Inorganic Fungicides (Phosphites) Instead of Organic Fungicides in Winter Wheat—Consequences for Nitrogen Fertilizer Productivity

Round 1

Reviewer 1 Report

The manuscript entitled „Inorganic Fungicides (Phosphites) Instead of Organic Fungicides in Winter Wheat – Consequences for Nitrogen Fertilizer Productivity“ describes the possibility of substitution of organic fungicides with inorganic in winter wheat. The Authors showed that phosphites are not efficient substitutes of organic fungicides in winter wheat.

The introduction describes in detail key issues regarding the use of fungicides in winter wheat. Study design and methodology are appropriate for this type or research. It is a pity that the authors did not introduce a control variant without the use of fungicides in the experiment. This would allow us to assess whether phosphites protect wheat against fungi at all.

Overall, this is an interesting manuscript that presents the findings of a well designed and executed research. This manuscript delivers a clear answer that phosphites can not effectively replace classic fungicides.

 I find some misspellings in lines:

Line23: indices of wheat infestation with pathons aloowed for a reliable prediction of both TGW

 

Line 90: Plant nutrients, by affecting the metabolic and pjysiological processes of the plant

 

Line 121: Both factors are singificantly interrelated.

 

Conclusion, line 119: The nitrogen gap, qunatifying the inefficiency of the applied fertilizer

 

Author Response

Review report 1_response

The manuscript entitled „Inorganic Fungicides (Phosphites) Instead of Organic Fungicides in Winter Wheat – Consequences for Nitrogen Fertilizer Productivity“ describes the possibility of substitution of organic fungicides with inorganic in winter wheat. The Authors showed that phosphites are not efficient substitutes of organic fungicides in winter wheat.

The introduction describes in detail key issues regarding the use of fungicides in winter wheat. Study design and methodology are appropriate for this type or research. It is a pity that the authors did not introduce a control variant without the use of fungicides in the experiment. This would allow us to assess whether phosphites protect wheat against fungi at all.

These studies did not address whether phosphirones have a fungicidal effect. The essence was the possibility of replacing organic fungicides. The yield gap of almost 5 t ha^-1 indicates that this is not possible.

Overall, this is an interesting manuscript that presents the findings of a well designed and executed research. This manuscript delivers a clear answer that phosphites can not effectively replace classic fungicides.

 I find some misspellings in lines:

Line23: indices of wheat infestation with pathons aloowed for a reliable prediction of both TGW

It has been corrected.

Line 90: Plant nutrients, by affecting the metabolic and pjysiological processes of the plant

 It has been corrected.

Line 121: Both factors are singificantly interrelated.

  It has been corrected. This sentence now reads as follows:

 „Both factors are significantly but negatively interrelated.”

Conclusion, line 119: The nitrogen gap, qunatifying the inefficiency of the applied fertilizer

  It has been corrected. This sentence now reads as follows:

„The nitrogen gap, which quantified the ineffectiveness of the applied nitrogen fertilizer, was 55 and 16 kg N ha^-1, respectively, in favorable and unfavorable conditions for the growth of winter wheat.”

 

On behalf of the authors

Witold Grzebisz

Author Response File: Author Response.pdf

Reviewer 2 Report

this work about Inorganic Fungicides (Phosphites) Instead of Organic Fungicides  in Winter Wheat – Consequences for Nitrogen Fertilizer 4 Productivity. they found Phosphites do not substitute organic fungicides in reduction pathogen pressure in winter 26 wheat. Moreover, increased pressure of pathogens significantly reduces Nf productivity. i think the idea is done well but i still have some comments

1- rewrite your abstract ( you have many important data you didnt add

2- introduction

line 31- 33  (the reference number one is too old i am against it totaly (we must reduce N nowdays and we do that

plz organize your introduction

M&M

applying fertilizers, was variable for P; in the high range for K; medium range for Mg and 150 low range for Cu. where the value?

make table 2 as figure it will be easier to read

Results

the quality of figures is bad

why you didnt measure chlorophyll content with spectrophotometer not by SPAD . Carorrteniod content is important why not measure

 

Disscusion

plz make it deep

 

