Review Reports

Reviewer 1 comment

Authors’ response

Introduction

1. Silicon is classified as an essential micronutrient under Brazilian agricultural regulation, which justifies the use of "Si fertilization" in the Brazilian context. However, the international scientific consensus treats Si as a beneficial rather than essential element, and under EU Regulation 2019/1009 it is not recognized as a plant nutrient. The Introduction should briefly acknowledge this distinction so that the terminology is transparent to international readers.

The Introduction has been revised to clarify that Si is generally considered beneficial rather than essential and to distinguish its regulatory classification in Brazil and the European Union, with references to the relevant legislation. We also clarified that Brazilian legislation includes Si among micronutrients without specifically establishing its essentiality. “Si fertilization” has been replaced with “Si application” throughout the manuscript.

2. Literature scope in Introduction relies primarily on studies from the same research group. Independent studies on silicon effects in passion fruit or comparable perennial fruit crops from other agroecological contexts (e.g., Australia, sub-Saharan Africa) would strengthen the rationale. The mulch materials used - Cenchrus ciliaris and Megathyrsus maximus - are not contextualized in the Introduction; no references discuss their properties, decomposition rates, or previous use as mulch materials in comparable systems.

The Introduction now includes independent studies on Si in passion fruit and avocado from different geographical contexts. We also contextualized the selected mulch species using studies on residue composition, decomposition, and persistence, acknowledging differences between those systems and our residue mixture.

Methods

3. Silicon source and dose specification The commercial product Sifol® contains 92% SiO₂, but the Keywords list "silicic acid" and doses are reported in grams of product per plant without specifying elemental Si content. Total elemental Si supplied per plant per treatment should be reported.

Section 2.2 now clarifies the Si source, distinguishes elemental Si doses from commercial product quantities, and specifies that the doses used in the regression analyses represent cumulative elemental Si supplied over six equal applications. The original manuscript incorrectly stated that two applications were performed; six applications were actually made. This reporting error has been corrected, and we apologize for the confusion.

4. The mulch is described only by layer thickness (8 cm). Species proportions, fresh or dried state, applied dry mass per unit area, and frequency and total amount of replenishment are not stated. Plant residues may supply nutrients upon decomposition; treatment effects cannot be attributed exclusively to moisture conservation without acknowledging this contribution. No comparison with other mulch types is provided, making it difficult to assess the generalizability of the results.

The Methods now specify the species proportions, dried state of the residues, and thickness-based replenishment criterion. The initial dry mass, number of replenishments, and cumulative mass added were not quantified. Despite these limitations, the findings remain relevant to agricultural management under the evaluated conditions. The Discussion acknowledges that nutrient release during decomposition may have contributed to treatment responses and that our experiment does not establish comparisons with other mulch materials. Such comparisons were beyond the study’s objective, which was to evaluate a low-cost mulch made from locally available plant residues.

5. Irrigation management and Kc source Crop coefficients are cited from a study on salinity and organomineral fertilization (Ref. 16), which is not a standard agrometeorological source for Kc values. The appropriateness of these coefficients for the local conditions and cultivar should be justified. Figure 1 caption does not identify irrigation depth, though the Methods state it is presented there.

The Kc reference has been replaced. The Methods now clarify that these coefficients were adopted as the available reference for local growing conditions, given the absence of coefficients specifically validated for the accession and experimental site.

6. Physiological measurements - sampling details are missing, Assessment date (DAT), number of plants and leaves measured per plot, and how subsamples were combined are not stated. Chamber conditions (CO₂ concentration, vapor pressure deficit, stabilization criteria) are not reported for gas exchange. The timing of measurements relative to the second Si application — both occurring at flowering - should be clarified.

The suggestions have been addressed, and the requested information has been incorporated into the Methods section.

