Ultramafic geoecology of South and Southeast Asia

Globally, ultramafic outcrops are renowned for hosting floras with high levels of endemism, including plants with specialised adaptations such as nickel or manganese hyperaccumulation. Soils derived from ultramafic regoliths are generally nutrient-deficient, have major cation imbalances, and have concomitant high concentrations of potentially phytotoxic trace elements, especially nickel. The South and Southeast Asian region has the largest surface occurrences of ultramafic regoliths in the world, but the geoecology of these outcrops is still poorly studied despite severe conservation threats. Due to the paucity of systematic plant collections in many areas and the lack of georeferenced herbarium records and databased information, it is not possible to determine the distribution of species, levels of endemism, and the species most threatened. However, site-specific studies provide insights to the ultramafic geoecology of several locations in South and Southeast Asia. The geoecology of tropical ultramafic regions differs substantially from those in temperate regions in that the vegetation at lower elevations is generally tall forest with relatively low levels of endemism. On ultramafic mountaintops, where the combined forces of edaphic and climatic factors intersect, obligate ultramafic species and hyperendemics often occur. Forest clearing, agricultural development, mining, and climate change-related stressors have contributed to rapid and unprecedented loss of ultramafic-associated habitats in the region. The geoecology of the large ultramafic outcrops of Indonesia’s Sulawesi, Obi and Halmahera, and many other smaller outcrops in South and Southeast Asia, remains largely unexplored, and should be prioritised for study and conservation.

Depauperate ultramafic soils may generate selective pressures promoting speciation and the evolution of ultramafic endemism (Anacker 2014;Kay et al. 2011;Rajakaruna 2004), often leading to distinctive plant communities worldwide (Anacker 2011;Brooks 1987). The biota of ultramafic soils has contributed greatly to the development of ecological and evolutionary theory (Harrison and Rajakaruna 2011;Strauss and Cacho 2013) and to the study of the genetics of adaptation and speciation (Brady et al. 2005;Palm and Van Volkenburgh 2014;von Wettberg and Wright 2011). Ultramafic floras are, however, threatened by deforestation, agricultural Open Access *Correspondence: nrajakaruna@gmail.com 5 Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407, USA Full list of author information is available at the end of the article development, mining, and climate change-associated stressors Harrison et al. 2009;Rajakaruna and Boyd 2008;Vallano et al. 2012). These threats to ultramafic biota provide opportunities for conservation and restoration-oriented research (Elam et al. 1998;O'Dell and Claassen 2011;Weiss 1999;Whiting et al. 2004;Wolf 2001).
South and Southeast Asia contain several globally significant biodiversity hotspots (Mittermeier et al. 2005), including areas in Indo-Burma, Philippines, Sundaland (western half of the Indo-Malayan archipelago), and Western Ghats and Sri Lanka. The Borneo lowlands is the only ecoregion globally to surpass 10,000 plant species (Kier et al. 2005) and North Borneo is one of the top five biodiversity centres in the world (Barthlott et al. 2007). Despite South and Southeast Asia harboring several important biodiversity hotspots, the influence of edaphic factors on biodiversity is largely unknown (van der Ent et al. 2015a). Compared to research on ultramafic outcrops in temperate and Mediterranean regions (Alexander et al. 2007;Rajakaruna et al. 2009), ultramafic geoecology in this part of the world is also substantially understudied (Proctor 1992(Proctor , 2003. In terms of tropical regions, most research related to ultramafic floras to date has focussed on New Caledonia (Isnard et al. 2016;Jaffré et al. 2010Jaffré et al. , 2013Pillon et al. 2010;Pillon 2012). Although ultramafic outcrops of New Caledonia are of a similar latitude and general climate to South and Southeast Asia, the evolutionary histories of its flora and fauna are distinct. New Caledonia is on the east of the Lydekker's Line, which separates the eastern edge of Wallacea from the Australian Region (which lies on the Sahul Shelf ), marking a distinct change in floristic affinities. In this review, we also exclude New Guinea (Indonesian West Papua and Papua New Guinea) for the same reason, but note that despite the concomitant occurrence of ultramafic outcrops and exceptionally high biodiversity, virtually nothing is known about the ultramafic geoecology of this island. Research on the floristics and ecology of the understudied ultramafics of South and Southeast Asia is critical to provide a comprehensive assessment of the ultramafic geoecology of tropical Asia.
This review examines the literature on the geoecology of ultramafic areas in South and Southeast Asia, covering India, Pakistan, and Sri Lanka to the west, Myanmar and Cambodia to the north, and Malaysia, Indonesia (excluding West Papua), and the Philippines to the east ( Fig. 1; Table 1); all of which lie on the western side of Lydekker's line and share a similar climate. We focus on (i) soil-plant relations, including studies on floristic diversity, soilplant elemental relations, and soil microbes; (ii) ecological aspects, including studies on vegetation structure and composition and plant endemism; (iii) cross-kingdom interactions, including studies on herbivory, mycorrhizal associations, and invertebrate diversity; (iv) evolutionary aspects; (v) physiology and genetics; (vi) phytotechnologies; and finally, (vii) threats and conservation. We conclude the review by highlighting countries within South and Southeast Asia requiring further study, drawing attention to major gaps in knowledge.

Soil-plant relations
Ultramafic soils worldwide share a distinct suite of chemical and physical features ); however, tropical ultramafic soils may differ in elemental content, moisture, organic matter content, and soil pedology (Kierczak et al. 2007;Vithanage et al. 2014), compared to those in temperate and Mediterranean regions (Alexander 2009;Alexander et al. 2007). Table 2 lists key soil properties of ultramafic soils from South and Southeast Asia, focusing on pH, Ca:Mg molar ratio, Ni, Cr, and the major nutrients, P and K. Plants growing on ultramafic soils have to contend with a suite of edaphic stressors, including low nutrient content, high levels of phytotoxic elements, and, at times, water stress (Brady et al. 2005). Plants and soil microbes of ultramafic soils tolerate these edaphic stressors via efficient uptake of essential nutrients, and exclusion of, or conversely accumulation and localization of high concentrations, of certain phytotoxic elements, among other adaptations (see Palm and Van Volkenburgh 2014 for a discussion).

