Skip to main content

Salt tolerance is evolutionarily labile in a diverse set of angiosperm families



Salt tolerance in plants is rare, yet it is found across a diverse set of taxonomic groups. This suggests that, although salt tolerance often involves a set of complex traits, it has evolved many times independently in different angiosperm lineages. However, the pattern of evolution of salt tolerance can vary dramatically between families. A recent phylogenetic study of the Chenopodiaceae (goosefoot family) concluded that salt tolerance has a conserved evolutionary pattern, being gained early in the evolution of the lineage then retained by most species in the family. Conversely, a phylogenetic study of the Poaceae (grass family) suggested over 70 independent gains of salt tolerance, most giving rise to only one or a few salt tolerant species. Here, we use a phylogenetic approach to explore the macroevolutionary patterns of salt tolerance in a sample of angiosperm families, in order to ask whether either of these two patterns – deep and conserved or shallow and labile - represents a common mode of salt tolerance evolution. We analyze the distribution of halophyte species across the angiosperms and identify families with more or less halophytes than expected under a random model. Then, we explore the phylogenetic distribution of halophytes in 22 families using phylogenetic comparative methods.


We find that salt tolerance species have been reported from over one-third of angiosperm families, but that salt tolerant species are not distributed evenly across angiosperm families. We find that salt tolerance has been gained hundreds of times over the history of the angiosperms. In a few families, we find deep and conserved gains of salt tolerance, but in the majority of families analyzed, we find that the pattern of salt tolerant species is best explained by multiple independent gains that occur near the tips of the phylogeny and often give rise to only one or a few halophytes.


Our results suggest that the pattern of many independent gains of salt tolerance near the tips of the phylogeny is found in many angiosperm families. This suggests that the pattern reported in the grasses of high evolutionary lability may be a common feature of salt tolerance evolution in angiosperms.


Only 1 - 2 % of angiosperm species are known to be halophytes, able to live and reproduce in saline soils [1, 2]. The rarity of salt tolerance is unsurprising considering it is a costly and complex ecological strategy; halophytes may have modifications to many parts of their physiology and anatomy in order to combat the damaging effects of osmotic and metabolic stress, which can cause impaired growth and reproduction [24]. However, halophytes are found in a wide range of angiosperm families and they occupy diverse habitats worldwide.

Salt tolerance has also clearly evolved multiple times in angiosperms [5, 6]. The evolutionary patterns of salt tolerance in plants have been studied in detail in only a few taxonomic groups, and these studies have revealed two very different patterns of salt tolerance evolution. In one well-studied group, the chenopods (Chenopodiaceae), salt tolerance appears to be phylogenetically conserved [7], arising only once or twice in the history of the group, then being retained in a large proportion of species in the family. In contrast, a study on the grass family (Poaceae) estimated that there have been at least 70 origins of salt tolerance within the family [8]. Most of these inferred origins were near the terminal taxa (tips) of the phylogeny, suggesting multiple shallow origins, each giving rise to only one or a few salt tolerant species. The pattern of salt tolerance evolution inferred in the Poaceae is interesting because it suggests that, at least in the grasses, salt tolerance has evolved repeatedly in a range of lineages, despite the complexity of salt tolerance adaptations. However, the observation that salt tolerance does not persist over long evolutionary timescales in the grasses may indicate that while salt tolerance is easy to gain, it is also frequently lost through trait reversal or extinction, implying that there are costs associated with the adoption of salt tolerance.

These two different phylogenetic patterns suggest very different macroevolutionary dynamics. Salt tolerance is highly conserved in the chenopods, with a large number of salt tolerant species arising from only a few independent origins. But in the grasses, salt tolerance is highly labile, in the sense that it is gained and lost relatively frequently. Which, if either, of these patterns is observed in other families of angiosperms? To answer this question, we use a phylogenetic comparative approach to investigate and characterize patterns of halophyte diversity and evolution among angiosperm families.

Halophytes use a variety of physiological and anatomical traits to survive in saline habitats, and these traits can vary between species. Some halophytes exhibit complex anatomical modifications like salt glands or hairs, but most halophytes rely on osmotic regulation, modifying existing physiological mechanisms to mitigate salinity levels within the plant [9, 10]. These strategies can also vary amongst closely related halophytes and among halophytes that occupy similar habitats, for example the differential presence of succulence among closely related chenopods [7] or salt glands among phylogenetically diverse mangrove species [11]. Instead of identifying specific environmental or physiological differences between halophytes, we focus on the broad distribution of salt tolerance as an ecological strategy amongst angiosperms, at the family and species levels. We first examine how halophytes are distributed among the angiosperm families, identifying any families that have more or less halophytes than expected by chance. Then we use a number of phylogenetic measures to analyze the observed evolutionary patterns of salt tolerance in a sample of 22 angiosperm families. This sample includes large families with many known halophytes, including families with both more and less halophytes than expected.


