This is an image that says University of Nevada, Las Vegas


M. p. var. glaberrima near HAVO

Metrosideros polymorpha var. glaberrima near Hawaii’s Volcanoes National Park, Hawai’i Island.

My research program centers on the problem of how differential adaptation of tropical woody species across heterogeneous environments leads to phenotypic divergence and the accumulation of reproductive isolating barriers (i.e., speciation).  Through numerous collaborations, my work also includes studies of phylogeny, physiology, population genomics, DNA barcoding, and the description of new plant species.  Though the core theme of my research is evolution, the results produced have clear applications to the conservation of biodiversity.  This multi-disciplinary program involves studies in the field, greenhouse, environmental chamber, and microscopy and molecular labs as well as a broad range of modern analytical approaches.  My work is increasingly focused on the integration of genomic analyses with the growing data set on phenotypes, ecology, and reproductive isolating barriers in my primary study system, Hawaiian Metrosideros.  This group is a rare example of an incipient adaptive radiation in trees with relatively tractable life histories and thus provides an exciting opportunity to understand speciation in trees.

Divergence and Speciation in Trees – There are an estimated 100,000 species of trees, yet we know little about how new tree species arise.  Speciation in trees is an especially intriguing problem because their long life spans, often large, continuous populations, and propensity for outcrossing and long-distance gene flow should counteract divergent selection and the evolution of reproductive isolating barriers.  Moreover, the long life spans, delayed maturity, and large body size of trees preclude the traditional experimental approaches of evolutionary biology for most species.

Hawaiian Metrosideros (Myrtaceae) is a hyper-variable woody species complex that dominates the islands’ native forests, occurs in massive, continuous populations that span a broad range of climates, and appears to capture multiple stages of the speciation process.  The group comprises numerous mostly conspecific, vegetatively distinct forms that are non-randomly distributed across Hawaii’s famously heterogeneous landscape.  Metrosideros possesses several characteristics that make it unusually amenable to studies of speciation: a) adult trees are short in many areas where conditions are limiting, allowing access to the canopy for hand-pollinations; b) in a greenhouse, individuals can begin annual flowering at 3-4 years of age, depending on taxon; c) air-layering of branches permits the collection of large numbers of independent flowering “trees” in a greenhouse for controlled crossing studies; and d) seeds can be dried for long-term storage, allowing the staggering of experiments with seeds and seedlings.  Studies in my lab have: (1) identified 25 predominantly single-island-endemic taxa or morphotypes of Metrosideros across the island chain, (2) established the group as a rare case of an incipient adaptive radiation in trees that allows examination of divergence before speciation is complete, (3) revealed that the strength of isolation of taxa is decoupled from time, (4) suggested that persistent, sharp ecotones may be required for speciation in trees, and (5) suggested that speciation may come at the cost of increased inbreeding depression within recently isolated taxa.

Population genetic divergence: Analyses of over 1,400 microsatellite (SSR) genotypes from 24 purported taxa across the Hawaiian Islands and Pacific support the group as a rare case of incipient radiation in trees (Stacy et al. 2014; Stacy & Sakishima, in review).  SSR results support the progression rule with Pacific Island populations connecting with the older islands of Oahu and Kauai, a significant pattern of isolation by distance across the island chain, and island-based clustering of populations with a few exceptions that suggest multiple colonizations and back-colonizations of individual islands.  Results also reveal a peak number of genetic clusters on islands of intermediate age, and increased genetic isolation and inbreeding of some taxa with island age.  The notable exceptions were the two most broadly distributed taxa, which were weakly but significantly differentiated only on volcanically active Hawaii Island where they partition early- and late-successional environments.  This latter result suggests that persistent disruptive selection in a stable environment is required for speciation in this group (i.e., the temporal and spatial scales of disruptive selection by lava flows are insufficient for ecological speciation of successional forms).  On a broader scale, the long time required for the evolution of reproductive isolating barriers in trees relative to the rate of geologic change of volcanic islands may limit speciation in this group.   Lastly, results suggest that the generalist, wet-forest variety of M. polymorpha may act as a flexible stem in the diversification of Metrosideros in Hawaii.

Intensive genotyping on two islands of contrasting age and taxonomic richness, Hawaii Island (DeBoer & Stacy 2013, Stacy et al. 2014) and Oahu (Stacy et al., in review), reveals a broad range of genetic distances among their 4 and 12 taxa, respectively, and with a single exception, no pattern of isolation by distance within taxa within either island.  Results from Hawaii Island reveal the colonization and diversification history of the island, including the probable recent origin of the island-endemic riparian variety (newellii) from a bog form on the oldest volcano (Kohala).  On Oahu, which hosts among the most extreme phenotypes within the entire Metrosideros genus (four species, including nine varieties or morphotypes of M. polymorpha), results demonstrate the heritability of phenotypes of at least eight taxa and strongest isolation of taxa associated with highest elevations or otherwise extreme environments.  These findings suggest a principal role for selection in the origin and maintenance of the exceptional diversity that occurs within continuous Metrosideros stands on Oahu.  Currently, we are carrying out: a) demographic analyses of > 3.5 million SNPs from full genome sequences (~312 Mb) of several Hawaiian Metrosideros taxa to better illuminate the evolutionary histories of these taxa, and b) population genomic analyses of M. p. var. newelli and its progenitor, M. p. var. glaberrima to examine how genomes diverge during sympatric speciation through disruptive selection along a sharp forest-riparian ecotone.

