The Prevalence of Egg Parasitoids of Two Cobweb Spiders in a Tropical Urban Gradient

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I have with great interest read the paper entitled 'IMPACT OF URBANIZATION ON THE INCIDENCE OF EGG PARASITOIDS OF TWO COBWEB SPIDERS', which looks at how egg parasite incidences on a native and an invasive cobweb spider different in area of different urbanisation. The study is generally very well carried out including sound data-analysis, and the results are of general interests to the readers of Arthropod. I only have a couple of relatively minor comments and suggestions as outlined below:

 

Introduction 

L56-58: I would like to see a bit more justification for why more eggs/egg sacs would be parasitised in more urban areas. I thought that predation risk (and therefore presumably also parasite risk) was lower in more urban areas (see for example Eötvös et al. 2018. Landscape and Urban Planning 180, 54-59).

 

Materials and Methods

L76-78: This sentence does not make sense. Please rephrase.

L79-81: Was there any justification for using a 2.5 km radius (rather than a 1km or a 5km for example)? It would be good to relate this to general foraging flight distance of parasitic wasps if known...

L99-100: Were these coffee plantations regularly treated with pesticides?

Figure 2: Please label the figures A, B and C.

 

Results:

L234-236: This sentence does not make sense. Please rephrase.

 

Discussion:

L292:296: Have you considered that it could be your rural site that has abnormally few parasitic wasps? If these coffee plantations are heavily treated with pesticides, perhaps that has lead to a local reduction in flying insects including parasitic wasps?

Comments on the Quality of English Language

Generally fine, but a few instances of sentences not fully making sense (high-lighted above).

Author Response

For research article

 

Response to Reviewer 1 Comments
1. Summary
Thank you very much for taking the time to review this manuscript. Please find detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.  2. Questions for General Evaluation We acknowledge the opinion for the general evaluation by the reviewer and the kind feedback. We consider the opinion to improve our manuscript.
3. Point-by-point response to Comments and Suggestions for Authors
L56-58: I would like to see a bit more justification for why more eggs/egg sacs would be parasitized in more urban areas. I thought that predation risk (and therefore presumably also parasite risk) was lower in more urban areas (see for example Eötvös et al. 2018. Landscape and Urban Planning 180, 54-59). Response: Thank you very much for your comments and the references suggested. We have added references for the justification. The text was changed as follows:  “ In previous studies conducted in Costa Rica, the prevalence of parasitoids in egg sacs increased at the same sampling site but with a higher level of urbanization [20,23] Based on this, we predicted that the number of egg sacs attacked by parasitoids or egg predators, and the proportion of parasitized eggs per egg sac, would increase with urbanization for both spider species due to anthropogenic factors that, in some cases, may favor an increase in parasitoids and predators of spiders and insects due to habitat connectivity and ecological matrix complexity [16,47].” L76-78: This sentence does not make sense. Please rephrase. Response: Thanks for the suggestion, we rephrase the sentence. L79-81: Was there any justification for using a 2.5 km radius (rather than a 1km or a 5km for example)? It would be good to relate this to the general foraging flight distance of parasitic wasps if known... Response: The dispersion capability and ecological requirements of egg parasitoids of spiders are unknown; therefore, we do not have a biological criterium that can be used to define a buffer zone due to the lack of information related to the natural history and egg foraging of spider parasites considering the current knowledge (small-sized parasitoids, and apterous females in Baeus spp.). L99-100: Were these coffee plantations regularly treated with pesticides? Response: In Costa Rica, coffee plantations are treated with pesticides one or two times per year according to the local farmers and previous studies mentioned in the manuscript. We have added references in the manuscript. The site with a low urbanism level, it is composed of coffee, sugar cane plantations, and forest patches (Figure 3). In general,  coffee plantations in Costa Rica and our study site are treated mainly with fungicides/pesticides (to control the coffee berry borer) [65,66], but also some farms practice organic agriculture [67,68]. Pesticides' potential effects on spider populations are un known, and there is also limited information about the specific substances used and how they are applied [68]. In the case of Other common parasitoids, like the braconid Psyttalia concolor, they are affected by pesticides and to a lesser extent by fungicides and herbicides [69].  The organophosphates and pyrethroids were the most toxic for the populations [69].  The possible effect of fungicides and pesticides on spider parasitoids could not be discarded, but due to the lack of information on the dispersion behavior, and natural history of the spider parasitoids [21,20,22,64], we need more studies.  65. Rojas, M.. Manejo sostenible de la broca del café (Hypothenemus hampei) mediante poda sistemática del cafeto en Costa Rica. Agron. Costarricense. 2012, 36 (2), 71-79. 66.Monge, P; Partanen, T; Wesseling, C; Bravo, V; Ruepert, C; Burstyn, I. Assessment of pesticide exposure in the agricultural population of Costa Rica. Annals of occupational hygiene. 2005, 49(5), 375-384. 67. Blackman, A;  Naranjo, M. A. Does eco-certification have environmental benefits? Organic coffee in Costa Rica. Ecological Economics 2012, 83, 58-66. 68. Guido Cruz, F.; Castro Sánchez, S. Viviendo con la crisis cafetalera: perspectivas futuras de pequeños y medianos productores de café en San Ramón de Alajuela, Costa Rica. InterSedes. 2011, 7(12),  69. Pinheiro, L. A.; Dader, B.; Wanumen, A. C.; Pereira, J. A.; Santos, S.  A. P.; Medina, P. Side Effects of Pesticides on the Olive Fruit Fly Parasitoid Psyttalia concolor (Szepligeti): A Review.Agronomy.2020, 10(11), 1755. Figure 2: Please label the figures A, B and C. Response: Thank you for the observation, we also think it is better with the labels, and we change the figure with the labels. L234-236: This sentence does not make sense. Please rephrase. Response: Thanks for the suggestion, we rephrased the sentence, we consider now it is better. L292:296: Have you considered that it could be your rural site that has abnormally few parasitic wasps? If these coffee plantations are heavily treated with pesticides, perhaps that has lead to a local reduction in flying insects including parasitic wasps. Response
4. Response to Comments on the Quality of English Language The coauthor Paul Hanson, who is a native English speaker, review again the manuscript. Language
5. Additional clarifications