Halophyte Diversity

We found that the observed distribution of halophytes across angiosperm families was significantly nonrandom (p < 0.001). Of the 411 families included in the taxonomic analysis, 146 families have one or more known halophytes (See Methods and Additional file 1). We found that 51 of the 411 families have significantly more halophytes than expected by chance; examples include Amaranthaceae, Poaceae and Rhizophoraceae (see Additional file 1: Table S1). 68 families have significantly fewer halophytes than expected by chance, for example Acanthaceae, Lamiaceae and Fabaceae.

Evolutionary Patterns

For each family analyzed, we created a family subtree that included all tips from a large angiosperm phylogeny [12] belonging to each family according to GenBank taxonomy (See Methods). In the 22 family subtrees analyzed, we observed a range of evolutionary patterns of salt tolerance (see Fig. 1 and Additional file 1: Figure S1). In general, evolutionary gains of salt tolerance appeared close to the tips, across the family subtrees. One measure used to assess the evolutionary patterns of salt tolerance across families was the number of tips per origin (NoTO), the average number of taxa descending from each inferred evolutionary origin of salt tolerance in a family. The median value of NoTO across all 22 families analyzed was 1.3 and nineteen of the family subtrees had a NoTO value less than two. This observation indicates that the inferred gains of salt tolerance in these family subtrees typically give rise to less than two descendant halophyte tips. In contrast, a few families (Rhizophoraceae, Amaranthaceae and Tamaricaceae) had higher NoTO values than the other families, meaning that each gain of salt tolerance in these families is deeper in the subtrees and leads to comparatively larger clades of halophytes than observed in the other families analyzed. Tamaricaceae was the only family in our sample with significantly fewer salt tolerance gains than expected given the number of known halophytes and halophytic taxa were significantly clustered.

Fig. 1
figure 1

Family subtrees for a sample of six of the families analyzed with significant NoTO and/or SSCD values. Inferred gains of salt tolerance (see Methods) are marked on each family with black circles. Tips in the subtrees identified as halophytes are marked in black in the ring around the subtree. The subtrees represent a) Amaranthaceae, b) Apiaceae, c) Brassicaceae, d) Cucurbitaceae, e) Cyperaceae, and f) Euphorbiaceae. All 22 subtrees included in the analysis are presented in Additional file 1: Figure S1. Subtree plots were created using the “Diversitree” package [58]

Over half of the families analyzed had a similar phylogenetic distribution to the pattern found in the grasses [8]. In these families, given the observed number of halophytes in each of the family subtrees, either 1) salt tolerance has evolved more times than expected under a Brownian motion model (significantly lower NoTO) and/or 2) clades of halophytes are less clustered than expected under Brownian motion model of trait evolution (significantly higher sum of sister clade differences (SSCD), see Methods) (Table 1). When comparing with the results from the angiosperm diversity analysis, we found that this labile evolutionary pattern is found in families with varying proportions of halophytes, including those with more, fewer, and within the expected number of halophytes based on family size.

Table 1 Results of taxonomic and phylogenetic analyses for a sample of 22 angiosperm families

Correlation tests suggest that there is no significant association between taxon sampling proportion and estimates of SSCD (p = 0.660, τ = 0.071) or NoTO (p = 0.728, τ = 0.056) estimates. The proportion of halophytic taxa in the subtrees is not correlated with SSCD (p = 0.158, τ = 0.230) or NoTO p-values (p = 0.087, τ = 0.270).


Using a list of known halophytes assembled from a range of published sources, we find that one-third of angiosperm families contain species reported as being able to live in saline conditions. We show that the distribution of salt tolerant species among angiosperm families is not consistent with a random distribution, and that some families have significantly more halophytes than expected given the family size, while others have significantly fewer salt tolerant species.

Not only does the proportion of halophytes differ between families, but the phylogenetic distribution of salt tolerance within families also varies. Specifically, we set out to test whether the pattern of salt tolerance in grasses (Poaceae) – with many, shallow, scattered gains – was also found in other families. A few families show the opposite pattern, where salt tolerance has been gained deep in the family and retained by a large proportion of descendants [7]; examples include Tamaricaceae which contains the highly salt tolerant salt cedars and many species with specialized anatomical traits like salt glands [13]; Amaranthaceae, which includes the halophyte-rich groups formerly classified under Chenopodiaceae; and Rhizophoraceae, which contains many mangrove species. However, over half of the families analyzed show a pattern like the grasses, consistent with many, shallow gains of salt tolerance. The fact that we find this labile pattern of salt tolerance evolution in a phylogenetically diverse set of families with different proportions of halophytes suggests that the observed pattern of many independent gains of salt tolerance is not simply explained by the proportion of halophytes in a group.

One limitation of broad comparative analyses like this one is that we can only gain information from data on known halophytes and sequenced angiosperm taxa. Specifically, the limitations in data used in this study come from two main sources. One problem is the incidence of false negatives in the halophyte list. Most published lists of halophytes are based on observational data, and there are likely to be other salt tolerant species that have not been described in the literature or included in published lists. For example, there are likely many species living in non-saline habitats that have the capacity for salt tolerance but have not yet been formally tested. One solution to improve future analyses is to move away from list-based methods drawn from single species experiments and observational data. For example, it may be possible for phylogenetic studies to contribute to the identification of salt tolerance lineages, for example by using use phyloprediction [14] or geochemical modeling to identify lineages that are likely to be salt tolerant [15, 16].