Author Response

Review report 2_response

Comments and Suggestions for Authors

this work about Inorganic Fungicides (Phosphites) Instead of Organic Fungicides  in Winter Wheat – Consequences for Nitrogen Fertilizer 4 Productivity. they found Phosphites do not substitute organic fungicides in reduction pathogen pressure in winter 26 wheat. Moreover, increased pressure of pathogens significantly reduces Nf productivity. i think the idea is done well but i still have some comments

• rewrite your abstract ( you have many important data you didnt add)

A correctly constructed abstract, in accordance with the instructions for authors, should contain: research hypothesis, basic elements of the methodology, basic results and the final conclusion that summarizes the research. The abstract contains all the basic components, including the size of the yield gap caused by substitution of organic fungicides by phosphites. Also included are nutritional monitoring tools with BBCH 75 stage as indicative.

The length of the abstract cannot exceed 200 words. If the editors of Agronomy increase the size of the abstract, the requirements of the reviewer will be fully met.

2- introduction

line 31- 33  (the reference number one is too old i am against it totaly (we must reduce N nowdays and we do that

The cited reference has been changed.

plz organize your introduction

This is not a review article. The introduction accounts for 14.4% of all characters and 17.3% of characters without cited leterature. This is in line with the general rules of Article type manuscripts.

I don’t know what the reviewer had in mind whean proposing changes in the organization  of this chapter. The presented issues take into account the key components of the conducted research.

The lack of reviewer’s comments regarding the purpose of this manuscript proves its correct organization.

M&M

applying fertilizers, was variable for P; in the high range for K; medium range for Mg and 150 low range for Cu. where the value?

All these details are in Table 1. There are included for each soil characteristic the mean value plus standard deviation.

make table 2 as figure it will be easier to read

Table 2 has been changed to a figure.

Results

the quality of figures is bad

The phrase „the quality of figures is bad” requires justification from the reviewer.

This emotional phrase should not be included in the review of a scientific paper.

The figures are legible and contain all the necessary information for the reader.

why you didnt measure chlorophyll content with spectrophotometer not by SPAD . Carorrteniod content is important why not measure

Determinantion of pigments in wheat leaf was not the aim of this study.

Let me include here two paragraphs from the article by Uddling et al. (2007): Evaluating the relationship between leaf chlorphyll concentration and SPAD-502 chlorophyll meter readings. Photosynth. Res. 91, 37-46.

„The leaf chlorophyll concentration, [chl], is usually determined by extraction from leaf samples and subsequent spectrophotometric measurements (e.g., Arnon 1949; Porra et al. 1989). Such in vitro determinations are destructive, expensive, and time consuming, and may therefore not be applicable for all purposes. There are also more rapid methods for estimating the leaf [chl] non-destructively, in vivo. These methods exploit the optical properties of leaves and are based on the reflectance and/or absorbance of radiation by chlorophyll”.

„The discrepancy between the absorption spectra of a leaf in vivo and an in vitro solution of chlorophyll-protein complexes is well-known and mainly caused by the non-uniform distribution of chlorophyll and multiple scattering in the intact leaf (Richter and Fukshansky 1996). So far, no study has quantitatively investigated how such deviations from the assumptions underlying Beer’s Law can be expected to influence the relationship between [chl] and the SPAD readings”.

Disscusion

plz make it deep

The Discussion accounts for 13.9% of all characters and 16.7% of characters without cited leterature. This is in line with the general rules of Article type manuscripts.

The discussion was significantly deepended, especially in terms of explaining the yield-forming effects of replacing organic fungicides with phosphites.

 

On behalf of the authors

Witold Grzebisz

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Main points:

Although it is stated in line 174 that the field experiment was a two-factor split-plot design, the Results section states that the effects were assessed by a two-way ANOVA. This suggests that the analysis was not done correctly. In a two-factor split-plot design, in which one factor (presumably the plant protection factor, although this doesn't seem to be stated) is allocated to the main plots, and the other factor (presumably the foliar phosphite) is allocated to the  subplots, there are two error terms, and the tests for significance of the two factors should be carried out using different error terms. There is no indication from the results presented in Tables 4 & 5 that these tests have been done correctly. The statement that the effects were assessed using a two-way ANOVA suggests otherwise.