Plant diversity and soil-plant elemental profiles
In Sukinda, India, chromite mine spoils composed of ultramafic substrates have Ni ranging from 187 to 215 µg g −1 and Ca:Mg molar ratios of 1.69-2.27; from which, in total, 113 plant species belonging to 51 families have been recorded (Samantaray et al. 2001). Some species which colonize the substrate exhibit traits typical of plants adapted to ultramafic soils, including sclerophyllous and microphyllous leaves (Brady et al. 2005), but individual plants also show chlorosis, leaf curling, and necrosis.
On the Andaman Islands, India, ultramafic soils with high Ni concentrations (2700-10,100 μg g −1 ) harbor eight Ni hyperaccumulator plant species belonging to eight different genera and seven different families . Of these, Dichapetalum gelonioides subsp. andamanicum (Dichapetalaceae) and Rinorea bengalensis (Violaceae) accumulated up to 30,000 μg g −1 Ni. There is substantial potential for using remote sensing tools to examine the vegetation communities on the ultramafics of the Andaman Islands, where the ultramafic outcrops are mostly inaccessible and the vegetation deserves more intensive exploration .
Research suggests that Sri Lanka's ultramafic flora is impoverished with respect to the total number of plant species and percent proportion of endemic species. To date, 67 plant species belonging to 61 genera and 30 families have been identified from Ussangoda (Samithri 2015). Combined with an additional 40 taxa reported from three other sites surveyed by Rajakaruna and Bohm (2002), the total ultramafic flora of Sri Lanka stands at a mere 107 species, compared to many-fold more documented from other sites in Southeast Asia (van der Ent et al. 2015a). Of the species documented from ultramafic soils, only Vernonia zeylanica (Asteraceae) is endemic to Sri Lanka (MOE 2012), although the taxon is not restricted to the substrate.

Soil microbes
Several recent studies, conducted in temperate and Mediterranean regions of the world, explore the roles microbes play in the ecology of ultramafic habitats as well as in the remediation of metal-contaminated soils (Batten et al. 2006;Ma et al. 2015;Schechter and Branco 2014). Although studies on microbial ecology of ultramafic soils in South and Southeast Asia are minimal, Pal et al. (2004Pal et al. ( , 2005Pal et al. ( , 2006Pal et al. ( , 2007 and Pal and Paul (2012) have carried out a series of studies on microbial diversity and ecology of ultramafic soils on the Andaman Islands, India. In one of these studies, Pal et al. (2005) compared physicochemical and microbial properties of ultramafic soils with those from adjacent non-ultramafic localities. The elemental profiles were characteristic of ultramafic soils, with high concentrations of Mg, Ni, Cr, and Co. Furthermore, the ultramafic soils showed low microbial density (6.2-11.3 × 10 6 colony forming unit/g soil) and activity (1.7-3.5 µg fluorescein/g dry soil/h) relative to non-ultramafic soils. The ultramafic-associated microbial population (including bacteria and fungi) was dominated by bacteria and was more resistant to metals than populations from non-ultramafic soils. Among the ultramafic isolates, 8 and 11 bacteria tolerated >12.0 mM Ni and >16.0 mM Cr, respectively, while six fungal isolates showed a minimum inhibitory concentration (MIC) value >8.0 mM Co. The ultramafic strains also showed co-resistance to Cu, Zn, and Mn. Pal et al. (2007) also examined the soil microflora associated with the rhizosphere of two known Ni hyperaccumulators from the Andaman Islands, R. bengalensis and D. gelonioides subsp. andamanicum. Of the total 123 microbes (99 bacteria and 24 fungi) that were isolated, bacteria were more tolerant of Ni than fungi, showing their greater potential for Ni tolerance.
In a study focusing on medicinal qualities of wild-harvested plants, 32 plant species collected from ultramafic outcrops of Sri Lanka were screened for antimicrobial properties . Of these, 29 species belonging to 12 families proved effective against at least one microorganism. Photoactivity was also observed from extracts of 10 species belonging to 10 families. There was no observed correlation between trace element hyperaccumulation (Rajakaruna and Bohm 2002) and antimicrobial activity.

Ecological aspects
Ultramafic outcrops have long-provided model settings for studies on the ecology of plant species and plant communities. Studies range from those investigating aspects of the ecology of edaphically specialized plant populations and plant-plant interactions to those exploring factors and mechanisms driving the assembly of plant communities (see Harrison and Rajakaruna 2011). Compared to other regions of the world, ecological studies on ultramafics of South and Southeast Asia are mostly limited to those examining floristics, plant community structure, and edaphic-floristic associations.