A second potential source of error is incomplete phylogenetic sampling. The phylogeny used in this study includes about 20 % of known angiosperm species, so there are some known salt tolerant taxa that are not included in this tree (see Table 1 for details). Correlation tests did not indicate any consistent effect of the proportion of total species sampled in a family or of the proportion of halophytes in the subtrees on the results of NoTO and SSCD, suggesting that sampling proportion does not significantly influence the results of our analysis. Increased sampling is unlikely to change the overarching pattern because salt tolerant taxa in most family subtrees with significant NoTO and SSCD are sparsely distributed and many of the inferred gains are distantly related on the trees (see Fig. 1 and Additional file 1: Figure S1). This pattern suggest that in most cases adding more salt tolerant taxa is likely to either increase the total number of inferred salt tolerance gains (if adding species that are not closely related to known halophytes) or maintain the number of inferred gains (if adding to a clade of known halophytes). However, there are some clades, for example in Cyperaceae and Amaranthaceae, with denser groups of halophytes, where the number of inferred gains relative to the number of halophytes is more likely to be reduced by adding more halophytes (Fig. 1a,f). Similarly, removing identified halophytes with only low or seasonal tolerance to salinity could in some cases increase phylogenetic clustering of halophytes, reducing the number of inferred gains (if removing species that are not closely related to known halophytes), or break up some clades, possibly increasing the estimated number of gains. Based on the extant pattern of halophytes, our analysis suggests that salt tolerance has been gained at least 600 times in the 22 families analyzed. And if we assume that each of the other angiosperm families that contain halophytes also represents at least one independent gain of salt tolerance, there are likely to be 124 additional gains or more across the angiosperms.

Our results are consistent with the findings of two previous group-specific studies on phylogenetic patterns of salt tolerance [7, 8]. We infer over 100 gains of salt tolerance within the grass family subtree, and confirm that halophytes are more phylogenetically dispersed than expected under Brownian motion. We also demonstrate that the Amaranthaceae has relatively high numbers of species per inferred gain, indicating a more conserved pattern of salt tolerance evolution compared to the other families in the analysis. While we estimate that salt tolerance is significantly less clustered than Brownian motion in Amaranthaceae, this result appears to be driven by about one-third of the family with notably fewer halophytes than the rest of the tree (Fig. 1a).

Given that salt tolerance may involve many anatomical, physiological and life history modifications, it may seem surprising that it has evolved so many times in such a wide range of lineages. However, it has been suggested that other stress tolerance strategies involving complex sets of ecophysiological traits have also evolved multiple times [1720]. One explanation for how salt tolerance has evolved multiple times is that the required physiological or anatomical changes can build on precursor traits acquired earlier in the history of the lineages. A well-studied example of how complex physiological traits can build on precursor traits is C4 photosynthesis in the grasses [21]. In a few angiosperm families researchers have inferred many independent evolutionary origins of C4 photosynthesis, a specialized form of photosynthesis often associated with arid-adapted lineages [19, 22, 23], which requires many biochemical and anatomical modifications. Lineages with a higher proportion of vascular bundle sheath cells have a higher frequency of evolution of C4 photosynthesis, suggesting that some types of foliar anatomy facilitate the transition to C4 [23]. Similarly, if salt tolerance builds on existing physiological or anatomical traits, then a lineage with these traits may have a higher likelihood of giving rise to halophytic species. For example, C4 grass lineages are more likely to contain halophytes than C3 lineages, possibly because C4 photosynthesis allows more efficient water use and therefore limits the impact of salinity by reducing the uptake of ions and limiting the effects of osmotic stress [24].

Although the idea of evolutionary precursors may explain why some lineages develop salinity tolerance more often than others, the question remains: why are salt tolerant lineages often found as singletons on the phylogeny or in small clades? There are several broad explanations for this pattern, which are not mutually exclusive. One explanation is that the observed distribution of halophytes could reflect patterns of change in land salinity over time. Although some saline areas are long lived (e.g., coastal habitats), in some areas salinity can vary over small spatial scales or shift on a seasonal basis [25]. If lineages are rapidly responding to changing salinity, this could partly explain why we infer mostly shallow gains of salt tolerance that give rise to only one or a few extant halophytes. However, recent origins of saline habitats is unlikely to provide a general explanation for the multiple recent gains of salt tolerance in many families, because some saline habitats are stable over long evolutionary time periods, so should provide persistent habitat for saline specialists.

