Volume 101, Issue 2 p. 219-224
AJB Centennial Review
Free Access

The nature of serpentine endemism

Brian L. Anacker

Brian L. Anacker

Department of Evolution and Ecology, University of California-Davis, One Shields Avenue, Davis, California 95616 USA

E-mail: [email protected]

Search for more papers by this author
First published: 01 February 2014
Citations: 82

Freely available online through the AJB open access option.

The author thanks two anonymous reviewers for helpful comments and suggestions during the preparation of this manuscript.

Abstract

Serpentine soils are a model system for the study of plant adaptation, speciation, and species interactions. Serpentine soil is an edaphically stressful, low productivity soil type that hosts stunted vegetation and a spectacular level of plant endemism. One of the first papers on serpentine plant endemism was by Arthur Kruckeberg, titled “Intraspecific variability in the response of certain native plant species to serpentine soil.” Published in the American Journal of Botany in 1951, it has been cited over 100 times. Here, I review the context and content of the paper, as well as its impact. On the basis of the results of reciprocal transplant experiments in the greenhouse, Kruckeberg made three important conclusions on the nature of serpentine plant endemism: (1) Plants are locally adapted to serpentine soils, forming distinct soil ecotypes; (2) soil ecotypes are the first stage in the evolutionary progression toward serpentine endemism; and (3) serpentine endemics are restricted from more fertile nonserpentine soils by competition. Kruckeberg's paper inspired a substantial amount of research, especially in the three areas reviewed here: local adaptation and plant traits, speciation, and the interaction of climate and soil in plant endemism. In documenting soil ecotypes, Kruckeberg identified serpentine soils as a potent selective factor in plant evolution and helped establish serpentine soils as a model system in evolution and ecology.

Serpentine soils are a model system for the study of evolution, ecology, and conservation (Harrison and Rajakaruna, 2011). Serpentine soils are edaphically stressful for plant growth, due to nutrient deficiencies, especially Ca, low water-holding capacity, and high levels of heavy metals and Mg (Kruckeberg, 1984). They are derived from weathered ultramafic rocks that have been uplifted on the margins of former crustal plates (Alexander, 2006). Serpentine soils are typically recognized on the landscape as patchily distributed rock outcrops with stunted vegetation.

Serpentine soils host a spectacular level of plant endemism in many regions; for example, California has 215, Cuba 854, and New Caledonia 1150 taxa found growing only on serpentine (Anacker, 2011). The remarkable levels of plant endemism and the unique chemistry of serpentine soils have stimulated a substantial amount of research on plant adaptation, plant speciation, species invasion, interspecific interactions, disturbance, plant diversity, and climate change (see the following books, reviews, and special issues: Proctor and Woodell, 1975; Kruckeberg, 1984, 2002; Brooks, 1987; Baker et al., 1992; Roberts and Proctor, 1992; Jaffré et al., 1997; Balkwill, 2001; Boyd et al., 2004; Brady et al., 2005; Alexander, 2006; Chiarucci and Baker, 2007; Kazakou et al., 2008; Rajakaruna and Boyd, 2009, 2014; Rajakaruna et al., 2009; Harrison and Rajakaruna, 2011; Damschen et al., 2012).

Here, I describe one of the first and most important papers on serpentine plant endemism, “Intraspecific variability in the response of certain native plant species to serpentine soil,” published in the American Journal of Botany by Arthur Kruckeburg in 1951. This paper has been cited 119 times in ecology and evolution journals (Web of Science [WOS; Thomas Reuters, New York, New York, USA], accessed 26 November 2013). Two other works are often cited interchangeably: Kruckeberg's Ph.D. dissertation (Kruckeberg, 1950, not indexed by WOS) and a 1954 symposium paper in Ecology (Kruckeberg, 1954: 188 cites). During his career, Kruckeberg proceeded to address serpentine adaptations with 40 different plant species from California and Washington (Wright et al., 2006) and write several indispensable research papers, review papers, and books (Kruckeberg, 1967, 1984, 1991, 2002; Kruckeberg and Rabinowitz, 1985; Brady et al., 2005).

