Stress alters the role of silicon in controlling plant water movement

1. One function of plant Si is ameliorating stress, including drought and salinity stress, which can induce active Si uptake in addition to passive uptake via transpiration. However, the interactions and feedbacks between stress, water movement and Si uptake remain unknown. 2. To examine this gap, we compiled papers reporting transpiration and/or stomatal conductance of plants exposed to stresses while varying Si availability. 3. Our meta-analysis (34 studies, excluding rice) showed that stress alters the role of Si in controlling water movement across diverse plant groups. Increased Si availability significantly increased water movement in

to help plants withstand environmental stress is increasingly important to understand, given current rates of climate change are rapidly altering environmental conditions (Bokor et al., 2021;Johnson et al., 2018;Song et al., 2018;Thorne et al., 2020) and Si fertiliser is predicted to be an important tool for increasing food security (Cooke et al., 2016;de Tombeur et al., 2021;Wang et al., 2021).
All plants accumulate some Si in their tissue, but concentrationswithin and across species-are more variable than any other element found in plants (Epstein, 1994).Plants take up Si from the soil solution through two main pathways: (1) passive uptake, as a nonselective flow of silicic acid up through transpiration stream, and (2) active uptake, which involves plant energy use and relies on a metabolic process that the plant controls (Piperno, 2006).Interestingly in some species, a thin layer of fatty substances on roots excludes most Si (Parry & Winslow, 1977;Piperno, 2006).However, species have an innate maximum Si accumulation capacity, such that some taxa (e.g.legumes) have consistently low accumulation and others (e.g.Cyperaceae) have much higher Si accumulation (Hodson et al., 2005).Active uptake occurs through Si transporters produced in root epidermal cells, identified in numerous plant taxa, including Poaceae, Cucurbitaceae and Fabaceae (Chiba et al., 2009;Deshmukh et al., 2013;Ma et al., 2006;Ma & Yamaji, 2008;Mitani et al., 2009;Mitani-Ueno et al., 2011;Montpetit et al., 2012;Piperno, 2006;Wang, Yu, et al., 2015;Yamaji & Ma, 2009).Hence, total plant Si accumulation depends on local environmental conditions (Si availability in the soil solution, and water availability in the soil and air) and plant physiology (the production of transporters and water movement in the plant), both associated with water availability.
Not surprisingly, the relationship between Si accumulation and water movement has received considerable research attention (reviewed in Wang et al., 2021) with Si flux from soil solution to xylem occurring via symplastic and apoplastic pathways (Raven, 1983).
Transpiration is considered more important for Si movement compared to other elements in plants (Epstein, 1999).When Si is passively taken up, transpiration (in addition to silicic acid concentrations in soil solutions) controls plant Si content (Ma & Yamaji, 2006;Piperno, 2006;Sangster & Parry, 1971).In contrast, in situations when Si accumulation involves active uptake, in addition to passive (Liang et al., 2006), transpiration is not the main driver of Si accumulation (Piperno, 2006), resulting in differential modes of Si accumulation uncoupling the relationship between water loss and Si accumulation (Figure 1, row A).However, the relationship between Si accumulation and water movement is more complicated because of how Si moves around the plant, where it can be deposited and the implications of deposition for subsequent water movement.
Irrespective of uptake method, high rates of evapotranspiration correlate with high Si accumulation in the aerial structures of some plants, especially in grasses (Fernández Honaine & Osterrieth, 2012;Jones & Handreck, 1965).However, in some plants, there is more Si in roots than shoots (e.g.Frew et al., 2017;Maguire et al., 2017), and large concentrations can be found in  Once deposited in plants, Si does not translocate (Raven, 1983), which may be due to the uncharged nature of the amorphous silica molecule (Epstein, 1999).Deposition as silica occurs when cellular silicic acid concentration exceeds 1.67 mM (Currie & Perry, 2007).
Precipitation of silica is also initiated by proteins and facilitated by cell wall polymers (Zexer et al., 2023).There are two groups of cells in many plants that become fully silicified with age: (1) short cells, which silicify early in life; and (2) other cells, including bulliforms, long cells and stomatal complexes and trichomes, which increasingly silicify with age (Fernández Honaine et al., 2013, 2017).