Another explanation for why there are so many small clades of halophytes is that salt tolerance may be a costly ecological strategy that is relatively easy to gain but difficult to maintain. For example, high plasticity could enable some lineages to transition into harsh or novel habitats over evolutionarily short time scales [26, 27]. However, maintaining a strategy like salt tolerance could be physiologically costly, for example due to the cost of producing osmoprotectants or increasing investment in reactive oxygen species scavenging and antioxidant production (reviewed in [28]). The high physiological cost of salt tolerance could lead to increased extinction rates in halophytes, or high reversal rates if lineages that invest less in salt tolerance mechanisms have a competitive advantage. This scenario could lead to an extant pattern of many shallow gains of salt tolerance dispersed across the phylogeny. Some research suggests that the more salt tolerant a species, the less competitive it is in less saline or non-saline environment [29, 30], although the generality of these claims are disputed [31]. Reduced competitive ability may threaten the persistence of halophytes if land salinity subsides, and halophytes may not be ecologically competitive when transitioning back into a non-saline environment [32], which could lead to local extinction or the loss of salt tolerance. However, the lower competitive ability may not always be a direct result of salt tolerance [31], and high salinity tolerance may even confer a competitive advantage for some species in non-saline habitats. For example, salt cedars (Tamarix, Tamaricaceae) are highly salt tolerant, yet they are invasive in some non-saline and low-saline riparian habitats. Salt cedar populations are capable of displacing natives by using more water and excreting salt into the soil, creating a toxic environment for non-salt tolerant native species [33].

It is unlikely that either changes in land salinity patterns or the cost of salt tolerance can fully explain why salt tolerance has evolved many times and why halophytes are often found as singletons and in small clades. And it has been suggested that, in general, the transition into different habitats and the evolution of ecological traits may be highly context dependent [21]. Identifying the phylogenetic patterns of salt tolerance represents an important step towards understanding salt tolerance evolution. We hope that reporting results in the context of angiosperm families will be useful for more detailed studies in future on the environmental and physiological aspects of salt tolerance evolution in different lineages. In future it would be interesting to explore the role that related traits, order of trait acquisition, and climatic history have played in the observed patterns of salt tolerance evolution, as has been examined for C4 photosynthesis [23] and freezing tolerance [34].


Salt tolerance in plants is an interesting case study in macroevolution [35]. Salt tolerance is an ecological strategy that often involves complex physiological features. Halophytes are rare, yet they are found in a diverse set of taxonomic groups. Our analysis shows that in a range of angiosperm families, salt tolerance has been gained a surprising number of times and that these transitions are shallow and spread out near the tips of the phylogeny. This suggests that while the evolutionary pattern of salt tolerance varies across angiosperm families, it seems that salt tolerance can evolve frequently in many different genetic backgrounds. This result is intriguing, given how difficult it has to been to manipulate the salt tolerance of commercial crop varieties [31, 36], but the frequent evolution of salt tolerance may give hope that many plant lineages can build on existing physiological and anatomical traits to develop increased tolerance of environmental salinity.


Halophyte database

Our aim was to broadly investigate the patterns of salt tolerance distribution as an ecological strategy across angiosperms. We first compiled a list of known halophytes. Instead of differentiating halophytes based on specific traits or environmental conditions, we analyzed salt tolerance as a binary trait, categorizing plants as reported to tolerate salt (labeled as 1) or not reported as salt tolerant (labeled as 0). Analyzing salt tolerance as a binary character is the only practical approach for a broad scale comparative study since there are relatively few species for which we have information on specific levels of salt tolerance, and this approach also allowed us to study a wide variety of salt tolerant species. We started with a published list of approximately 2600 taxa observed in saline habitats [37]. We then searched the literature and added taxa from five additional halophyte lists that were published more recently [3842] (See Additional file 1 for details). These published lists included halophytes identified from field surveys and observational data. It is possible that some taxa included in these lists have low salinity tolerance, are only tolerant to limited exposure to salinity (e.g., seasonal salinity), or have experienced acclimation to salinity [43, 44]. For this study, we consider that these species have an underlying propensity for developing salt tolerance, and so their inclusion is useful in a broad study on the evolution of salt tolerance. The resulting list contained 4515 taxa reported to be salt tolerant (including infraspecific taxa). We then searched for synonyms and accepted names of each taxon in this list according to The Plant List [45] using the R package ‘taxonstand’ [46]. Because the taxonomic and phylogenetic analyses had different aims, we created separate lists for each analysis, which are described below.

Halophyte Diversity

Our first aim was to investigate the taxonomic distribution of halophytes across angiosperm families. Although families may represent lineages of different ages or evolutionary patterns, here they are used simply as a convenient taxonomic division of angiosperm diversity into defined groups. We identified 411 unique angiosperm families by checking the 413 families recognized by the Linear Angiosperm Phylogeny Group (LAPG III) [47] against the APG III website [48, 49] (two were found to have equivocal names, see Additional file 1). We then collected mean estimates of species numbers for each of the 411 families from the APG III website [49], totaling 276,000 species. Since these family size estimates are reported at the species level, we needed to compile a list of halophytic species. We selected only the unique set of accepted halophyte species names according to The Plant List [45]. We collapsed the names of accepted infraspecific taxa to the species level, counting a species as salt tolerant if one of its varieties or subspecies was listed as a halophyte. The resulting list contained 2852 unique halophyte species (see archived data for halophyte list). We then counted the number of known halophyte species in each angiosperm family.