For his 1951 paper, Kruckeberg conducted greenhouse experiments to test for soil ecotypes, in the context of establishing a link between natural selection and speciation in four taxa: Achillea borealis Bong. (now Achillea millefolium L.), Gilia capitata Sims, and two subspecies of Streptanthus glandulosus Hook., Streptanthus glandulosus subsp. typicus Morris. and Streptanthus glandulosus subsp. secundus (Greene) Kruckeb. Each taxon has both nonserpentine and serpentine populations. Today, we refer to such taxa interchangeably as soil generalists, serpentine tolerators, or bodenvag. For each taxon, seeds collected from nonserpentine and serpentine soils were sown into nonserpentine and serpentine field soils in greenhouse trays and performance was measured. For Streptanthus, Kruckeberg only provided the results for plants grown on serpentine soils, somewhat confusingly; for the other two taxa, results are given for both soils. The results of the greenhouse experiments, summarized into three key points below and in Fig. 1, reveal a surprising amount about the nature of serpentine endemism. I will now review each point and discuss the status of related contemporary research.

Details are in the caption following the image

Performance of putative soil-specific ecotypes of two plant species when grown on nonserpentine and serpentine soils in the greenhouse. * Plant height data were not presented in the paper; I measured them on the photograph in fig. 5 of Kruckeberg (1951). The plant height data are presented here for illustrative purposes only.

(1) STRONG SELECTION BY SERPENTINE SOIL LEADS TO LOCAL ADAPTATION AND THE FORMATION OF SOIL ECOTYPES

Serpentine race plants grew larger than nonserpentine race plants when grown in serpentine soils (Fig. 1). This result established serpentine soils as a “potent selective factor” (Kruckeberg, 1951; p. 415) that drives differentiation and local adaptation on par with the selective strength of mine tailings (Jain and Bradshaw, 1966; Snaydon, 1970; Antonovics et al., 1971).

While reports of ecotypes are relatively common today, many midcentury ecologists viewed the natural world through a lens of continuous variation along environmental clines (see Lowry, 2012 and references therein). The demonstration of local adaptation across discrete soil boundaries likely made a strong impression on adherents to ecoclines. Clausen, Keck, and Hiesey (1948) would also largely reject the idea of clines, calling them “oversimplified abstractions not commensurate with natural entities” (Clausen, 1951: p. 28). These authors also focused on ecotypes, but at the much larger scale of climatic zones. Interestingly, both Clausen et al. and Kruckeberg used Achillea borealis in their experiments. Kruckeberg found soil ecotypes within a single climatic ecotype of A. borealis and said “edaphic races appear to be superimposed upon climatic races” (Kruckeberg, 1951: p. 415). This result was important because it suggested that ecotypes might form in response to more than one environmental factor, possibly operating at different scales. The interacting roles of climate and soils on endemism are discussed in more detail below.

Since Kruckeberg (1951), ecotypic differentiation to serpentine has been reported in several studies (Wright et al., 2006; O'Dell and Rajakaruna, 2011; Wright and Stanton, 2011). These findings suggest some generality to Kruckeberg's results, despite the fact that Kruckeburg's experiments were limited in a number of ways. For example, Kruckeburg used only a few sources for plant seeds and sampled soils from just one location. Another criticism is that Kruckeburg's transplant studies were conducted in a controlled greenhouse environment. However, Wright et al. (2006) suggested that the local adaptation that Kruckeburg measured in the greenhouse would only be stronger if measured in natural field conditions. In their study of Collinsia sparsiflora, ecotypic variation across soil boundaries was absent in the greenhouse, but strong in the field (Wright et al., 2006). Another advance since Kruckeberg (1951) has been to show that there is substantial variation of serpentine soils at both large and small scales; i.e., the edaphic properties of serpentine soils vary among and even within serpentine outcrops. For example, in the Lasthenia californica-L. gracilis complex, ecotypes have been documented within a single serpentine outcrop, reflecting small-scale variation in soil chemistry and soil structure (Rajakaruna and Bohm, 1999; Rajakaruna, 2003; Yost et al., 2012).

