Review Reports - Soil Invertebrate Communities as Indicator of Ecological Conservation Status of Some Fertilised Grasslands from Romania

Round 1

Reviewer 1 Report

The study is worthy to be published because of its sampling size (efforts), i.e. in total 250 soil samples collected, given that in the country such data are still limited. However, the way of presentation of the study needs further efforts.

I could not find well the knowledge gaps clearly written in the present manuscript. Hypotheses are not presented.

Table 1 is missing.

A map showing study sites is recommended.

In the result section, the term "biotic" and "abiotic" factors should be exchanged.

In the figure 3 it should be NMDS not NMSD.

Author Response

Reviewer: The study is worthy to be published because of its sampling size (efforts), i.e. in total 250 soil samples collected, given that in the country such data are still limited. However, the way of presentation of the study needs further efforts.

Answer: Thank you very much for your consideration.

Reviewer: I could not find well the knowledge gaps clearly written in the present manuscript.

Answer: Please, see the lines 94-98. “In the Europe, as well as in Romania, this multi-taxa approach was not so used. In Europe, many researches are focused only of one taxonomic group or few, without taking into consideration the complex interrelations between edaphic communities and their relations with soil abiotic and biotic parameters, in experimental grasslands, organically and chemical fertilized, situated on high altitude, in a mountain area [4, 7, 22, 26, 34, 37].”

Reviewer: Hypotheses are not presented.

Answer: Please, see the penultimate paragraph from the introduction: “Considering all these informations, the objectives of the present paper are: 1) to investigate the structural communities’ differences between soil fauna from five fertilised experimental plots and 2) to demonstrate the relation between five environmental factors and the community structures of soil invertebrates.”

I consider that the objectives of the present study are the hypotheses of the research.

Reviewer:Table 1 is missing.

Answer: The table 1 was inserting.

Reviewer:A map showing study sites is recommended.

Answer: The map for this research is similar with those published in the paper from references with no. 37: Manu, M., Băncilă, R.I., Mountford, O.J., Maruşca, T., Blaj, V.A., Onete, M., 2022. Soil Mite (Acari: Meso-stigmata) Communities and Their Relationship with Some Environmental Variables in Experimental Grasslands from Bucegi Mountains in Romania. Insects, 13(3), 285; https://doi.org/10.3390/insects13030285.

This was the reason that I don’t insert any map, but only the references!

Reviewer:In the result section, the term "biotic" and "abiotic" factors should be exchanged.

Answer: The terms “biotic” and “abiotic” were exchanged

Reviewer:In the figure 3 it should be NMDS not NMSD.

Answer: According to another reviewer request, the figure 3 was deleted.

THANK YOU VERY MUCH FOR YOUR ADVICES AND YOUR EFFORT!

Reviewer 2 Report

Dear authors, unfortunately, I need to recommend a major revision to your article. I detected numerous mistakes related to English writing and grammar. Ecology is a subject that needs to be very well explained and without clearness in your sentences, it is impossible to understand your results. Also, I think you can reduce the number of figures and tables in the main text, and bring the supplementary table to the main text, It has important information that should be easily accessed by the readers.

Author Response

Reviewer: Dear authors, unfortunately, I need to recommend a major revision to your article. I detected numerous mistakes related to English writing and grammar. Ecology is a subject that needs to be very well explained and without clearness in your sentences, it is impossible to understand your results. Also, I think you can reduce the number of figures and tables in the main text, and bring the supplementary table to the main text; it has important information that should be easily accessed by the readers.

Answer: Thank you very much for your recommendations!

The English language checking was made by a native speaker, Mr. Owen John Mountford, which is also one of the coauthor of the paper (twice).

According to another reviewer requests the figures 1 and 3 were deleted. The Table 4 was moved to the Supplementary Material.

Significant modifications were made in the results, discussions and conclusions chapters. Please, see the manuscript!

THANK YOU VERY MUCH FOR YOUR ADVICES AND YOUR EFFORT!

Reviewer 3 Report

Overall, I feel that this research has the potential to make important contributions to our understanding of the role of soil invertebrates as bioindicators. With major revisions this study could shown insights into the use of soil invertebrates as bioindicators, however as currently written, this does not come across as the primary point of this paper. I think that refocusing this study around the point of bioindicators (especially the discussion section) would be very interesting.

Line 16 – “from” should be “for”

Lines 15-31 – Abstract can be streamlined, and some methodological details cut to make it more impactful. Should expand on last sentence which is the take-away of the paper.

Lines 116-117 – It would be beneficial to at least mention the management treatments in the study area briefly, as some individuals may not be able to access the other papers referenced.

