No escape: The influence of substrate sodium on plant growth and tissue sodium responses

Abstract As an essential micronutrient for many organisms, sodium plays an important role in ecological and evolutionary dynamics. Although plants mediate trophic fluxes of sodium, from substrates to higher trophic levels, relatively little comparative research has been published about plant growth and sodium accumulation in response to variation in substrate sodium. Accordingly, we carried out a systematic review of plants' responses to variation in substrate sodium concentrations. We compared biomass and tissue‐sodium accumulation among 107 cultivars or populations (67 species in 20 plant families), broadly expanding beyond the agricultural and model taxa for which several generalizations previously had been made. We hypothesized a priori response models for each population's growth and sodium accumulation as a function of increasing substrate NaCl and used Bayesian Information Criterion to choose the best model. Additionally, using a phylogenetic signal analysis, we tested for phylogenetic patterning of responses across taxa. The influence of substrate sodium on growth differed across taxa, with most populations experiencing detrimental effects at high concentrations. Irrespective of growth responses, tissue sodium concentrations for most taxa increased as sodium concentration in the substrate increased. We found no strong associations between the type of growth response and the type of sodium accumulation response across taxa. Although experiments often fail to test plants across a sufficiently broad range of substrate salinities, non‐crop species tended toward higher sodium tolerance than domesticated species. Moreover, some phylogenetic conservatism was apparent, in that evolutionary history helped predict the distribution of total‐plant growth responses across the phylogeny, but not sodium accumulation responses. Our study reveals that saltier plants in saltier soils proves to be a broadly general pattern for sodium across plant taxa. Regardless of growth responses, sodium accumulation mostly followed an increasing trend as substrate sodium levels increased.

1. As an essential micronutrient for many organisms, sodium plays an important role in ecological and evolutionary dynamics. Although plants mediate trophic fluxes of sodium, from substrates to higher trophic levels, relatively little comparative research has been published about plant growth and sodium accumulation in response to variation in substrate sodium. Accordingly, we carried out a systematic review of plants' responses to variation in substrate sodium concentrations.
2. We compared biomass and tissue-sodium accumulation among 107 cultivars or populations (67 species in 20 plant families), broadly expanding beyond the agricultural and model taxa for which several generalizations previously had been made. We hypothesized a priori response models for each population's growth and sodium accumulation as a function of increasing substrate NaCl and used Bayesian Information Criterion to choose the best model. Additionally, using a phylogenetic signal analysis, we tested for phylogenetic patterning of responses across taxa.
3. The influence of substrate sodium on growth differed across taxa, with most populations experiencing detrimental effects at high concentrations. Irrespective of growth responses, tissue sodium concentrations for most taxa increased as sodium concentration in the substrate increased. We found no strong associations between the type of growth response and the type of sodium accumulation response across taxa. Although experiments often fail to test plants across a sufficiently broad range of substrate salinities, non-crop species tended toward higher sodium tolerance than domesticated species. Moreover, some phylogenetic conservatism was apparent, in that evolutionary history helped predict the distribution of total-plant growth responses across the phylogeny, but not sodium accumulation responses. 4. Our study reveals that saltier plants in saltier soils proves to be a broadly general pattern for sodium across plant taxa. Regardless of growth responses, sodium accumulation mostly followed an increasing trend as substrate sodium levels increased.