Often Si deposition occurs at locations of transpiration termini or intense water loss (Raven, 1983), with higher concentrations of Si in leaf edges, outer epidermal cell walls, trichomes, bulliform cells and hairs of leaves and stems compared to other plant tissues (Epstein, 1999;Fernández Honaine et al., 2017;Gao et al., 2018).
Moreover, elevated Si deposition has been found within a plant species in tissues more exposed to solar radiation and higher temperatures, likely driven by high water losses at these sites (Fernández Honaine et al., 2017;Ma & Takahashi, 2002;Sangster et al., 2001).However, some cells not associated with the transpiration stream must rely on transporters to be silicified (e.g.idioblasts or short cells) (Sangster et al., 2001).Therefore, cell type and the location of transporters in above-ground plant parts also influence the rate of Si accumulation and explain high quantities of plant Si observed in places not associated with water loss (Figure 1, row C).
The guard cells of stomates, the main sites of transpiration, open and close to regulate gas exchange, including the loss of water vapour from tissue.Silicification of these cells has been reported in multiple species to greater or lesser extents (Gao et al., 2018;Gao et al., 2006;Verma, Song, et al., 2020).Partly or completely closed silicified stomate guard cells and complexes (such as those shown in broadleaf leaves by Gao et al., 2018) could limit their function and reduce transpiration, as suggested by Sangster and Hodson (2007).Hence, while transpiration is involved in Si uptake, there may be a balancing feedback loop in which increased silicification reduces transpiration, which in turn limits subsequent Si uptake.In contrast to complete silicification, smaller silica deposits in stomatal guard cells of a C3 grass have been shown to increase the sensitivity of stomates, enabling them to close more quickly (Vandegeer et al., 2021) or keep guard cells open (Camargo et al., 2021).Therefore, in addition to transpiration only being partly responsible for moving Si through a plant, there is conflicting evidence about how Si deposition may impact further water movement-and in turn, Si accumulation (Figure 1, row D).
Another likely factor driving differential relationships between transpiration and plant Si content is plant stress level.With the bestknown functions of Si for plants associated with the alleviation of abiotic (e.g.drought, salinity, heavy metal, wind) and biotic stress (e.g.herbivory), it follows that the impacts of Si availability are clearest-and at times only visible-in stressed plants.The impact of experimental Si fertilisation on stress alleviation correlates with the severity of stress (Cooke & Leishman, 2016).Increased Si uptake is an inducible response to stress; plant species capable of active Si uptake are plastic in response to stress (Brightly et al., 2020;Johnson & Hartley, 2018), and can alter their relative reliance on passive and active uptake to accumulation plant Si in order to mitigate plant stress (Kumar et al., 2017).Increased Si uptake has been shown in response to herbivore attack, with Si deposition reducing palatability through physical abrasion and digestibility by reducing cell crushing (Hartley & DeGabriel, 2016 as examples;McNaughton & Tarrants, 1983;Soininen et al., 2013).However, abiotically stressed plants do not routinely increase Si accumulation (Cooke & Leishman, 2016), despite demonstrated functions of Si in the stress alleviation associated with diverse physical and physiological mechanisms (Liang et al., 2007).
Experimental Si additions allow for increased hydraulic conductance of roots (Hattori, Sonobe, Araki, et al., 2008) and maintenance of water use efficiency (Rios et al., 2017).Also, Si addition to drought stressed plants have been recorded to increase stomatal sensitivity (Vandegeer et al., 2021).Nevertheless, while Si additions are routinely shown to increase water movement in plants exposed to drought and salinity stress (Abbas et al., 2015;Hattori, Sonobe, Araki, et al., 2008;Rios et al., 2017;Sonobe et al., 2009), evidence of the opposite effect in lowered transpiration and/or conductance exists (Gao et al., 2006;Street, 1974;Yeo et al., 1999).As such, the mechanisms for water stress alleviation remain incompletely defined (but see Coskun et al., 2016;Malik et al., 2021;Rios et al., 2017;Wang et al., 2021).To date, studies on the amelioration of drought and salinity stress on plant water movement with Si addition are either qualitative examinations across species (e.g.Rios et al., 2017;Wang et al., 2021) or quantitative studies on one or few species (e.g.Romero-Aranda et al., 2006;Vandegeer et al., 2021).These studies suggest that stress exposure can alter the relationship between Si and water movement in plants; hence, stress should be considered an important mechanism contributing to the diversity of Si accumulation within and across plant taxa (Figure 1, row E).
