Can Nitrogen Fertilizer Management Improve Grain Iron Concentration of Agro-Biofortified Crops in Zimbabwe?

Round 1

Reviewer 1 Report

The current manuscript investigates if similar effects are observed for grain Fe concentration and grain yield in two years, in two contrasting agro-ecologies in Zimbabwe, on maize, cowpea and two finger millet “seed pools” collected during previous farmer surveys and representing “high” and “low” Fe concentrations. They found that grain Fe concentration increased according to species: maize < finger millet<cowpea but varied widely according to treatment. Significant effects of N-form on grain Fe concentration were observed in the low finger millet “seed pool”, for which mineral N fertilizer application increased grain Fe concentration to a greater extent than other N forms, but not for the other species. Co-application of N and Fe fertilizer significantly increased maize, cowpea, and finger millet grain yields. They also demonstrated that there was site-specific effects (agro-ecology) on grain yields and grain Fe concentration.

The manuscript is well written and discussed, and I think it is an interesting contribution to how nitrogen fertilizer management can improve grain iron concentration and grain yield for agro-biofortified crops such as maize, cowpea and millet in Zimbabwe.

Author Response

We are glad that the reviewer likes this paper, and has no comments requiring our response.

Reviewer 2 Report

The paper titled “Can nitrogen fertilizer management improve grain iron concentration of ago-biofortified crops in Zimbabwe?” discusses nitrogen fertilizer management of two region and 3 crops of Zimbabwe in conjunction with iron concentrations.   Below are both minor and major revisions that should be consider:

Major concerns:

• This study is complicated with 7 treatments and 4 crops (if you categorize the two millet seed pools as separate crops) and two locations. That issue in and of itself isn’t a problem with the correct statistics.   Currently the reading of the methods and results make it very hard for the reader to understand what how and why of statistics were conducted and what they mean.   I recommend a running a mix model with treatment, location and crop.  Yes, there are some obvious results (i.e. crop is going to be significant, but then you have a justification to run separate ANOVA tests).  This will also support the argument that environmental effects are impacting the results.   Also, there is also no analysis of means conducted (i.e. student T test or Tukey).  Adding this would strengthen your comparisons of all the treatments and you wouldn’t need to run separate ANOVAs for each comparison.
• Based on current stats, line 215 -217 are misrepresented. Lines 261-263 contradict each other of for data interpretation.
• The methods are incomplete. There should be a detailed paragraph on how iron was measured in the laboratory.  Those methods should be cited as well.  If samples were conducted by another laboratory who they are and what methods were used should be stated.

Minor concerns:

• Line 22 – Either missing a period or sentence is unclear.
• Lines 63-73 – Rewrite paragraph for clarity. Currently confusing to reader.
• Line 77 – In formal writing “This” must be followed by a subject.
• Line 98-104 – Description of methods and should not be in the introduction.
• Line 132 – Table 1 does not support the sentence that it is cited for.
• Line 139- 141 – Confusing, please rewrite for clarity.
• Line 143 – Put treatments in a table not equation formatting
• Lines 145- 158 – Very confusing. Please reorganize or use transition words to help the reader follow along.
• Line 171- If you are using methods outlined in another paper, still write a quick summary of those details. Readers are not going to want to chase down references to see what your actual methods are.  (i.e. In brief…)
• Lines 173-194 – no need to repeat information that is in table form (Table 3).
• Line 198 – R must be cited (company) and if any packages were used, they need to also be cited. If no additional packages are used it should be cited “base R (XXXXX).
• Line 200 – “ANOVA tables…” This sentence is redundant, please remove.
• Lines 332 & 337 – “Data not shown” should be placed in supplementary section.
• Fe concentrations should be in a table.
• TABLE 1 – Move crops grown into a row instead of a column. Recommend adding number of plots per crop and location.  Also find a way to differentiate Mutasa location 1 and Mutasa location 2 on the table.
• FIGURES 2-4 – Add error bars to the bar graph not just above in the legend. Readers should be able to quickly determine if there is a difference.  Also, add letters or symbols to show significance between bars (i.e. Tukey comparison of means test).
• FIGURE 4 – Merge locations with different colored bars to differentiate seed pools for millet. Also, Mutasa is hard to read because the bars are very small; either reduce the size of y-axis for both locations or just Mutasa and note the change in axis in the figure legend.
• Overall the discussion is well written, however discussion points may change based on better statistical analysis.