 

Reviewer 2 Report

Comments and Suggestions for Authors

Review for arthropoda-3247379

Title: IMPACT OF URBANIZATION ON THE INCIDENCE OF EGG PARASITOIDS OF TWO

COBWEB SPIDERS

Authors: Natalia Jiménez-Conejo *, Paul E. Hanson, Eduardo Chacón-Madrigal,

Geovanna Rojas-Malavasi, Gilbert Barrantes

 

Summary: This study investigates the prevalence of egg parasites and predators in the egg sacs of Parasteatoda tepidariorum and Latrodectus geometricus across three locations in Costa Rica that differed in both large- and small-scale urbanization. The study also occurred across several months and spanned two seasons (dry and wet).  The investigators found that parasitism was more common in the native species P. tepidariorum than L. geometricus. The investigators claim that intermediate urbanization influenced the prevalence of parasitism, being more prevalent than the high urbanization sites.  

 

General comments: While the study is well conceived, i.e., the idea is a good one, I have some concerns about the methods. First, I believe the study needs more data to support the interpretation that large-scale urbanization had any effect. The study only had three locations, one for each large-scale urbanization category, and while these three sites were sampled repeatedly throughout the year, this kind of sampling represents pseudo-replication and is a nested design. Site and urbanization categories are confounded and should be considered random effects in the GLMM. Thus, the observations presented in this study do not necessarily indicate urbanization's effects on the large scale and may merely reflect the sites selected. The authors need to provide more information, raw data, or an explanation of the methods to evaluate how they constructed their models and if the construction was appropriate. 

Second, I wonder if using AIC scores for selecting model parameters is warranted in this case. For example, the authors believed that their measure of small-scale urbanization was important enough to calculate from the panoramic photos. So, they believed they had a biologically justifiable reason to include it as a parameter in their model. I agree with them on this. However, PC1 was never reported in the summary tables of the model results, only in the large-scale urbanization categories. This undermines the conclusion that urbanization is an important predictor. Again, given that the structure of the data appears to be nested (three locations with unique urbanization categories, with 19 samples from each location taken across two seasons, with 302 egg sacs collected in total across two species), it is not clear that the GLMM constructed were appropriate for the data.

I am also curious about the possible collinearity between large-scale and small-scale measures of urbanization. If I understood the methods correctly, within each large-scale site a specific 10m² area was chosen. This area was photographed, the photo was categorized using imageJ, and then the 4-5 variables that were measured in each photo were reduced to a single PC score. However, the methods do not clearly explain if the same 10m² area was sampled each of the 19 times, as the building maintenance schedule permitted, or if it varied and by how large a distance. A more detailed explanation of the methods would help clarify this. If the same area was sampled repeatedly, then it is possible that the large-scale scores correlate with the small-scale scores, leading to collinearity. Thus, including both in the same model might lead to erroneous results. And, since the models created were not reported, I can not evaluate this.

Overall, I believe that the data needs to be reanalyzed. I do not think that the large-scale measure of urbanization can be tested with the data this study has collected at present. I highly recommend that the data be reanalyzed with a focus on small-scale urbanization using a nested GLMM model. I do not believe that the use of AIC scores for model selection is warranted here. I know this is debated, but I believe that ecologically relevant models should be created prior to any use of AIC for parameter evaluation. The authors would be better served by creating models that they believe are relevant based on ecological and biological justifications. The models created need to be specified more clearly and both the data and models should be more fully reported. A more thorough reporting of the data and models would enhance the transparency and reproducibility of the study. These recommendations will necessitate substantial revision of the manuscript. It is for this reason that I recommend that the manuscript be resubmitted.

Specific comments:

Incidence is a technical term in statistics and epidemiology and was not actually measured in this study. The study measured prevalence, not incidence. This needs to be fixed throughout the manuscript.

P. tepidariorum was misspelled in several places throughout the manuscript. Please ensure that all instances are spelled correctly and in the same way.

The figures are not labeled. The figure captions refer to A or B, but A and B do not appear in the figure.

Given that the dependent variable is the number of egg sacs in a 10 m² area parasitized and the number of eggs in a sac that were affected, it would be better to use a Poisson distribution (for count data) instead of a binomial distribution (dichotomous) for the GLMM.

I prefer to see 95% CI instead of SE for the coefficients reported in the tables.

The conclusion should focus on the implications of the research, not a summary of the results.