Lines 179-189 – The methods only mention that a Berlese-Tullgren extraction was conducted on the soil samples, which is appropriate for most of the soil fauna measured. However, this extraction technique is not accurate for measuring nematodes. On line 184 you indicate that another extraction was conducted, but it is not clear what that extraction was for?

Further, the soil sampling method used is not ideal for getting accurate estimates of many of the megafauna (Opiliones, Araneae, Isopoda, Coleoptera, etc.).

Generally, section 2.3 needs further clarification and to address these considerations mentioned above.

Lines 196-199 – There is a table caption here, but no table? It reads as if this table should be in the results section? I see Table 1 is mentioned in the text on line 210, but it is not present.

Line 216 – Formatting of table 2 can be improved with an additional heading to indicate that the letters represent the different grassland (especially since the CG does not follow the general naming pattern of the grasslands). Within table 2 you list Acaridae as a taxon group, and then you also list Oribatida as another taxon group. Acaridae is a family within Astigmata mites, which is now grouped within Oribatida. It is unclear how this grouping occurred? Or why one taxon of mites was potentially identified to family level, but the others were identified to order level?

Lines 229-233 – I question whether this information adds any value to this paper. While the data from the Mao-Tau curves can be useful to inform the number of appropriate soil samples for future research, you do not make this point in the paper. Further, you do not use this data to make a point about using soil fauna as bioindicators either.

Line 244 – Figure 2 looks compressed vertically, which makes it more difficult to read.

Lines 249-251 – NMDS ordination is typically not good at separating treatments when using high level taxonomic resolution as done in this study. Since there are no treatment differences, I question the value of reporting this in this paper since it does not contribute an important point to the discussion [lines 322-326 mention these results, but it does not really add anything to the storyline of the paper].

Lines 255-256 – I question whether the ANOSIM results were interpreted correctly based on the wording of this sentence. An ANOSIM tests whether distances between groups are greater than within groups. PERMANOVA tests whether distance differ between groups. The wording of this sentence sounds more like the ANOSIM was interpreted as a PERMANOVA result.

Line 266 – Typically model selection results are not shown like this in the results section, I would more this to the Supplementary materials and just state the results of the best model.

Lines 287-321 – This is a lot of details about other studies with no clear point to the paragraph until the last sentence. This paragraph can be reworked to be more impactful and clearer.

Lines 450-471 – This is a summary of the entire paper, not a strong conclusion section with take-away points.

Lines 473-475 – As it is currently written, this discussion does not connect the results to the importance of soil invertebrate communities as indicators.

Overall, the writing can be made more direct and concise to improve the clarity of the paper, especially in the introduction section. Many of the sentences throughout the paper were unclear, most notable on lines 40-42, 52-53, 60-61, 92, 104-107, 201-203, 219-220, 331, 361-367.

There are a lot of figures and tables that could be condensed or removed to improve the clarity of this paper and make it more direct and impactful.

There were many results statements reiterated within the discussion section, but the discussion did not really elaborate on these results as much as desired [lines 327, 331-335, 343, etc.]. Each paragraph within the discussion read as isolated points, and the overall point of the paper and the main take-aways were unclear. It could be beneficial to consolidate the discussion section under central themes [i.e. effects of soil hardness (resistance to penetration), bioindicators, etc.].

Author Response

Reviewer: Overall, I feel that this research has the potential to make important contributions to our understanding of the role of soil invertebrates as bioindicators. With major revisions this study could show insights into the use of soil invertebrates as bioindicators, however as currently written, this does not come across as the primary point of this paper. I think that refocusing this study around the point of bioindicators (especially the discussion section) would be very interesting.

Reviewer: Line 16 – “from” should be “for”

Answer: The change was made.

Reviewer: Lines 15-31 – Abstract can be streamlined and some methodological details cut to make it more impactful. Should expand on last sentence which is the take-away of the paper.

Answer: Thank you for your recommendation. The modifications in abstract were made. Please, see line 21, 32 and 33.

Reviewer: Lines 116-117 – It would be beneficial to at least mention the management treatments in the study area briefly, as some individuals may not be able to access the other papers referenced.

Answer: A briefly presentation of the management treatments of the experimental plots was inserted. Please, see the last paragraph from the subchapter 2.1 Study area.