Author Response

Major concerns:

 

This study is complicated with 7 treatments and 4 crops (if you categorize the two millet seed pools as separate crops) and two locations. That issue in and of itself isn’t a problem with the correct statistics.   Currently the reading of the methods and results make it very hard for the reader to understand what how and why of statistics were conducted and what they mean.  

 

We have added some additional text at lines 216 onward to clarify the reason for the analysis that we have done. The design has also been used/published for previous zinc study (Manzeke et al., 2020. Agronomy Journal).

 

I recommend a running a mix model with treatment, location and crop.  Yes, there are some obvious results (i.e. crop is going to be significant, but then you have a justification to run separate ANOVA tests).  This will also support the argument that environmental effects are impacting the results.  

We do not think it appropriate to undertake a single mixed model because we thought it very unlikely that treatment effects would be additive with the substantial differences between the Natural Regions. This expectation was justified by the differences between the sites in our final results.  For example, there was much stronger evidence for a N-rate effect on cowpea yield (Table 5) at Mutasa than at Hwedza, while at Hwedza there was stronger evidence for differences between forms of nitrogen fertilizer than at Mutasa. On this basis more informative results would be obtained from separate analyses for data from the different sites.  There is a comment on this in the revised text at lines 200-205.

Within any one site a linear mixed model with crop as a factor would be incorrect because there is no genuine replication of this factor (there was one set of experimental plots with each crop, rather than independent randomization of crops among available plots, or multiple replicates of crops at main-plot level in a split plot design).

Also, there is also no analysis of means conducted (i.e. student T test or Tukey).  Adding this would strengthen your comparisons of all the treatments and you wouldn’t need to run separate ANOVAs for each comparison.

As explained in the revised text we have chosen to examine prior contrasts among the means of treatments, or groups of treatments, to test our hypotheses.  This is the most robust way to analyse an experiment, making best use of experimental power.  Contrasts among means, with appropriate control for multiple hypothesis testing as in Tukey’s method, is appropriate as a “wash-up” activity after the primary hypotheses have been evaluated, but not as the main means of inference.  This is well established in the statistical literature (e.g. Welham et al., 2015; Webster and Lark, 2018).

Welham, S.J., Gezan, S.A., Clark, S.J., Mead, A. 2015.  Statistical methods in biology.  CRC Press, Boca Raton.

Webster, R., Lark, R.M.  2018.  Analysis of variance in soil research: let the analysis fit the design.  European Journal of Soil Science 69, 126 – 139.

 

Based on current stats, line 215 -217 are misrepresented. Lines 261-263 contradict each other of for data interpretation.

Lines 215-217 in the original manuscript describe the analysis as we undertook it.  We have added some additional words to ensure that it is clear.

At Lines 261-263 we have clarified this by revisions to the text.  The second inference was not about N rate but the form of nitrogen applied.

The methods are incomplete. There should be a detailed paragraph on how iron was measured in the laboratory.  Those methods should be cited as well.  If samples were conducted by another laboratory who they are and what methods were used should be stated.

The Methods section on laboratory analyses have now been presented in detail and methods used cited in text and in the reference list (Lines 174-188) as follows:

 

Grain samples were collected from a net plot of 3.0 x 1.8 m² at physiological maturity (BBCH principal growth stage 9) and analysed according to standard laboratory protocols for Fe and other multi-elements as described in Watts et al. [36] and Joy et al. [37]. About 0.2 g of milled grain sample was digested using a microwave heating system for 90 minutes and at a pressure of 2 megapascals (MPa) in 2 mL of 70% trace analyses grade HNO₃, 1.0 mL of H₂O₂ and 1 mL of milli-Q water. Digested samples were then analysed using an inductively coupled plasma mass spectrometer (ICP-MS, Agilent 8900 Triple Quad, Santa Clara, CA). Two samples of a certified wheat reference material (National Institute of Standards and Technology-NIST 1567b), two blank and three random sample duplicates were included in the analyses of every 41 samples to ensure accuracy of measured Fe.