Given the contradictory information in the literature about the relationship between water movement in plants and Si uptake and accumulation, the objective of our research was to quantify the role of Si additions in influencing water movement a variety of plant taxa.Specifically, we address the following questions: (1) How does stress affect any relationship between Si and water uptake?and (2) How is the above relationship influenced by stress type, taxonomy and photosynthetic pathway (i.e.C3 vs. C4 plants)?A consistent approach was taken to identify relevant papers, data and analysis (following Koricheva & Gurevitch, 2014, see Table S1).
To systematically identify papers testing or reporting the impact of Si availability on plant conductance and transpiration (Figure 2), we searched the Web of Science database, for TOPIC (silicon OR silica) AND TOPIC (transpiration OR stomatal conductance) AND TOPIC (plant OR leaf OR leaves OR foliar) and NOT TOPIC (rice).We excluded rice from our analyses, as studies on Oryza sativa heavily dominate published experiments.The literature search (see Figure 2) was conducted in March 2020 and repeated in November 2021 and the new papers identified added to the existing database.We then excluded papers that did not measure conductance or transpiration, did not compare different levels of Si addition or did not provide measures of variance in results.We extracted values for stomatal conductance, transpiration and tissue Si concentration for treatments with and without Si addition (+/− Si) and with or without stress (+/− stress), where available, from the remaining 34 papers (but see note below) that we used in our analyses (Ahmed et al., 2014;Al-Huqail et al., 2019;Altuntas et al., 2018;Amin et al., 2018;Bianchini & Marques, 2019;Chen et al., 2016;Chung et al., 2020;Dorneles et al., 2019;Farooq et al., 2013;Gong et al., 2005;Gunes et al., 2007;Hajiboland et al., 2017;Hattori et al., 2007;Liu & Guo, 2013;Liu et al., 2015;Maghsoudi et al., 2020;Mateos-Naranjo et al., 2013;Muneer & Jeong, 2015;Murillo-Amador et al., 2007;Resende et al., 2012;Romero-Aranda et al., 2006;Saja-Garbarz et al., 2021;Sattar et al., 2017;Savvas et al., 2007;Siddiqui et al., 2014;Silva et al., 2012;Verma, Anas, et al., 2020;Verma, Song, et al., 2020;Viciedo et al., 2019;Wang, Liu, et al., 2015;Yin et al., 2014;Zhang et al., 2019;Zhang et al., 2020;Zhu et al., 2015).In summary, our study only includes papers reporting water movement (transpiration or conductance) in plants exposed to varying levels of Si availability, with and without stress.The majority of the stress treatments imposed were either drought or salinity (over 75% of observations), but also included other stresses (Al stress, ammonium toxicity, Fe deficiency, heavy metal stress, anthracnose infection, K deficiency and B toxicity).The term 'Si addition' is used in this manuscript to describe artificially added Si, contrasted with either ambient or, more often, reduced Si in control treatments.
Data for both stomatal conductance (mmol H 2 O m 2 s −1 ) and transpiration rate (mol H 2 O m 2 s −1 ) were collected, often from the same study, as they provide complementary information: A plant can have high stomatal conductance enabling large amounts of water loss, but humid conditions could mean the realised transpiration rate is low.We converted all reported values to the units listed above, and all measures of variance to standard deviations.We also collated plant Si concentration (mg g −1 DW) measured most frequently on a leaf tissue, although sometimes measured at the whole plant (above-ground) level from these papers.However, we saw no significant relationship between plant Si concentration and water movement across or within diverse taxa (see Figure S1).