      The other problem with the results given in Tables 4 & 5 is that the means and subsequent post hoc Tukey tests are not informative when interactions are significant. Consider Table 4, where the interaction Y x PPMs is significant for several response variables, e.g. grain yield GY. The means are given only for the main effects, but it has to be recognized that these are averages taken over the other factor. These means do not take into account  the fact that the interaction was significant. Rather than present means for PPMs which are averages over the two years, it would be more informative to have additional rows in Table 4, so that then means can be presented for each of the two years separately.

        A similar comment can be made about Table 5, which is even more complicated because there is not only a highly significant Y x PPMs interaction for most response variables, but also a highly significant Phi x PPMs  interaction and a highly significant 3-factor interaction. This presents a very real challenge to tabulate and interpret the results. The authors may need to provide additional tables in order to achieve that.

         Lines 258-259. Consider the sentence “In contrast, HI responded only to years and PPMs, but not to the interaction of both factors.” This sentence makes no sense to a trained applied statistician. Response variables do not “respond” to an interaction. A significant interaction tells the reader that the assumed additive model specified in the ANOVA needs to be rejected. An example would be that significant differences between PPMs for one phi variant were not observed for one (or both) of the other variants. There are, however, many other ways in which interactions can be significant and the reader should be given tables of means to show the effects properly.

Other points:

            There are an extraordinary number of misspellings in the manuscript. In the Abstract, line 24, there is  “pathons aloowed”

In the Introduction, the following misspellings occur:

Line 47. “regardles”

Line 51. “controvery”

Line 57. “qustions”

Line 62. “metioned”

Line 72. “immediatley”

Line 80. “dispite”

Line 118 “lenght”

Line 123. “attact”

Line 129. “phostosynthetic”

Line 137 “pahogens”

Many other misspellings appear in the later sections.

Line 125. The authors of the fungal species are incorrect. Use Index Fungorum (online). They should read

 Zymoseptoria tritici (Roberge ex Desm.) Quaedvl. & Crous  

and Puccinia recondita Roberge ex Desm.

In addition, there are many grammatical errors. For example, Line 62 has “which has been allowed interest”, Lines 75-76 has “replaces of one”, Line 96 should have “elements”, not “element”, and so forth.

Line 149. The words “was variable for P” tells the reader very little. More informative would be to state that P was very high in 2016/17 but low in 2017/18.

Table 2. The reader should not have to guess which month IX (and others) signifies. Replace the Roman numerals by Jan, Feb, Mar, etc.  or use the initial letter J F M A etc.

Line 211. “as shown for the variant F”. Do you mean method F? It would be clearer if you used ‘variant’ consistently for foliar applied phosphite and ‘method’ for the six PPMs.

Tables A1, A2 & A3, n=36. Is this obtained from the product of 2 years, 3 variants and 6 methods (=2 x 3 x 6)? But there were 4 replicates of each treatment combination, so were mean values used for the regression analysis and calculation of the correlations? If so, this is non-standard and needs to be clarified and justified.

Line 256. “clearly clearly”. One clearly is enough.

Lines 263 & 269. Is Table 4 meant rather than Table 1? Or should it be Table A1?

Figures 1-5. “standard deviation value of the mean”. Standard deviations are usually used to characterize populations, not samples. Did you really intend “standard error (s.e.) of the mean”? That would make sense, as s.e.’s can be used to make rapid approximate assessments as to which pairwise means significantly differ from other pairs.

 

 

Author Response

Review report 3_response

 

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Are all the cited references relevant to the research?	(x)	( )	( )	( )
Is the research design appropriate?	( )	( )	(x)	( )
Are the methods adequately described?	(x)	( )	( )	( )
Are the results clearly presented?	( )	( )	(x)	( )
Are the conclusions supported by the results?	( )	( )	( )	(x)

Comments and Suggestions for Authors

Main points:

Although it is stated in line 174 that the field experiment was a two-factor split-plot design, the Results section states that the effects were assessed by a two-way ANOVA. This suggests that the analysis was not done correctly. In a two-factor split-plot design, in which one factor (presumably the plant protection factor, although this doesn't seem to be stated) is allocated to the main plots, and the other factor (presumably the foliar phosphite) is allocated to the  subplots, there are two error terms, and the tests for significance of the two factors should be carried out using different error terms. There is no indication from the results presented in Tables 4 & 5 that these tests have been done correctly. The statement that the effects were assessed using a two-way ANOVA suggests otherwise.

