COMMENTARY Multiple origins of serpentine-soil endemism explained by preexisting tolerance of open habitats W. Scott Armbrustera,b,c,1 be restricted to greenstone (an ultramafic aSchool of Biological Sciences, University of Portsmouth, Portsmouth PO1 2DY, United mineral) outcrops (12). Later observations Kingdom; bInstitute of Arctic Biology, University of Alaska, Fairbanks, AK 99775; and c (13, 14) and experiments (14) showed that Department of Biology, Norwegian University of Science and Technology, N-7491 all putative greenstone endemics were found Trondheim, Norway also on a variety of other parent materials. The taxa are indeed rare and disjunct but endemic to south-facing slopes so steep Plant specialization on soils derived from be expected from the diversity of plant line- and dry that they are largely open habitats unusual parent materials is an important ages in which serpentine tolerance and ende- with sparse vegetation. Thus, plants adap- contributor to regional biodiversity. These mism have evolved, each with different ted to the stress of dry, open habitats can stressful substrates include serpentine, gab- genetic and physiological starting points. Re- have similar ecological behavior as edaphic bro, and other ultramafic rocks rich in heavy gardless of the physiological causes, there are endemics. This could lead one to ask if ad- metals. The effect of substrate on plant di- well-defined serpentine floras and vegetation aptation to open habitats might sometimes versity is illustrated by serpentine soils in types recognized in North America, including be the first step toward tolerance of ultra- California: they comprise less than 1% of the Cuba; Europe, including Britain; South mafic soils. surface of the state (1), but serpentine endem- Africa; and Oceania, among other regions. CachoandStrauss(10)addressthisand ics (species restricted to serpentine soils) The floras are characterized by very high en- related questions by looking at the evolution make up about 10% of the flora (2). How demism, and the vegetation is usually sparser of tolerance to open habitats and to various “ ” such edaphic endemics (plants restricted and often more xeromorphic (appearing elements associated with serpentine soils in to stressful soils) evolve is a long-standing drought adapted) than nearby vegetation on relation to soil shifts across the phylogeny of “ ” A B question that remains largely unresolved. normal soils (Figs. 1 and ). a group of wild mustards (Streptanthus and For example, plant tolerance of serpentine Theopennessofthehabitatmayitselfbe relatives; “streptanthoid mustards”)inwhich soils may often involve tradeoffs in competi- a source of stress contributing to endemism. serpentine endemism has evolved four to five tive ability, and restriction to serpentine soil Several authors have found that the balance times (15). They use a powerful, but rarely may reflect poor competitive ability on less between competition (for light, water, and/or realized, approach in evolutionary ecological stressful soils rather than obligate associa- nutrients) and facilitation between neighbor- research: that of integrating comprehensive – tion (3 5), although possible counter-exam- ing plants in a landscape depends on the level sets of measurements and/or experiments ples exist (6). A common feature of plant of stress: in a benign landscape competition into a molecular phylogeny (16) to gain communities on stressful soils is the wide dominates, but in a stressful environment, insights into the evolutionary history of spacing of plants and openness of the hab- facilitation dominates (8, 9). As the habitat soil specialization. itat. Openness may itself be stressful for a gets more open, especially in hot, dry climates Intriguingly, Cacho and Strauss (10) find – variety of reasons (detailed below) (7 9). (e.g., summer in regions with Mediterranean evidence that tolerance to open habitats In PNAS, Cacho and Strauss (10) use a climates), the paucity of neighbors means appears to have evolved before tolerance of novel comparative experimental approach greater exposure to the drying effects of wind serpentine soils (Fig. 1C) and hence was a to explicitly assess the role of openness vs. and high direct-radiation loads, as well as the preaptation (preadaptation). Cecchi et al. soil chemistry as factors in the evolution of stress of high leaf temperatures and potential (17) also concluded recently that preaptations plant tolerance of, and endemism to, ser- UV damage. Recent research has also shown (e.g., for drought tolerance) have been critical pentine soils. that plants in sparsely vegetated sites are also in the origins of serpentine tolerance, but had Plants that tolerate and are endemic to more apparent to herbivores (e.g., more