| INTRODUC TI ON
Plants are key conduits in many, especially terrestrial, biogeochemical cycles (Elser & Bennett, 2011;Farago, 1995;Neubauer et al., 2005;Yuan & Chen, 2015). As intermediaries between soils and higher trophic levels, they often control, limit, or enhance the availability of elements to consumers. Plant phytochemistry varies substantially in elemental composition, stoichiometry, and concentration of essential micronutrients for animals and decomposers (Farago, 1995;Sterner & Elser, 2002). Hunter (2016) envisioned the geographic patterning of phytochemistry as the phytochemical landscape. The phytochemical landscape of micronutrients has considerable effects on plant-herbivore interactions, as well as community and ecosystem dynamics across landscapes that vary in soils, climate, etc. Kaspari et al., 2008;Moore et al., 2010).
Nonetheless, the composition, formation, and intermediary function of the phytochemical landscape remains poorly characterized and understood (Hunter, 2016), especially for certain elements such as sodium (Kaspari, 2020).
Sodium is the seventh most abundant element in the Earth's crust (Kaspari, 2020). However, its presence in terrestrial ecosystems is highly heterogeneous, but spatially correlated with xeric conditions, certain geological formations and proximity to a marine coast or source of marine aerosols (Kaspari, 2020;Martin et al., 2010;Smith, 2013;Stallard & Edmond, 1981). Sodium is unusual as a nutrient for life because although it is a nonessential element for most plants, it is a key and essential element for animals and decomposers (Kaspari, 2020). Although sodium requirements vary among organisms, the availability and intake of sodium are tightly linked to organismal performance across ecosystems and form fundamental components of ecological and evolutionary dynamics (Baxter & Dilkes, 2012;Kaspari et al., 2009;Sterner & Elser, 2002).
Plant populations and communities are exposed to a wide range of saline substrates across terrestrial landscapes. Many plants actively avoid or limit sodium intake, and most plants tolerate sodium in soils to remarkably high levels (at millimolar levels) before they show signs of growth defects compared to many other nonessential or toxic cations such as lithium or many heavy metals that induce toxicity symptoms at micromolar levels (Nawaz et al., 2017;Pantha & Dassanayake, 2020;Shahzad et al., 2016;Vithanage et al., 2019;van Zelm et al., 2020). Most plants can tolerate or can be acclimated to survive up to 200 mM NaCl in their growth media, but those plants that can complete their life cycles at salinity levels higher than 200 mM NaCl are generally identified as halophytes (Cheeseman, 2015;Flowers et al., 1986Flowers et al., , 2010. Unlike most plants, many halophytes need sodium to thrive and suffer growth defects under limited sodium (Bose et al., 2017;Wang et al., 2012). However, only about 1% of the global flora are considered halophytes; they are distributed in multiple plant clades that reflect their convergent evolution to saline environments (Flowers & Colmer, 2008).
Even though most plants do not need sodium, they cannot necessarily avoid it, nor escape having to cope with it. As sodium concentration increases in the substrate, its concentration in plant tissue also generally increases, and in turn affects plant fitness, especially in plants highly sensitive to salt stress (Greenway & Munns, 1980;Pantha & Dassanayake, 2020;Yang & Guo, 2018;Zhu, 2001). With increasing sodium, plants have been shown to decrease biomass accumulation; increase osmotic, oxidative, and ionic stress responses; and arrest growth due to changes in cellular biochemistry (Maathuis, 2014;Zhao et al., 2020). Furthermore, variation in soil concentration of sodium salts has direct links to variation in foliar sodium, which in turn influences plant-herbivore interactions and higher trophic-level performance (Bravo, Harms, & Emmons, 2010, 2012Cheeseman, 2015;Kaspari, 2020;Kaspari et al., 2014;Snell-Rood et al., 2014).
Decades of physiological, biochemical, and genetic studies have contributed to our current understanding of how plants respond to salt stress. Even so, these studies have primarily targeted salt stresssensitive model plants such as Arabidopsis, salt-sensitive crops, or extremely tolerant halophytes. For example, most crops or Arabidopsis ecotypes will show signs of salt-stress at 100 mM NaCl (0.58%) treatments, whereas some halophytes can survive salinities exceeding seawater strengths (3.5%) (Debez et al., 2010;Flowers, 2004;Kazachkova et al., 2018;Zhu, 2000). However, these two extremes in the plant salt-tolerance spectrum represent less than 2% of all angiosperm diversity. Therefore, it is unclear how plants with varying degrees of salt-stress responses growing in diverse salinity conditions conform to general expectations of how sodium accumulates in plants and how this accumulation affects their growth.
We conducted a systematic review of 49 published studies that included 67 species and 107 cultivars or populations, to identify broad-scale patterns of salt accumulation and growth responses across terrestrial angiosperms. Employing a priori response models that we could test against experimental data, we surveyed the relationships between plant biomass growth and substrate NaCl concentration from controlled experiments across taxa. We also characterized relationships between plant-tissue sodium accumulation and substrate NaCl concentration across taxa and examined how biomass growth responses associate with sodium accumulation.
Finally, we assessed phylogenetic patterning of growth and sodium accumulation responses to reveal the role that evolutionary history has played in the distribution of these traits.