Here are some clarifications to the reviewer’s comments:

1. Field layout of the experiment:

Phosphirone (Phi)	Plant protection methods (PPMs, acronyms)	Plant protection methods – detailed description of tested variants
Cu	A	3 × FP + 3 × Phi
	B	3 × FP + 4 × Phi
	C	FP1+ FP2 + FP3 + Phi1
	D	FP1+ FP2 + Phi1 + Phi4
	E	FP1 + Phi1+ Phi + Phi4
	F	4 × Phi
Mg	A	3 × FP + 3 × Phi
	B	3 × FP + 4 × Phi
	C	FP1+ FP2 + FP3 + Phi1
	D	FP1+ FP2 + Phi1 + Phi4
	E	FP1 + Phi1+ Phi + Phi4
	F	4 × Phi
Cu/Mg	A	3 × FP + 3 × Phi
	B	3 × FP + 4 × Phi
	C	FP1+ FP2 + FP3 + Phi1
	D	FP1+ FP2 + Phi1 + Phi4
	E	FP1 + Phi1+ Phi + Phi4
	F	4 × Phi

Legend:  A ─ full fungicide protection + 3 × phoshite application; B ─ full fungicide protection + 4 × phoshite application; C ─ full fungicide protection + phosphite at BBCH 21; D ─ fungicide protection at BBCH 30 and BBCH 39-45 + phosphite at BBCH 21 and BBCH 55; E ─ fungicide protection at BBCH 30 + phosphite at BBCH 21, BBCH 32, BBCH 55; F – phosphite alone at BBCH 21, BBCH 29, BBCH 32, BBCH 55.

      The other problem with the results given in Tables 4 & 5 is that the means and subsequent post hoc Tukey tests are not informative when interactions are significant. Consider Table 4, where the interaction Y x PPMs is significant for several response variables, e.g. grain yield GY. The means are given only for the main effects, but it has to be recognized that these are averages taken over the other factor. These means do not take into account  the fact that the interaction was significant. Rather than present means for PPMs which are averages over the two years, it would be more informative to have additional rows in Table 4, so that then means can be presented for each of the two years separately.

Unfortunately, I cannot agree with the reviewr’s opinion. I attach the basic facts:

1. Excerpt from an ANOVA, regarding GY. This analysis clearly shows that the main factors determining the yield were years (Y) and methods of wheat protection (PPMs).
2. The interaction of both factors clearly indicates that the yield of wheat regardless of the weather, in both years showed the same response to the tested methods of wheat canopy protection. The relevant figure is attached (see below).
3. As presented in the manuscript, the main yield-forming factor was weather in a given growing season. Despite this, each year wheat protected with phosphirones yielded worse. There is no formal justification (ANOVA) to discuss grain yield separately for years.

	Degree	GY	GY	GY	GY
Free word	1	11755,70	11755,70	25777,80	0,000000
Y - year	1	329,62	329,62	722,80	0,000000
Phi	2	1,94	0,97	2,13	0,124296
PPM	5	107,52	21,50	47,15	0,000000
Y*Phi	2	0,48	0,24	0,53	0,592282
Y*PPM	5	48,77	9,75	21,39	0,000000
Phi*PPM	10	4,38	0,44	0,96	0,482427
Y*Phi*PPM	10	4,18	0,42	0,92	0,519961
Error	108	49,25	0,46
Total	143	546,15

 

Figure 1. Effect of plant protection methods on grain yield of winter wheat in two growing  seasons

        A similar comment can be made about Table 5, which is even more complicated because there is not only a highly significant Y x PPMs interaction for most response variables, but also a highly significant Phi x PPMs  interaction and a highly significant 3-factor interaction. This presents a very real challenge to tabulate and interpret the results. The authors may need to provide additional tables in order to achieve that.

1. The primary goal of crop production for the farmer is yield.
2. These studies clearly demostrated, as discussed above, that grain yield was determined by the interaction of Y × Therefore, any three-factor interactions, i.e. Y × Phi × PPMs, are of secondary importance. These cases were clearly emphasized when discussing these characteristics.
3. The role of weather (years), especially for SPAD 75, was discussed in detail both in the results and in the discussion and conclusion.