The search for plant traits associated with serpentine adaptation and endemism is very active. Whole-plant and leaf-level traits underlying serpentine tolerance include traits associated with resource-use efficiency and water retention, such as low specific leaf area, low growth rates, and perenniality (Chapin, 1980; Chapin et al., 1993; Anacker et al., 2011a; Anacker and Harrison, 2012b). At the physiological level, studies have investigated how serpentine tolerance is a function of the cellular regulation of Ca, Mg, and heavy metals (Brady et al., 2005). The subject of heavy metals, in particular, has received much attention, including studies on uptake, exclusion, and sequestration, extensive field reconnaissance for putative metal hyperaccumulators (van der Ent et al., 2013), and research on how hyperaccumulation may deter herbivores (Boyd and Martens, 1998; Proctor, 1999; Boyd and Jaffré, 2001; Davis et al., 2001; Boyd, 2012). Belowground studies have also revealed the selective power of serpentine soils. Serpentine plants tend to have more developed rooting systems on serpentine and may, based on a limited number of studies, form belowground mutualisms with a unique biota found in serpentine soils (Rune, 1953; Tadros, 1957; Schechter and Bruns, 2008, 2013; Porter et al., 2011). Little is known about the genetics that underlie serpentine tolerance, but tools to determine whether specific genes are associated with serpentine adaptations are now available (Brady et al., 2005; Turner et al., 2010).

(2) SOIL ECOTYPES MAY BE A STAGE IN THE PROGRESSION TOWARD SERPENTINE ENDEMISM

Plant endemics are conventionally considered to have one of two contrasting historical origins. In the first treatment of the subject, Cain classified endemics as being either “new” or “old.” New endemics are recently formed species confined to a region by geographic barriers; old endemics are formerly widespread species capable of spreading again (Cain, 1944). The principal factors determining the balance between new and old endemics are the age and isolation of an area. This definition would be later refined for the California flora by Stebbins and Major (1965), who classified species as neo- or paleoendemics. Neoendemics originate after a single founder event and have had inadequate time to spread; such taxa are also referred to as “insular endemics.” Paleoendemics form via a pathway of biotype depletion, where environmental changes may lead to altered competitive regimes and habitat-specific population extirpation. As a result, paleoendemics now occupy a small subset of their former range.

Kruckeberg's 1951 experiments were designed to apply the concept of paleoendemism to soil endemism. Kruckeberg viewed the formation of serpentine plant ecotypes as the first stage in the evolution of serpentine endemic species through the biotype depletion process. In stage one, local soil ecotype populations are formed by strong selection and local adaptation. In stage two, nonserpentine populations are extirpated, leaving only serpentine ecotypes intact.

Kruckeberg (1951; p. 418) suggested that the ecotypes of S. glandulosus that he identified in the greenhouse were precursors of endemic species and that we are observing the “closing stage in the elimination of all but the serpentine-adapted biotypes”. Kruckeberg provided no evidence to support this assertion, although later work in Streptanthus would largely support the claim (Mayer and Soltis, 1994; Mayer et al., 1994). Moreover, Kruckeberg does not say what would actually cause the loss of nonserpentine populations. We can speculate that an environmental change that negatively impacted the nonserpentine populations might cause their extirpation. For example, increased aridity might differentially impact the ecotypes, if serpentine plants can tolerate more drought than nonserpentine plants (Hughes et al., 2001; Wu et al., 2010). Alternatively, decreased aridity may increase competition levels on nonserpentine soils and cause nonserpentine populations to decline (Rune, 1953; Moore et al., 2013).

While Kruckeberg was mostly concerned with the idea that endemics might arise via a biotype depletion scenario, it is often assumed that the majority of Californian endemics have arisen via a neoendemic pathway (Harrison, 2013). Several studies to date have shown that endemics appear to arise from nearby progenitor taxa (Kay et al., 2011). For example, in Layia, a serpentine endemic was derived from a sympatric progenitor nontolerator relative (Baldwin, 2005; see also Leptosiphon in Kay et al., 2011). These examples are remarkable because they do not involve large-scale geographic isolation, which is typically thought to be required for the completion of reproductive isolation (Kay et al., 2011). Instead, strong directional selection by serpentine soils appears to be able to overcome the homogenizing effects of gene flow, demonstrating that speciation can be relatively rapid and local (Clausen, 1951; Lowry, 2012). Shifts in reproductive attributes, such as flowering time, flower morphology, and mating system, can also serve as important prezygotic barriers that limit gene flow between progenitor and nascent derivative species (Kay et al., 2011; O'Dell and Rajakaruna, 2011; Anacker and Strauss, 2014).

Geographic isolation does, however, likely play a role in speciation of serpentine endemics. In Streptanthus, studies of genetic diversity of Streptanthus glandulous and its derivative endemics offer support for speciation via biotype depletion, but also for the subsequent formation of insular taxa (Kruckeberg, 1957; Mayer and Soltis, 1994, 1999; Mayer et al., 1994). In Lasthenia, there is support for the geographic isolation of serpentine outcrops playing a role in the allopatric divergence of two species (Kay et al., 2011; Yost et al., 2012).