“The control plot (CG) is a semi-natural pasture, unimproved, grazed in summer with cows; grassland A is a semi-natural pasture, fertilized with chemical fertilizers in periods 2000-2002, 2010-2012, 2014-2016, with an average 100 kg ha‑1 N + 50 kg ha‑1 P2O5 + 50 kg ha‑1 K2O; grassland B is a semi-natural pasture, chemically fertilized in 1996-1998 with 150 kg ha‑1 N + 75 kg ha‑1 P2O5 + 75 kg ha‑1 K2O and organically fertilized in 2004, 2010 and 2016 by paddocking with dairy cows; grassland C is a semi-natural pasture, calcium liming in 1995, chemically fertilized similar to the B variant and paddocked in 2003, 2009 and 2015; and grassland D is a reseeded pasture in late summer of 1995, after herbicide Roundup at 5 l ha‑1, calcium liming and chemically fertilized similar to C variant and paddocked in 2002, 2008 and 2014 [38].

Reviewer: Lines 179-189 – The methods only mention that a Berlese-Tullgren extraction was conducted on the soil samples, which is appropriate for most of the soil fauna measured. However, this extraction technique is not accurate for measuring nematodes. On line 184 you indicate that another extraction was conducted, but it is not clear what that extraction was for?

Further, the soil sampling method used is not ideal for getting accurate estimates of many of the megafauna (Opiliones, Araneae, Isopoda, Coleoptera, etc.).

Generally, section 2.3 needs further clarification and to address these considerations mentioned above.

Answer: I am agreeing with the reviewer, but the main reason of usage Tullgren-Berlese method, was to identify especially the mesofauna (soil mites). After I counted and identified on species level (Mesostigmata- for other purpose), I realized that it is a lot of biological material of other soil fauna groups, and this material must be presented to others and published. The specialists and data on soil fauna are less and less in Europe and I considered that any information in this direction is precious, even if the used collecting method is limited.

On line 184 I indicated that the “Berlese–Tullgren method, using natural light and heat”. According to the Berlese method, it is used the light bulb burning, the light and heat that were produced created gradients of three different physical conditions. All three gradients caused animals to move downward. We used natural light and heath, in order to not create sweat on the funnels, being possible that soil fauna (especially immatures) remaining agglutinated of the funnels. And it is cheaper.

Reviewer: Lines 196-199 – There is a table caption here, but no table? It reads as if this table should be in the results section? I see Table 1 is mentioned in the text on line 210, but it is not present.

Answer: The table 1 was inserting!

Reviewer: Line 216 – Formatting of table 2 can be improved with an additional heading to indicate that the letters represent the different grassland (especially since the CG does not follow the general naming pattern of the grasslands). Within table 2 you list Acaridae as a taxon group, and then you also list Oribatida as another taxon group. Acaridae is a family within Astigmata mites, which is now grouped within Oribatida. It is unclear how this grouping occurred? Or why one taxon of mites was potentially identified to family level, but the others were identified to order level?

Answer: The heading of the table 2 was improved with: “(CG= control plot; A, B, C, D = fertilized experimental plots)”.

Regarding the Acaridae family, I admit that the taxonomical expertise was limited. The Acaridae, Trombidiidae, Glycyphagidae are families that were easy to identify them. The Oribatida order (except Acaridae), contains many other families, that request a specialized expertise in taxonomical identification and I want to avoid the wrong identification. We have experience in soil fauna groups taxonomy only for few years.

Reviewer: Lines 229-233 – I question whether this information adds any value to this paper. While the data from the Mao-Tau curves can be useful to inform the number of appropriate soil samples for future research, you do not make this point in the paper. Further, you do not use this data to make a point about using soil fauna as bioindicators either.

Answer: We agree with the reviewer and we deleted the figure and the information about the taxa accumulation curve.

Reviewer: Line 244 – Figure 2 looks compressed vertically, which makes it more difficult to read.

Answer: We replaced the figure.

Reviewer: Lines 249-251 – NMDS ordination is typically not good at separating treatments when using high level taxonomic resolution as done in this study. Since there are no treatment differences, I question the value of reporting this in this paper since it does not contribute an important point to the discussion [lines 322-326 mention these results, but it does not really add anything to the storyline of the paper].

Answer: We removed the NMDS from data analysis and results sections

Reviewer: Lines 255-256 – I question whether the ANOSIM results were interpreted correctly based on the wording of this sentence. An ANOSIM tests whether distances between groups are greater than within groups. PERMANOVA tests whether distance differ between groups. The wording of this sentence sounds more like the ANOSIM was interpreted as a PERMANOVA result.

Answer: We reformulated as follows; The ANOSIM tests showed significant differences among treatments.

Reviewer: Line 266 – Typically model selection results are not shown like this in the results section, I would more this to the Supplementary materials and just state the results of the best model.

Answer: The Table 4 was moved to the Supplementary Material. Remain only the state of the best model.

Reviewer: Lines 287-321 – This is a lot of details about other studies with no clear point to the paragraph until the last sentence. This paragraph can be reworked to be more impactful and clearer.