 

Line 46: When reference 19 is used, there are a few other papers that the authors should consider using. Vetter was involved in two articles that specifically examined the parasitism of brown widow egg sacs. Marie and Vetter 2015 and Vetter et al. 2012

Comments on the Quality of English Language

English is pretty good overall; some minor language editing is necessary. 

Author Response

For research article

Response to Reviewer 2 Comments
1. Summary
Thank you very much for taking the time to review this manuscript. Please find detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. 2. Questions for General Evaluation We acknowledge the opinion for the general evaluation by the reviewer and the kind feedback. We consider the opinion to improve our manuscript.
Reviewer 2 To properly answer the comments of this reviewer, which we highly appreciate because allowed us to clarify some points, and provide a clearer view of the information we provide in the paper, we further dissected the comments to provide more specific answers.  General comments:  Rev. comment: While the study is well conceived, i.e., the idea is a good one, I have some concerns about the methods. First, I believe the study needs more data to support the interpretation that large-scale urbanization had any effect. The study only had three locations, one for each large-scale urbanization category, and while these three sites were sampled repeatedly throughout the year, this kind of sampling represents pseudo-replication and is a nested design.  Response: I understand the concern of the reviewer. To address this issue, we reorganized and provided more information in the section “Urban gradient” that we expect will clarify the criteria used to define the urban gradient: “ 2.2. Urban gradient. We conducted this research in three urban locations in the Costa Rican Central Valley, the most urbanized and populated region of the country. Each location is surrounded by a matrix affected to varying degrees by human intervention (Figure 1). The Central Campus of the University of Costa Rica (Campus Rodrigo Facio, CC), San José province (9°54'N, 84°03'W) is in the most urbanized area. The campus includes large areas with buildings and large gardens and small patches of second-growth forest [6]. The Occidental Campus of the University of Costa Rica (Sede de Occidente, SO), Alajuela province (10°08'N, 84°47'W) is within an area with an intermediate level of urbanization, with small buildings and sparce patches of second-growth vegetation. The Occident campus includes some buildings separated by large green areas with abundant trees. The third location was Concepción of San Ramón (CSR), Alajuela province (10°11'N, 84°44'W). This is a small rural village dominated by coffee plantations and forest patches (Figure 1; Table S1). The sampling locations were in the western section of the Costa Rican Central Valley (Figure 2) at an elevation of 1200 m, with an average temperature of 20.5°C, and a dry season from December to April. To further characterize the urban locations, we estimated the percentage of green vs gray areas in a buffer zone of 2.5 km around each location, using the polygon selection tool in the QGIS software [39]. The green area consisted of gardens, secondary forests, patches of early-growth vegetation, and agricultural patches, and the gray area included human-made constructions. The dispersion capability and ecological requirements of egg parasitoids of spiders are unknown; therefore, we do not have a biological criterium that can be used to define a buffer zone. Considering the current knowledge, we arbitrarily selected 2.5 km as the potential area could affect the biology (e.g., dispersion, reproduction) of parasitoids.  Within each urban location, we sampled three groups of buildings (one in CSR) several times during the study period. The prevalence of parasitoids of P. tepidariorum and L. geometricus at each location is likely affected by the urban matrix surrounding them, as well as by the immediate environment near each group of sampled buildings.”  Locations are distributed along an urban gradient and within each site we expect to evaluate the effect of the surrounding matrix. Rev. comment: “these three sites were sampled repeatedly throughout the year, this kind of sampling represents pseudo-replication and is a nested design”. Response:We included date of sampling as a (intercept) random factor to account for the effect of multiple visits. Rev. comment: Site and urbanization categories are confounded and should be considered random effects in the GLMM. Thus, the observations presented in this study do not necessarily indicate urbanization's effects on the large scale and may merely reflect the sites selected. The authors need to provide more information, raw data, or an explanation of the methods to evaluate how they constructed their models and if the construction was appropriate.  Response: I appreciate your comment, the information included in the ms is insufficient to have a clear panorama of the structure of the models and how they were used. We expect that the urban category and group of buildings characterized by the PC1 were included as fixed factors. The description of each of these factors was very confusing in the ms, and we apologize for that. In this fixed version of the ms, we provide a clear description and definition of each one, and maintained both in the models despite the fact that buildings had a negligible effect. This is because we hypothesized that could have an important effect on the prevalence of eggs parasitized.  We run the models again to include the PC1 in all cases, and clarified the information in the ms accordingly.  We agreed that the word site is confusing in the ms. Site refers to the cover features (e.g., vegetation cover) around each sampling site (e.g. buildings), and such features are included in the PC1. We maintained both site (PC1) and urban category in the models because these are the two factors that had biological meaning for our study system and thus for the hypotheses. In addition, these two variables are not correlated (no collinearity) (r = 0.18,; 0.48). “2.4. Data analyses. To evaluate the effects of the urban matrix (urban location) and the local environment (the immediate surroundings of the sampling sites) on the number of egg sacs parasitized and the proportion of parasitized eggs, we first created a model to compare these factors between the two species. This model included either the number of egg sacs parasitized and the proportion of eggs parasitized as response variable and urban category, PC1 (representing the immediate surroundings), season, building, and sampling date as random factors. In consecutive models, we evaluated the effects of urban category and immediate surroundings (PC1) on the number of egg sacs parasitized and the proportion of parasitized eggs for each species, with season, building, and sampling date included as random factors. In all cases, we used General Linear Mixed Models (GLMMs) with a Poisson error distribution for the number of egg sacs parasitized and a binomial error distribution for the proportion of eggs parasitized (library Ime4), [45]. The two predictor factors were not collinear (correlation between both factors varied between 0.18 and 0.46), therefore, PC1 was maintained despite having a negligible effect, since both factors were essential to test the predictions. In addition, we correlated the proportion of parasitized eggs and the number of emerging spiderlings using Spearman’s rank correlation. We performed all analyses in the R statistical language (R Development Core Team) [46].” Rev. comment: Second, I wonder if using AIC scores for selecting model parameters is warranted in this case. For example, the authors believed that their measure of small-scale urbanization was important enough to calculate from the panoramic photos. So, they believed they had a biologically justifiable reason to include it as a parameter in their model. I agree with them on this. Response: We agreed with the reviewer concern and modified the information on the basis of this and previous comments. Rev. comment: However, PC1 was never reported in the summary tables of the model results, only in the large-scale urbanization categories. This undermines the conclusion that urbanization is an important predictor.  