It is clear that serpentine tolerance has evolved independently many times, both among and within species, based on the number and taxonomic distribution of endemics and tolerators as well as molecular phylogenetic studies (Kruckeberg, 1967; Kruckeberg and Rabinowitz, 1985; Rajakaruna et al., 2003; Berglund et al., 2004; Rajakaruna, 2004; Rajakaruna and Whitton, 2004; Safford et al., 2005; Anacker, 2011; Anacker et al., 2011; Kay et al., 2011). For example, the 215 endemic species of California come from 39 different families (Cuba's 854 endemics come from 24 families, New Caledonia's 1150 endemics from 64). Preliminary estimates suggest that the number of serpentine tolerant species in the California flora is over 1000 (Safford et al., 2005). Each of these 1000+ bodenvag taxa offers a chance to repeat studies like those of Kruckeberg (1951), thereby revealing the contingencies that are likely involved in the evolution of serpentine adaptation and ensuing reproductive isolation.

Interestingly, despite the fact that the strong selection and spatial isolation of serpentine appears to promote adaptation and speciation, serpentine endemism is not associated with adaptive radiations along insular island outcrops, unlike oceanic islands. It is possible that the edaphic stress and/or absence of empty niche space limits subsequent lineage radiation (Anacker et al., 2011b). Moreover, the observation that tolerators outnumber endemics four to one indicates constraints on the transition from serpentine tolerance to serpentine endemism. For example, serpentine soils might represent sink habitats that cannot support their own populations without dispersal from more productive habitats. A second possibility is that gene flow across soil boundaries could prevent selection from driving complete reproductive isolation. Even if reproductive isolation is complete, nascent lineages are likely to have small, isolated populations that have relatively high levels of inbreeding, increased sensitivity to stochastic events, and increased extinction risk.

In the future, we must expand our studies outside of California to include the other serpentine exposures in the world, each of which offers an independent look at adaptation and speciation on serpentine under different climate conditions and biogeographic histories. In addition, comparison of the serpentine floras with endemic floras on other unusual bedrock and soil types, such as granite rock outcrops in the southeastern United States, can help establish the strengths and weaknesses of serpentine as a model system for ecology and evolution (Shure, 1999). Unfortunately, the data required to make such a meaningful comparison is scarce. Finally, we should expand our studies to understudied groups, such as cryptograms, where ultramafic endemism appears to be uncommon, possibly because selection by physical factors is stronger than selection by chemical factors (Rajakaruna et al., 2012 and references therein).

(3) SERPENTINE ENDEMICS ARE RESTRICTED FROM MORE FERTILE SOILS BY COMPETITION

The experimental results showed that serpentine plants do better on nonserpentine soils than on serpentine soils (Fig. 1). Why then are serpentine-adapted plants not found on nonserpentine soils? One explanation is that competition in matrix habitats plays an important role in the restriction of serpentine-adapted plants to serpentine soils. If true, serpentine soils may be viewed as a refuge from competition, rather than a source of unique resources. In Kruckeberg (1951; p. 418) words, “Serpentine endemics are not restricted to serpentine merely because of some specific requirement uniquely provided by serpentine.”

Many of Kruckeberg's contemporaries agreed that competition may restrict taxa to serpentine (Rune, 1954; Stebbins and Major, 1965), and this notion is generally accepted today. However, the evidence that competition is lower on serpentine soils and that serpentine endemics are competitively inferior is not substantial. A review on neighbor removal studies found several examples where competition was lower on serpentine soils (Moore and Elmendorf, 2011), but the experimental evidence is mixed (describe in more detail below; Fernandez-Going and Harrison, 2013).

Underpinning this discussion is the idea that serpentine tolerance comes at a cost, though this idea requires further study. For example, there is little evidence for a cost of tolerance of low Ca to Mg ratios or heavy metals (Brady et al., 2005; Kay et al., 2011). Species-level traits, however, indicate that endemics are typically slow-growing stress tolerators, rather than fast-growing competitive dominants (Anacker and Harrison, 2012b; Fernandez-Going et al., 2012). The stress tolerance traits of endemics are consistent with a trade-off of competitive ability for serpentine tolerance traits, by which drought adaptations may come at the cost of fast growth rate (Grime et al., 2008).