K E Y W O R D S
biomass accumulation, fitness, halophytes, model selection, plant growth, plant salt stress responses, sodium, sodium accumulation 2 | MATERIAL S AND ME THODS

| Article search and selection protocol
To determine the effects of experimentally controlled, laboratory-or greenhouse-based substrate sodium chloride (NaCl) treatments on plant biomass and sodium accumulation in their tissues, we searched for peer-reviewed studies using Web of Science in December 2017 and May 2019 following the PRISMA protocol (Moher et al., 2009).
We performed an initial search in December 2017 using the search criteria: "sodium AND biomass AND plant AND growth;" a timespan of "All years;" and indexes "Sci Expanded." These criteria yielded 6,503 articles. For a second search in May 2019, we used the keywords: "sodium AND biomass AND plant OR sodium AND growth AND plant OR sodium accumulation AND shoot AND root AND plant OR sodium AND plant AND halophytes AND biomass;" a timespan of "All years;" and indexes "Sci Expanded." This search yielded 6,654 articles. Subsequently, 6,387 duplicates were removed from the dataset, which produced a total of 6,770 non-duplicate articles from the two searches.
The articles grouped into five unique categories: effects of sodium on growth, biomass, and tissue sodium accumulation in plants (1,305); salt-related responses involving other taxa (animals, fungi, bacteria, protists, etc.) (906); transcriptomics, genomics, proteomics, or other molecular responses (627); influences of other elements and/or compounds (1,750); and other miscellaneous articles (2,183). We used the 1,305 articles that provided data for growth (biomass accumulation) and sodium accumulation in plant tissues.
Although these controlled experiments were conducted by different groups, in different controlled environments, and at different time scales, each used specific NaCl treatments between control and salt-treated plants for a uniform duration specific to each study, keeping all other macronutrients and micronutrients constant.
The plant material subjected to NaCl treatments was mostly seedlings (80.37%), with the remaining studies conducted on cuttings (13.08%), rootstocks/grafts (3.74%), and bulbs (2.80%). Prior to analysis, we updated nomenclatural changes for all species considered in this study using Tropicos (www.tropi cos.org) and NCBI taxonomical databases (Table S2).

| Data extraction and compilation
Articles differed substantially in their data representation, ranging from tables to graphical illustrations. We directly extracted data from tables, whereas measurements in figures were extracted using WebPlotDigitizer (Rohatgi, 2019). Treatments of NaCl were converted when necessary to mM. We focused on the mean responses of plants across treatments compared to their relevant control group as defined in each published study.  Table S3.
Using the same methods described above, we extracted sodium concentrations per dry mass of aboveground, belowground, or total tissues. It is important to note that some plants may have expelled sodium, by means of salt glands or other adaptations.
Tissue sodium concentration was considered as reported by each study. Acceptable sodium concentration measurements included weight-by-weight basis (i.e. mg/g, mg/kg), molarity (i.e. µM, mM or M (mol/L)), molality (i.e. mol/g), percentage (%), or parts per million (ppm). We converted all measurements, when necessary, to percentage (%) values. Measurements of electrical conductivity (S/m or psu) were excluded because, unless stated, they do not necessarily reflect sodium concentrations accurately since conductivity RBD = Treatment biomass Control biomass − 1.
Aboveground (Na A ) and belowground (Na B ) tissue sodium concentrations (%) were used to calculate total plant sodium concentration (Na T , %) using the formula: All extracted raw data for sodium accumulation have been organized in Table S4.