Vegetation structure and composition
Mount Silam in Sabah, Malaysia, has been extensively studied, including the general floristics, forest structure, hydrology and chemical analysis of tree foliage and leaf litter (Proctor et al. 1988a(Proctor et al. , b, 1989Bruijnzeel et al. 1993). The study plots on Mount Silam range from 280 to 870 masl in elevation, documenting a broad spectrum of vegetation changes with altitude. The site is extremely species-rich in terms of its tree flora, ranging between 19 species in a 0.04-ha plot at 870 masl to 104 species in a 0.4-ha plot at 480 masl (Proctor 1992). Ultramafic-associated rainforests on Mount Guiting-Guiting, Sibuyan Island, Philippines (Proctor et al. 1998) and those of Mount Silam, Sabah (Proctor et al. 1988a, b) are similar in their soil features (Ni, Ca:Mg, and depth) and lack of stunted lowland forests. At these locations, small-statured forests are associated with higher elevations.
On Mount Bloomfield in the western Philippines (Palawan), Proctor et al. (1999) described a very different forest structure from those of Mount Silam and Mount Guiting-Guiting. The soil depths on Mount Bloomfield are much less compared to these other sites; Bruijnzeel (1990) suggested that drought in the shallow soils is a major cause of forest stunting on ultramafics, perhaps in association with fire . Mount Bloomfield lacks tall forests and instead is characterised by trees less than 18 m tall. No statistical relationship could be established between tree height and soil chemistry, although Proctor et al. (1999) did find a direct proportional relationship between maximum tree height and soil water retention. The authors indirectly linked soil water to fire susceptibility in establishing the particular vegetation pattern on Mount Bloomfield, one that superficially resembles fire-dependent vegetation of New Caledonia. Proctor et al. (2000a, b) compared vegetation on ultramafic soils to those on non-ultramafic (greywackederived) soils in Palawan and found that the species richness and diversity of ultramafic and greywacke sites were similar. However, the individual species and familial composition were rather different, with only members of the Saxifragaceae occurring on both ultramafic and greywacke soils. Trees on the serpentinized peridotite had a high proportion of microphyllous leaves, which is not a general feature of ultramafic forests in the region. Differences in water supply and fire frequencies, in combination with edaphic difference, may contribute to the distinct forests overlying these soils (Proctor et al. 1999(Proctor et al. , 2000a. Sulawesi and Halmahera in Indonesia have 15,400 and 8000 km 2 of ultramafic outcrops, respectively (van der Ent et al. 2013a). Lateritic soils overlaying the bedrock harbor both sclerophyllous ultramafic vegetation and more cryptic tropical rainforest, which are nonetheless inhabited by a high proportion of endemic flora. Proctor et al. (1994) examined the ultramafic soil-plant relations of Mount Piapi on Karakelong part of the Talaud Islands, North Sulawesi, Indonesia and reported that the short stature of the local vegetation is a result of low water-holding capacity of the soil, while the unusual species assemblage likely results from the soil chemistry typical of ultramafic soils. They also documented an undescribed Ni-hyperaccumulating species of Rinorea from their study site.
Kinabalu Park, Sabah, one of the world's most species-rich hotspots with more than 5000 plant species recorded in an area of just 1200 km 2 , is also home to extensive ultramafic exposures (van der Ent et al. 2014). Plant diversity on ultramafics of the Park decreases with elevation, with a mid-elevation (circum 1500 masl) 'hump' occurring for some plant groups (Orchidaceae, Pteridophytes) resulting from the presence of cloud forests (van der Ent et al. 2016a). Six main vegetation classes with associated soil types are described by van der Ent et al. (2016a), including Sub-Alpine Scrub and Graminoid Scrub, both associated with Hypermagnesic Cambisols ('hypermagnesian soils'), Montane Cloud Forest, associated with Cambisols often with accumulation of humus, Mixed Dipterocarp Forest, associated with deep Ferralsols ('laterites'), and Pioneer Casuarina Scrub and Mature Mixed Casuarina Forest, both associated with Hypermagnesic Leptosols. The 'adverse' soil chemistry exacerbates vegetation stunting but no clear correlation between elevation, soil chemistry and plant diversity was found, as some of the most 'adverse' soils on the summit of the entirely ultramafic Mount Tambuyukon (2359-2534 masl) had up to 132 species per 250 m 2 (van der Ent et al. 2016a).
Samithri (2015) examined the vegetation community composition and patterns at Ussangoda, Sri Lanka's most intensively studied ultramafic outcrop. She found a higher diversity of plant species in 'forest islands' compared to the extensive 'plains' characterizing the site (Fig. 2c). Although the plains make up over 90% of the outcrop area, they only harbor 18 herbaceous species belonging to 17 genera and 11 families compared to 49 tree, shrub, herb and climber species belonging to 44 genera and 27 families found in the 'forest islands. ' Although the soil chemical features did not differ significantly between sites on the 'plains' versus those in the 'forest islands, ' soil features such as depth and resulting water holding capacity in 'forest islands' may favor the growth of a wide range of species than on the exposed and shallow soils of the 'plains. ' Studies on bryophytes, lichens, and epiphytes on ultramafic outcrops are sparse worldwide (but see Briscoe et al. 2009;Favero-Longo et al. 2004;Rajakaruna et al. 2012). In South and Southeast Asia, such studies are mostly non-existent. However, one study from the Philippines (Proctor et al. 2000b) documents epiphytic plants on trees of ultramafic and adjacent greywacke soils. The trees on the greywacke had fewer lianas and much less bole bryophyte cover than those on the serpentinized peridotite. Forty-one percent of trees on peridotite had >10% bryophyte cover, while none of the trees on greywacke soils had >10% bryophyte cover. The greywacke soils also hosted significantly higher densities of ferns, Cyperaceae spp., rattans (Arecaceae: Calamoideae), and Pandanaceae spp. compared to ultramafic soils, while ultramafic soils harbored significantly more herbaceous and bamboo (Poaceae: Bambusoideae) species. Floristic differences between the sites were attributed to differences in geochemistry, hydrology, and fire-frequencies (Proctor et al. 1999(Proctor et al. , 2000b.