Response: We addressed this point and reported the PC1 in the summary tables of the models. Rev. comment: Again, given that the structure of the data appears to be nested (three locations with unique urbanization categories, with 19 samples from each location taken across two seasons, with 302 egg sacs collected in total across two species), it is not clear that the GLMM constructed were appropriate for the data. Response:We agree that a nested design could have been used in this case, which would have provided more detailed information for each specific sampling site (e.g., building); however, we preferred a more general approach and included sampling site (e.g., building) as a random factor to account for the multiple sites sampled in each location. This approach aligns better with one of our main questions of how surrounding vegetation affects the occurrence of parasitized eggs. Rev. comment: I am also curious about the possible collinearity between large-scale and small-scale measures of urbanization.  Response:There is not collinearity between these two factors, the correlation between these two factors varied between 0.18 and 0.46. Rev. comment: If I understood the methods correctly, within each large-scale site a specific 10m2 area was chosen. This area was photographed, the photo was categorized using imageJ, and then the 4-5 variables that were measured in each photo were reduced to a single PC score. However, the methods do not clearly explain if the same 10m2 area was sampled each of the 19 times, as the building maintenance schedule permitted, or if it varied and by how large a distance. A more detailed explanation of the methods would help clarify this. If the same area was sampled repeatedly, then it is possible that the large-scale scores correlate with the small-scale scores, leading to collinearity. Thus, including both in the same model might lead to erroneous results. And, since the models created were not reported, I can not evaluate this. Response:We appreciate the reviewer comment. The building maintenance are scheduled once a month; therefore, there is not expected a significant change in the sites’ surroundings over a year period. Thus, we took the decision of characterized the surrounding of each building only once during the study period. We explain this issue in detail in the corrected ms. Rev. comment: Overall, I believe that the data needs to be reanalyzed. I do not think that the large-scale measure of urbanization can be tested with the data this study has collected at present. I highly recommend that the data be reanalyzed with a focus on small-scale urbanization using a nested GLMM model.  Response:It was desirable to have a larger sample size (e.g., urban localities); however, we still consider that the comparisons made among the tree urban categories permitted evaluating the effect of the surrounding matrix on the prevalence of eggs parasitized. As we mentioned above, we have detected differences in behavior, reproduction, and genetics in birds using a similar spatial design, so we expected to find an effect on these spiders. We are planning to compare the effect of urbanization on behavior, population dynamics and genetics of other spiders, and we will surely follow your advice to include more sites that allow us to evaluate the effect of urbanization more precisely. We also reanalyzed the data modifying the structure of the models as indicated in the fixed version of the ms. Rev. comment: I do not believe that the use of AIC scores for model selection is warranted here. I know this is debated, but I believe that ecologically relevant models should be created prior to any use of AIC for parameter evaluation. The authors would be better served by creating models that they believe are relevant based on ecological and biological justifications. The models created need to be specified more clearly and both the data and models should be more fully reported. A more thorough reporting of the data and models would enhance the transparency and reproducibility of the study. These recommendations will necessitate substantial revision of the manuscript. It is for this reason that I recommend that the manuscript be resubmitted. Response: After checking the structure of the models in relation with the hypotheses proposed, we agree with your suggestion. In the new models we included the factors biologically relevant that permitted us evaluating both the large and the small-scale effects on the prevalence of eggs parasitism. Specific comments: Incidence is a technical term in statistics and epidemiology and was not actually measured in this study. The study measured prevalence, not incidence. This needs to be fixed throughout the manuscript. Response:This was corrected in the ms. P. tepidariorum was misspelled in several places throughout the manuscript. Please ensure that all instances are spelled correctly and in the same way. Response: This was corrected in the ms. The figures are not labeled. The figure captions refer to A or B, but A and B do not appear in the figure. Response: This was corrected in the ms. Given that the dependent variable is the number of egg sacs in a 10 m2 area parasitized and the number of eggs in a sac that were affected, it would be better to use a Poisson distribution (for count data) instead of a binomial distribution (dichotomous) for the GLMM. Response: We run the models using a Poisson error distribution as suggested, and rewrote part of the results accordingly. I prefer to see 95% CI instead of SE for the coefficients reported in the tables. Response: We included the 95% CI as the reviewer suggested. The conclusion should focus on the implications of the research, not a summary of the results. Response: We rewrote the conclusions following the reviewer suggestion. Line 46: When reference 19 is used, there are a few other papers that the authors should consider using. Vetter was involved in two articles that specifically examined the parasitism of brown widow egg sacs. Marie and Vetter 2015 and Vetter et al. 2012 Response: Thak you, we included the new references.  4. Response to Comments on the Quality of English Language The coauthor Paul Hanson, who is a native English speaker, review again the manuscript. Language
5. Additional clarifications