Kruckeberg and others speculated that competition plays a key role in endemism by restricting such slow-growing, stress-tolerator plants from high-fertility nonserpentine soils (Kruckeberg, 1951; Stebbins and Major, 1965). Kruckeberg also acknowledged that climate plays a key role in plant adaptation and ecogeographic isolation. However, what is little recognized is that climatic and edaphic factors might interact to control endemism (Anacker and Harrison, 2012b; Fernandez-Going et al., 2013).

Across a rainfall gradient, increasing precipitation will have the direct effect of increasing competition in fertile, nonserpentine soils and the indirect effect of restricting bodenvag species to serpentine. Several observational studies are consistent with such an indirect effect of climate on serpentine. Serpentine endemism peaks in wet regions, both regionally and globally (Anacker, 2011; Fernandez-Going et al., 2013). In addition, some soil generalists become specialists in the more climatically favorable parts of their ranges (“regional endemics”: Rune, 1953; Kruckeberg, 1984; Safford et al., 2005). At the community level, there is more species, functional, and phylogenetic turnover across soil boundaries in the mesic northwest of California than the arid south (Anacker and Harrison, 2012b; Fernandez-Going et al., 2013).

The evolution of serpentine endemism also bears the strong imprint of climate. Endemics occupy wetter regions than even their closest relatives (Anacker and Harrison, 2012a), but only in the case of endemics that have arisen via a neoendemic pathway. For neoendemics, nascent lineages that colonize serpentine in regions with benign and stable climate are more likely to persist and achieve reproductive isolation than nascent lineages in regions with harsh and fluctuating climates. Paleoendemics, on the other hand, do not occupy different climates than their closest relatives, suggesting that range contractions were equally common in wetter and drier regions.

Experiments are required to mechanistically determine how climate influences serpentine endemism. Such tests should simultaneously address whether competition is lower on serpentine, whether serpentine endemics are relatively poor competitors, whether climate can alter competitive environments and the soil-specific performance of endemics, and whether paleo- and neoendemics differ in their sensitivities to competition. The only experimental study to date on this topic provides mixed evidence for the interaction of water availability and soils on serpentine restriction (Fernandez-Going and Harrison, 2013). The survival of three serpentine endemic plant species was unaffected by competition removal or water addition, but endemic biomass was lowest in uncleared, nonserpentine plots. These results are consistent with competition restricting endemics from relatively fertile, nonserpentine soils. Studies with more endemic plant species and factorial manipulations of soil fertility and water availability are still required to make conclusions about the interacting effects of climate and soils on serpentine plant endemism.

If mesic climates do promote endemism, then what will be the fate of serpentine endemics under a drier future? In California, many climate models predict a drier future with rainfall falling in fewer, bigger storms (Pierce et al., 2013). Under future climate scenarios, species distribution models predict that the new climate range for a set of serpentine endemic species will include a large amount of habitat, but these areas are far from known occurrences, requiring dispersal events on the order of tens of kilometers (Damschen et al., 2012). However, several lines of evidence suggest that the stress-tolerant traits of endemics may prevent increased mortality under drying conditions, preventing the need for them to colonize new outcrops (Grime et al., 2000, 2008; Damschen et al., 2012; Fernandez-Going et al., 2012). Interestingly, a drier future could lead to the loss of endemics (i.e., they expand onto nonserpentine), if competition in matrix habitats decreases. For example, Arenaria norvegica and Minuartia laricifolia subsp. ophiolitica both expanded out of serpentine soils with the retreat of forest from matrix habitats during climatic cooling in glacial periods (Rune, 1953; Moore et al., 2013).

CONCLUSIONS

Kruckeberg's 1951 paper was among the first to show how serpentine soils are important selective factors in plant evolution. The three key ideas described here stimulated much additional research on serpentine plant adaptation, speciation, and plant restriction. Moreover, studies of serpentine endemism have revealed much about the nature of plant endemism in general. In particular, serpentine research has highlighted the role of geology as a major environmental determinant of endemism through direct effects on topography and soil properties and indirect effects on habitat availability, degree of spatial isolation, and microclimate (Kruckeberg and Rabinowitz, 1985). A second insight is that plant endemics may have contrasting origins, arising from widespread progenitor taxa or existing as relicts of formerly widespread, generalist species.

Kruckeberg would later describe an equation relating plant diversity to five processes: climate, organism, topography, rock type, and time (Kruckeberg, 1986). We are indebted to Kruckeberg for establishing serpentine as an excellent model system for disentangling these often-interacting components.