Air-dried and ground (0.5 mm) cattle manure samples were analysed for total Zn and Fe using the aqua regia method. Total N was determined using the Kjeldahl procedure described in Okalebo et al. [38]. Total P was determined using wet digestion with aqua regia by Anderson and Ingram [39] and analysed as described by Murphy and Riley [40]. Exchangeable bases (sodium-Na, potassium-K, magnesium-Mg and calcium-Ca) were analysed following extraction with 1 M acidified ammonium acetate (NH₄OAc) as described by Anderson and Ingram [39]. Organic carbon was determined using the Walkley-Black method [38]. The nutrient composition of cattle manure used is shown in Table 2 and has been presented earlier as Manzeke et al. [19]. 

 

Minor concerns:

 

1. Line 22 – Either missing a period or sentence is unclear.

Sentence has now been revised for clarity to read “The two finger millet “seed pools” were collected during previous farmer surveys to represent “high” and “low” Fe concentrations”.

2. Lines 63-73 – Rewrite paragraph for clarity. Currently confusing to reader.

Paragraph has now been revised for improved readability as suggested by the Reviewer (Lines 64-77).

3. Line 77 – In formal writing “This” must be followed by a subject.

The sentence has been rewritten for clarity.

4. Line 98-104 – Description of methods and should not be in the introduction.

Text has now been moved to lines 134-35 and lines 155-159.

5. Line 132 – Table 1 does not support the sentence that it is cited for.

Deleted the word “Table 1” in the main text.

6. Line 139- 141 – Confusing, please rewrite for clarity.

Sentence has now been rewritten for clarity (Lines 149-152).

7. Line 143 – Put treatments in a table not equation formatting

Treatments for maize, cowpea and finger millet have now been presented in Tables.

8. Lines 145- 158 – Very confusing. Please reorganize or use transition words to help the reader follow along.

Paragraph has now been revised to include comments from Reviewer 2 and 3.

9. Line 171- If you are using methods outlined in another paper, still write a quick summary of those details. Readers are not going to want to chase down references to see what your actual methods are. (i.e. In brief…)

The orthogonal contrasts have been detailed in Table 1 and text briefly describe. The reference was just included to indicate that the contrasts have been presented in earlier work. Text has now been revised from “…as detailed in 19…” to “…as detailed earlier in 19…”.

 

10. Lines 173-194 – no need to repeat information that is in table form (Table 3).

Text in the paragraph referred to by the reviewer explicitly explains the hypotheses we were testing. Authors think presenting the hypotheses would assist readers in understanding the tested contrasts which are presented in Table 3.

11. Line 198 – R must be cited (company) and if any packages were used, they need to also be cited. If no additional packages are used it should be cited “base R (XXXXX).

R is now cited as base R (R Core Team, 2014). The reference is number 38 in the list.

 

12. Line 200 – “ANOVA tables…” This sentence is redundant, please remove.

We have now deleted the sentence as suggested by the Reviewer.

 

13. Lines 332 & 337 – “Data not shown” should be placed in supplementary section.

We have now placed the ANOVA output Tables in the supplementary material as suggested by the Reviewer.

14. Fe concentrations should be in a table.

We presented the grain Fe in a graph to indicate the trend in changes with N addition. We also wanted to highlight the initial seed Fe content (in red) on the graph. We do not think the Fe concentrations should be presented in a Table given that the means are all shown as plots.

15. TABLE 1 – Move crops grown into a row instead of a column. Recommend adding number of plots per crop and location. Also find a way to differentiate Mutasa location 1 and Mutasa location 2 on the table.

Table has been revised as suggested by the Reviewer.

16. FIGURES 2-4 – Add error bars to the bar graph not just above in the legend. Readers should be able to quickly determine if there is a difference. Also, add letters or symbols to show significance between bars (i.e. Tukey comparison of means test).

 

No.  It is better to put a single bar, emphasizing that the SE is pooled.  Not adding multiple comparisons for reasons given.

 

17. FIGURE 4 – Merge locations with different colored bars to differentiate seed pools for millet. Also, Mutasa is hard to read because the bars are very small; either reduce the size of y-axis for both locations or just Mutasa and note the change in axis in the figure legend.

The option of merging the plots within location was considered, but it makes it harder to compare the treatments within a single “seed pool”. We think that, with common scales on the y axis for any location, and with the plots for the seed pools adjacent, the comparison can be made on the plots as they stand.

We have considered the Reviewer’s comment to change the y-axis for Mutasa District to make treatment differences clearer. We also included a caption which draws the reader‘s attention to the difference in the y-axis scales between the two plots.

18. Overall the discussion is well written, however discussion points may change based on better statistical analysis.

Based on our responses t the Reviewer on statistics, authors did not consider revising the discussion.

Author Response File: Author Response.docx

Reviewer 3 Report

Review of the publication: „ Can nitrogen fertilizer management improve grain  iron concentration of agro-biofortified crops in  Zimbabwe?”

           

The iron deficiency problem is very big all over the world. The research concerns an important social and economic problem of plant biofortification into iron in Zimbabwe

 

The quality of the research and the organization of the manuscript are excellent.

It is a pleasure to review such articles.

 

 I haven't found any serious shortcomings - that's why my opinion is positive.

 

I have only a few suggestions for way improving the of manuscript.

 

Materials and Methods:

L 153 was adjuvant used for foliar application? Please write what water was used from the foliar application - tap water, well water, or maybe distilled? If possible, write down the iron content in the water.

L 155 and the following lines. Also use BBCH or other numbering for the plant growth phases.

160. 159-160. I suggest writing a few sentences on the details of the analysis methods.

 

Table 2. What was the difference between Wet digestion with Aqua Regia "A" and wet digestion with Aqua Regia and Murphy Riley "C". What methods were used to analyze K, Ca, Mg and Na? Complete and explain in the publication.

 

In the discussion, please mention the possible impact of nitrogen fertilization on the nodulation process on the roots of cowpea Vigna unguiculata? In addition, how can nodulation or the lack of it affect iron uptake by this plant from the soil. I think this is an important point. I think this is an important point.

 

In addition, I am curious whether bacterial vaccines are used in Zimbabwe to stimulate the development of nodulation - I just need information in the response to the review.

Author Response

Comments and Suggestions for Authors

Review of the publication: „ Can nitrogen fertilizer management improve grain iron concentration of agro-biofortified crops in  Zimbabwe?”   

1. The iron deficiency problem is very big all over the world. The research concerns an important social and economic problem of plant biofortification into iron in Zimbabwe
2. The quality of the research and the organization of the manuscript are excellent.
3. It is a pleasure to review such articles.
4. I haven't found any serious shortcomings - that's why my opinion is positive.
5. I have only a few suggestions for way improving the of manuscript.

 

Materials and Methods:

 

L 153 was adjuvant used for foliar application? Please write what water was used from the foliar application - tap water, well water, or maybe distilled? If possible, write down the iron content in the water.

We did not apply any adjuvant to assist/ enhance absorption of the foliar Fe fertilizer. However, we ensured absorption of the Fe fertilizer by applying it early in the morning, later in the day and/ or on a cloudy day when temperatures were not too high and leaves not too dry (i.e. when the stomata pores will be open). High temperatures would potentially dry out the foliar fertilizer before it gets absorbed into the leaves/ plant. Well water was sourced from the host farmers and would be water the farmer uses for house chores and drinking. We value the comment from the reviewer that an adjuvant could have had added benefits on Fe absorption and a water sample should have been collected for determination of its Fe concentration. Authors acknowledge the oversight of not collecting a water sample for analysis and will consider the Reviewer’s comment in future related research. The source of water used for foliar application and the time of day when foliar fertilizer was applied has now been included in the text (Lines 160-164).