         Lines 258-259. Consider the sentence “In contrast, HI responded only to years and PPMs, but not to the interaction of both factors.” This sentence makes no sense to a trained applied statistician. Response variables do not “respond” to an interaction. A significant interaction tells the reader that the assumed additive model specified in the ANOVA needs to be rejected. An example would be that significant differences between PPMs for one phi variant were not observed for one (or both) of the other variants. There are, however, many other ways in which interactions can be significant and the reader should be given tables of means to show the effects properly.

I agree with the reviewer on this point, but the ANOVA is pretty clear for HI (see Figure 2). This sentence was removed from the text.

	Degree	HI	HI	HI	HI
Free word	1	293772,2	293772,2	24023,24	0,000000
Y - year	1	390,4	390,4	31,93	0,000000
Phi	2	2,4	1,2	0,10	0,905878
PPM	5	261,4	52,3	4,27	0,001382
Y*Phi	2	5,7	2,8	0,23	0,793402
Y*PPM	5	96,5	19,3	1,58	0,172097
Phi*PPM	10	74,4	7,4	0,61	0,803950
Y*Phi*PPM	10	188,7	18,9	1,54	0,134130
Error	108	1320,7	12,2
Total	143	2340,1

Figure 2. Effect of plant protection methods on harvest index of winter wheat in two growing  seasons.

Other points:

            There are an extraordinary number of misspellings in the manuscript. In the Abstract, line 24, there is  “pathons aloowed”

It has been corrected.

In the Introduction, the following misspellings occur:

Line 47. “regardles”

Line 51. “controvery”

Line 57. “qustions”

Line 62. “metioned”

Line 72. “immediatley”

Line 80. “dispite”

Line 118 “lenght”

Line 123. “attact”

Line 129. “phostosynthetic”

Line 137 “pahogens”

Many other misspellings appear in the later sections.

All spelling errors in the text have been corrected.

Line 125. The authors of the fungal species are incorrect. Use Index Fungorum (online). They should read

 Zymoseptoria tritici (Roberge ex Desm.) Quaedvl. & Crous 

and Puccinia recondita Roberge ex Desm.

It has been corrected.

In addition, there are many grammatical errors. For example, Line 62 has “which has been allowed interest”, Lines 75-76 has “replaces of one”, Line 96 should have “elements”, not “element”, and so forth.

It has been corrected.

Line 149. The words “was variable for P” tells the reader very little. More informative would be to state that P was very high in 2016/17 but low in 2017/18.

It has been corrected.

Table 2. The reader should not have to guess which month IX (and others) signifies. Replace the Roman numerals by Jan, Feb, Mar, etc.  or use the initial letter J F M A etc.

Table 2 has been replaced by a figure.

 

Line 211. “as shown for the variant F”. Do you mean method F? It would be clearer if you used ‘variant’ consistently for foliar applied phosphite and ‘method’ for the six PPMs.

It has been corrected. The term ”variant” referes to the form and combination of the applied phosphites and the term „method” to the plant protection treatments. The terms used are explained in detail in the MM chappter.

Tables A1, A2 & A3, n=36. Is this obtained from the product of 2 years, 3 variants and 6 methods (=2 x 3 x 6)? But there were 4 replicates of each treatment combination, so were mean values used for the regression analysis and calculation of the correlations? If so, this is non-standard and needs to be clarified and justified.

The realiability of a given relationship is assessed on the basis of the coefficient of determinantion (R²) and probability (p) for a given number of observations (n).

In agricultural, field studies with a specified number of blocks (repitations), the correlation analysis is carried out on averages for given combination. The rule of assessing the significance of a given correlation is simple, the lower the „n” number, the greater the critical values of R² and p.

This requirement is met and presented for each matrix table.

Line 256. “clearly clearly”. One clearly is enough.

It has been corrected.

Lines 263 & 269. Is Table 4 meant rather than Table 1? Or should it be Table A1?

It has been corrected.

Figures 1-5. “standard deviation value of the mean”. Standard deviations are usually used to characterize populations, not samples. Did you really intend “standard error (s.e.) of the mean”? That would make sense, as s.e.’s can be used to make rapid approximate assessments as to which pairwise means significantly differ from other pairs.