Plant endemism
Ultramafic soils, often with disproportionately high numbers of endemic species (Anacker 2011), are prime settings to explore the nature of edaphic endemism (Rajakaruna 2004). In New Caledonia, 2150 species occur on ultramafic soils of which 83% are restricted to these soils (Jaffré 1992;Jaffré and L'Huillier 2010), whereas in Cuba, 920 species (approximately one-third of the taxa endemic to Cuba) are found exclusively on ultramafic soils (Borhidi 1992). Similar restrictions and notable floristic associations are also found on ultramafic outcrops of Mediterranean climates (including California; Alexander et al. 2007;Safford et al. 2005), as well as in South Africa/Zimbabwe and Australia (Anacker 2011;Brooks 1987).
The restriction of habitat specialists to ultramafic soils is generally considered a consequence of their inherent slow growth rates that leads them to being outcompeted on more favorable soils (Anacker 2014;Anacker et al. 2011;Kay et al. 2011). Although some growth experiments have shown that habitat specialists can grow faster on more nutrient-rich soils (Kruckeberg 1954), species from the ultramafic maquis in New Caledonia have inherently slow growth, albeit becoming larger under more fertile conditions (Jaffré 1980). Table 3 lists the countries within the South and Southeast Asian region with ultramafic floras, including the number of ultramafic-associated species documented and the number of ultramafic endemics described in each country.
In Sabah, Malaysia, Borneodendron aenigmaticum (Euphorbiaceae) is one of the few large rainforest trees restricted to ultramafic soils (Proctor et al. 1988a). Wood (2012, 2013) describing orchid species associated with ultramafics in Sabah, Malaysia, documented many endemic species (Orchidaceae) restricted to narrow valleys with steep slopes, dominated by Gymnostoma sumatranum (Casuarinaceae) and Ceuthostoma terminale (Casuarinaceae). Further, van der Ent et al. (2015b) show habitat partitioning among ultramafic endemic Nepenthes species (Nepenthaceae) of Mount Kinabalu and Mount Tambuyukon, with distinct habitats and elevation ranges for the different Nepenthes taxa. Eriobotrya balgooyi (Rosaceae) was described as a new species restricted to ultramafic soils on a hill near the eastern ridge of Mount Kinabalu and on the nearby Mount Tambuyukon in Sabah, Malaysia (Wong and van der Ent 2014). The importance of scientific exploration of the ultramafics of Southeast Asia cannot be stressed enough; a survey on the ultramafic Mount Guiting-Guiting, Philippines (Argent et al. 2007) also led to the discovery of a new species, Lobelia proctorii (Campanulaceae).
Sri Lanka's ultramafic outcrops and their flora, compared with ultramafic floras of Southeast Asia and Australia-Pacific region (van der Ent et al. 2015c, d), have received relatively little attention partly because they do not harbor any endemic species nor many metal hyperaccumulators (Chathuranga et al. 2015). All species so far documented from the ultramafic outcrops of Sri Lanka also have non-ultramafic populations, and it is unclear whether the ultramafic populations are physiologically distinct (i.e. ecotypes).

Cross-kingdom interactions
Edaphically stressful substrates, like ultramafic soils, present plants with challenges that differ from more 'benign' substrates. Growing under such stress, ultramafic plants will likely encounter other organisms (herbivores, pathogens, beneficial insects and pathogens) that are also able to tolerate some of the same stressors affecting the plants (Strauss and Boyd 2011). There is evidence to suggest that pressures from enemies will be greater on edaphically stressful substrates than on normal soils (Strauss and Cacho 2013). Additionally, the enriched concentrations of certain trace elements, such as nickel, found in ultramafic soils may provide plants with opportunities for elemental defence (Boyd 2014). A significant body of research exists on plant-other biota interactions on ultramafic soils from temperate and Mediterranean climes, including studies on elemental defence ), defence against pathogens (Hörger et al. 2013;Springer 2009), herbivory (Lau et al. 2008), mycorrhizal associations (Southworth et al. 2014), plant-pollinator interactions (Meindl et al. 2013;Wolf and Thorp 2011), and seed dispersal (Spasojevic et al. 2014). However, such studies are minimal in tropical Asia.

Herbivory
In the only known published study on herbivory in ultramafic ecosystems in the region, Proctor et al. (2000a) found that the percentage of leaf area consumed was similar for plants found on and off of ultramafic soils on Mount Bloomfield, Palawan (Philippines), although the actual leaf area consumed was greater for the ultramafic forest as it had plants with larger leaves. There was no relationship between herbivory and leaf elemental chemistry; even the metal-accumulating taxa were attacked by herbivores. Proctor et al. (2000a) speculate that the gall-forming and leaf-mining insects must be tolerant of nickel as they spend their entire juvenile stage in the leaf tissue.
Recent work by van der Ent and Mulligan (2015) show Ni accumulation in various parts of Ni hyperaccumulator plants occurring in Sabah, Malaysia, with the highest Ni concentration recorded in the phloem tissue (up to 7.9% in R. bengalensis) and phloem sap (up to 16.9% in Phyllanthus balgooyi); Ni localization in phloem tissue is visible by the bright green coloration in field-collected samples (Fig. 3b, f ). The discovery of toxic levels of Ni in the phloem tissue suggests that the increased Ni in the phloem provides a defence against phloem-sap feeding insects, pathogens, and other herbivores (Boyd 2014;Hanson et al. 2004). However, Geometric moth larvae (Erebidae: Erebinae:Poaphilini) were found feeding on the leaves of the Ni hyperaccumulator P. balgooyi, furthermore aphids were found feeding on Phyllanthus cf. securinegioides (van der Ent et al. 2015f ).

Mycorrhizal associations
Pisolithus tinctorius (Sclerodermataceae), an ectomycorrhizal fungus, is found in the rhizosphere of Eucalyptus urophylla (Myrtaceae) from ultramafic soils in the Philippines, New Caledonia, and Western Australia (Aggangan et al. 1998). Pisolithus tinctorius was cultured with E. urophylla to determine the effects of Cr and Ni on the fungal growth rate. The fungus concentrates metals in the extramatrical hyphae and extra-hyphal slime and is particularly tolerant of high concentrations of Ni and Cr. There was geographic variation in terms of metal tolerance in the fungus, with the New Caledonian isolate outperforming both the Australian and the Philippines isolates. The Philippines isolate grew well on agar in the presence of Cr up to 2000 µmol L −1 and Ni up to 200 µmol L −1 , but formed fewer mycorrhizae in vitro and in vivo than its counterparts from New Caledonia and Western Australia.

Soil invertebrates
A study comparing termite assemblages on ultramaficderived forest soils to those on non-ultramafic soils in Borneo, Malaysia shows that ultramafic sites have low species density (<35%), low relative abundance (<30%), a virtual absence of soil-feeders, significantly fewer wood-feeders, and a near-absence of species of Rhinotermitidae, Amitermes-group, Termes-group, Pericapritermes-group and Oriensubulitermes-group (Jones et al. 2010). The authors suggest that metal toxicity, high pH disrupting gut physiology, metal poisoning of essential microbiota in the termite gut, and metal bioaccumulation by gut microbes with subsequent poisoning of the termite host, as possible reasons for the depauperate termite communities on ultramafic soils.
A study on the patterns of Oribatid mite communities in relation to elevation and geology on the slopes of Mount Kinabalu, Sabah, Malaysia, shows that the density and morphospecies richness of Oribatid mites are greater in non-ultramafic soils than in the ultramafic soils at each of the same elevations (Hasegawa et al. 2006). The density and richness of Oribatid mites decreased with elevation on both substrates, but the effects of elevation on their density in non-ultramafic soil were less significant than in the ultramafic substrate.
An investigation of the invertebrate communities in forest litter and soil on Mount Guiting-Guiting in the Philippines, shows that ultramafic soils, even at higher elevations, were not poor in soil invertebrates, including Oligochaeta (Thomas and Proctor 1997), similar to earlier findings on Mount Silam, Sabah (Leakey and Proctor 1987).

Physiology and genetics
There is considerable interest in understanding the physiology and the underlying genetic basis for traits conferring adaptation to ultramafic soils (Bratteler et al. 2006;Palm and Van Volkenburgh 2014;von Wettberg and Wright 2011;Wu et al. 2008). The advent of novel molecular methods has provided unique approaches to exploring stress tolerance (Selby et al. 2014;Visioli and Marmiroli 2013) and ultramafic-associated plants will continue to provide model systems for such investigations (Arnold et al. 2016;von Wettberg et al. 2014). While these advances have not yet been made in tropical Asia, the region provides numerous opportunities for investigating the physiological and genetic aspects of adaptation to ultramafic soils. To date, much of the research in South and Southeast Asia has focused on discovering new hyperaccumulating plant species from ultramafic soils in the region.

Trace element hyperaccumulation
Plants found on ultramafic soils have long-been recognized as model systems to explore trace element hyperaccumulation (Gall and Rajakaruna 2013). There are well over 450 Ni hyperaccumulator plant species globally, all occurring on ultramafic soils (van der Ent et al. 2013c). Ultramafic associated plants are known to hyperaccumulate cobalt (Co) and Cu (>300 μg g −1 in their dry leaf tissue), and Ni (>1000 μg g −1 in their dry leaf tissue). For recent reviews of trace element hyperaccumulation, see Reeves (2003), Krämer (2010), van der Ent et al. (2013c and Pollard et al. (2014). Table 4 lists documented hyperaccumulator plants from the South and Southeast Asia region, listing the element hyperaccumulated, country of discovery, and relevant references. Figure 3 documents some of the nickel hyperaccumulator plants discovered from ultramafic soils in parts of South and Southeast Asia.
In one of the earliest geoecological studies of the region, Wither and Brooks (1977) and Brooks et al. (1977b) analysed herbarium samples of plants originating from Obi Island (North Moluccas). They identified Myristica laurifolia var. bifurcata (Myristicaceae), Planchonella oxyhedra (Sapotaceae), and Trichospermum kjellbergii (Malvaceae) as hyperaccumulators of Ni. The authors then analysed Ni concentrations in herbarium specimens of T. kjellbergii and P. oxyhedra from  In a study describing the general influence of the ultramafic geochemistry on growth patterns of plants overlying two Malaysian massifs, the Bukit Rokan and Petasih along the Bentong-Raub suture zone on the Peninsula, Tashakor et al. (2013) document that the serpentinite of the area is strongly weathered and gives rise to characteristic red lateritic soils (Ferralsols). They point out that the greatest physiological stress experienced by plants growing on ultramafic soils is due to the low Ca: Mg ratio and the generally low available nutrients, and not due to potentially phytotoxic elements present in the soil, which are, for the most part, not in a plant-available form.