Then, we tested if halophytes were distributed randomly across families. We tested whether the number of halophytes in each family followed a binomial distribution parameterized by family size and the probability of being salt tolerant equal to the observed proportion of halophytes over all the angiosperm families. We applied a G-test of independence to estimate the overall fit of the model to the angiosperm families, using the likelihood.test function in R package “Deducer” [50].

We then calculated the probability of observing the number of known halophytes for each family based on the binomial distribution. This probability allowed us to identify families with significantly more or fewer halophytes than expected by chance. If the probability of a family having the same number or more halophytes than observed is lower than 0.05, the family is considered to have significantly more halophytes than expected by chance. If the probability of a family having the same number or fewer halophytes than observed is lower than 0.05, the family is considered to have significantly fewer halophytes than expected by chance.

Evolutionary patterns


Our second aim was to investigate the phylogenetic patterns of halophyte distribution. Because halophytes are rare, we did not analyze the distribution of salt tolerance across the entire angiosperm phylogeny, but focused on a subset of families that each contained many halophytes. Our main aims in family selection were to: (1) collect the largest families containing a sufficient number of halophytes to provide sufficient power for the analysis of evolutionary patterns; and (2) select families that were found to have more, fewer or within the expected range of halophytes in the taxonomic analysis, so as not to bias the analysis to families with a higher representation of salt tolerant taxa. All our analyses were conducted using phylogenetic information from the largest available phylogeny for angiosperms [12]. This phylogeny includes all appropriate angiosperm sequences on GenBank, including infraspecific taxa, which covers approximately 20 % of angiosperm species.

We first assigned each tip in the phylogeny to a family according to GenBank taxonomy using the TaxoGB function in the R package “BoSSA” [51], and recorded the total number of tips (terminal taxa in the phylogeny) associated with each family. We then determined which tips in the phylogeny were halophytes, based on whether the tip name was included in either the halophyte names presented in the original publications or the accepted names found in the synonymy search (total 5030 halophyte names, see archived data files). We chose this identification method because the published phylogeny includes all angiosperm sequences on GenBank and is not restricted to the taxonomic classification on The Plant List [45]. We did not collapse infraspecific taxa to the species level for this analysis since the published phylogeny includes infraspecific taxa. We also identified whether the tips associated with each family were monophyletic in the phylogeny. In order to restrict the analysis to families with sufficiently large phylogenies and more than a few halophytes, we considered only families with 25 or more taxa included in the Smith et al. [12] phylogeny, of which at least six were recognized halophytes.

Under these criteria, we selected 22 families, including families with more halophytes than expected by chance, families with fewer halophytes than expected, and also families that fell within the expected number of halophytes for the family size. We first extracted sixteen families that were monophyletic in the Smith et al. [12] phylogeny. For each monophyletic family we extracted the family subtree from the phylogeny, which included all tips in the Smith et al. [12] phylogeny belonging to that family according to GenBank taxonomy. Next we extracted subtrees for families that met our selection criteria but were not strictly monophyletic in the Smith et al. [12] tree. For these families, we extracted a monophyletic family subtree by removing a small number of taxa that were assigned to the target family in GenBank taxonomy but did not fall into that family clade in the published phylogeny. In some cases we also excluded a small number of taxa within the target family clade that were assigned to a different family (see Additional file 1 for details on the names of taxa excluded from each family subtree). We then removed tips from the family subtrees that were not identified with standardized genus and species epithets. We excluded any tips with labels that included the taxonomic epithets “af”, “aff”, “cf” or “sp”. We also removed tips that represented hybrid taxa by identifying tip labels that included one genus and two specific epithets, as well as the word “hybrid” or where the two species names were separated by “x”. All family subtrees were extracted and analyzed with equal branch lengths. Polytomies in the family subtrees were randomly resolved using the multi2di function in the R package ‘ape’ [52].

Metrics for analyzing phylogenetic patterns

For each family subtree we used two metrics to assess the phylogenetic pattern of halophytes within the family. Specifically, we aimed to test whether any of these families showed the same evolutionary pattern of salt tolerance as the grass family, having (1) many shallow inferred origins of salt tolerance, near the tips of the phylogeny, which gave rise to small clades of halophytes, and (2) origins that were spread across the phylogeny, occurring in many different lineages [8]. To detect these patterns in our sample of families, we used two metrics: the number of tips per origin (NoTO) and the sum of sister clade differences, a measure of phylogenetic clustering (SSCD) [53, 54].