 

Reviewer 3 Report

Comments and Suggestions for Authors

This ms addresses the impact of urbanization on the effect of urbanization on the incidence of parasitoids in egg sacs and proportion of eggs parasitized in each egg sac of two synanthropic cobweb spider species along an urban gradient (from highly urbanized to rural sites). More parasitized egg sacs were found at the intermediate urbanized site for both species. Therefore, urbanization has an effect on the incidence of parasitoids, but they seem to be favored by the environmental matrix prevailing in the site with intermediate urban development. More parasitized egg sacs were found in the native species than the invasive species, suggested as a consequence of native parasites not being adapted to a new host.

This study is generally well done and the manuscript is satisfactorily presented. The conclusions are novel and align with the results. I have only minor criticisms, as follows:

It would be nice to add a figure with photos of the two spiders, their egg sacs, and the parasitoids, to place the study in context.

The map (Fig 1) could be improved (it is rather dark).

Some minor editorial corrections (mostly spelling mistakes that are repeated many times) were noted; this list may not be exhaustive:

L 47: “Baeus” needs to be italicized

LL 48, 53: It is recommended to give the taxon author (and date of publication) of each taxon, e.g., Scelionidae, Theridiidae, on first mention in the text.

L 183: Start sentence with full word Parasteatoda not abbreviation “P.”

L 186: correct spelling “tepidariorum” not “tepidariourm”

LL 192, 196, 198, Table 1, Table 2, Table 3: correct spelling “Latrodectus” not “Lactrodectus”

LL 196, 198, Table 1, 204, 218, Table 2, Table 3: correct spelling “tepidariorum” not “tepidadiorum”

L 248: insert space “[4,12].But”

L 282: correct spelling “tepidariorum” not “tepideariorum”

Author Response

For research article

For research article

 

Response to Reviewer 3 Comments
1. Summary
Thank you very much for taking the time to review this manuscript. Please find detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.  2. Questions for General Evaluation We acknowledge the opinion for the general evaluation by the reviewer and the kind feedback. We consider the opinion to improve our manuscript.
3. Point-by-point response to Comments and Suggestions for Authors
L56-58: I would like to see a bit more justification for why more eggs/egg sacs would be parasitised in more urban areas. I thought that predation risk (and therefore presumably also parasite risk) was lower in more urban areas (see for example Eötvös et al. 2018. Landscape and Urban Planning 180, 54-59). Response: Thank you very much for your comments and the references suggested. We have added references for the justification. The text was changed as follows:  “ In previous studies conducted in Costa Rica, the prevalence of parasitoids in egg sacs increased at the same sampling site but with a higher level of urbanization [20,23] Based on this, we predicted that the number of egg sacs attacked by parasitoids or egg predators, and the proportion of parasitized eggs per egg sac, would increase with urbanization for both spider species due to anthropogenic factors that, in some cases, may favor an increase in parasitoids and predators of spiders and insects due to habitat connectivity and ecological matrix complexity [16,47].” L76-78: This sentence does not make sense. Please rephrase. Response: Thanks for the suggestion, we rephrase the sentence. L79-81: Was there any justification for using a 2.5 km radius (rather than a 1km or a 5km for example)? It would be good to relate this to general foraging flight distance of parasitic wasps if known... Response: The dispersion capability and ecological requirements of egg parasitoids of spiders are unknown; therefore, we do not have a biological criterium that can be used to define a buffer zone due to the lack of information related to the natural history and egg foraging of spider parasites     considering the current knowledge (small-sized parasitoids, and apterous females in Baeus spp.). L99-100: Were these coffee plantations regularly treated with pesticides? Response: In Costa Rica, coffee plantations are treated with pesticides one or two times per year according to the local farmers and previous studies mentioned in the manuscript. We have added references in the manuscript. Figure 2: Please label the figures A, B and C. Response: Thank you for the observation, we also think it is better with the labels, and we change the figure with the labels. L234-236: This sentence does not make sense. Please rephrase. Response: Thanks for the suggestion, we rephrase the sentence, we consider now is better. L292:296: Have you considered that it could be your rural site that has abnormally few parasitic wasps? If these coffee plantations are heavily treated with pesticides, perhaps that has lead to a local reduction in flying insects including parasitic wasps? Response: we included information.   Costa Rica, coffee plantations are treated with pesticides one or two times per year according to the local farmers and previous studies mentioned in the manuscript. We have added references in the manuscript. The site with a low urbanism level, it is composed of coffee, sugar cane plantations, and forest patches (Figure 3). In general,  coffee plantations in Costa Rica and our study site are treated mainly with fungicides/pesticides (to control the coffee berry borer) [65,66], but also some farms practice organic agriculture [67,68]. Pesticides' potential effects on spider populations are un known, and there is also limited information about the specific substances used and how they are applied [68]. In the case of Other common parasitoids, like the braconid Psyttalia concolor, they are affected by pesticides and to a lesser extent by fungicides and herbicides [69].  The organophosphates and pyrethroids were the most toxic for the populations [69].  The possible effect of fungicides and pesticides on spider parasitoids could not be discarded, but due to the lack of information on the dispersion behavior, and natural history of the spider parasitoids [21,20,22,64], we need more studies.  65. Rojas, M.. Manejo sostenible de la broca del café (Hypothenemus hampei) mediante poda sistemática del cafeto en Costa Rica. Agron. Costarricense. 2012, 36 (2), 71-79. 66.Monge, P; Partanen, T; Wesseling, C; Bravo, V; Ruepert, C; Burstyn, I. Assessment of pesticide exposure in the agricultural population of Costa Rica. Annals of occupational hygiene. 2005, 49(5), 375-384. 67. Blackman, A;  Naranjo, M. A. Does eco-certification have environmental benefits? Organic coffee in Costa Rica. Ecological Economics 2012, 83, 58-66. 68. Guido Cruz, F.; Castro Sánchez, S. Viviendo con la crisis cafetalera: perspectivas futuras de pequeños y medianos productores de café en San Ramón de Alajuela, Costa Rica. InterSedes. 2011, 7(12),  69. Pinheiro, L. A.; Dader, B.; Wanumen, A. C.; Pereira, J. A.; Santos, S.  A. P.; Medina, P. Side Effects of Pesticides on the Olive Fruit Fly Parasitoid Psyttalia concolor (Szepligeti): A Review.Agronomy.2020, 10(11), 1755.
4. Response to Comments on the Quality of English Language The coauthor Paul Hanson, who is a native English speaker, reviewed again the manuscript. Language
5. Additional clarifications