sus- no direct evidence to support their hypothe- stressful ultramafic soils have become a well- ceptible to attack by herbivorous insects), and sis. Cacho and Strauss (10) also provide evi- developed system for the study of adapta- such plants may have had to evolve expensive dence that tolerance of stressful soil elements tion, speciation, endemism, competition, and counter adaptations (e.g., increased chemical evolved in streptanthoid mustards coincident community ecology (2, 6, 11). Classic studies defenses and/or more cryptic foliage) (7). with,orafter,switchestoserpentinesoils, have shown the probable role of elemental A series of studies from a remote corner of rather than before switches to serpentine, as imbalances (e.g., high concentrations of northeast Alaska illustrates a possible biolog- expected. In addition, in a novel, common heavy metals, including Ni, Co, Cu, Cr, ical connection between edaphic endemism Pb; low Ca:Mg ratios; and low concentra- and open habitats. Studies in the region have tions of macronutrients), although there is noted the narrow distribution of possible Author contributions: W.S.A. wrote the paper. no consensus on the main causes of stress relict plant species, disjunct from nearest The author declares no conflict of interest. and species exclusion (2, 6). There may in- relatives by hundreds of kilometers (12–14). See companion article on page 15132. deed be a diversity of causal factors, as might Some of these endemics appeared, at first, to 1Email: [email protected]. 14968–14969 | PNAS | October 21, 2014 | vol. 111 | no. 42 www.pnas.org/cgi/doi/10.1073/pnas.1417242111 Downloaded by guest on September 24, 2021 COMMENTARY Fig. 1. Evolution of serpentine-soil endemism from occupation of normal soils is apparently contingent on prior adaptations to occurring in open habitats. (A) Dense vegetation char- acteristic of normal soils in Napa County, CA. Pink flowers are Collinsia sparsiflora Fisch. and C.A. Mey, a species with normal and serpentine ecotypes. (B) Open, sparse vegetation on a serpentine outcrop in nearby Lake County, CA. Pink flowers in foreground are Clarkia gracilis subsp. tracyi (Jeps.) Abdel-Hameed and R. Snow, a subspecies largely restricted to serpentine soils. (C) Maximum-credibility tree of the streptanthoid mustards, showing inferred evolution of open-habitat adaptations (maximum-likelihood ancestor-state reconstruction; shading on branches, with darker indicating occurrence in more open habitats; reprinted with permission from ref. 10. Arrows indicate inferred origins of serpentine endemism (>87% of records from serpentine soils; red) and serpentine tolerance (>11%, <87% of records from serpentine soils; violet), evolving from occupation of normal soils (branches to left of the violet or red arrows; data from ref. 10). Soil-type optimization on branches used ordered parsimony, where serpentine tolerance is assumed to be a precursor of serpentine endemism. (To simplify the diagram, inferred coincident origins of tolerance and endemism are not shown.) environment competition experiment, the answer in both cases may be that critical by chance of one or more preaptations (19– authors show that plants from more open preexisting features that enable an ecological 21). Once these preaptations are in place, it habitats have lower competitive abilities. shift are already in place. A recent study by becomes much more likely that multiple Importantly, this relationship is similar Christin et al. (18) found evidence for multi- independent parallel shifts (parallelisms) will whether the plants are from bare habitats in ple parallel origins of C4 photosynthesis in occur, hence explaining the commonness of serpentine-soil landscapes or from bare hab- grasses and that these shifts were contingent parallel evolution as is often observed in itats in normal soil landscapes. on appropriate preexisting leaf anatomy adaptive traits (15, 18, 21). Additional detailed The multiple origins of serpentine toler- (exaptation). The parallels to the present phylogenetic comparative studies, such as this ance in the streptanthoid mustards raise a study (10) are striking. Thus, there seems to one (10), may show evolution by exaptation long-standing question in evolutionary be growing support for the idea that major (preaptations acquiring new functions) to be biology: how do new ecological adaptations adaptive transitions in evolution often, or even the dominant theme wherever parallel evolu- originate and why do parallelisms occur? The usually, occur through the prior establishment tion is observed in groups of related species. 1 Kruckeberg AR (1984) California Serpentines: Flora, Vegetation, 8 Bertness MD, Callaway R (1994) Positive interactions in 15 Cacho NI, Burrell AM, Pepper