The number of tips per origin (NoTO) metric is used to test whether, given the number of halophyte taxa in the tree (tips), there are significantly more inferred origins of salt tolerance, and thus smaller clades of halophytes, than we would expect under a Brownian motion model of trait evolution. It is possible that salt tolerance has been gained and lost multiple times and that salt tolerant lineages have since gone extinct [35]. Here we only infer gains of salt tolerance that lead to extant salt tolerant species, as our aim is to infer the minimum number of independent gains needed to explain the extant phylogenetic distribution of halophytes. Our aim was to compare the observed taxonomic and phylogenetic distribution of halophytes to a null model of trait evolution using trait reconstruction techniques. To estimate the NoTO for each family subtree, we inferred the minimum number of gains and losses of salt tolerance required to explain the observed topological distribution of halophyte tips, and then calculated the average number of halophyte tips per inferred gain. A shallow, scattered distribution of halophytes, where most halophyte species arise from gains near the tips of the phylogeny and most gains lead to only a few halophyte species, will have a low value for NoTO.

To generate a null distribution of the expected number of gains given a number of known halophytes, we used established methods to simulate salt tolerance as a continuous trait on each family subtree using a Brownian motion model [54, 55]. We then used an appropriate threshold to convert the continuous trait to a binary one, such that the number of halophyte tips in the simulated tree was equal to the number of identified halophytes in the family subtree. We repeated this process 1000 times to generate a null distribution of NoTO values, specific to the observed number of halophytes and the size of the subtree for comparison with the observed NoTO value for each family subtree. To generate a p-value for each family, we calculated the proportion of simulated trees that had a NoTO value lower than or equal to the observed. P-values less than or equal to 0.05 represent significantly smaller clades of halophytes than expected under Brownian motion.

The sum of sister clade differences (SSCD) metric describes the degree of phylogenetic clustering of halophytes. We used the method of calculating SSCD described by Fritz and Purvis [54]. Each tip was coded as 1 if it was on the halophyte list and 0 if it was not. Each internal node in the family subtree was assigned a trait state using the mean of the descendant node or tip states (e.g., if one descendant was state 1 and the other was 0, the node value was 0.5). The SSCD was calculated as the sum of the absolute difference between trait states of each pair of sister nodes or tips over the whole tree. If gains of salt tolerance were scattered across the phylogeny, each giving rise to one or few halophyte species in a small clade of salt tolerant taxa, we would expect a large SSCD value. We compared the observed SSCD of each family subtree to the SSCD for 1000 traits generated under Brownian motion on the subtree, using the same method for generating the null distribution for the NoTO value. The p-value was the proportion of simulated trees that have higher SSCD values than the observed. P-values less than or equal to 0.05 indicated that salt tolerance is significantly scattered on the phylogeny compared to a Brownian motion trait.

We conducted correlation tests to assess whether estimates of NoTO and SSCD were influenced by incomplete sampling in the family subtrees or by the proportion of halophytic taxa in the family subtrees. Since NoTO, SSCD, and sampling values represent proportions, we used a non-parametric correlation test, Kendall’s tau [56].

Availability of supporting data

The data set supporting the results of this article is available in the Dryad repository [57].


  1. Glenn EP, Brown JJ, Blumwald E. Salt tolerance and crop potential of halophytes. Crit Rev Plant Sci. 1999;18:227–55.

    Article  Google Scholar 

  2. Flowers TJ, Colmer TD. Salinity tolerance in halophytes. New Phytol. 2008;179:945–63.

    Article  CAS  PubMed  Google Scholar 

  3. Greenway H, Munns R. Mechanisms of salt tolerance in nonhalophytes. Annu Rev Plant Phys. 1980;31:149–90.

    Article  CAS  Google Scholar 

  4. Flowers TJ. Improving crop salt tolerance. J Exp Bot. 2004;55:307.

    Article  CAS  PubMed  Google Scholar 

  5. Flowers T, Troke PF, Yeo AR. The mechanism of salt tolerance in halophytes. Annu Rev Plant Phys. 1977;28:89–121.

    Article  CAS  Google Scholar 

  6. Flowers TJ, Galal HK, Bromham L. Evolution of halophytes: multiple origins of salt tolerance in land plants. Functional Plant Biol. 2010;37:604.

    Article  Google Scholar 

  7. Kadereit G, Ackerly D, Pirie MD. A broader model for C4 photosynthesis evolution in plants inferred from the goosefoot family (Chenopodiaceae s.s.). Proc R Soc Lond B Biol Sci. 2012;279:3304–11.

    Article  Google Scholar 

  8. Bennett TH, Flowers TJ, Bromham L. Repeated evolution of salt-tolerance in grasses. Biol Lett. 2013;9:20130029-20130029.

    Article  Google Scholar 

  9. Flowers TJ. Physiology of halophytes. Plant Soil. 1985;89:41–56.

    Article  CAS  Google Scholar 

  10. Tester M, Davenport R. Na+ tolerance and Na+ transport in higher plants. Ann Bot. 2003;91:503–27.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  11. Shi S, Huang Y, Zeng K, Tan F, He H, Huang J, et al. Molecular phylogenetic analysis of mangroves: independent evolutionary origins of vivipary and salt secretion. Mol Phylogenet Evol. 2005;34:159–66.

    Article  CAS  PubMed  Google Scholar 

  12. Smith SA, Beaulieu JM, Stamatakis A, Donoghue MJ. Understanding angiosperm diversification using small and large phylogenetic trees. Am J Bot. 2011;98:404–14.