We analysed data in R (R Core Team, 2022) using the 'metafor' package (Viechtbauer, 2010) and prepared figures using the 'orchard' package (Nakagawa et al., 2021).We used pairs of treatments (e.g +Stress-Si and +Stress+Si) and one response (e.g.conductance or transpiration) at a time to calculate the effect size, Hedges' d, for individual studies.This effect size measure compares two means using a pooled standard deviation and bias correction and is a measure of the number of standard deviations by which the means differ (Hedges & Olkin, 1985).For example, a resulting positive value indicates that the response measure was higher in the experimental compared to control treatment.We checked the data for outliers and normality of residuals, which led to the exclusion of one further study and one cultivar from another, as the measures of variance were extremely low and effect sizes >6 (Verma, Song, et al., 2020 and 'Super Marmate' excluded from Silva et al., 2012, respectively, Figure 2).We analysed models with the rma.mv function in the metaphor package.Overall effect sizes across studies were determined using a (null) model including publication (reference) as a random factor, which accounted for non-independence of results from the same study.Including species as a random factor did not have additional explanatory power and hence, was not included in models.
To further explore the sources of heterogeneity among responses, a combination of fixed and random effects, in mixed models, was used.In these models, manuscript ID as a random factor was re- not overlap zero, then there is a significant overall positive or negative effect) or each other.
We could not test for interactions between order, stress type, substrate and photosynthetic pathways within the same model, as the data compiled did not sufficient replication of different combinations.Instead, we tested the importance of stress type and photosynthetic pathway as a moderator in a single order, the Poales, for which the most data had been collected.When these moderators explained significant heterogeneity in analysis of both the full data set and the Poales subset, we could be more confident of a true effect of the moderator.

| RE SULTS
Across the studies combined, when Si was added to unstressed (control) plants, we found no significant impact or consistent pattern on transpiration or stomatal conductance (Figure 3a,b).However, when Si was added to stressed plants, we observed significant increases in water movement, including both transpiration (p < 0.01, k = 88) and conductance (p < 0.001, k = 84) (Table 1 null models, Figure 3c,d).
Different stress types resulted in some consistent and some dif- Similarly, substrate also explained variation in conductance versus transpiration response to treatments, with Si addition significantly increasing transpiration in hydroponic plants (p < 0.05, k = 22) but not soil-grown plants (k = 66), while conductance was significantly increased with Si addition in both hydroponic (p < 0.05, k = 20) and soil-grown experiments (p < 0.01, k = 64; Table 1; Figure 4c,d).
Photosynthetic pathway was also an impacting factor (Table 1).
Si addition significantly increased both transpiration (p < 0.01, k = 44) and conductance (p < 0.001, k = 41) in stressed C4 plants to a much more pronounced degree than C3 plants.Although conductance was still significantly increased in stressed C3 plants exposed to Si addition (p < 0.05, k = 43), the signal was weaker, and we found no significant impact on transpiration in stressed C3 plants when Si was supplied compared to when it was not (Figure 4e,f).
Finally, taxonomic order also explained some of the response to Si addition in conductance but not transpiration in stressed plants (Table 1), as the Poales showed significant increased conductance with Si addition (p < 0.001, k = 47).This signal was not observed in any other taxonomic order (Figure 5  size of non-Poales orders, as the Poales order alone accounted for 55%-60% of the data and 40% of publications (Table 1).When models with stress type, substrate type and photosynthetic pathway as fixed factors were repeated using only Poales, they showed the same significant explanatory power of these factors on conductance and transpiration as for the whole data set (Table 1, last four panels).
The maintenance of within-order patterns provides confidence of a true effect of the moderators (i.e.photosynthetic pathway, stress type and substrate), rather than being patterns driven by plant order.
While there was no significant impact of Si addition to the transpiration and conductance of unstressed plants overall (Figure 3a,b), we found that in roughly half of the studies collated, Si addition reduced water movement in the unstressed plants.Although not significant, this trend was most pronounced in unstressed plants within the non-Poales orders, which showed a general reduction in water movement with Si addition.In contrast, unstressed plants within the Poales order showed a trend for increased transpiration and conductance when Si was added, although again this trend was not significant (Figure 6, Table 1).
Our results presented so far highlight that Si additions increase water movement in stressed plants.Next, we sought to determine whether this increase in water movement allows plants to return to baseline levels of unstressed water movement.To answer this question, we first examined how stress alone (without Si addition) impacted water movement (Figure 7).Unsurprisingly, stress (salinity, drought and other types together) notably reduced both Plants also resist stress by adjusting ion transport to maintain nutrient balance with both micro-and macronutrients (Rios et al., 2017;Verslues et al., 2006).

Several physical mechanisms have been identified by which
Si can improve water relations in plants.Si can increase cuticle thickness (Hattori et al., 2005) and membrane stability (Agarie  et al., 1998) or deposits in the cuticular layer, albeit predominantly recorded in rice (Yoshida et al., 1962), which can reduce transpiration as a way to conserve water use.Together, these physical changes often increase the water potential in stressed plants supplied with Si (e.g.Gong & Chen, 2012).However, Si deposition in bulliform cells can also reduce leaf wilting through the deposition (Amin et al., 2016;Fernández Honaine & Osterrieth, 2012), potentially increasing transpiration.Salinity and drought stress alleviation in plants by Si also includes physiological and biochemical responses that lead to an increase of water movement, though the pathways which Si is involved remains an area of active research (see Thorne et al., 2020 for discussion).For droughtstressed plants, these include the modification of osmolytes and hormones (including proline, abscisic acid, jasmonic acid, salicylic acid and gibberellic acid), increase in the antioxidants which decreases oxidative stress and modification of gas exchange attributes (Liang et al., 2007;Rizwan et al., 2015).In addition to these, for salt-stressed plants, Si causes decreases in Na uptake and increases in K uptake, which reduces the Na ion toxicity (Rizwan et al., 2015).Together, these responses enabled by Si availability mitigate many of plant responses to salinity and drought (for a thorough description, see Thorne et al., 2020), and consequently enable more water movement.The stronger response to Si addition in increasing water movement in plants exposed drought and salinity, versus other stresses (Figure 4a,b), is likely a consequence of these stresses particularly reducing water flow, either because of availability or to maintain cellular homeostasis under stress conditions.Si addition may also increase water availability in some soils (Kuhla et al., 2021).However, we found more water  The effect of Si addition to stressed plants stronger on conductance than transpiration, illustrated by the lower significance values for the latter in Table 1, when all stressed plants were considered (Figure 3c,d), but also for drought-stressed plants (Figure 4a,b) and in both hydroponic and soil-grown plants (Figure 4e,f).The differences between conductance and transpiration response may be an artefact of experiments, rather than a differential impact of Si, as most studies used glasshouses or growth cabinets where humidity can be high, limiting transpiration to a greater degree than conductance.There appeared to be a stronger response in Poales, especially for conductance (Figure 5), to Si addition to stressed plants in terms of water movement.This may be function of statistical power, as Poales was the largest order represented in the meta-analysis, or because the Poales are a high Si-accumulating order of plants (Hodson et al., 2005) Etesami et al., 2022).
While the benefits of Si for plants are increasingly well understood, understanding of the costs of plant Si accumulation has not kept pace, hampering understanding of plant-Si use (de Tombeur et al., 2023).In this meta-analysis, both conductance and transpiration were reduced wood and bark (e.g.Clymans et al., 2016).Silicon transporters are produced throughout some plant groups, enabling control of Si F I G U R E 1 Schematic diagram of the accumulation, allocation and stress alleviation mechanisms associated with and independent from the transpiration stream.

13652435, 0 ,
Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14447by The Open University, Wiley Online Library on [19/10/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License movement and accumulation location.Taken together, silicification can be controlled by the plant, irrespective of the transpiration stream, again decoupling the relationship from water for some species (Figure 1, row B).