 This has been corrected.

Are the conclusions supported by the results?

This opinion is puzzling in the light of the content of the conclusions presented. I fully disagree with it.

The conclusions contain all the basic components, including the size of the yield gap caused by substitution of organic fungicides by phosphites. Also included are nutritional monitoring tools with BBCH 75 stage as indicative.

This chapter was rewritten so that there is no doubt about the conclusions of these studies.

On behalf of the authors

Witold Grzebisz

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

The author's response (Grzebisz acting on behalf of all the authors) does not answer my question about the field layout, that is, how the treatment combinations (3 phosphite variants and 6 plant protection methods) were physically laid out in the field plots. As I described in detail in my review of the original manuscript, a split-plot design, which the authors still maintain was used, has plots of two different sizes, the main plots being used for one of the factors and the subplots (smaller units into which the main plots are subdivided) used for the other factor. The authors have not addressed this. Instead, they still maintain that the statistical analysis was a "two-way ANOVA". This is contradicted by the response to my original review in which he (Grzebisz) presents a table (which has 143 degrees of freedom for Total) giving the ANOVA for grain yield. This is clearly a three-way ANOVA. That is, there are 3 factors here, not 2, namely Year, Phi and PPM. This is not the way a split-plot design should be analysed; in a split-plot design, there are two error terms, one for the large plots and one for the sublots. The F-test for the main plot factor has a different denominator than the F-test for the interaction and the subplot factor. The addition of Year as an extra factor complicates the analysis even further.

             The interactions involving GY and TGW are not difficult to understand, in view of the graphs presented in Figures 2 and 3. But the same cannot be said for many of the response variables given in the original Table 5, where there were not only highly significant two-factor interactions but also a highly significant 3-factor interaction. To properly present the results so that readers can comprehend their importance can be rectified by the presentation of a small number of additional graphs. The authors should stop being recalcitrant and agree to doing this.

                I stick to my original recommendation that a major revision is required. They must clarify whether the design really was a split-plot design, and if so, the layout (main plots and subplots) should be specified. Additional graphs should be added to illustrate the significant interactions for the important response variables.

Author Response

Review Report 2 – response

The author's response (Grzebisz acting on behalf of all the authors) does not answer my question about the field layout, that is, how the treatment combinations (3 phosphite variants and 6 plant protection methods) were physically laid out in the field plots. As I described in detail in my review of the original manuscript, a split-plot design, which the authors still maintain was used, has plots of two different sizes, the main plots being used for one of the factors and the subplots (smaller units into which the main plots are subdivided) used for the other factor. The authors have not addressed this. Instead, they still maintain that the statistical analysis was a "two-way ANOVA". This is contradicted by the response to my original review in which he (Grzebisz) presents a table (which has 143 degrees of freedom for Total) giving the ANOVA for grain yield. This is clearly a three-way ANOVA. That is, there are 3 factors here, not 2, namely Year, Phi and PPM. This is not the way a split-plot design should be analysed; in a split-plot design, there are two error terms, one for the large plots and one for the sublots. The F-test for the main plot factor has a different denominator than the F-test for the interaction and the subplot factor. The addition of Year as an extra factor complicates the analysis even further.

             The interactions involving GY and TGW are not difficult to understand, in view of the graphs presented in Figures 2 and 3. But the same cannot be said for many of the response variables given in the original Table 5, where there were not only highly significant two-factor interactions but also a highly significant 3-factor interaction. To properly present the results so that readers can comprehend their importance can be rectified by the presentation of a small number of additional graphs. The authors should stop being recalcitrant and agree to doing this.

                I stick to my original recommendation that a major revision is required. They must clarify whether the design really was a split-plot design, and if so, the layout (main plots and subplots) should be specified. Additional graphs should be added to illustrate the significant interactions for the important response variables.

Response

1. I confirm that field experimental setup, which was the source of the results, was a two-factor  split-plot design.
2. The main factor was three forms/combinations of phosphirones.
3. The second order factor was the fungicide/phosphite combinations (subplots).

This information has been added to the MM chapter.