| Model design, selection, and population classification
We postulated a set of a priori potential response models for both RBD (Table 1) and sodium accumulation (  (Table 2), the quadratic function accounted for hump-shaped and nonlinear increase. In these quadratic-function cases, we used the vertex value (a) to classify cases as hump-shaped (when a was negative) or nonlinear decrease and nonlinear increase (when a was positive).
We used an Information Criterion (IC) approach to select the model that best fit the data extracted for each population, using three different ICs: Akaike Information Criterion (AIC), the AIC small-sample corrected version (AICc), and Bayesian Information Criterion (BIC). We used the R package 'AICcmodavg' to calculate AIC, AICc, and BIC values (Mazerolle, 2020). Although we examined results from all three metrics, we based our conclusions on BIC, since this metric gave consistent results across the data sampled, it is more specific (reduced Type-I error or lower false-positive rate), and it is considered a more conservative test, as advocated by Dziak et al. (2020). AIC is mainly recommended for larger datasets and does not account for sample size. Furthermore, for AICc, the penalization that is given to the AIC formula increases the chances of overfitting the data due to the extremely small sample sizes for the data analyzed (Bolker, 2008;Dziak et al., 2020). The models from Tables 1 and 2 that best fit each response (i.e. the smallest BIC value) were used to designate a response shape for each population's aboveground, belowground, and total plant biomass growth and sodium accumulation, respectively. Since we based our conclusions on BIC, we provide the corresponding likelihood values, ΔBIC and BIC weights for each model chosen; we also share results from the other two IC metrics for comparison (Tables S5-S8).
Fisher's exact test contingency analysis with simulated p-values in R-Studio following recommendations from Broman and Caffo (2003)

| Phylogenetic patterns among responses
We performed a phylogenetic signal analysis to assess whether phylogenetic relationships may have influenced growth and sodium accumulation responses in the diverse set of taxa used in our systematic review. The phylogenetic signal is the tendency of closely related species to resemble each other more in trait values than species drawn at random (Blomberg et al., 2003;Münkemüller et al., 2012). We used a subset of the rooted and dated ALLMB phylogeny from Smith and Brown (2018)  We tested for phylogenetic signals for the discrete characters of aboveground, belowground, and total plant growth and sodium accumulation responses using the Maddison and Slatkin (1991) TA B L E 1 A priori response predictions for relative biomass growth and models used to classify populations in plants exposed to increasing concentrations of NaCl in the substrate

Equation Classification
A priori representation estimates the minimum trait transitions at each node and compares this to a distribution sampled from a null model (Head et al., 2018;Paleo-López et al., 2016). We used 1,000 randomizations to infer a significant result if the number of observed trait changes was significantly (α = 0.05) less than the median of the null model distribution. All data were analyzed using R software version 3.6.3 (R Core Team, 2020).

| Increasing substrate NaCl has varied effects on total plant growth responses
Using model selection for each of our chosen 107 populations, we classified relative total plant growth responses as shown in Table 1 ( Table S1). Growth was negatively affected as sodium increased in the substrate for most taxa. However, 12 taxa grew better in at least one treatment ≥200 mM NaCl. Growth was severely reduced in all populations that were exposed to NaCl concentrations >500 mM as compared to 0 mM of NaCl ( Figure 1). None of the populations that we classified as having linear increase or zero slope biomass responses were exposed to treatments >360 mM NaCl.
Plant growth based on relative biomass difference showed similar trends in response to increased salinity regardless of the tissue sampled from aboveground or belowground ( Figure S1a,b). The overall growth patterns of aboveground or belowground tissue mirrored the patterns observed at the total plant level, as visualized by similarity in the alluvial plot (Figure 2a).

| Total plant sodium increases as substrate sodium increases
Using model selection for each of the 107 populations, we classified total plant sodium accumulation responses into six groups shown in Table 2 (Table S1). The total sodium concentration within a plant generally increased as the substrate concentration of sodium increased (Figure 1b). However, the level of sodium accumulation was highly variable among populations and between aboveground and belowground tissues ( Figure S1c,d). Notably, the aboveground sodium concentrations were generally higher than in belowground tissues for most populations in saline treatments ( Figure S1c,d). Additionally, regardless of the variation observed, both relative aboveground and belowground responses F I G U R E 1 Populations' responses to increasing substrate NaCl concentrations. Total relative biomass growth responses (a) across NaCl treatments for each population sampled in the study. Negative and positive values represent a growth inhibition or an increase, respectively, in growth relative to control NaCl substrate concentrations. Also, the effect of NaCl treatments on total plant sodium accumulation (b) across increasing NaCl substrate concentrations for each population. The main data shown cover the range from 0 to 600 mM treatments of NaCl. An inset with the complete dataset and treatments is included with each panel. Colors represent the responses that describe biomass growth and sodium accumulation responses, as in Tables 1 and 2 were similar to relative total sodium accumulation responses ( Figure 2b).