7. Two central plants were used for physiological assessments and 10 fruits per plot for fruit characterization. Please confirm that the plot was the experimental unit, explain how subsamples were averaged before analysis, and clarify whether pulp chemical attributes were determined on individual or pooled fruits.

Sections 2.5.1 and 2.5.3 now clarify that plants and fruits were treated as subsamples, with measurements averaged within each plot before statistical analysis. The plot was the experimental unit.

8. Results state only that significant effects were detected for "most variables" with "data not shown." An ANOVA table with F-values, degrees of freedom, and p-values for all main effects and interactions must be provided for each variable, even as supp file.

Table S1 has been added as supplementary material, reporting F-values, degrees of freedom, and p-values for all main effects and interactions for each variable.

9. Regression analysis - models were fitted to individual plot observations or treatment means? or how blocks were accounted for? Point estimates such as 58.06 g Si plant⁻¹ imply precision that five dose levels cannot support. Confidence intervals for predicted optima should be provided, or estimates presented as model-based approximations.

Section 2.6 now clarifies that regression models were fitted to treatment means across the four blocks, whereas block effects were considered in the factorial ANOVA. Doses associated with maximum or minimum responses are now explicitly described as model-based approximations rather than experimentally validated optima, acknowledging the five dose levels evaluated.

Results

10. Figure 2a - unit error and formatting inconsistency 

The unit in Figure 2a has been corrected to mmol H₂O m⁻² s⁻¹, and the formatting inconsistency has been resolved.

Discussion

11. Literature scope in Discussion The Discussion contextualizes results primarily through studies by the same group and Brazilian journals. Independent studies from comparable agroecological systems outside Brazil should be included to assess whether the observed Si and mulch responses are consistent with broader evidence. The absence of a comparison with other mulch materials limits interpretation of the mulch.

The Discussion has been expanded to include independent studies from outside Brazil and literature-based comparisons with wheat straw and plastic mulch. Differences in experimental conditions and the absence of direct comparisons with other mulch materials in our study are now acknowledged.

12. The Conclusions present dose recommendations without acknowledging that the experiment covered only one production cycle.

The Conclusions now acknowledge that the experiment covered a single production cycle at one location. The estimated dose is presented as a model-based approximation, and validation across additional cycles and locations is required before making dose recommendations.

Reviewer 2 comment

Authors’ response

1. Lines 141–145 and 263–267: The authors report significant effects of silicon, mulching, and their interaction without presenting the corresponding ANOVA results. A complete ANOVA table, including F-values and p-values, should be provided.

The ANOVA results have been added as supplementary material (Table S1), reporting degrees of freedom, F-values, and p-values for all main effects and interactions for each variable.

2. Lines 146–149: The silicon application procedure is unclear. Please clarify whether the reported doses represent the amount applied at each application or the cumulative amount across both applications, and specify the actual elemental silicon input.

Section 2.2 now clarifies that the reported doses represent cumulative elemental Si inputs over six applications and provides the corresponding commercial product quantities. We apologize for the error in the original description, which incorrectly stated that two applications were performed.

3. Lines 127–139 and 542–546: Neither initial soil-available silicon nor plant tissue silicon concentrations were determined. This limits the interpretation of silicon uptake and its contribution to the observed responses.

Initial soil-available Si and plant tissue Si concentrations were not determined, as these measurements were outside the original analytical scope of the experiment. We acknowledge that their absence limits conclusions regarding Si uptake and accumulation, and this limitation has been explicitly acknowledged in the revised manuscript. Nevertheless, we believe the field evaluation provides useful evidence of physiological, yield, and fruit quality responses to Si application and mulching. We respectfully ask the reviewer to consider these findings within this limitation.

4. Lines 175–189 and 283–296: Despite daily irrigation based on crop evapotranspiration, leaf relative water content was approximately 64% in the zero-silicon treatments. The authors should clarify the irrigation conditions and explain these relatively low values.