13652435, 0 ,
Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14447by The Open University, Wiley Online Library on [19/10/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License tained and fixed effects were plant order, photosynthetic pathway or stress type.Null model results indicated if there were significant overall effects of Si addition on water movement.Mixed-effect model results indicated if another factor type explained a significant proportion of variation (heterogeneity) among studies.If the heterogeneity explained by the model including a moderator (QM) was significant, we considered the moderator likely to be an important factor.The 5%-95% confidence intervals in figures were used to determine whether overall effect sizes for each factor (plant family or stress type) were significantly different from zero (i.e. if 95% confidence intervals do F I G U R E 2 A schematic diagram of the systematic approach to identifying and collating data for analyses. fering responses of transpiration and conductance exposed to Si additions.While Si addition significantly increased both transpiration (p < 0.01, k = 19) and conductance (p < 0.001, k = 20) in plants exposed to salinity stress, we found that only conductance (not transpiration) was significantly increased with Si addition in plants exposed to drought stress (p < 0.01, k = 45).Moreover, we found that only transpiration (not conductance) was significantly increased in plants exposed to other stressors beyond drought and salinity (p < 0.05, k = 21; Figure 4a,b).
), potentially due low sample F I G U R E 3 Orchard plots showing effects of Si addition on transpiration and stomatal conductance for both unstressed (a and b) and stressed (c and d) plants.Each dot represents the effect size of a comparison from a study.Black-ringed circles with whiskers indicate the overall effect size with 5%-95% (thick line) and 1%-99% confidence intervals (thin line).p-values indicate where an overall effect size is significant in a null model with publication as a random factor.TA B L E 1 Meta-analysis results for plant responses to Si supply in unstressed and stressed plants.In each row, the null models (no fixed factor) are reported with manuscript ID is included as a random factor, and the number of papers given.A z-score with a significant result (p-value, in bold) indicates an overall effect on the response to Si addition.The number of data points in each analysis is given (k) and Q E is the amount of unexplained heterogeneity.Subsequent models with Poales/Non-Poales, plant order, stress type, substrate or C3/C4 as fixed factors are reported, and Q M is indicative of the heterogeneity explained by the structured model, with a significant p-value (in bold) showing the amount of variation explained that is more than the null model.

F
Orchard plots showing effects of Si availability on transpiration and stomatal conductance for stressed plants with and without Si addition by stress type (a and b), substrate (c and d) and photosynthetic pathways (e and f).Key as in Figure 3. p-values indicate where an overall effect size is significant in a model with the categories as fixed factors and publication as a random factor.transpiration and stomatal conductance (indicated by nearly all values below the 1:1 dashed line on Figure 7a,b).Next, to determine the degree of recovery from stress due to Si addition, we compared water movement in Si-supplied stressed plants with unstressed plants without Si additions (Figure 7c,d).While we found evidence of recovery from stress with Si addition, this recovery was not complete (as the lines of best fit for Si-treated stressed plants on bottom row are closer to the 1:1 dashed lines compared to the figures on the top row; Figure 7), Thus, although Si supply correlates with an increase in transpiration and conductance in stressed plants, full recovery of water movement with Si addition was not obtained overall.4| DISCUSS IONStress exposure, including drought, salinity and other stressors included in the studies examined here, resulted in a reduction in water movement in plants, and this meta-analysis demonstrated that Si addition increased water movement (stomatal conductance and transpiration) in stressed plants, including but not limited to high Si-accumulating Poales.Reduction in water movement with plant stress is both a function of simply less water available for the plant to transpire in drought conditions, but also combinations of highly complex physiological responses of plants to both avoid and tolerate water stress(Chaves et al., 2009;Verslues et al., 2006).Plants avoid water stress by restricting water loss through stomata closures and altering cell wall properties to resist dehydration, such as solute accumulation and the stiffening of cell walls(Verslues et al., 2006).

F
Orchard plots showing effects of Si availability on transpiration and stomatal conductance for stressed plants with and without Si addition by plant order.Key as in Figure 3. p-values indicate where an overall effect size is significant in a model with the categories as fixed factors and publication as a random factor.F I G U R E 6 Orchard plots showing effects of Si availability on (a) transpiration and (b) stomatal conductance for unstressed plants with and without Si addition in Poales and non-Poales.Key as in Figure 3. p-values indicate where an overall effect size is significant in a model with the categories as fixed factors and publication as a random factor.
movement (transpiration) in responses to Si addition in hydroponic systems compared to soil substrates (Figure4c,d), although the total amount of water available to plants, and the experimental conditions (lighting and humidity) may differ with experimental set-ups associated with hydroponic, pot and field experiments.