    Article  PubMed  Google Scholar 

  13. Waisel Y. The biology of halophytes. London: Academic; 1972.

    Google Scholar 

  14. Joseph S, Bhave M, Miller JT, Murphy DJ: Rapid identification of Acacia species with potential salt tolerance by using nuclear ribosomal DNA markers. Sustainable Agriculture Research 2013;2(4).

  15. Saslis-Lagoudakis CH, Hua X, Bui E, Moray C, Bromham L. Predicting species’ tolerance to salinity and alkalinity using distribution data and geochemical modelling: a case study using Australian grasses. Ann Bot. 2015;115:343–51.

    Article  PubMed  Google Scholar 

  16. Bui EN, Thornhill A, Miller JT. Salt- and alkaline-tolerance are linked in Acacia. Biol Lett. 2014;10:20140278-20140278.

    Article  Google Scholar 

  17. Crayn DM, Winter K, Smith JAC. Multiple origins of crassulacean acid metabolism and the epiphytic habit in the Neotropical family Bromeliaceae. Proc Natl Acad Sci U S A. 2004;101:3703–8.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  18. Anacker BL, Whittall JB, Goldberg EE, Harrison SP. Origins and consequences of serpentine endemism in the California flora. Evolution. 2010;65:365–76.

    Article  PubMed  Google Scholar 

  19. Sage RF, Christin PA, Edwards EJ. The C4 plant lineages of planet Earth. J Exp Bot. 2011;62:3155–69.

    Article  CAS  PubMed  Google Scholar 

  20. Cappa JJ, Pilon-Smits EAH. Evolutionary aspects of elemental hyperaccumulation. Planta. 2014;239:267–75.

    Article  CAS  PubMed  Google Scholar 

  21. Edwards EJ, Donoghue MJ. Is it easy to move and easy to evolve? Evolutionary accessibility and adaptation. J Exp Bot. 2013;64:4047–52.

    Article  CAS  PubMed  Google Scholar 

  22. Edwards EJ, Smith SA. Phylogenetic analyses reveal the shady history of C4 grasses. Proc Natl Acad Sci U S A. 2010;107:2532–7.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  23. Christin P-A, Osborne CP, Chatelet DS, Columbus JT, Besnard G, Hodkinson TR, et al. Anatomical enablers and the evolution of C4 photosynthesis in grasses. Proc Natl Acad Sci. 2013;110:1381–6.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  24. Bromham L, Bennett TH. Salt tolerance evolves more frequently in C4 grass lineages. J Evol Biol. 2014;27:653–9.

    Article  CAS  Google Scholar 

  25. Bui EN. Soil salinity: A neglected factor in plant ecology and biogeography. J Arid Environ. 2013;92(C):14–25.

    Article  Google Scholar 

  26. Antonovics J, Bradshaw AD, Turner RG. Heavy metal tolerance in plants. Adv Ecol Res. 1971;7:1–85.

    Article  Google Scholar 

  27. Richards CL, Walls RL, Bailey JP, Parameswaran R, George T, Pigliucci M. Plasticity in salt tolerance traits allows for invasion of novel habitat by Japanese knotweed s. l. (Fallopia japonica and F.xbohemica, Polygonaceae). Am J Bot. 2008;95:931–42.

    Article  PubMed  Google Scholar 

  28. Bose J, Rodrigo-Moreno A, Shabala S. ROS homeostasis in halophytes in the context of salinity stress tolerance. J Exp Bot. 2014;65:1241–57.

    Article  CAS  PubMed  Google Scholar 

  29. Egan TP, Ungar IA. Competition between Salicornia europaea and Atriplex prostrata (Chenopodiaceae) along an experimental salinity gradient. Wetl Ecol Manag. 2001;9:457–61.

    Article  Google Scholar 

  30. Greenwood ME, MacFarlane GR. Effects of salinity on competitive interactions between two Juncus species. Aquat Bot. 2009;90:23–9.

    Article  CAS  Google Scholar 

  31. Rozema J, Schat H. Salt tolerance of halophytes, research questions reviewed in the perspective of saline agriculture. Environ Exp Bot. 2013;92:83–95.

    Article  CAS  Google Scholar 

  32. Ungar IA. Are biotic factors significant in influencing the distribution of halophytes in saline habitats? Bot Rev. 1998;64:176–99.

    Article  Google Scholar 

  33. Ladenburger CG, Hild AL, Kazmer DJ, Munn LC. Soil salinity patterns in Tamarix invasions in the Bighorn Basin, Wyoming, USA. J Arid Environ. 2006;65:111–28.

    Article  Google Scholar 

  34. Zanne AE, Tank DC, Cornwell WK, Eastman JM, Smith SA, FitzJohn RG, et al. Three keys to the radiation of angiosperms into freezing environments. Nature. 2014;506:89–92.