Although irrigation was applied daily based on estimated ETc, this management does not, by itself, ensure high leaf hydration throughout the day. RWC reflects tissue water status at the time of sampling, and its interpretation depends on species characteristics and assessment conditions; there is no universal threshold for adequate hydration. Values near 64% correspond to model estimates at zero Si and indicate lower relative hydration, but they do not independently establish its cause. Evaporative demand and uncertainties associated with crop coefficients that were not locally validated may have contributed to these values; however, these possibilities were not demonstrated in the experiment. Therefore, we interpret the increase in RWC with Si application as a comparative response among treatments, without assuming that daily irrigation eliminated potential water limitations.

As requested, this possible explanation and its limitations have been added to the Discussion (Lines 491–496).

5. Lines 191–198: Gas exchange was measured only once at full flowering. Please justify the representativeness of these measurements and provide the missing measurement conditions, particularly leaf temperature, CO₂ concentration, and vapor pressure deficit.

Gas exchange was assessed at full flowering, a commonly used stage for evaluating leaf physiological responses because of its relevance to reproductive development. Standardized leaf selection and a common measurement window supported comparisons among treatments. These measurements provide a physiological snapshot at this stage and are not intended to represent the entire production cycle. The requested measurement conditions, including leaf temperature, CO₂ concentration, and vapor pressure deficit, have been added to the revised manuscript.

6. Lines 250–256 and 297–304: Optimum silicon doses were estimated from quadratic regressions using only five dose levels, yet values are reported to two decimal places. The uncertainty of these estimates should be evaluated, and excessive numerical precision should be avoided.

The use of five dose levels supports the evaluation of quadratic responses and is consistent with the methodological recommendation of at least four quantitative factor levels discussed by Cavalcante and Costa [1]. Polynomial regression with a limited number of dose levels has also been used in recent MDPI studies, including Barbosa et al. [2] and Lopes Sobrinho et al. [3]. Additionally, Palmeira et al. [4], in Horticulturae, fitted linear and quadratic regressions to responses across four levels of available water in yellow passion fruit seedlings. We therefore hope that these examples adequately support the analytical approach adopted.

The reported doses were calculated from the first derivative of the fitted quadratic equations, using x = −b/(2a). Their decimal places represent numerical outputs of the fitted models rather than experimentally demonstrated precision. Section 2.6 now explicitly identifies these estimates as model-based approximations rather than experimentally validated optima. The Conclusions present the yield-associated dose as approximately 58 g Si plant⁻¹, providing a more practical expression with fewer decimal places and avoiding the implication of excessive precision.

References:

1. Cavalcante, M.; Costa, J.G. Considerações sobre planejamento experimental e adequabilidade do uso de testes estatísticos em Ciências Agrárias. Diversitas Journal 2021, 6, 3706–3723. https://doi.org/10.48017/dj.v6i4.1931.

2. Barbosa, C.F.; Correa, D.A.; Carneiro, J.S.d.S.; Melo, L.C.A. Biochar Phosphate Fertilizer Loaded with Urea Preserves Available Nitrogen Longer than Conventional Urea. Sustainability 2022, 14, 686. https://doi.org/10.3390/su14020686.

3. Lopes Sobrinho, O.P.; dos Santos, L.N.S.; Soares, F.A.L.; Teixeira, M.B.; Reis, M.N.O.; Bessa, L.A.; Vitorino, L.C. Adjusting Irrigation and Phosphate Fertilizer to Optimize Tomato Growth and Production. Agronomy 2024, 14, 1616. https://doi.org/10.3390/agronomy14081616.

4. Palmeira, I.V.S.; Pereira, R.F.; Sousa, C.S.; Alves, S.B.; Izidro, F.F.S.; Pires, J.P.M.M.; Sousa, F.S.; Fernandes, A.R.G.; Silva, A.C.Z.; Melo, A.S.; Brito Neto, J.F.; Gomes, P.M.A.; Mesquita, E.F. Combined Application of Bacillus aryabhattai and Silicon Enhances Membrane Stability, Biochemical Attributes, and Soil Biological Quality in Yellow Passion Fruit Under Water Deficit. Horticulturae 2026, 12, 707. https://doi.org/10.3390/horticulturae12060707.