F
Scatter plots comparing different treatments within studies.The first row (a and b) shows the impact of the stresses on transpiration and stomatal conductance by comparing unstressed versus stressed plants with no Si addition.The second row (c and d) depicts the recovery in water movement with Si addition, comparing Si-supplied stressed plants with unstressed and untreated plants.Lines of best fit (solid) have been fitted through each cloud of data and can be compared with the 1:1 dashed line.red = non-Poales, turquoise = Poales, ○ = C3, ∆ = C4.

13652435, 0 ,
Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14447by The Open University, Wiley Online Library on [19/10/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License in 50% of unstressed plants when Si was added compared to no Si addition.A reduction of water movement caused by Si accumulation when plants are not stressed may indicate an important, previously unrecognised, cost of Si accumulation.These data are from control treatments of many experiments and, thus, represent little publication bias (see Supporting Information), providing confidence that the reduction of water movement it is likely a genuine pattern.Si accumulation has been found to be highest in plants with shorter lived leaves compared to long-lived leaves (Cooke & Leishman, 2011); this suggests that the cost of Si accumulation is high relative to transpiration and conductance reductions for long-lived leaves, compared to annual species that are shorter lived where reductions in water movement associated with Si accumulation in unstressed plants are of less consequence.de Tombeur et al. (2023) noted that although the energy costs of Si accumulation are well below those for C, acquisition (direct) costs may include root exudates, mycorrhizal and bacterial associations and Sitransporter production and use costs.In addition, ecological costs of Si accumulation include the high density of Si, poorer biomechanical properties and reduced defence against some types of herbivores compared with equivalent functions provided through C-based compounds (summarised in de Tombeur et al., 2023).Multiple authors have shown negative correlations betweengrowth rate and leaf Si among Poaceae(Massey et al., 2007;Simpson et al., 2017;Thorne et al., 2022), which suggests trade-offs associated with Si accumulation.In addition, the inducibility of Si accumulation in Poales following herbivore attack is also suggestive of costs, given induced responses are considered a mechanism for saving on costly responses unless needed, as shown for other responses to herbivory(Karban et al., 1999).However, the Poales are a high Si-accumulating group(Hodson et al., 2005), and although not significant, unstressed non-Poales to show a bigger reduction in water flow with Si addition than unstressed Poales (Figure 6a,b), indicating that Poales have some adaptations to overcome costs of high Si accumulation.Building on this meta-analysis, experimental studies are needed comparing uptake and allocation of Si in Poales and non-Poales to better understand the costs of Si accumulation on water movement in plants, how Poales may mitigate these costs more than other plant orders, and how these costs may have driven the evolution of plant Si use.AUTH O R CO NTR I B UTI O N S Julia Cooke and Joanna C. Carey jointly conceived the idea for this work, identified and extracted data (with input from Jason Shackleton), conducted the analyses and wrote the manuscript.
Tully et al., 2019)ic adaptations.However, it has bStreet-Perrott & Barker, 2008)at the benefits of Si to stressed plants are not limited to those with high Si accumulation capacity(Katz, 2014), and indeed, recovery from multiple stresses has been shown in lowSi-accumulators (Cooke & Leishman, 2016).Tully et al., 2019).The availability of Si for plants varies with soil type and age, and the rate of Si recycling in the vegetation(Cornelis & Delvaux, 2016; de Tombeur et al., 2020;Street-Perrott & Barker, 2008).Soils with higher Si availability for plants may en- (Bhattachan et al., 2018;ress is increasing in many soils, partially due to a build-up of salts that accompany increasing evapotranspiration rates and decreased freshwater inputs(Minhas et al., 2020)and saltwater intrusion, where increasing sea levels combined with freshwater groundwater withdrawals and land subsidence result in salt contamination of groundwater and soil solutions(Bhattachan et al., 2018; Fulweiler, 2016).Given the high Si requirements for many important crops, and increasing water stress, our results provide strong support for the use of Si fertilisers in improving agricultural yields, especially when crops are exposed to salinity and drought stress.