Answer: Thank you for you advices.

The paragraph was changed, and some information was added, as following: “For highlighting the similarities or differences between Romanian and European data, we compared the results of studies developed on soil fauna from different types of grassland ecosystems from temperate regions of Europe. Thus, we stress the importance of soil fauna as bioindicators for the conservation status of this type of ecosystems in Europe, considering the geographical position and the climate. In the view of these considerations, many European studies focused on soil invertebrate communities from grasslands, revealed the high diversity and abundance of soil fauna taxa.”

Reviewer: Lines 450-471 – This is a summary of the entire paper, not a strong conclusion section with take-away points.

Answer: the modification was made, as following: “In this study, we gained detailed knowledge about the soil fauna communities in experimental grasslands from Romania. The communities’ structure of soil fauna, evaluated through some parameters (as numerical abundance, taxa richness, dominance, Shannon diversity index of taxa and equitability) were influenced by the type of fertilization and by the local biotic and abiotic drivers (as vegetation coverage, soil temperature, soil acidity, soil resistance at penetration, soil moisture content). The composition of investigated grasslands revealed the presence of seventeen soil invertebrates groups, with different spatial distribution. Natural grassland (control plot) was characterized by an equilibrate distribution of the numerical abundance, of the soil fauna. Chemical fertilization has a positive effect on numerical abundance and dominance of soil taxa, with a major contribution of Collembola, Acaridae and Oribatida. Organic and chemical fertilization, together with calcium liming, has positive effects on taxa diversity. In the same time, this study revealed that soil fauna groups are sensitive to the abiotic and biotic factors, confirming their quality as bioindicators. Oribatida and Enchytraeidae were sensitive to the vegetation coverage, Nematoda and Coleoptera to the soil resistance at penetration.”

Reviewer: Lines 473-475 – As it is currently written, this discussion does not connect the results to the importance of soil invertebrate communities as indicators.

Answer: The change was made, as following: “A responsible usage of fertilizers, in recommended limits, could influence in a positive way the soil fauna groups. The present study demonstrates that soil invertebrate communities could be precious indicators of ecological conservation status of some fertilized grasslands.”

Reviewer: Overall, the writing can be made more direct and concise to improve the clarity of the paper, especially in the introduction section. Many of the sentences throughout the paper were unclear, most notable on lines 40-42, 52-53, 60-61, 92, 104-107, 201-203, 219-220, 331, 361-367.

Answer:

Lines 40-42, were changed as following: “Numerical abundance, community structure/composition, taxa diversity or relations between biotic - abiotic factors are some important parameters, which can be used to evaluate the role of soil fauna as bioindicators”.

Lines 52-53, were changed as following: “They are more abundant in the first 20 centimeters of soil. “

Lines 60-61, were changed as following:” In many ecological studies, the soil fauna taxa have been used as biological indica-tors. Higher taxa are widely used as surrogates for species-level identification in invertebrates, being a cost-effective approach to monitoring impacts of anthropogenic disturbance.”

Line 92, was changed as following: “In Europe, as well as in In Romania, this multi-taxa approach in soil biodiversity assessment of the experimental grasslands, was not so used.”

Lines 104-107, were changed as following:” The present study provides valuable information regarding the influence of grasslands management practices (fertilization) on the soil invertebrate communities. This study can constitute an assessment instrument of the conservation status of grasslands, using soil fauna as bioindicators”.

Lines 201-203, were changed as following: “Considering the environmental factors, we quantified one biotic factor (vegetation cover) and four abiotic parameters (soil temperature, soil pH, soil moisture content and soil resistance at penetration).”

Lines 219-220, were changed as following: “The highest value of Shannon diversity index was recorded in plot D (Table 2).”

Line 331, was changed as following: “Our research team up with similar studies, confirming that the soil fauna groups as Enchyraeidae, Lumbricidae, Collembola, Mesostigmata, Acaridae, Oribatida and insect larva are bioindicators for different types of grasslands from Europe, being used in monitoring programs [1, 6, 10, 11, 21, 58, 62-64].”

Lines 361-367, were changed as following: “Taking into consideration the number of identified taxa and their abundance, significant differences were recorded between plots B and CG (control plot). The number of taxa was higher and their abundance was lower in plot B, in comparison with plot CG. This experimental plot (B) was characterized by the highest average values of soil resistance at penetration (1.39 Map) and by the most acid soil (4.55), in comparison with the other investigated grasslands.”

Reviewer: There are a lot of figures and tables that could be condensed or removed to improve the clarity of this paper and make it more direct and impactful.