L 155 and the following lines. Also use BBCH or other numbering for the plant growth phases.

Authors have now included the BBCH stages for maize, cowpea and finger millet. A reference and weblinks have been included to support the text (Lines 162-166).

 

 

159-160. I suggest writing a few sentences on the details of the analysis methods.

Additional details on grain Fe analyses has now been presented in the manuscript (Lines 175-180).

 

Table 2. What was the difference between Wet digestion with Aqua Regia "A" and wet digestion with Aqua Regia and Murphy Riley "C". What methods were used to analyze K, Ca, Mg and Na? Complete and explain in the publication.

Method used for total P determination was wet digestion with aqua regia as described by Anderson and Ingram, (1993) and analysed as described by Murphy and Riley (1962). The text in manuscript has been revised for clarity and additional methods for organic carbon, total N and exchangeable bases described as follows:

“Air-dried and ground (0.5 mm) cattle manure samples were analysed for total Zn and Fe using the aqua regia method. Total N was determined using the Kjeldahl procedure described in Okalebo et al. [38]. Total P was determined using wet digestion with aqua regia by Anderson and Ingram [39] and analysed as described by Murphy and Riley [40]. Exchangeable bases (sodium-Na, potassium-K, magnesium-Mg and calcium-Ca) were analysed following extraction with 1 M acidified ammonium acetate (NH₄OAc) as described by Anderson and Ingram [39]. Organic carbon was determined using the Walkley-Black method [38]. The nutrient composition of cattle manure used is shown in Table 2 and has been presented earlier as Manzeke et al. [19]”. Lines 181-188 .

In the discussion, please mention the possible impact of nitrogen fertilization on the nodulation process on the roots of cowpea Vigna unguiculata? In addition, how can nodulation or the lack of it affect iron uptake by this plant from the soil. I think this is an important point.

We did not look at nodulation but is an interesting area for future study. We are not aware of any study which looked at effects of N on nodulation of cowpea grown in smallholder farming systems. Authors value this comment from the Reviewer and have included a discussion point to show this as an area of future investigation (Lines 494-495).

“Effects of soil inoculants on uptake and loading of micronutrients into legume grains could be a potential area for investigation to determine contribution of nitrogen- (N₂) fixing bacteria (i.e. Rhizobium) to improved grain Fe concentration. While this study did not focus on cowpea nodulation, possible impacts of N fertilization on the nodulation process on the roots of cowpea could also be investigated.”

 

In addition, I am curious whether bacterial vaccines are used in Zimbabwe to stimulate the development of nodulation - I just need information in the response to the review.

The Zimbabwean Government’s Marondera Soil Productivity Research Laboratory (SPRL) has the mandate to produce rhizobia for use in legume production. Rhizobia is mostly used in soyabean (Glycine max) and pigeon pea (Cajanus cajan) production often grown for commercial use and has been widely promoted (see research work by Mpepereki and Giller, Mapfumo and Mafongoya). However, while farmers have a fair understanding on importance of rhizobia in legume production, it is not often used in cowpea production. This would be a potential area for future study to determine effects of rhizobia on inoculation and uptake of micronutrients (i.e. zinc and iron). We have added the following sentence to the discussion (Lines 491-493) to reflect this:

“Effects of soil inoculants on uptake and loading of micronutrients into legume grains could be a potential area for investigation to determine contribution of nitrogen- (N₂) fixing bacteria (i.e. Rhizobium) to improved grain Fe concentration. While this study did not focus on cowpea nodulation, possible impacts of N fertilization on the nodulation process on the roots of cowpea could also be investigated.”

 

Author Response File: Author Response.docx