| Crop species do not adequately represent general plant responses
In our study, crop species represented 43.3% (29)

| Plant growth responses do not predict sodium accumulation responses
Total plant biomass growth responses were largely independent of the type of sodium accumulation response, which we illustrate using an alluvial plot (p = .43; Figure 5). Furthermore, irrespective of the growth response, tissue sodium concentrations increased monotonically in the majority of populations, i.e. increase in plant sodium continued at a steady positive rate as sodium in the substrate increased or increased to a plateau for 77% of the populations (Figures 1 and 5).
Only those populations with hump-shaped growth responses differed significantly in sodium accumulation between aboveground and belowground tissues across saline treatments (Wilcoxon Test: n = 17, Z = 1.9, p > .046). There were no statistically significant differences for any other biomass growth responses between sodium F I G U R E 2 Alluvial plot describing the association between above-and belowground phenotype responses to total plant biomass (a) and sodium accumulation (b). Thickness of each connector indicates the proportion of populations in each response group accumulation of aboveground versus belowground tissues across saline treatments. Additionally, for nonsaline treatments, there was no statistically significant difference for any biomass growth response groups when aboveground and belowground sodium accumulation was compared ( Figure 6).

| Phylogenetic relationships predict biomass growth but not sodium accumulation responses
Biomass growth, both aboveground and belowground, showed significant phylogenetic signal (i.e., phylogenetic relationships help explain the distribution of the trait across the phylogenetic tree in our dataset; p = .031 and p = .046, respectively; Figure 7). We

| D ISCUSS I ON
Understanding the influence of sodium in the substrate on plant performance (growth, fitness) and tissue sodium accumulation is essential to comprehend ecological and evolutionary dynamics of plants across terrestrial environments. Our study emphasizes that plant adaptations to substrate sodium vary substantially across taxa in terms of growth and sodium accumulation with a degree of phylogenetic conservatism. However, regardless of growth responses, sodium accumulation mostly followed an increasing trend and did not have any apparent association to growth responses as substrate sodium levels

F I G U R E 6
Mean log-transformed tissue sodium concentration (%) (and SE) for above-and belowground tissues across biomass growth responses for non-saline (0 mM NaCl) and saline treatments (30-300 mM NaCl). Significant differences (p < .001, Wilcoxon Test) for above-and belowground mean response comparisons are indicated by asterisks (***). Sample sizes for each growth response for above-and belowground tissues were the same for non-saline and saline treatments: hump-shaped = 17; linear decrease = 36; linear increase = 3; non-linear decrease = 32; threshold decline = 11; and zero slope = 3 increased. Additionally, we note that for the taxa for which we had data, domesticated plant species present a narrower range of variation among salt stress responses as compared to non-crop species. In order to understand how substrate sodium influences plant growth, we must consider non-crop species in our studies. Moreover, we advocate the importance of characterizing responses using a systematic approach, and we provide recommendations on experimental designs to reach a broader understanding of plant-salt stress.