Answer: The figures 1 and 3 were deleted. Table 4 was moved to the Supplementary Material.

Reviewer: There were many results statements reiterated within the discussion section, but the discussion did not really elaborate on these results as much as desired [lines 327, 331-335, 343, etc.]. Each paragraph within the discussion read as isolated points, and the overall point of the paper and the main take-aways were unclear. It could be beneficial to consolidate the discussion section under central themes [i.e. effects of soil hardness (resistance to penetration), bioindicators, etc.].

Answer: from the line 327 to the 343 the following modification was made: “Our research team up with similar studies, confirming that the soil fauna groups as Enchyraeidae, Lumbricidae, Collembola, Mesostigmata, Acaridae, Oribatida and insect larva are bioindicators for different types of grasslands from Europe, being used in monitoring programs [1, 6, 10, 11, 21, 58, 62-64]. We also identified, these seven edaphic communities in all five investigated experimental plots (control plot, organically and chemical fertilized grasslands) from Bucegi Mountains, Romania.

Analyzing the invertebrates’ fauna from the five investigated plots, we found dif-ferent patterns in soil community structure and composition, shaped by the biotic and abiotic factors shaped these patterns. Plot A, was characterized by the highest percentage of vegetation coverage, the highest average value of soil moisture content and the lowest soil resistance at penetration, which means higher porosity. These environmental condi-tions created favorable habitats for Lumbricidae, Acaridae and Oribatida, which recorded the highest values of numerical abundance. Earthworms recorded greatest number and biomass in the soil of the northern exposition with a highest soil moisture content, bur recorded significant decrease in their population density (63%) as a response to the reduction of plant species richness. Other studies revealed that a low abundance and diversity of Oribatida reflects low soil porosity [19, 65-67].”

THANK YOU VERY MUCH FOR YOUR ADVICES AND YOUR EFFORT!

Round 2

Reviewer 1 Report

As said, such extensive study is valuable for the region regarding its faunal documentation and results and data definitely deserve publication. However, the writing and the analyses of the study are not satisfying yet. In this revised manuscript there are several faults that the authors certainly have to deal with:

(1) One oblivious mistake is in the description of the results concerning abundance and species richness of the treatment B. In Figure 3 (line 329) Treatment B is the lowest in abundance among the others. Figure 4 shows the same pattern, i.e., the lowest in species richness (line 332). However, in lines 322-326 the authors have written “… abundance was significantly ‘higher’ only in treatment… taxon richness was significantly ‘higher’ also only in treatment B …”.

(2) The second one is description and interpretation for the CCA result. That was fine to select only two environmental variables if the other variables are correlated with the two (…although I did not see the rationale for only two variables included. It may be simply because that the VIF told these two the most sufficient ones. But, if the “meaningful (line 285)” variables are “only two”, it is probably not necessary to do a “multivariate” analysis.). However, interpretation of the taxa correlated with each of the axis (thus environmental variables) needs cautions. The authors presented here that Oribatida and Enchytreidae significantly correlated with VegCovr. However, the taxa in the opposite direction along the first axis and with the highest scores (i.e., Glycpyhg, Isopoda and Opiliond) are actually those the most negatively correlated with the VegCovr, and Formicod (with the highest score of CCA1) and potentially also Acaridae are the most positively correlated taxa with VegCovr. The Enchytrd, however, posits in the central point of the figure, with scores at CCA1 and CCA2 both equal to zero (i.e. not correlated at all). That is to say, what the authors concluded about taxa correlation with environmental variables were inadequate.

Base on the issues raised in the above-mentioned points, the manuscript could have been rejected, but I would love to see such nice study and the efforts that the authors have made give some impacts to soil zoological research communities. Therefore, I recommend a major revision of the manuscript.

Some other concerns are provided below:

Exclusion of NMDS plot from previous version of manuscript. I don’t know why the authors removed the NMDS but not ANOSIM in their revised manuscript. I would either keep the NMDS, or to exclude the ANOSIM, although I found that there was some information provided by the NMDS in their previous version of the manuscript, i.e. distance of community compositional similarity of between treatments. They could also redraw the NMDS by adding the mean position of each treatment and the size of variation in a form of ellipse (variance, SD, SE or 95% CI) for each treatment.

The unit of y-axis in both Figure 2 and 3 is wrong. It looks like that it should be multiplied by 100. Otherwise, the abundance (number of individuals) collected in each treatment was not consistent with that written in Table 2.

Figures 2, 3, and 4. Are the values shown here the mean “per soil sample” or “per treatment”?