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Please check the attachment.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The new sections need to be edited, and there are still many typos throughout the manuscript. 

Author Response

For research article

Response to Reviewer 2-Round 2 Comments

1. Summary
Thank you very much for taking again the time to review this manuscript. Please find detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. 2. Questions for General Evaluation We acknowledge the opinion for the general evaluation by the reviewer and the kind feedback. We consider the opinion to improve our manuscript.

 

General comments

Spirit in which my comments are given: It is my sincere hope that this research is published. I believe that it has value and will expand our knowledge of spider/wasp parasitism in these species. However, I do feel that the interpretations provided in the manuscript are not strongly supported by the data presented. The easiest path toward publishing this data would be to soften and expand the interpretations of what is causing the fluctuations in parasitism observed. Given the limited sample size across all scales, I would strongly recommend that the authors take a much more tentative tone and propose other possible explanations for the variation observed. If the authors wish to focus on urbanization, then they would need to collect more data from a wider range of locations i.e., additional locations within each urbanization category. Without this additional data, I do not believe that the focus of their discussion and conclusions are justified. I hate to be reviewer #2 here and hold this paper up, but I do not believe that it is ready for publication at this time; however, I do believe it could be made ready following either of the options described above. General Comments: While the revision is much improved, considerable changes should still be made. 

 

1. The authors have not changed all instances of the word "incidence" to "prevalence." The authors should ensure that studies they site as measuring incidence actually measure incidence. For example, I believe that some of the references listed in line 53 do not measure incidence. Incidence still appears at least four times in the introduction and the authors must ensure that in each case it is being used appropriately.

R/ Thanks for the observation; we have changed all instances of the word “incidence” to “prevalence”. We use references in line 53 specifically to support the statement, “Urbanization provides new habitats for some invasive and tolerant native species, whose populations often increase rapidly.” We did not use the references to support any statement about prevalence or incidence.

2. Formatting issues. The PDF provided has several formatting issues, including paratheses around Koch 1841 in lines 49 and 50. There are also many extra tabs and spaces. 

R/ Thanks for the observation. We have been more careful with all typo errors. We check all the text in the manuscript again. 

3. Spelling and grammar. The additions are not as polished as the original text and contain many typos and grammatical mistakes. Please use either software like Grammarly or a professional to find and correct these mistakes. 

R/ Thanks for the observation. We have used Grammarly to find grammar errors. Paul Hanson, a coauthor, and a native English speaker, reviewed the orthography and grammar in the original manuscript. 

4. At the time of writing this review, the editor's office did not provide the authors’ written responses to my previous comments, so I am unaware of any counterargument the authors may have made. (NOTE: I have now been provided with their responses: my opinion below is unchanged.) However, I still believe that this study measured the effects of urbanization to the exclusion of other possible interpretations like simple site differences. Because there are only three large-scale locations, arbitrarily categorized as low, intermediate, and high, the study design cannot differentiate between the effect of location and urbanization on a large scale. However, the language used in the article gives the impression that urbanization is the cause of the fluctuations observed in parasitism. This study's design does not allow for this conclusion. It could be that urbanization is a cause. However, it could also be one of many other factors that are site-specific but unrelated to urbanization or random fluctuations between sites that this study's small sample size happened to observe. The data does not strongly support the arguments made in this article, yet no mention of these limitations appears in the discussion section. Thus, the paper's tone and focus on urbanization are inappropriate. Because of this, I recommend that the introduction and discussion be altered to match the limitations of the studies design and the limited data collected. If the authors want to refrain from softening and broadening their interpretations, they must collect more data and expand their sampling efforts. 4b.