Region
In a study of the Bela Ophiolite in the Wadh area of Balochistan, Pakistan, Naseem et al. (2009) discovered Pteropyrum olivieri (Polygonaceae) in a localized population over ultramafic soils. Although the plant did not hyperaccumulate, it had moderate concentrations of Ni, Co, and Cr in its tissues, typical of most plants growing on ultramafic soils.
The ultramafics of Malaysia and Indonesia have received considerable attention with regard to taxa with high metal-accumulating behavior. A chemical analysis of leaf litter from trees growing on ultramafics in Sabah, Malaysia (Proctor et al. 1989) confirmed that trees grow at low foliar nutrient concentrations and can concentrate Ca in their leaf tissue. Leaf litter showed an average Ca:Mg ratio as well as a high level of Ni, suggesting that senescence may act as a way of excreting excess Ni. From analysis of leaf litter, they found that Shorea tenuiramulosa (Dipterocarpaceae) and an unidentified species of Syzygium (Myrtaceae) accumulated Ni and Mn, respectively, with 1000 µg g −1 Ni and 13,700 µg g −1 Mn dry leaf weight. Proctor et al. (1994) also reported a yet to be named Ni-hyperaccumulating species of Rinorea from Mt Piapi on Karakelong Island, northeast of Sulawesi in Indonesia with up to 1830 µg g −1 foliar Ni.
In an effort to understand the factors contributing to Ni hyperaccumulation in Sabah, Malaysia, van der Ent et al. (2016b) examined the soil chemistry associated with 18 Ni hyperaccumulator plant species, comparing the chemistry of ultramafic soils where Ni hyperaccumulators were absent. The results showed that Ni hyperaccumulators are restricted to circum-neutral soils with relatively high phytoavailable Ca, Mg, and Ni. They hypothesized that either hyperaccumulators excrete large amounts of root exudates, thereby increasing Ni phytoavailability through intense rhizosphere mineral weathering, or that they have extremely high Ni uptake efficiency, thereby severely depleting Ni and stimulating re-supply of Ni via diffusion from labile Ni pools. Their results, however, tend to favor the latter hypothesis.
Nuclear microprobe imaging (micro-PIXE) shows that in P. balgooyi collected from ultramafic soils in Sabah, Malaysia, Ni concentrations were very high in the phloem of the stems and petioles, while in the leaves Ni was enriched in the major vascular bundles (Mesjasz-Przybylowicz et al. 2015). The preferential accumulation of Ni in the vascular tracts suggests that Ni is present in a metabolically active form. This research is important as the elemental distribution of P. balgooyi differs from that of many other Ni hyperaccumulators from temperate and Mediterranean regions where Ni is preferentially accumulated in leaf epidermal cells (Bhatia et al. 2004;Broadhurst et al. 2004;Tylko et al. 2007;Baklanov 2011).

Evolutionary aspects
Ultramafic outcrops often harbor populations which are morphologically and physiologically distinct from those found on non-ultramafic soils. Such intraspecific variation, especially with respect to functionally important traits, is common in many ultramafic taxa worldwide (O'Dell and Rajakaruna 2011). Such variation can result from both local adaptation (i.e., ecotypic differentiation; Sambatti and Rice 2006;Turner et al. 2010) or phenotypic plasticity (Murren et al. 2006;Wu et al. 2010), and must be examined on a case-by-case basis. Suitable methods of examination include reciprocal or unilateral transplant experiments and common garden studies (Wright and Stanton 2011), as well as functional genomic and proteomic approaches (Selby et al. 2014;von Wettberg et al. 2014;von Wettberg and Wright 2011). Detecting intraspecific variation is the first step toward any investigation on the causes and consequences of adaptive evolution. Populations exhibiting intraspecific variation on ultramafic and non-ultramafic soils have led to numerous studies of speciation (Anacker 2014;Kay et al. 2011) and phylogenetic investigations (Anacker 2011;Anacker et al. 2011;Anacker and Harrison 2012), advancing our understanding of evolutionary and ecological theory (Harrison and Rajakaruna 2011). Molecular phylogenetic methods provide a unique protocol for testing and establishing species relationships, helping to shed light on how ultramafic endemics evolve (Baldwin 2005). The analysis of phylogenies for 23 genera from California shows that ultramafic endemics exhibit few transitions out of the endemic state (Anacker et al. 2011), suggesting that adaptation to ultramafics and subsequent diversification can lead to an evolutionary "dead end". But ultramafic lineages may not always represent evolutionary "dead ends" and may have the potential to further diversify via independent polyploidization and hybridization, even providing a pathway to radiate off ultramafic soils (Kolář et al. 2012).
Compared to these studies from other regions of the world, there is little information on evolutionary aspects of plants associated with ultramafic soils in South and Southeast Asia. A recent study from Sri Lanka shows that the ultramafic and non-ultramafic populations of Fimbristylis ovata (Cyperaceae) may be locally adapted to their respective soils (Chathuranga et al. 2015). The ultramafic population translocated significantly more Ni from its roots to shoots (translocation factor 0.43) than the non-ultramafic population (translocation factor 0.29). However, additional studies are required to determine whether the populations of F. ovata, or other species, including those hyperaccumulating metals such as Ni and Cu, deserve ecotypic recognition. Several ultramafic-associated taxa in Sri Lanka might benefit from  further observations and additional greenhouse studies to determine whether the ultramafic-associated populations are genetically distinct and are worthy of ecotypic recognition (Rajakaruna and Bohm 2002). These taxa include several Ni-accumulating and -hyperaccumulating species, particularly Hybanthus enneaspermus (Violaceae), Evolvulus alsinoides (Convolvulaceae), Crotalaria sp. (Fabaceae), Desmodium triflorum (Fabaceae) and  Proctor et al. (1994) Fimbristylis sp. (Cyperaceae), all of which show detectable phenotypic differences between ultramafic and nonultramafic populations. Studies exploring causes and consequences of phenotypic differences between populations found on and off ultramafic soils can add much to our understanding of the origins of ultramafic specialists in the South and Southeast Asia region.

Phytotechnologies
The use of trace element hyperaccumulators to clean up polluted sites, i.e. phytoremediation, is gaining recognition as a viable green technology (Neilson and Rajakaruna 2014). Phytoremediation is based on the premise that plants which remove selected pollutants from the soil and translocate them to their above-ground biomass can then be harvested and disposed of through incineration or elemental recovery, a process known as phytomining (Chaney et al. 2014;van der Ent et al. 2015g). Ultramafic plants in the genera Alyssum (Brassicaceae), Streptanthus (Brassicaceae), Noccaea (Brassicaceae), and Berkheya (Asteraceae) have been used in phytoremediation and phytomining of Ni-enriched ultramafic sites in temperate and Mediterranean regions (Ho et al. 2013;Morel et al. 2006;Gall and Rajakaruna 2013;Sheoran et al. 2009;van der Ent et al. 2015g). Given the large number of hyperaccumulator species currently known from tropical Asia (Gall and Rajakaruna 2013;Reeves 2003), there should be considerable interest in using these unique plants in the remediation of regional sites contaminated with metal and metalloid pollutants. Bandara et al. (2017) investigated the effect of biochar and fungal-bacterial co-inoculation on soil enzymatic activity and immobilization of heavy metals in soil collected from an ultramafic outcrop in Sri Lanka. The addition of biochar to ultramafic soil immobilized heavy metals and decreased soil enzymatic activities while the addition of microbial inoculants improved plant growth by mitigating heavy metal toxicity and enhancing soil enzymatic activities. Additional studies from Sri Lanka confirm the importance of (i) bacterial-fungal inoculation as a soil-quality enhancer and a plant-growth promoter in the presence of heavy metals found in ultramafic soils (Seneviratne et al. 2016a, b), and, (ii) biochar as a soil amendment to immobilize Cr, Ni, and Mn in ultramafic soil, thereby reducing metal-induced plant toxicities (Herath et al. 2014).

Phytoremediation and phytomining
The potential for microbial remediation (reduction) of Cr(VI) by indigenous microbial populations from the ultramafic soils of Sukinda mines in Jaipur, Orissa, India, was investigated by Mishra et al. (2009). The best reducer of Cr (V1) was Staphylococcus aureus, a gram-positive bacterium whose thick layer of peptidoglycan acts as a strong absorbent. The taxon tolerated a Cr concentration of 250 mg L −1 and was resistant to Ni up to 1000 mg L −1 . The bacterium was recommended for the bioremediation of both Cr and Ni, showing complete Cr(VI) to Cr(III) degradation in 22 h, and Ni 2+ degradation to 90% in 22 h. Similarly, Bohidar et al. (2009) explored the possibility of Ni recovery from chromite tailings at the Sukinda mines by using three fungal strains.
In another study, Mohanty et al. (2011) utilized phytoremediation in South Kaliapani, a chromite mining ultramafic area in Orissa, India. Chromium was extracted by growing Oryza sativa cv. Khandagiri (rice; Poaceae) in contaminated soil and irrigating with mine wastewater. Chromium levels were reduced (70-90%) after 100 days, with accumulation levels ranging from 125 to 498 µg g −1 in leaves, 25 to 400 µg g −1 in stems, and 5 to 23 µg g −1 in the grain. Absorption into roots was higher by two orders of magnitude than into any aerial part of the plant. Mohanty et al. (2012) also investigated the phytoremediation potential of O. sativa, Brachiaria mutica (Poaceae), and Eichhornia crassipes (Pontederiaceae) to reduce levels of Cr(VI) in mine waste-water. Eichhornia crassipes was most successful with 25-54% reduction while B. mutica contributed to an 18-33% reduction. Kfayatullah et al. (2001), in a study of plants and soils of the Malakand chromite-rich ultramafic area and Mardan non-ultramafic areas of the North-West Frontier Province, Pakistan, focused on enzyme-bound metal accumulation in plant tissue. Verbascum thapsus (Scrophulariaceae), an edible plant, accumulated greater than 100 µg g −1 of several metals, including Ni and Cr, but was not recommended for phytoremediation efforts.
Indonesia (Sulawesi and Halmahera Islands) has some of the largest surface exposures of ultramafic bedrock in the world. Lateritic Ni-mining operations have continued in the region since the early twentieth century, setting the stage for exploring the use of native plants for phytoremediation and phytomining. Twelve native species known to hyperaccumulate Ni are recommended by van der Ent et al. (2013a) for use in phytotechnologies in Indonesia.

Threats and conservation
Ultramafic areas are a high priority for biodiversity conservation because of the relatively large numbers of endemic species, ecotypes, and rare species that they harbour ). The conservation and restoration of these naturally fragmented, edaphically unique, and biodiverse habitats require special attention (Baker et al. 2010;O'Dell 2014;Thorne et al. 2011;Whiting et al. 2004). It is unclear how stressors, such as atmospheric N deposition (Vallano et al. 2012), suppression of fire (Arabas 2000;Safford and Harrison 2004) and climate change (Damschen et al. 2012;Anacker and Harrison 2012) documented for temperate and Mediterranean ultramafics, impact tropical Asia's ultramafic ecosystems.
The combined forces of forest clearing, agricultural development and mining contribute to unprecedented habitat loss in South and Southeast Asia (Duckworth et al. 2012;Hughes 2017;Sodhi et al. 2004). In fact, Southeast Asia has a higher annual rate of deforestation than Meso-America, South America, or sub-Saharan Africa, and that rate has continued to increase between 1990 and 2005 (Giam et al. 2010;Sodhi et al. 2010). This is especially of concern as Southeast Asia has a higher proportion of its vascular plant, reptile, bird, and mammal species categorised as globally threatened on the Red List compared to Meso-and South America and sub-Saharan Africa (Sodhi et al. 2010). With such limited study of ultramafics in South and Southeast Asia, it is unclear how increasing habitat loss is impacting biodiverse ultramafic outcrops in the region.
Malaysia has one of the most species-rich ultramafic floras in the world. The over 3500 km 2 of ultramafic outcrops in Sabah (4.6% of the total landmass of the state) on the island of Borneo harbor a total of 4252 plant species (van der Ent et al. 2015a). Over 2542 plant species have been documented on ultramafic outcrops in Kinabalu Park alone, of which a large percentage is endemic to either Kinabalu Park or to Borneo (van der Ent et al. 2015a; Fig. 4). Despite the existence of this species-rich flora, the plant diversity and ecology of many ultramafic outcrops in Sabah remain largely unknown because of a lack of focused research. Furthermore, plant diversity in many areas of Sabah is severely threatened by land-use conversion and, because often plant species occur only at a single or a few ultramafic sites, and hence impacts on the ecosystems that support them could eventually result in their extinction. While it is necessary to identify stressors impacting ultramafic habitats of South and Southeast Asia for their proper management, it is even more critical that basic geoecological surveys of ultramafic outcrops, including the extensive exposures in Sulawesi and Halmahera, are prioritised for cataloguing plant diversity and other biota. This is especially critical as many of these outcrops likely harbor rare and endemic species in need of urgent conservation attention.
Although Sri Lanka's ultramafic flora appears to be impoverished with respect to endemic species or hyperaccumulator taxa, the ultramafic sites harbor several taxa worthy of conservation. For example, Ussangoda, the site that has received the most research attention, is home to: four near-threatened species, Striga angustifolia (Orobanchaceae), Maerua arenaria (Capparaceae), Salvadora percia (Salvadoraceae), and Olax imbricata (Olacaceae); two vulnerable species, Cyanotis adscendens (Commelinaceae), Pachygone ovata (Menispermaceae); and one data deficient species, Alysicarpus monilifer (Fabaceae; MOE 2012). Therefore, it is critical that Sri Lanka's ultramafic outcrops receive regional and national recognition and are declared as ecologically sensitive sites (i.e. geoecological preserves) to be set aside for future investigations. In 2010, Ussangoda was declared as a National Park with approximately 350 hectares, including areas overlaying ultramafic rock, set aside for conservation purposes (Department of Wildlife Conservation 2015). Without such conservation, proper management, and research, these unique habitats and their physiologically distinct biota are extremely vulnerable. Rinorea bengalensis (Violaceae) offers an example of why such efforts are urgently needed. Brooks et al. (1977a, b) conducted a survey of herbarium specimens from the entire range of this species, encompassing Sri Lanka, the Malay Archipelago, New Guinea, the Solomon Islands and Queensland, Australia, and found that Ni hyperaccumulation is a constitutive trait in this species when growing on ultramafic soil. The herbarium specimen analysed from Sri Lanka contained 10,000 µg g −1 and the locality indicated on the map presented by Brooks et al. (1977a) suggests a collection in the central part of the island (see Fig. 1 in Rajakaruna and Baker 2004). However, it was not encountered in field exploration by Rajakaruna and Bohm (2002) and was presumed extinct in Sri Lanka (Ministry of Environment and Renewable Energy 2012). Interestingly, the taxon was recently recollected in southwestern Sri Lanka (Siril Wijesundara, National Institute of Fundamental Studies, Sri Lanka, pers. comm.), however, soil and plant tissue elemental concentrations have yet to be determined.

Information gaps and future directions
Ultramafic outcrops are natural laboratories for experimental and applied research in a wide range of disciplines. They provide numerous opportunities for collaborations among geologists, pedologists, botanists, zoologists, microbiologists, and land managers focusing on conservation and restoration research. However, research on the ultramafic outcrops in South and Southeast Asia has been limited, with most effort to date focused on Malaysia, the Philippines, the Andaman Islands (India), and Sri Lanka (Table 1). We were unable to find any published literature on ultramafic geoecology of other South (Afghanistan, Bhutan, Nepal) and Southeast Asian (Myanmar, Laos, Thailand, Vietnam) countries despite the known occurrences of ultramafic lithologies in these locales. The limited number of published studies we found for Myanmar, Thailand, and Vietnam (Table 1) focused on geological, mineralogical, or geochemical research.
Throughout South and Southeast Asia, detailed and systematic surveys will likely reveal numerous species new to science, including trace element hyperaccumulators. Recent research conducted in Sabah, Malaysia by van der Ent et al. ( , 2015a which led to the discovery of 24 new hyperaccumulator species, is a case in point. Detailed floristic surveys should be undertaken across the region and species showing unusual physiological behavior (such as trace element accumulation) or exhibiting distinct morphological traits relative to populations on non-ultramafic soils may be further studied under laboratory and greenhouse conditions. Additionally, species showing intraspecific variation between ultramafic and non-ultramafic populations may be evaluated via population genetic studies to determine whether ultramafic populations are genetically distinct from those found on non-ultramafic soils. For those species showing intraspecific variation with respect to morphological or physiological features, including flowering times between ultramafic and nonultramafic populations, common garden and reciprocal transplant experiments can be undertaken to examine whether populations are locally adapted to their substrate. Such types of experimental studies are currently lacking entirely from the region.
In addition to detailed studies of vascular plants, it is important to pay close attention to non-vascular plants such as bryophytes, cryptogamic species such as lichens, soil algae and cyanoprokaryotes, and belowground microbes and soil invertebrates. Such investigations will likely reveal species that are endemic to the substrate or show a high affinity to ultramafic soils, as shown for such research conducted in South Africa (Venter et al. 2015) and California, USA .
Species documented as trace element hyperaccumulators may be investigated under controlled conditions for their suitability for phytoremediation or phytomining and tested under field conditions for their effectiveness in site reclamation and restoration. The resulting information can be added to the global database of metal hyperaccumulating species (Global Hyperaccumulator Database 2016: http://www.hyperaccumulators.org). Finally, it is critical that tropical Asia's ultramafic outcrops receive regional, national, and global recognition and that key sites receive appropriate statutory protection so that future scientific research is possible.
One of the options for protection at a national level by the state is the inclusion of ultramafic sites in the Global Geopark Network (GGN). Conservation and protection of landscapes of geological significance at a national and international level is promoted by UNESCO under its Global Geoparks Scheme (UNESCO 2016). At a national level, relevant authorities should pursue this option as a long-term conservation strategy, which would provide a holistic approach to protection by incorporating a management strategy including education and sustainable development. The latter would mobilize the local population for economic benefits by participating in the conservation efforts through local and international ecotourism. This, however, also requires meeting the stringent guidelines laid out by UNESCO to be included in the GGN. Currently, ultramafic sites in South and Southeast Asia are not in the GGN but would meet the basic requirements laid out by UNESCO.