| Increasing substrate sodium influences plant growth and sodium accumulation in variable ways
Saline soils are known to hinder plant growth, and crop losses are reported when soil salinity is above a crop-specific threshold (Bernstein, 1975;Zhao et al., 2020;Zörb et al., 2019). Whereas our analysis is aligned with this general consensus on the negative impact of soil salinity on plant growth, it sheds light on how plant F I G U R E 7 Total (T), above-(A) and belowground (B) plant biomass growth responses mapped onto a phylogeny. Tips represent species pruned from rooted and dated ALLMB phylogeny from Smith and Brown (2018). Plant orders are indicated to the right of the phylogeny growth varied in response to substrate NaCl levels across plant taxa that ranged from highly studied crops to scarcely examined wild species (Table 1 and Figure 2a,b). Despite the overall trend of decreased biomass concurrent to increasing substrate NaCl levels, several taxa in the order Caryophyllales (e.g., families Amaranthaceae, Plumbaginaceae and Portulacaceae) showed a hump-shaped or linear increase in biomass growth to increasing substrate NaCl (Figures 1a   and 7). Most halophytes are non-randomly distributed, and the order Caryophyllales holds the greatest number of recorded halophytes among angiosperms (Flowers et al., 2010). Halophytes not only are tolerant of high NaCl, but also use Na + and Clions for osmotic adjustment in an energetically favorable manner and are equipped with structural and physiological traits that aid the compartmentalization of salts to promote growth while avoiding ionic or osmotic F I G U R E 8 Total (T), above-(A) and belowground (B) plant sodium accumulation responses mapped onto a phylogeny. Tips represent species pruned from rooted and dated ALLMB phylogeny from Smith and Brown (2018). Plant orders are indicated to the right of the phylogeny stress until threshold NaCl levels are reached (Munns et al., 2020;Slama et al., 2015). This set of characteristics would account for the positive growth in saline substrates that we observed within the Caryophyllales taxa (Figures 1a and 7). Furthermore, plants that follow these hump-shaped or linear increase growth responses to increasing substrate sodium follow a subsidy-stress gradient, i.e.
at low substrate sodium levels overall plant growth is subsidized, reaching a threshold leading to growth inhibition due to salt stress as sodium in the substrate becomes toxic (Odum et al., 1979). All plants that followed these trajectories in our analyses (Figure 7) are considered salt tolerant, as classified in the eHALOPH database (Santos et al., 2016) and by the respective authors in each study (Table S1).
Regardless, even among those salt-tolerant taxa, plant biomass eventually decreased at the highest NaCl concentrations (Figure 1a). The use of sodium as an inexpensive osmolyte has convergently evolved in many halophytes as well as other plants adapted to water deficit stress and is found in multiple orders of plants. For example, even at low sodium levels in the soil, the xeric adapted plant, Zygophyllum xanthoxylum (Zygophyllaceae), accumulates high concentrations of sodium in shoots, resulting in large mesophyll cells leading to leaf succulence (Xi et al., 2018).
The taxa that showed linear or nonlinear decreases (Figures 1a   and 7) as NaCl increased in the substrate are non-halophytes highly sensitive to salt stress where growth is inhibited by excess salts (Munns et al., 2020;van Zelm et al., 2020). Moreover, we found that closely related lineages resembled each other with respect to biomass growth responses (i.e. significant phylogenetic signal indicating shared physiological responses within clades); thus, the patterns observed in this trait are at least somewhat explained by shared evolutionary history (Figure 7). However, phylogenetic patterns do not account for sodium accumulation responses (Figure 8).
In plants, tissue sodium concentrations are generally linked with increasing substrate sodium concentrations ( Figure 1b). However, plant sodium accumulation seemed to be uncoupled from biomass growth responses and any discernible phylogenetic signal among taxa ( Figures 5 and 8). Similar patterns were observed when aboveground sodium accumulation was compared in the species Plantago maritima and Plantago media as NaCl in the substrate was increased (Maathuis, 2014; note that these populations -among others in the literature -were not included in the current study since they did not meet the criteria for our selection). The variation in responses by each species was mainly due to differential and discrete tolerance thresholds and external sodium concentrations (Maathuis, 2014), which might explain the idiosyncratic variation that was observed among taxa used in this study in terms of sodium accumulation responses (Figures 1b and 8).
Additionally, the accumulation of higher amounts of sodium in aboveground ( Figure S1c) than belowground ( Figure S1d) tissues is apparent when comparing sodium accumulation responses for each population across increasing treatments of substrate NaCl ( Figure 1b). This observation agrees with the current understanding that sodium, once in the transpiration stream, is retained in the shoots as phloem recirculation to roots and is considerably less then xylem loading from roots to shoots (Munns, 2002;Munns & Tester, 2008).
Sodium accumulation in the shoots is dependent on the local tissue and species-specific tolerance capacity. Plants are known to store excess sodium in older leaves to protect younger growing tissue from salt toxicity, and sustain growth until species-specific tolerance levels are reached (Munns & Tester, 2008). Alternatively, a few halophytes have developed salt glands to remove sodium from shoots against a concentration gradient -an adaptation that is found in several plant orders (Dassanayake & Larkin, 2017).
Once sodium enters the roots, plants have transporters that preferentially export sodium back to the soil at an energy cost. However, this capacity to export sodium at the soil-root interphase is easily exceeded even among halophytes, and accumulation of sodium inside the plant is unavoidable when external sodium concentrations increase (Zhao et al., 2020). Therefore, other sodium transporters that facilitate ionic balance throughout the plant organs play critical roles in sustaining growth or survival during salt stress (Apse & Blumwald, 2007;Yamaguchi et al., 2013). Our systematic review agrees with previous studies investigating single or small groups of taxa subjected to salt stress to highlight that almost all plants accumulated sodium monotonically (or nearly monotonically) as sodium increased in the substrate (Figures 1b and 5). Plants that expressed the biomass growth hump-shaped response accumulated significantly higher concentrations of sodium in aboveground than belowground tissues. Alternatively, populations characterized by the other growth responses did not differ significantly in aboveground versus belowground sodium accumulation in saline treatments but not in nonsaline treatments ( Figure 6). We discussed earlier that the humpshaped response was preferentially represented by taxa in the order Caryophyllales and that this clade is an evolutionary hotspot for halophytes, but this response is not confined to the order (Figure 7).
Furthermore, Caryophyllales species often are shoot sodium hyperaccumulators; they are enriched in plants that develop salt glands and have a higher tolerance to higher tissue sodium levels than predominantly salt-sensitive orders (Dassanayake & Larkin, 2017;Flowers et al., 2010;White et al., 2017).