 The terms "gradient" and "matrix" are still unclear. We cannot truly call something a gradient when only three observations exist. First, what changes to a habitat do and do not constitute “urbanization” are not clear. This makes the classification of the three locations as low, intermediate, and high seem arbitrary. These are just three ill-defined categories based on convenience sampling. The categorization of low, intermediate, and high are relative to each other, do not represent a gradient, and do not have clear external validity (applicability, i.e., can I use the technique employed here to categorize sites elsewhere in the world? Is that is high in CR also high in Melborn Australia or Hong Kong?). 

I recommend that the authors avoid using jargon laden terms like “gradient”, or use an externally validated dataset that has previously been used to quantify/classify landscape disturbance or urbanization (for example, data on land use exists as Raster files that could be used to provide a score or classification for the three locations sampled in this research, and could serve as an independent validation for the QGIS green/grey method.

 Incidentally, I have no issue with the green/grey methodology, I simply have questions about what a gradient is and if one was measured here. At a minimum, The authors should be more mindful of the limited generalizability of their method and data and specifically bring up these limitations in their discussion.

Likewise, the use of "matrix" is not well defined. It seems to simply describe the polygon maps they used to categorize urbanization. Please define what an urban matrix means in a technical sense and cite appropriate literature, or find a less jargon-heavy term to describe the authors' work. Matrix is a jargon term in many other fields of science but does not appear to be one here. Would mosaic or patchwork assemblage… work better?

R/ Thanks for the recommendation. We understand the limitations of our conclusions with only three sites. We have rewritten several parts of the title, abstract, introduction, and discussion to emphasize the data about parasitoidism and tone down the conclusions about the effect of urbanization. We have removed jargon as “matrix” from the paper. However, we prefer to maintain our analysis and our results including habitat measurements because it is part of our experimental design. We acknowledge the validation of the other data.

5. in lines 58 and 59, the authors "estimated the proportion of green area surrounding the collecting sites." However, the authors used a PCA to summarize their collected variables into a single variable, interpreted as measuring "green area." While this interpretation is warranted, it is an interpretation, and a description of how the variables were reduced has not occurred yet; therefore, I do not feel that the quote above has been adequately justified. Please rewrite the description to be more accurate. Something like, "We quantified several immediate habitat features, e.g., trees, bare earth, concrete, surrounding the collection sites." This is more accurate to what the authors did and still allows them to describe and justify the PCA in the methods.

R/ We acknowledge the observation. We have rewritten the text regarding the use of the PCA to summarize data.

6. The introduction states that one of the dependent variables was the "proportion of eggs parasitized per egg sac." This seems like it would be a continuous variable, e.g., 40% or 0.40. However, the GLMM modeled this variable using a binomial error distribution, which is more appropriate for categorical (parasitized or not) variables. Why was a binomial Generalized linear mixed model used instead of a continuous Linear Mixed Model? By proportion, did the authors mean presence/absence?

R/ We used a binomial distribution because proportions adjust well to this distribution.  In discrete counts, we can, for instance, measure the number of presence (parasitoidism) in relation to the total of eggs. We will thus obtain a proportional number of “success” by dividing the counts with parasitoids by the total counts (Dunn &  Smyth 2018).

Dunn, P.K., Smyth, G.K. (2018). Chapter 9: Models for Proportions: Binomial GLMs. In: Generalized Linear Models With Examples in R. Springer Texts in Statistics. Springer, New York, NY. https://doi.org/10.1007/978-1-4419-0118-7_

7. As mentioned above, what is “urbanization” in a technical sense? How does it differ from the human-mediated habitat alteration the farmland location (low category) experienced? This distinction is not clear, but it is an important one, given that this location also had an altered natural plant diversity and experienced periodic pesticide use (not present at the other locations). 

R/ We have defined a higher degree of urbanization as a higher proportion of buildings over the natural surface, which means a higher proportion of “gray landscape” over “green landscape. It is explained in lines 163-178.

8. The figure numbers in the text need to be updated. In many places in the manuscript, the figure numbers still need to be updated to reflect the new figures and the resulting changed numbers. 

R/ Thanks for the observation. We checked again the numbers of the figures and corrected those with mistakes.

9. line 95: What is an "arge garden"? Do the authors mean "large," or is this a technical term?

R/ It was “large” but now we have changed the sentence, and we only used “garden”.

10. Am I understanding correctly that the authors standardized the size of the area searched? So, the 10 m2 area in the high category may have been on a single wall, but the 10 m2 area in the intermediate area may have been composed of more than one wall on a smaller building. Is this accurate? If not, this is another reason why urbanization may not be the only factor influencing parasitism. 

R/ We standardize our unit sampling to 10 m2. All the units were in buildings or houses in the rural zone. The complexity of the units was taken into account with our measurements on the local scale and included in the models.

11. line 117: "Within each urban location, we sampled three groups of buildings (one in CSR)." Please clarify what one in CSR means. It seems like it means that the authors collected samples from 3 separate buildings in two locations but only from one building in the last location. How far apart were the buildings? Why did they do this? What effect could this difference have on the outcome of the study? Please address this more clearly and thoroughly. 

R/ We remove abbreviations as CSR, which do not support any in the text.