Table 2. If the column “Total nr.ab.” is the total abundance per taxon, it should be the sum across all 250 soil samples. However, if the numerical abundances written in the columns CG, A, B, C and D are the “mean”, rather than the sum, there should be no standard error written together. Either write the mean (calculated based on 50 samples per treatment) ± standard error, or the sum (across 50 samples per treatment) without standard error. Otherwise, the information provided is misleading.

The dominance in the Table 2. Usually we calculate dominance for the “taxa”, not for treatment or site. The dominance written here is uninformative, and I do not know how the authors got the values for treatments, rather for taxa.

For readability, the authors may replace the names of treatment “A, B, C and D” with the other informative names that suggest the managements of the pastures.

The rationale of using GLMM to compare models with different combinations of environmental variables is unclear, and redundant. For sure from the abundance and richness figures (Figures 3 and 4) and the ANOSIM result, we have known that “treatment” has certain effects. In the GLMMs the author could have been interested in knowing what environmental factors driving such differences in abundance and richness, but in the model they erroneously take the treatment as a fixed factor but the other environmental variables as covariates and sample points as random factor. I would recommend to use environmental variables as fixed factors but treatment, that is actually each pasture from which the authors took 50 samples, as random factors. This will help to decipher which of the environmental factors predict the most to the abundance and richness, thereby to complement the results in Figure 3 and 4.

The structure of the section Materials and methods. Put the Data analysis to the last paragraph of the section.

I am looking forward to a fully revised manuscript “without contradiction throughout the paper”.

Author Response

Reviewer 1

Answer: The modifications were made in the manuscript, as following: “The Tukey multiple comparisons test indicated that abundance was significantly higher only in treatment C compared to the control (Table 5 and Figure 3). The taxon richness (df = 4, Chi square = 55.243, P < 0.001) was significantly differed among treatments (df = 4, Chi square = 22.613, P < 0.01). The Tukey multiple comparisons test showed that taxon richness was significantly higher also only in treatment D compared to the control (Table 5 and Figure 4).”

Answer: We agree with the reviewer that the interpretation of the CCA biplot was inadequate. Thank you for this comment and the opportunity to correct this. Vegetation cover was highly negatively corelated with the first axis (-0.943) and RPs was highly negatively correlated with the second axis (-0.947). Therefore, indeed the Glycpyhg, Isopoda and Opiliond were the most negatively correlated taxa with the VegCovr while the Formicod and Acaridae were the most positively correlated taxa with VegCovr. The Diplura, Lumbricd and Aranea were the most negatively correlated taxa with RPs and the Coleoptr, Formicod and Stphylnd were the most positively correlated taxa with RPs.

Some other concerns are provided below:

Answer:

We also removed ANOSIM. The NMDS was removed from the previous version of the manuscript because it did not provide significant information. Adding the ellipse (95 % CI) highlighted the overlap between treatments (please see the graph in attached file).

Answer:

Because we used GLMM to plot the effects we did not use the means and SE from the data but we used the effects R package to convert the parameter estimates into condition mean and SE estimates and then plot them. We explained this in the data analysis section.

Figures 2, 3, and 4. Are the values shown here the mean “per soil sample” or “per treatment”?

Answer: The values are means per treatment.

Answer: the standard errors were deleted! We have the sum across the 50 samples/treatment.

Answer: The dominance for the table 2 was calculated taking into consideration the identified taxa from each type of grassland.

For readability, the authors may replace the names of treatment “A, B, C and D” with the other informative names that suggest the managements of the pastures.

Answer: Thank you very much for the suggestion, but I described the type of pastures management in Material and methods chapter and more detailed in Supplementary material Table S1.

Answer: We aimed to investigate weather abundance and richness were affected only by treatment, only by environmental factors or by the additive effects of treatment and environmental factors. We entered the environmental variables as fixed factors. We referred to them as covariates because a factor is usually a categorical variable and a covariate is a continuous variable. However, since we used the term covariate in a GLMM not in an ANCOVA we believe that the formulation is correct and as well the construction of the models.

The structure of the section Materials and methods. Put the Data analysis to the last paragraph of the section.

Answer: The change was made!

I am looking forward to a fully revised manuscript “without contradiction throughout the paper”.

Thank you very much for your effort and for your time!

