Effect of Beauveria bassiana Fungal Infection on Survival and Feeding Behavior of Pine-Tree Lappet Moth (Dendrolimus pini L.)
Round 1
Reviewer 1 Report
The paper presented is very well structured, however the presentation results are not clear and oblige the reader to move across the manuscript in order to understand the idea of the author.
It is very important to refer in the text the number of the table or the number of the figure and present it immediately. This principle is not suited in the present manuscript. I suggest reformulating the manuscript with attention to this point.
As example:
Line 127 – 133: Should be moved to line 122
Lines 139 -145: Should be moved to line 134
Line 180: Should start with Fig 1.
Line 189: Should start with actual Table I (which must be renumbered to Table III).
The information given in Experience I and Experience 2 is given as Table. However, is not presented in the Formal Format. I suggested do it and enumerating the information as follow:
Table I – Experience 1: Pipetting directly Beauveria bassiana on the back of larvae.
Table II – Experience 2: Using sporulating cadavers with Beauveria bassiana.
Of course, the other tables must be renumbered.
Other comments:
Line 167. Explain the mean of LD100.
Line 202 – Actual Table 2 (renumber Table 4) is confused. I suggest present only the results of ANOVA and Kruskal Wallis test between all the treatments.
Author Response
Response to Reviewer 1
Line 127 – 133: Should be moved to line 122
We agree and made de changes
(see L: 126)
Lines 139 -145: Should be moved to line 134
We agree and made de changes
(see L: 142)
Line 180: Should start with Fig 1.
We made a mistake and instead Table 1 (now it is Table 3) we wrote Figure 1. This was changed and now it is logical to start with the table and move than to Figure 1. In the meantime, the Figure 1 was changed to dose-response chart as suggested by reviewer 2
(see L: 215-218)
Line 189: Should start with actual Table I (which must be renumbered to Table III).
Please see the comment above (on Line 180)
The information given in Experience I and Experience 2 is given as Table. However, is not presented in the Formal Format. I suggested do it and enumerating the information as follow:
Table I – Experience 1: Pipetting directly Beauveria bassiana on the back of larvae.
Table II – Experience 2: Using sporulating cadavers with Beauveria bassiana.
Thank you for this suggestion; We made de changes
(see L: 139 and L154)
Of course, the other tables must be renumbered.
We checked the new version as requested and made the changes.
Other comments:
Line 167. Explain the mean of LD100.
We add: lethal dose (LD100 -dose required to kill 100% of D. pini larvae)
(see Line: 181)
Line 202 – Actual Table 2 (renumber Table 4) is confused. I suggest present only the results of ANOVA and Kruskal Wallis test between all the treatments.
The term “dieback speed“ was replaced by „lethal time“ considering the reviewers suggestion. We agreed and adjust the terminology throughout the manuscript.
(see Line: 228)
Author Response File: 
Author Response.docx
Reviewer 2 Report
This is a well-designed study of the effect of dosage of a strain of B. bassiana on an important forest pest. However, the statistical analysis is not very relevant and can be greatly improved. I recommend deleting all the multiple comparisons and replacing them with fitting dose-response curves and probit analysis. The important results are to know and be able to predict the level of response to a given dosage. Presenting the results as curves and their equations would enable anyone to make such predictions. Note that Abbott's formula (Abbott 1925) is usually applied to correct for control mortality.
It would be more correct to us the conventional 'lethal time' (LT) than "die back speed". Speed is normally distance/time, whereas the authors have measured time until death. It appears that the authors measured lethal time for all the insects in each treatment, which is a useful measure regarding population level responses (e.g., from the point of view of pest control). However, to measure the biological response, I think that it would be more appropriate to include only those individuals that ultimately sporulated. The same comments apply to 'feeding disruption'.
The authors should define the strain of B. bassiana and deposit an accession in a repository.
There are minor grammatical and typographic errors (that I have not marked) that should be corrected.
Other specific comments follow.
Introduction
Briefly describe the life history of D. pini. What is the approximate development time of larvae? Are they multivoltine? Do infected larvae usually remain in trees or fall to the ground (which would be important for horizontal transmission)? What are the humidity requirements for sporulation of B. bassiana? How long are cadavers infective?
Methods
What was the approximate age or size of the larvae used? Were Sept. larvae the same size as those from Aug.? Were 3rd and 4th larval stages selected because of availability or preference for older larvae? If so, why? Would younger larvae be more susceptible to infection?
L 142. Please provide more complete description of the methods. What was the duration of exposure to cadavers, no. of larvae per cadaver, the container size, humidity level?
L 166. "... difference exists between observed mortalities and LD100." -- It is not clear to me why you want to test this hypothesis. Testing the significance of differences between the various treatment levels is less informative than defining the shape of the dose-response curve. Such a curve would enable one to predict mortality for any dose. Note that the X-axis in Fig. 1 could be spores/uL, which would allow fitting a curve from 0 to 7.9 x 10^4 spores per uL.
L 160. It is not clear to me what the utility of "interruption difference" is.
Results
Fig. 1 change from histograms to a dose-response curve.
L 204. A response curve would be much more informative than multiple comparisons.
Fig. 2 Change from histograms to fitting a dose-response curve. Why is it longer for E1T01 than E1C?
Fig. 3 The symbols of the boxplots should be defined in the legend. The treatments should be in order of increasing concentration (i.e. E1T20 to the right of E1T5). However, it would be better to replace this figure with a dose-response curve.
Table 4 No need for multiple comparisons. Define the shape of the dose-response curve.
Fig. 4 It would be better to replace this format with a dose-response curve.
Table 5 The legend should state that it refers to dieback speed. However, there is no need for multiple comparisons.
L 274-6. This text should be in the Discussion, not in the figure legend.
Discussion
It would be nice to estimate the number of spores per cadaver to relate to the dosage 7.9 x 10^4 spores per uL. This would help inform a discussion of what would be needed to produce a formulation that could be applied to insects in the field.
Author Response
Response to Reviewer 2
This is a well-designed study of the effect of dosage of a strain of B. bassiana on an important forest pest. However, the statistical analysis is not very relevant and can be greatly improved. I recommend deleting all the multiple comparisons and replacing them with fitting dose-response curves and probit analysis. The important results are to know and be able to predict the level of response to a given dosage. Presenting the results as curves and their equations would enable anyone to make such predictions. Note that Abbott's formula (Abbott 1925) is usually applied to correct for control mortality.
According to your comment we made changes in the manuscript, and used Abbotts formula as suggested.
It would be more correct to us the conventional 'lethal time' (LT) than "die back speed". Speed is normally distance/time, whereas the authors have measured time until death. It appears that the authors measured lethal time for all the insects in each treatment, which is a useful measure regarding population level responses (e.g., from the point of view of pest control). However, to measure the biological response, I think that it would be more appropriate to include only those individuals that ultimately sporulated. The same comments apply to 'feeding disruption'.
We agree and according to your comment we made changes in the manuscript (please see below).
The authors should define the strain of B. bassiana and deposit an accession in a repository.
This is not done yet, but it is a good suggestion thank you for that we will start the process.
There are minor grammatical and typographic errors (that I have not marked) that should be corrected.
We corrected such errors, Abstract was revised.
Other specific comments follow.
Introduction
Briefly describe the life history of D. pini. What is the approximate development time of larvae? Are they multivoltine? Do infected larvae usually remain in trees or fall to the ground (which would be important for horizontal transmission)? What are the humidity requirements for sporulation of B. bassiana? How long are cadavers infective?
We add the description.
(see: L 42-50)
Also, we add Sierpinska 1998 in the Reference list:
Sierpinska, A. 1998. Towards an integrated management of Dendrolimus pini L. In: McManus ML, Liebhold AM (eds) Proceedings: Population Dynamics, Impacts, and Integrated Management of Forest Defoliating Insects. USDA Forest Service General Technical Report NE-247, pp. 129-142
(see: L 384-386)
Methods
What was the approximate age or size of the larvae used? Were Sept. larvae the same size as those from Aug.? Were 3rd and 4th larval stages selected because of availability or preference for older larvae? If so, why? Would younger larvae be more susceptible to infection?
The approximate size of the larvae used was 5-7 cm in length. September larvae were in general maybe a bit bigger than August larvae, but all of them were chosen to be approximately similar size (bigger ones in August, and smaller ones in September). Assumption is that younger larvae would be more susceptible to infection, but 3rd and 4th larval stages were selected because of their availability in that time.
L 142. Please provide more complete description of the methods. What was the duration of exposure to cadavers, no. of larvae per cadaver, the container size, humidity level?
We add a sentence.
(see L: 145-146)
About humidity level: it was not measured, but in each Petri dish there was a cotton well soaked with distilled water, so we can only estimate that the level of humidity was high.
L 166. "... difference exists between observed mortalities and LD100." -- It is not clear to me why you want to test this hypothesis. Testing the significance of differences between the various treatment levels is less informative than defining the shape of the dose-response curve. Such a curve would enable one to predict mortality for any dose. Note that the X-axis in Fig. 1 could be spores/uL, which would allow fitting a curve from 0 to 7.9 x 10^4 spores per uL.
We change the Figure 1 see below.
(see L: 215-218)
L 160. It is not clear to me what the utility of "interruption difference" is.
We decided to delete this term from the text.
Results
Fig. 1 change from histograms to a dose-response curve.
We made the change
(see L: 215-218)
L 204. A response curve would be much more informative than multiple comparisons.
As we changed the figure 1, we find it rather informative to represent the results of multiple comparison because ANOVA demonstrates only the fact that there is a signifficant difference among groups in analysis, lacking to discriminate relationships between groups or to pin point which of the groups differe signifficantly from each other. Therefore, a multiple comparison is needed in order to discriminate the later.
We add paragraph on dose response curve;
(see L 210-214)
Fig. 2 Change from histograms to fitting a dose-response curve. Why is it longer for E1T01 than E1C?
We have changed the figure 1 and discarded table 3. However, we find it rather informative to represent the results of multiple comparison because ANOVA demonstrates only the fact that there is a signifficant difference among groups in analysis, lacking to discriminate relationships between groups or to pin point which of the groups differe signifficantly from each other. Therefore, in our opinion, a multiple pairwise comparison is needed in order to discriminate the later.
Why is it longer for E1T01 than E1C? We can not explain this observation distinctly. Perhaps it is the artefact of plain chance and natural life cycle difference between individual caterpillars as the effect of treatment was very small and practically negligible.
Fig. 3 The symbols of the boxplots should be defined in the legend. The treatments should be in order of increasing concentration (i.e. E1T20 to the right of E1T5). However, it would be better to replace this figure with a dose-response curve.
We made the change
(see L: 245-251)
We made the same changes to the Figure 5.
(see L: 276-282)
Table 4 No need for multiple comparisons. Define the shape of the dose-response curve.
We discard the table and keep the Figure as we find it important and iformative to clearly show which groups diferre signifficantly from one another.
Please see the caption of new Figure 1.
Fig. 4 It would be better to replace this format with a dose-response curve.
. As we have changed the figure 1 and discarded table 4., we find it rather informative to represent the results of multiple comparison because ANOVA demonstrates only the fact that there is a signifficant difference among groups in analysis, lacking to discriminate relationships between groups or to pin point which of the groups differe signifficantly from each other. Therefore, a multiple comparison is needed in order to discriminate the later.
Table 5 The legend should state that it refers to dieback speed. However, there is no need for multiple comparisons.
No, it refers to feeding disruption speed. We have changed the figure 1 and discarded tables 3 and 5.
L 274-6. This text should be in the Discussion, not in the figure legend.
We agree and change that in the text.
(see: 340-342)
Discussion
It would be nice to estimate the number of spores per cadaver to relate to the dosage 7.9 x 10^4 spores per uL. This would help inform a discussion of what would be needed to produce a formulation that could be applied to insects in the field.
Unfortunately, we don’t have that data. Estimation would be rather speculative.
Author Response File: Author Response.docx
Reviewer 3 Report
I have reviewed you article, and while it is an interesting and important topic, I feel that significant editing and formatting is necessary for further consideration. As well, I was wondering why a dose-response analysis where dose-response curves comparing LD50 values for different concentrations of fungus was not used. Such an analysis would allow for the results be better placed within the context of other similar studies. I found the experiment where cadavers were used to disseminate fungi to be interesting, as this is what would happen in the wild. I have attached further comments in a word document.
Comments for author File: Comments.docx
Author Response
Response to Reviewer 3
Abstract and Introduction
These sections are largely alright in terms of content; however, it must be heavily edited for the English language prior to any further consideration for publishing.
We made the editing of the grammar by a native speaker and replaced the Abstract with the new version
(see L: 10-30)
Methods
L86: Please provide latitude and longitude coordinates of collection sites.
We add te coordinates
(see L: 95)
L117: What age were the larvae at the time of treatment?
Treated larvae were at 3rd and 4th larval stage and their approximate size was 5-7 cm in length, but we are not familiar with how many days they were precisely old.
L122-L125: This explanation is listed, but should be written out in sentences within the methods section.
We rearranged the MM section as it as requested by reviewers. The specific explanation is now written in the text.
(see L: 135-136)
L125: The table does not have a title or caption. As well, the group names (E1T1, E1T5, etc.) are not defined. Also, all decimal points should be periods (‘.’) not commas (‘,’).
We change that and form Table 1 and Table 2. We change the decimal points to periods.
(see L: 139 and L:154)
L134: This explanation is listed, but should be written out in sentences within the methods section. This table could also be removed and written within the methods section.
We rearranged the MM section as it as requested by reviewers. The specific explanation is now written in the text.
(see L: 147-148)
L150: Insert ‘either’ between ‘with and ‘no’.
Done.
(see L: 160)
L157: The term ‘dieback speed’ is not clear. What are the units for this term?
We change the expression “dieback speed” to “Lethal time” and explain that units represent days.
(see L: 172-173)
L167: In most papers studying similar process, LD95 is used.
Based on LD100 the data shows very clear results, we would like to keep it this way.
L179, 180: Italicize B. bassiana.
Done.
(see L: 202 and L:204)
L180: I presume that this is over the 15 day observation period. If so, then this should be indicated. Also, (p) should be rewritten as P.
The 15 day observation is explained in MM section L: 135. We rewrite p to P.
I find this whole section confusing in that there are alternate ways to analyze this data that fit in with the current standards of measuring efficacy. When one wishes to find out how effective a potential treatment is on a pest population, they will test varying doses on separate groups and observe a) mortality, and b) when you have reached a target proportion killed. For example, one could test an isolate at 1%, 10%, 20%, and 50% dilutions, and measure mortality over a certain time period. From this data, you will have mortality curves for each respective dilution. You would then compares these curves and calculate LD50 and LD95, and identify when LD50 and LD95 occur. This information would also allow you to see if there’s any major differences between doses. Such an analysis would be simpler to understand.
Figure 1 is lacking a number of things, including: y-axis title, x-axis title, y-axis line. As well, the title of the graph should not be within the figure. Also, there is no explanation in the caption of what the treatment groups are, and no explanation of what ‘Bb’ and ‘m’ mean. Furthermore, if there are any significant difference in the mortality values, as was indicated in the table, then they should be indicated in the figure.
We have changed the figure 1. To dose-response chart as suggested by reviewers
(see L: 215-218)
L196: Dieback speed is again mentioned, but without any units it cannot be understood. As well, die speed here should be discussed as a mean, and not a set of values. As written, it implies that all values for 20ul are lower than all values for 5 and 1 (which themselves should be reversed in the sentence).
The ANOVA analysis conducted does not present a P-value.
We change the expression “dieback speed” to “Lethal time” and explain that units represent days.
(see L: 172-173)
The P-value for ANOVA is now included.
(Notification: Table 2 is now Table 4 as it requested by a reviewer)
(see L: 228-230)
L204: The explanation is unclear. Simply use the Tukey test to do pairwise comparisons.
We accept the suggestion;
A Tukey Multiple pairwise comparison grouped treatments in two groups; group A) – 1:100 diluted spore solution treatment; group B) 1 µL, 5 µL and 20 µL dose treatment; while the control group could not be discriminated between groups (Figure 4).
Furthermore, an ANOVA test is largely resistant to problems associated with non-normality, and is probably ok to use with the KW-test. I would have attempted to transform the data using log10 or ln then reanalyzed. If you do decide to use the KW test, then all ANOVA statistics can be deleted from this section.
We have performed both ln and Log10 transformations and repeated the ANOVA and in both cases the normality test failed again. We agree that non-parametric test is enough to be shown in this case, however, readers could be left wondering on what were the results of ANOVA and therefore we decided to keep both analyses.
Figure 2: Figure should not be in color, rather in greyscale (color does not translate well in black and white editions of the manuscript). Font and font sizing are not consistent with previous figure. These must be standardized between all figures. Bars also overlap with one another. Units for dieback speed should be included. No error bars on means are presented. These must be calculated and included. X-axis title can simply be ‘treatment group’, while y-axis can be ‘Mean dieback speed (unit) (+/- SE)’. Legend should also be placed in the top-right corner. Caption of the figure also does not explain the figure. The figure presents the mean dieback speeds of D. pini at different doses of B. bassiana.
We have made the changes according to suggestions.
(see L: 234-237)
The explanation of the statistical tests should be presented in the methods section. They should not be introduced in the results. When mentioning ‘parametric and non-parametric’ on L168, explanations of these tests should be made.
We add a new section 2.6. Statistical analysis where additional explanations are given as suggested:
(see L: 168-188)
We also add the reference in Line:
Abbott, W. S. 1925. A method of computing the effectiveness of an insecticide. Journal of economic entomology 18: 265-267
(see L: 440-441)
Table 3 requires an explanation of what exactly is being presented.
As we have changed the figure 1 as suggested by reviewers, we discarded table 3.
L229: Again, if data is non-normal, then it could be transformed and analyzed using an ANOVA. This analysis would be your best bet versus a non-parametric test. Also, if you decide to analyze it with the non-parametric test, then remove your results for the ANOVA test. There is no need to present both.
We have performed both ln and Log10 transformations and repeated the ANOVA and in both cases the normality test failed again. We agree that non-parametric test is enough to be shown in this case, however, readers could be left wondering on what were the results of ANOVA and therefore we decided to keep both analyses.
As with the introduction, significant editing of the grammar and sentence structure must be completed before this paper can be considered for Forests.
We made the editing of the grammar by a native speaker and replaced the Abstract with the new version (see L: 10-30)
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Although the Methods section states that Abbott's formula has been applied, this is not evident to me in the way that the results are presented. In Table 3 and Fig. 1, are the Bb mortalities the observed levels, or the corrected levels (mortality = 1 - survivorship, where Bb survivorship = 1- (total survivorship)/(Bb survivorship))? This should be made clear to the reader.
Fig. 1 is a nice addition to the paper. Clearly state whether the mortality rates are the observed rates, or those adjusted by Abbott's formula.
Please report the equations fitted in Fig. 1.
Author Response
Reply to Reviewer 2 (Second round):
Although the Methods section states that Abbott's formula has been applied, this is not evident to me in the way that the results are presented. In Table 3 and Fig. 1, are the Bb mortalities the observed levels, or the corrected levels (mortality = 1 - survivorship, where Bb survivorship = 1- (total survivorship)/(Bb survivorship))? This should be made clear to the reader.
We add L: 204:
… with Abbott’s correction
Fig. 1 is a nice addition to the paper. Clearly state whether the mortality rates are the observed rates, or those adjusted by Abbott's formula.
We add L 219-220:
Abbott's formula was used for calculation of relative mortality rates
Please report the equations fitted in Fig. 1.
We add L 220-221:
Equations and coefficients of determination for T and Bb dose response curves are y = ‑6.45x2 + 61.69x - 43.7; R2 = 0.9999 and y = -3.125x2 + 37.375x - 14.375; R2 = 0.9552 respectively.
Author Response File: Author Response.docx