4. The experimental setup was tested (repated) in two growing seasons, i.e. 2016/2017 and 2017/2018.
5. The environmental factor – the growing season – is a random factor. Thus it had to be included as the first factor in statistical analyses.
6. Statistical analysis of the results from two growing seasons clearly showed that the main factor factor was not signifificant for the final effect (end effect in crop production) which is yield (Table 3).
7. Thus, the interaction Y x PPMs was crucial for the yield. It has been also confirmed for the yield determinants, i.e. TGW and B89.
8. For a number of characteristics, such as NL-89, SPAD75, SEPPTR, PUCCRTm GREENT, the interaction analysis Y x Phi x PPMs was significant, but the Years were still the dominant factor (Tables 4 and 5).
9. The analysis of grain yield separately for years is neither fomally nor logically necessary. The farmer and his adviser expect information whether a given factor, in this case, PPMs, works or not. This is the purpose of repeating the same experimental variant for at least two years.
10. Reporting the results as shown in the figure below does not provide any information. The main factor did not work in the 2016/2017 growing season. The fact that it acted in the 2017/2018 season does not matter to the user of these results. It is just an artifact. The differences between extreme combinations in both growing seasons was 5.69 t ha^-1.

 

Figure 1. Grain yield of winter wheat in two growing seasons in response to interaction of experimental factors. 

 

On behalf of the authors

Witold Grzebisz

 

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

   Thank you for your email. I have downloaded the latest version (ver. 3) and have read the authors' response. I see that  authors have made some changes as a result of my comments about the split-plot design.  authors have revised the wording in the Experimental Design section of the Materials and Methods and made clear which were the main plots and which were the subplots. That's fine. But they still persist, in the Statistical Analysis section (line 245), in saying that the factors and their interaction were assessed using "two-way ANOVA". To a statistician and to experienced agricultural scientists who do experimental field work, a two-way ANOVA implies that there is a single error term against which all the sources of variation are assessed, encompassing the main effects and their interaction. But that is not the case with a split-plot design. Because the plots and the subplots are of different sizes, there are now two error terms. The 'main plot' error term is used to assess the effect of the main plot treatment factor, and the 'subplot' error term is used to assess the effect of the subplot treatment factor and the interaction between the two factors. It is a more efficient design than a so-called factorial design for assessing this interaction, but it sacrifices some efficiency in the test for the main plot treatment.

 

         Therefore, I would like to suggest that the authors modify the wording of line 245 of their revised manuscript and replace "two-way ANOVA" by "analysis of variance". Although this latter wording lacks detail, readers will then assume that the authors did the statistical analysis correctly (and hopefully authorsy did!).

Author Response

Review Report 3 ver.3 - Response

Thank you for your email. I have downloaded the latest version (ver. 3) and have read the authors' response. I see that  authors have made some changes as a result of my comments about the split-plot design.  authors have revised the wording in the Experimental Design section of the Materials and Methods and made clear which were the main plots and which were the subplots. That's fine. But they still persist, in the Statistical Analysis section (line 245), in saying that the factors and their interaction were assessed using "two-way ANOVA". To a statistician and to experienced agricultural scientists who do experimental field work, a two-way ANOVA implies that there is a single error term against which all the sources of variation are assessed, encompassing the main effects and their interaction. But that is not the case with a split-plot design. Because the plots and the subplots are of different sizes, there are now two error terms. The 'main plot' error term is used to assess the effect of the main plot treatment factor, and the 'subplot' error term is used to assess the effect of the subplot treatment factor and the interaction between the two factors. It is a more efficient design than a so-called factorial design for assessing this interaction, but it sacrifices some efficiency in the test for the main plot treatment.

         Therefore, I would like to suggest that the authors modify the wording of line 245 of their revised manuscript and replace "two-way ANOVA" by "analysis of variance". Although this latter wording lacks detail, readers will then assume that the authors did the statistical analysis correctly (and hopefully authorsy did!).

Response

I thank the reviewer for critical comments regarding the formal analysis of the results. A suggested correction has been made in the text of the manuscript. Statistical analysis of the results was carrierd out in accordance with the design of the experiment. The problem for the authors was the fact that taking into account the replication of the experiment (growing seasons) the main factor was not significant.

On behalf of the authors

Witold Grzebisz

Author Response File: Author Response.pdf