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear authors,

The manuscript was greatly improved after the review process. I have only a few considerations:

Page 1, line 31: This sentence is lost in this position, maybe it should be moved to the beginning;

Page 1, line 41: Delete space between 1- and 3;

Page 1, line 41: …, soil invertebrates’ communities have been used…;

Page 2, line 47: parameters;

Page 2, line 54: The soil fauna is classified … it comprises…

Page 2, line 71: The is no reference number 154;

Page 3, line 123: Please decide which English you want to use, sometimes you wrote “fertilized” and others “fertilised";

Page 4, line 166: … Vegan package v2.5-2…

Page 4, line 167: … Vegan package…

Page 5, line 214: “All the statistical...” Maybe this sentence can be moved to the beginning;

Page 5, line 217: “The present study was accomplished in July, 2017.” You already said that before;

Page 5, line 220: … five grasslands were collected…

Page 5, line 223: What do you mean by next extraction? I think it should be until the sorting process;

Page 5, line 225: Taxa identification was …

Page 5, line 252: … variables differentiated significantly between the five...

Page 10, line 311: Don’t need to write supplementary material, only Table S2.

Page 13, line 379: … affirm that…

Page 13, line 379: … the soil fauna from Romania is more diverse…

Page 13, line 395: … delete “shaped these patterns”.

Page 14, line 401: … but…

Page 14, line 436: … performed…

Page 15, line 458: … drastic reduction… will decrease…

Page 15, line 474: … explained…

Page 15, line 478: delete “a”

Author Response

Reviewer 2

The manuscript was greatly improved after the review process. I have only a few considerations:

Page 1, line 31: This sentence is lost in this position, maybe it should be moved to the beginning;

Page 1, line 41: Delete space between 1- and 3;

Page 1, line 41: …, soil invertebrates’ communities have been used…;

Page 2, line 47: parameters;

Page 2, line 54: The soil fauna is classified … it comprises…

Page 2, line 71: The is no reference number 154;

Page 3, line 123: Please decide which English you want to use, sometimes you wrote “fertilized” and others “fertilised";

Page 4, line 166: … Vegan package v2.5-2…

Page 4, line 167: … Vegan package…

Page 5, line 214: “All the statistical...” Maybe this sentence can be moved to the beginning;

Page 5, line 217: “The present study was accomplished in July, 2017.” You already said that before;

Page 5, line 220: … five grasslands were collected…

Page 5, line 223: What do you mean by next extraction? I think it should be until the sorting process;

Page 5, line 225: Taxa identification was …

Page 5, line 252: … variables differentiated significantly between the five...

Page 10, line 311: Don’t need to write supplementary material, only Table S2.

Page 13, line 379: … affirm that…

Page 13, line 379: … the soil fauna from Romania is more diverse…

Page 13, line 395: … delete “shaped these patterns”.

Page 14, line 401: … but…

Page 14, line 436: … performed…

Page 15, line 458: … drastic reduction… will decrease…

Page 15, line 474: … explained…

Page 15, line 478: delete “a”

Answer: All the requested modifications were made in the manuscript. Please, see the last version of it!

Thank you very much for your effort and for your time!

Reviewer 3 Report

I believe this manuscript is greatly improved, and is more direct and impactful. Most of the comments were addressed and the writing is clearer. I noticed several minor spelling and grammar mistakes throughout the manuscript, which can be fixed easily with another proof read.

Author Response

Reviewer 3

Answer: Modifications regarding the spelling and grammar were made in the manuscript. Please, see the last version of it!

Thank you very much for your effort and for your time!

Round 3

Reviewer 1 Report

After reading the revised manuscript, I was almost going to recommend a rejection for the manuscript, given that the authors did not pay sufficient attentions to the revision of the manuscript, especially in the descriptions of the results related to the figures 3 and 4. Also, they corrected only some parts following my last comments (e.g. results), but carelessly not the other parts related to the corrections (e.g. in discussions and abstract). However, I still believe that the data collected in such an extensive study deserve a publication and are valuable for soil fauna research of the region. Therefore, I would give the authors the last chance for a revision.

L291-293: This is method, not result.

L295-303: Such kinds of results can be better presented by principal component analyses together with a biplot showing differences in environmental factors among grasslands.

Table 2: It does not make any sense to calculate dominance for “grassland” instead for “each of the taxa” in a given grassland. If the values represented here are the “mean value” across all taxa in a given grassland, it reads even stranger and without any ecological relevance. Please delete the dominance from the table, as well as from the results, discussion and abstract.

L313: remove the phrase “In spatial dynamics,”.

L363-365: I am still not convinced by the authors’ argument about why they conducted the GLMM. The effects of treatments have been already validated in the other analyses of the study. Here only “what of the environmental factors that impacted the most” is needed for further investigation. Please, either re-do the analyses according to what has been suggested in the previous comments, or remove the GLMM (including both methods and results) from the manuscript.