| Domesticated plant species tend to occupy a narrow range of variation among salt stress responses
Our systematic review demonstrated a clear dichotomy between salt tolerance (deduced from growth responses) during increased external sodium in crops compared to wild species or plants that have not been subjected to domestication. Most wild species in our study tend to have a higher capacity to tolerate higher tissue sodium than crop or domesticated species (Figure 2a,b). The exception to this is seen with crops in Caryophyllales, such as Beta vulgaris, Salicornia bigelovii, and Spinacia oleracea (Choo et al., 2001;Wu et al., 2013;Yamada et al., 2016). Recent studies have illustrated how crop species have lost traits related to salt tolerance their ancestral wild relatives had before domestication (Quan et al., 2018;Rozema et al., 2015;Wang et al., 2020Wang et al., , 2021.

| Characterizing responses promotes our understanding of plant-salt stress
The modeling approach that we used in this study provides a useful way to quantify and categorize individual plant population responses to variation in NaCl in the substrate. These models describe the response trajectories of biomass growth and sodium accumulation responses and could be used extensively across taxa of interest. By using an Information Criterion approach, one can select the best-fit model for each population, given that our formulated models (e.g. linear decrease, hump-shaped, etc.) effectively describe natural patterns (Brewer et al., 2016), within and among species (Tables 1 and   2). For many purposes, it may be more useful to categorize plants by their responses across a range of sodium conditions, as opposed to performance above and below strict thresholds, as is often done with halophytic or salt-tolerant plants (see Grigore et al., (2014) for a review on definitions and descriptions related to halophytes).

| Experimental design to achieve broader understanding
Many studies have tested the effects of NaCl on plant growth and yield, especially in crop species (Cheeseman, 2015). However, because of differences in methodology, it is a challenge to make comparisons and contrasts of results across studies. We make several observations and recommendations for future studies: a. Often, there is a lack of enough replication and/or treatments.
For us, this prevented effective response pattern identification in some cases, especially in studies that presented only three treatments with few replicates.
b. The determination of treatments was often arbitrary. Limitations are imposed using independent categorical variables (ANOVAbased approach) instead of applying treatments as independent numeric discrete or continuous variables (regression-based approach). Experimental designs that cover a wide range of treatments may provide more accurate estimates. A regression-based approach allows one to better fit nonlinear responses, which encompasses most of the responses we measured in our study (Inouye, 2001;Whitlock & Schluter, 2014 Generally, a combination of copper, calcium, magnesium, and sodium is found in tap water on average at 1%, with some regional variation (Patterson et al., 2013 e. Studies generally focus on biomass to the exclusion of other fitness-related traits. Even though biomass is often an acceptable proxy for fitness measurements in plants (Younginger et al., 2017), observations on flower production, survivorship, seed set, and seed germination success should be quantified, to provide a more complete understanding of sodium's influence on whole-plant performance and fitness (Primack & Kang, 1989).
f. Studies also should consider that salt stress is often combined with water deficit and heat stress, or other nutrient stresses in natural habitats. Additionally, biotic stresses such as herbivory and disease can compound the overall plant response to salt stress, with special consideration of wild taxa.

CO N FLI C T O F I NTE R E S T
The authors have no competing interest to declare.