7. Lines 218–244: Ten fruits per plot were sampled for quality analysis, but the statistical treatment of these samples is unclear. Please clarify the experimental unit and whether fruit measurements were averaged at the plot level to avoid pseudoreplication.

Section 2.5.3 now clarifies that the 10 fruits sampled per plot were subsamples and that their measurements were averaged to obtain one value per plot for statistical analysis. The plot was the experimental unit, thereby avoiding pseudoreplication.

8. Lines 218–226 and 297–332: The relationships among fruit number, average fruit mass, and yield require further verification. Please provide the observed treatment means and clarify the calculations used to estimate yield per hectare.

We checked the plot-level calculations and confirmed that yield was consistent with fruit number per plant and average fruit mass. Yield (t ha⁻¹) was calculated as fruit number per plant × average fruit mass (g fruit⁻¹) × planting density (1666 plants ha⁻¹)/10⁶. The relevant clarifications have been added to the manuscript.

For Si doses of 0, 27, 54, 81, and 108 g plant⁻¹, the observed mean yields were 17.62, 24.93, 33.12, 29.09, and 20.74 t ha⁻¹ with mulch, and 9.305, 20.61, 20.17, 18.52, and 10.79 t ha⁻¹ without mulch, respectively. These observed means are provided here for transparency. Presenting fitted dose-response curves and model-derived maxima is a widely used approach in regression analyses of quantitative treatments. However, the equations provide fitted predictions rather than reproduce the observed means exactly; therefore, predicted maxima are distinguished from observed treatment values in the revised manuscript.

9. Lines 196–198 and 363–391; Figures 2 and 5: Some regression models show relatively low R² values, raising concerns about the reliability of their predicted extrema. Model adequacy should be verified. In addition, the unit of stomatal conductance in Figure 2a is inconsistent with the Materials and Methods section.

Thank you for this observation. All regression models in Figure 2 have R² values above 0.65, whereas only one equation in Figure 5 has an R² of 0.55. There is no universal R² threshold for model acceptance, and this statistic should be interpreted together with other evidence of model suitability, as emphasized in the NIST/SEMATECH e-Handbook of Statistical Methods [5]. Agricultural studies have also reported regressions with lower coefficients of determination; for example, Palmeira et al. [4] presented models with R² values of 0.50 and 0.64 in yellow passion fruit seedlings. These examples provide context rather than independently validate our models.

We reviewed the model-selection criteria again, considering the statistical significance of the regression, the coefficient of determination (R²), and the biological consistency of the responses, and retained the selected models on this basis. Nevertheless, we acknowledge the lower explanatory power of the equation with R² = 0.55 and interpret its predicted extremum as a model-based approximation with limited precision. We hope this clarification adequately addresses the concern. If the reviewer considers this support insufficient, we are willing to remove this fitted equation and its associated extremum estimate while retaining the observed data.

The stomatal conductance unit has also been corrected to mmol H₂O m⁻² s⁻¹ in Figure 2a and standardized with the Materials and Methods section.

Reference:

1. NIST/SEMATECH. e-Handbook of Statistical Methods. Section 4.4.4: How Can I Tell If a Model Fits My Data? Available online: https://itl.nist.gov/div898/handbook/pmd/section4/pmd44.htm?utm (accessed on 21 October 2026).

10. Lines 542–564: The conclusions are based on a single-location, single-season experiment. The authors should avoid broad recommendations regarding optimum silicon doses without further validation across growing conditions.

The Conclusions now acknowledge that the study covered one location and one production cycle. The estimated Si dose is presented as a model-based approximation, with validation across additional cycles and locations required before making recommendations.