Figures 2, 3, and 4: If the treatment here is equal to the grassland, the values here cannot be the means “per treatment”, because for each grassland 50 soil samples (= 1 grassland) were collected. There should be no SE for a single grassland as the unit to be presented in the figure. The authors should re-write the figure legends and specify the units in the y-axis of the figures 2 and 3, i.e. whether the values represent numbers per soil sample or for the all 50 soil sampled summed up in each treatment (grassland). I would also recommend to either write “treatment” or “grassland” throughout the manuscript for a better understanding. Please do not write two terms pointing out the same thing.

L377: higher or lower? Treatment C or B?

L380: higher or lower? Treatment B or D?

Please, revise the manuscript throughly and carefully and write the contents both precisely and concisely.

Author Response

L291-293: This is method, not result.

Answer: I delete the lines 291-293. The same information was already presented in subchapter 2.3 - Environmental variables.

L295-303: Such kinds of results can be better presented by principal component analyses together with a biplot showing differences in environmental factors among grasslands.

Answer:

The principal component analysis (PCA) of the environmental factors did not show a clear separation of grasslands (please see the figure from attached file).

Fig. X. Principal component analysis (PCA) of the environmental factors.

Table 2: It does not make any sense to calculate dominance for “grassland” instead for “each of the taxa” in given grassland. If the values represented here are the “mean value” across all taxa in given grassland, it reads even stranger and without any ecological relevance. Please delete the dominance from the table, as well as from the results, discussion and abstract.

Answer: I removed the dominance from the table, as well as from the results, discussion, conclusions and abstract.

L313: remove the phrase “In spatial dynamics”.

Answer: I remove it.

L363-365: I am still not convinced by the authors’ argument about why they conducted the GLMM. The effects of treatments have been already validated in the other analyses of the study. Here only “what of the environmental factors that impacted the most” is needed for further investigation. Please, either redo the analyses according to what has been suggested in the previous comments, or remove the GLMM (including both methods and results) from the manuscript.

Answer:

We removed treatment from all models and redone the analysis investigating only the environmental factors, as the reviewer suggested.

Table S2. Model selection results. Models are ranked in a decreasing order of the Akaike weights (w_i). For clarity, models with w_i< 0.02 are not shown. Statistics include: LL - log likelihood; K - number of parameters; the second-order Akaike information criterion corrected for small sample sizes AICc; ∆_i- AICc differences; w_i - Akaike weights.

Model structure	LL	K	AICc	∆_i	w_i
Abundance
VegCovr + pH	-1288.81	4	2585.62	0.00	0.35
VegCovr + Ts + pH	-1288.49	5	2586.97	1.35	0.18
VegCovr + Rhs + pH	-1288.49	5	2586.98	1.36	0.18
VegCovr + pH + RPs	-1288.81	5	2587.61	2.00	0.13
VegCovr + Ts + Rhs + pH	-1288.03	6	2588.07	2.45	0.10
Species richness
pH	-485.66	3	977.33	0.00	0.15
VegCovr + pH	-484.95	4	977.91	0.58	0.11
pH + RPs	-485.09	4	978.18	0.85	0.10
VegCovr + pH + RPs	-484.53	5	979.06	1.73	0.06
Rhs + pH	-485.60	4	979.21	1.88	0.06
Ts + pH	-485.66	47	979.33	2.00	0.06
RPs	-486.73	3	979.45	2.12	0.05
VegCovr + Rhs + pH	-484.75	5	979.51	2.18	0.05
VegCovr + Ts + pH	-484.95	5	979.91	2.58	0.04
Rhs + pH + RPs	-485.05	5	980.09	2.76	0.04
VegCovr	-487.06	3	980.13	2.80	0.04
VegCovr + RPs	-486.31	4	980.62	3.29	0.03

Figures 2, 3, and 4: If the treatment here is equal to the grassland, the values here cannot be the means “per treatment”, because for each grassland 50 soil samples (= 1 grassland) were collected. There should be no SE for single grassland as the unit to be presented in the figure. The authors should rewrite the figure legends and specify the units in the y-axis of the figures 2 and 3, i.e. whether the values represent numbers per soil sample or for the all 50 soil sampled summed up in each treatment (grassland). I would also recommend to either writing “treatment” or “grassland” throughout the manuscript for a better understanding. Please do not write two terms pointing out the same thing.

Answer:

Figures 2, 3 and 4 were redone and they changed due to different best models resulted from a new model design and selection (please see above). The values on the y axis are predicted values obtained using “effects” package in R. This is explained in the data analysis section.

The “treatment” was replaced with “grassland”.

L377: higher or lower? Treatment C or B?

Answer: This line was deleted.

L380: higher or lower? Treatment B or D?

Answer: This line was deleted.

Please, revise the manuscript thoroughly and carefully and write the contents both precisely and concisely.

Please, see the manuscript.

Author Response File: Author Response.pdf