Buildings were separated between them by 20-50m in the three sampling sites. All the buildings sampled varied, in terms of the surrounding vegetation, bushes, number of trees, and cemented areas. This analysis was done because we thought there could be a difference in the parasitism taking into account the characteristics of the immediate surrounding area besides the  landscape variables considered.

12. Lines 118-120: "The prevalence of parasitoids of P. tepidariorum and L. geometricus at each location is likely affected by the urban matrix surrounding them, as well as by the immediate environment near each group of sampled buildings." This is the central assumption of the paper. Thus, this is unknown. Given this is the assumption being tested, I am not sure how it can also be used as a justification.

R/ We acknowledge the observation, we have removed this part of the text. We have rewritten the section.

13. Lines 146-148: "From each photo, we obtained the proportion of bare soil, green area (e.g., grass and shrubs-trees), cemented area, and number of trees and bushes in front of each building." Why did they measure the number of trees and not an area of the photo that was tree/bush? All the other factors were measured as area.

R/The two variables were measured because they are not correspondent. A large tree area can result by just one big tree or by many small trees. 

14. Starting on line 151: "Building maintenance was scheduled once a month. Therefore, we characterized the surroundings of each building only once during the study period." If buildings were "maintained" (please specify the nature of the maintenance, is this the painting the first manuscript mentioned) once a month, why weren't photos taken after every maintenance? Unless the maintenance did not affect the immediate landscape, this could have significantly impacted parasitism. 

R/ The maintenance did not affect the immediate landscape, only the buildings. Maintenance consisted of domestic cleaning of the walls or sporadically painting. We have clarified in the text.

15. Please provide the model equation for each model constructed. This can be supplemental, but it would be very helpful. 

R/ Thanks for the suggestion. The model equations tested were included in a Supplementary Materials. Now, it was specified in the text.

16. Line 216: It would be helpful to add the specific predictors after "predictor factors" (??).

R/ We specify in the line, the specific predictors.

17. Figure 4 has several issues. The x-axis labels are reversed and misspelled. The figure currently makes it appear that L. geometricus experienced more parasitism and had more webs sampled than the native species. 

R/Thank you very much. We have already corrected the figure number 4.

18. figure 5 has a typo, "LatrodectusLatrodectus" 

R/ The typo was corrected.

19. Concerning the tables. I appreciate the addition of the lower and upper bounds! However, seeing these in a figure would be much easier, not as a table. I recommend making a figure like this

This would make significant and marginally significant results more obvious. The tables can be included in the supplemental information. NOTE: Since I believe the discussion section needs to be significantly rewritten, I will not comment on all the specific typos that appear in it. Here are a few comments. 

R/ We appreciate the reviewer's suggestion; however, we prefer to present the results in a table because we can use just one table for three different analyses; otherwise, we should make three different plots for each table, each one with very little information because for each analysis the meaning of the zero is a distinct factor.  To facilitate the reading, we have written with a boldface font the factors different from zero.

20. Starting on line 321: "For both species, P. tepidariorum and L. geometricus, egg sacs were more parasitized in the intermediate urban site than in sites with low and high urbanization (Figure 4)." As mentioned previously, it should be Figure 5, not 4. In addition, the sentence gives the impression that parasitism was higher in the intermediate than low or high for both species. This is both false and misleading. False: there was no L. geometricus at the low site, so it would be impossible to compare intermediate with low for this species. Misleading: The absolute occurrence of parasitism is not the most important thing to talk about here. Whether the difference was significant is much more important, and those results are mixed at best. Also, the results only compare the urbanization categories with high and not to each other. The authors could run some pairwise comparisons to evaluate intermediate with low.

R/ Thanks for the observation; we have corrected the sentence and the reference to the correct figure.  L. geometricus colonizes sites in urban spaces, it is not expected to find an L. geometricus in a forest or plantations (reference vetter Florida). This could be one of the reasons why we did not find this species in the third site: the countryside, with houses separated by more than 20 meters and many vegetation and bushes. The coefficients of the models in Table 4 and Table 5 bring the comparison between categories of urbanization. The text now refers to the tables.

21. Starting on line 332: "Although the effect of increasing heat in cities on the development and mortality of spider egg parasitoids has not been tested, it is likely that they are also affected. Hence, the prevalent conditions in highly urbanized sites likely affect the ecological requirements of parasitoids." This is an example of why the tone of this manuscript must be softened. First, just because something occurs in one species does not mean it will occur in the same way or to the same extent in another species, let alone in a completely different taxonomic group. The author can suggest that this is "conceivable," but they have not provided enough justification to say it is "likely." 

R/ Thanks for the observation, we have corrected the paragraph and we have added evidence to discuss how it could affect some parasitoids. We have rewritten the section.

22. line 381: How would the presence of latrotoxins in the egg sac prevent parasitism? Latrotoxins are proteinaceous; thus, short of being injected into the parasite, there is little reason to suspect that they would enter its body. Has this been tested?

R/ Neurotoxins are present in the egg silk of Latrodectus spp.; therefore, we could not discard the possibility that these toxins could reduce the prevalence of parasitoids in the Latrodectus spp. genus in general, compared to P. tepidadiorum. The toxicity in egg sacs has been demonstrated by Schmidt, J. O., Vetter, R. S., & Howe, A. K. (2017). Egg toxicity in diverse spider taxa. The Journal of Arachnology, 45(2), 209-212.

Author Response File: Author Response.pdf