    Article  CAS  PubMed  Google Scholar 

  35. Bromham L. Macroevolutionary patterns of salt tolerance in angiosperms. Ann Bot. 2015;115:333–41.

    Article  PubMed  Google Scholar 

  36. Rozema J, Flowers T. Crops for a salinized world. Science. 2008;322:1478–80.

    Article  CAS  PubMed  Google Scholar 

  37. Menzel U, Lieth H. Halophyte database version 2.0. In: Cash crop halophytes: recent studies. Dordrecht: Kluwer Academic Publishers; 2003. p. 221–50.

    Google Scholar 

  38. Khan MA, Qaiser M: Halophytes of Pakistan: characteristics, distribution and potential economic usages. Sabkha Ecosystems, Tasks for Vegetation Science 2006;42:129–153, Springer, Netherlands.

  39. Guvensen A, Gork G, Özturk M: An overview of the halophytes in Turkey. Sabkha Ecosystems, Tasks for Vegetation Science 2006;42:9–30, Springer, Netherlands.

  40. Dagar J, Gurbachan S: Biodiversity of saline and waterlogged environments : documentation, utilization and management. NBA Scientific Bulletin Number - 9, National Biodiversity Authority, Chennai, TamilNadu, India; 2007:78.

  41. Özturk M, Guvensen A, Sakçali S, Gork G. Halophyte plant diversity in the Irano-Turanian phytogeographical region of Turkey. In: Biosaline Agriculture and High Salinity Tolerance. Basel: Birkhäuser; 2008. p. 141–55.

    Chapter  Google Scholar 

  42. Zhao K, Song J, Feng G, Zhao M, Liu J. Species, types, distribution, and economic potential of halophytes in China. Plant Soil. 2011;342:495–509.

    Article  CAS  Google Scholar 

  43. Djanaguiraman M, Sheeba JA, Shanker AK, Devi DD, Bangarusamy U. Rice can acclimate to lethal level of salinity by pretreatment with sublethal level of salinity through osmotic adjustment. Plant Soil. 2006;284:363–73.

    Article  CAS  Google Scholar 

  44. Pandolfi C, Mancuso S, Shabala S. Physiology of acclimation to salinity stress in pea (Pisum sativum). Environ Exp Bot. 2012;84:44–51.

    Article  CAS  Google Scholar 

  45. The Plant List. 2010. Version 2 []

  46. Cayuela L, Granzow-de la Cerda I, Albuquerque FS, Golicher DJ. Taxonstand: An R package for species names standardisation in vegetation databases. Methods Ecol Evol. 2012;3:1078–83.

    Article  Google Scholar 

  47. Haston E, Richardson JE, Stevens PF, Chase MW, Harris DJ. The Linear Angiosperm Phylogeny Group (LAPG) III: a linear sequence of the families in APG III. Bot J Linn Soc. 2009;161:128–31.

    Article  Google Scholar 

  48. APG III. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Bot J Linn Soc. 2009;161:105–21.

    Article  Google Scholar 

  49. Stevens, PF. Angiosperm Phylogeny Website. 2001. Version 12, July 2012. [].

  50. Fellows I. Deducer: A data analysis GUI for R. J Stat Softw. 2012;49:1–15.

    Google Scholar 

  51. Lefeuvre P: BoSSA: a Bunch of Structure and Sequence Analysis. []

  52. Paradis E, Claude J, Strimmer K. APE: analyses of phylogenetics and evolution in R language. Bioinformatics. 2004;20:289.

    Article  CAS  PubMed  Google Scholar 

  53. Felsenstein J: Phylogenies and the comparative method. Am Nat 1985;125(1)1–15.

  54. Fritz SA, Purvis A. Selectivity in mammalian extinction risk and threat types: a new measure of phylogenetic signal strength in binary traits. Conserv Biol. 2010;24:1042–51.

    Article  PubMed  Google Scholar 

  55. Felsenstein J. Using the quantitative genetic threshold model for inferences between and within species. Philos Trans R Soc Lond B Biol Sci. 2005;360:1427–34.

    Article  PubMed Central  PubMed  Google Scholar 

  56. Kendall MG. A new measure of rank correlation. Biometrika. 1938;30:81–93.

    Article  Google Scholar 

  57. Moray C, Hua X, Bromham L: Data from: Salt tolerance is evolutionarily labile in a diverse set of angiosperm families. Dryad Digital Repository.

  58. FitzJohn RG. Diversitree: comparative phylogenetic analyses of diversification in R. Methods Ecol Evol. 2012;3:1084–92.

    Article  Google Scholar 

Download references


We would like to thank Thomas H. Bennett for assistance with data collection and project design, Robert Lanfear for methodological assistance and Timothy J. Flowers for his encouragement and helpful feedback.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Camile Moray.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

CM collected the data and performed the analyses. XH developed the methods. LB proposed the study, and LB and CM designed the study. All authors wrote, reviewed and approved the final manuscript.

Additional file

Additional file 1:

Supplemental material.

Rights and permissions

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Moray, C., Hua, X. & Bromham, L. Salt tolerance is evolutionarily labile in a diverse set of angiosperm families. BMC Evol Biol 15, 90 (2015).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: