higher? Is it significant? Why do you expect this difference in MA values to imply a higher degree of flexibility in the diet?

 

Response 14: We think that the best answer to these questions is the results of the analysis of variance and in the Figure 4b, where boxplots show clearly at what extent MA of P. macropterus is lower than those of other species (comprised N. ionah). The higher flexibility of N. ionah is inferred since the species can add to the possibilities give from SI, those from the values of MA, which cannot have P. macropterus.

 

Comments 15: Line 336. You did not mention in the discussion that you considered P. macropterus and N. ionah as one cluster. Would you consider them as on group due to similarities in suction capabilities or as different strategies? 

 

Response 15: They are clustered together because of similar morphological characteristics linked to suction index. And in the improved Discussion all the clusters are now indicated.

 

Comments 16: 352. Here you started to discuss about the third cluster? I think you can organize better the discussion to highlight your results, if you continue to talk about 3 (or less) clusters that you identified, the differences between them and what factors influence the clustering of different species. Also, I think it is not clear how your discussion benefits with the figure 5, since the relationship between C. wilsoni and C. hamatus is closer than the one you are trying to highlight, but this species exhibit different ecomorphology. 

 

Response 16: As suggested, the discussion has been better organised, following the cluster groupings, in order to highlight our results. Figure 6 has been included to highlight phylogenetic relationships among the species analysed.

 

Comments 17: Line 357. Is there any reason why you did not discuss phylogenetic relations among the other species? It seems you only take this information into account comparing the species in the cluster including C. wilsoni, C. dewetti and P. antarcticus.

 

 Response 17: As suggested, the discussion was developed on the basis of the groupings provided by the cluster analysis. We have found two case that can drive the clustering of a group of species: i) phylogenetic driver; ii) prey driver.

 

Comments 18: Line 381. It seems to me that you attempt to discuss about phylogenetic relationships, but not providing sufficient references or linking your result of each species (or clusters) described in your results section. If you really want to go that way I will suggest improving your discussion in that aspect. How do you give a phylogenetic explanation when you have more related species (e.g. C. wilsoni and C. hamatus) forming two different cluster and grouped with other species? Maybe this is not what you want to highlight, but the way is written in the current manuscript it is not clear. You mentioned the hypothesis of Parker et al. 2022, but there is much information that you could discuss, such as their finding of repeated ecomorphological convergences among distantly related species of icefishes.

 

Response 18: See response 17.

 

Comments 19: Table 2. Please change P. macropeterus by P. macropterus

 

Response 19: It has been corrected.

 

Comments 20: Figure 3. Where did you construct the boxplots?

 

Response 20: More details about the used software (R) were added in the Material and Methods section.

 

Comments 21: Figure 4. (a) Please change Dim 1 and Dim 2 for PC1 and PC2 respectively, (b) Please provide axis label (PC1 and PC2). Also for (b), you did not mention it in the materials and methods section. Please provide more information of the analyses performed and the software used. 

 

Response 21: The labels of the plot (a) have been corrected. Plot (b) was changed with a dendrogram. All the graphs were performed using R software, as it is now indicated in the Methods.

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This paper measures nine morphological traits across seven species of icefish and summarizes the ecomorphological differences using multivariate analysis. Additionally, it visualizes feeding modes with 3D animations, providing interesting results.

The paper adequately summarizes, organizes, and compares the findings of this study, drawing conclusions. While there are no unusual findings compared to previous results, it provides satisfactory results and discussions.

However, I would like to suggest a few minor revisions to more effectively communicate the results of this paper.

 

1.     In Figure 1(c), the parameters measured are not clearly visible. It would be beneficial if the measurement methods for the parameters presented in Table 3 were clearly indicated.

2.     Could you clarify what the coefficients in Tables 2 and 3 are related to?

3.     For Figure 3, it would be helpful if the results of the ANOVA post-hoc test (Tueky test) were indicated on the figure. For example, groups that do not significantly differ could be marked with the same letter (a, b, c….).

4.     In both Figure 4 (a) and (b), the markers are too small to distinguish by color. The figures should be modified to more clearly differentiate between species or clusters. Additionally, it would be beneficial if the figures also delineated the three identified clusters with lines, accompanied by information on what percentage of similarity was used as the criteria for their differentiation.

5.     Discussion : This study reproduced species-specific feeding behaviors using 3D animations, and due to developments in modeling and 3D tools, similar research approaches are expected to increase in the future. It would be beneficial if the authors could further elaborate in the discussion section of this paper on the results obtained through the 3D animation approach, the potential future applications, issues and constraints, and the limitations encountered.

Author Response

1. Summary
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files
2. Questions for General Evaluation	Reviewer’s Evaluation	Response and Revisions
Does the introduction provide sufficient background and include all relevant references?	Yes
Are all the cited references relevant to the research?	Yes
Is the research design appropriate?	Yes
Are the methods adequately described?	Can be improved	A figure has been added to clarify how the measures used for the analyses had been taken. Furthermore, some more details added and some parts have been re-written.
Are the results clearly presented?	Can be improved	The major comments from the authors have been taken into account. In particular, we have improved some of the  figures to better highlight the results of the statistical analyses
Are the conclusions supported by the results?	Yes
3. Point-by-point response to Comments and Suggestions for Authors This paper measures nine morphological traits across seven species of icefish and summarizes the ecomorphological differences using multivariate analysis. Additionally, it visualizes feeding modes with 3D animations, providing interesting results. The paper adequately summarizes, organizes, and compares the findings of this study, drawing conclusions. While there are no unusual findings compared to previous results, it provides satisfactory results and discussions. However, I would like to suggest a few minor revisions to more effectively communicate the results of this paper.   Thank you for your suggestions and comments.
1.       Comments 1: In Figure 1(c), the parameters measured are not clearly visible. It would be beneficial if the measurement methods for the parameters presented in Table 3 were clearly indicated.
Response 1: The figure was modified, the landmarks, which were used for 3D animation, were removed to avoid confusion with the measured morphological traits. Table 3 indeed do not refers to measures indicated in the figure that you have cited, but they refer to the new added Figure 1.
Comments 2: Could you clarify what the coefficients in Tables 2 and 3 are related to?
Response 2: In Table 2 are the values of suction index (SI) and mechanical advantage (MA) in the seven species, and in Table 3 there are all morphological measurements acquired from the samples. We hope that adding the new Figure 1 it is now clear how we obtained the measures and the indexes
Comments 3: For Figure 3, it would be helpful if the results of the ANOVA post-hoc test (Tueky test) were indicated on the figure. For example, groups that do not significantly differ could be marked with the same letter (a, b, c….)
Response 3: As suggested, in figure 4, the letters have been added to highlight significant differences among species.
Comments 4: In both Figure 4 (a) and (b), the markers are too small to distinguish by color. The figures should be modified to more clearly differentiate between species or clusters. Additionally, it would be beneficial if the figures also delineated the three identified clusters with lines, accompanied by information on what percentage of similarity was used as the criteria for their differentiation.
Response 4: The markers of Figure 4 (a) (now Figure 5 (a)) have been enlarged. In addition, to increase the visibility of species groupings, Figure 4 (b) (now Figure 5 (b)) has been changed with a hierarchical cluster.

Comments 5: Discussion: This study reproduced species-specific feeding behaviors using 3D animations, and due to developments in modeling and 3D tools, similar research approaches are expected to increase in the future. It would be beneficial if the authors could further elaborate in the discussion section of this paper on the results obtained through the 3D animation approach, the potential future applications, issues and constraints, and the limitations encountered.

 

 Response 5: Concerning the 3D technology used to create digital copies of the real specimens and animation sequences, we believe that they are a useful tool to support the conservation, study and dissemination of biological species. As mentioned, the models for animation were manually modelled on the gross shape of the digitised models. Indeed, due to the shape modifications induced by the pose on the rotating plane during the shooting stage and the deformation occurred after the long conservation in ethanol or IPA made the shape of the specimen far from the original aspect of the alive exemplar. However, this required a considerable manual work.

 

Conversely, if the digitisation happens as soon as the specimen is collected, we probably could directly use the acquired model for animation and for annotation and documentation as well, since metric measurements would be more reliable on an under formed smooth shape. However, fishes in general, and Antarctic species in particular, are particularly difficult to digitize, behind the deformable shape issue: they are typically very reflective and might be semi-transparent, facts that increase the noise in the resulting point cloud or reduces the coverage of points on the target surface. Furthermore, fins are very thin, causing points on the one side of the fin very close to points on the opposite, and consequently, poor meshing. However, results shown by Kano et al. are promising in this direction.

We have added the above concepts in the final section of the Discussion.

 

Author Response File: Author Response.pdf