Phenotypic and ecological divergence: Studies of ecological divergence within Hawaiian Metrosideros based on greenhouse, environmental chamber, and reciprocal transplant experiments demonstrate both heritable phenotypic variation and differential local adaptation among taxa over broad environmental gradients or sharp ecotones.  On Hawaii Island, we found significant divergence between the two successional varieties (M. p. vars. incana and glaberrima) in their use of light and nitrogen, consistent with their adaptation to either end of the successional gradient caused by recurring lava flows (Morrison & Stacy 2014).  This work demonstrated the presence of the classic plant life history trade-off of fast growth in high light and high survivorship in shade, but within a single species.  We then documented the presence of an incana–glaberrima hybrid zone on an intermediate-aged lava flow and used parent–offspring analysis to estimate heritabilities and reveal contrasting genetic architectures of leaf traits in these varieties (Stacy et al., 2016).  Field and greenhouse studies of divergence between two taxa along the forest-riparian ecotone and between two taxa along a steep elevation gradient revealed the effects of strong, persistent disruptive selection — differential adaptation of the Hawaii Island-endemic riparian M. p. var. newellii to the high-light, high-mechanical-stress environment of rivers (Ekar et al., in review), and contrasting tolerances of seedlings of M. p. vars. incana (low-middle elevations) and polymorpha (high elevation) to high uv light, high temperature, and freezing (Sakishima et al., in prep; Sakishima et al., unpub. data).  Our study of floral morphology of common-garden trees revealed that in addition to divergence in vegetative traits associated with adaptation to contrasting abiotic conditions, the varieties of M. polymorpha on Hawaii Island show heritable differences in floral characters that may be associated with biotic factors (pollinators and population density; Stacy & Dietrick, in prep).  Studies of eight Oahu taxa distributed in a predictable sequence along the island’s elevation gradient reveal variation in morphology and physiology consistent with differential adaptation along the gradient – from the dry low end to steep windy slopes and the wet volcano backbone, including root:shoot ratio, cuticular conductance, and saturation tolerance (Engelbrecht et al., in prep).  Further, using light and scanning-electron microscopy, significant variation was found in leaf micromorphology that facilitated taxonomic delineation among the island’s unnamed forms (Sur et al., 2018).  These results indicate that significant ecological variation can evolve within a long-lived tree species through divergent selection in the presence of gene flow, and they implicate Hawaii’s striking environmental heterogeneity in the emergence and maintenance of Metrosideros taxa.  Lastly, through collaboration with USDA-ARS pathologists, among-variety variation has been documented in resistance to Ceratocystis lukuohia, a fungal pathogen that is currently devastating Metrosideros forests on Hawaii Island (Luiz et al., in prep).

Four hybrid zone tree classes

Branches of four genotypes in an ephemeral, intraspecific Metrosideros polymorpha hybrid zone: var. incana (left), var. glaberrima (bottom), F1 hybrid (top), and backcross-incana hybrid (right).

Reproductive isolation (RI): Differential adaptation within a population across a heterogeneous environment should lead to the evolution of reproductive isolating barriers.  Fourteen within- and between-species crosses on Hawaii Island and Oahu reveal six partial reproductive isolating barriers with the most weakly diverged varieties isolated by a late-acting barrier only, and earlier-acting barriers observed in crosses between more genetically diverged taxa (Rhoades 2012; Stacy et al., 2017; Stacy & Ekar, unpub. data); in sharp contrast, the two most isolated taxa show strong evidence of inbreeding depression.  Combined, these results suggest that for continuously distributed, highly dispersible tree species, divergent adaptation across environmental gradients or ecotones may lead first to postzygotic isolation with earlier-acting barriers arising later and that speciation in trees may come at the cost of substantial inbreeding depression.  Determining the relative order of pre- and post-zygotic barrier evolution, however, will require quantification of all reproductive isolating barriers from immigrant inviability and flowering time divergence through to the fertility of backcross hybrids for several taxon pairs representing a range of genetic distances (i.e., strengths of isolation).  Currently, we are characterizing early post-pollination barriers through late-stage postzygotic isolation (through to backcross fertility) through greenhouse-based experiments with pure-taxon and F1 trees to yield near-comprehensive data on reproductive isolation for several taxon pairs representing a range of genetic distances and spanning both sides of the species boundary.  The ongoing study of RI will also yield novel data on fine-scale local adaptation within a landscape-dominant tree across moderate and extreme environments, data that will provide baseline information for the management of Hawaii’s forests in the face of climate change, and will set the stage for follow-on studies of the mechanisms underlying local adaptation and RI among taxa as well as their molecular bases.

Pollination at 2,400 m elevation

Hand-pollinations with maternal trees of high-elevation M. polymorpha var. polymorpha on Hawaii Island.

Metrosideros phylogeny: Through a phylogenetic analysis based on a subset of single-copy nuclear genes (Pillon et al., 2014), we recently proposed an expansion of genus Metrosideros to include Carpolepis of New Caledonia and Tepualia of South America (Pillon et al. 2015).  Subsequent analysis of 40 nuclear genes for Metrosideros and allied groups with an emphasis on Hawaiian taxa suggests that the genus arrived in Hawaii 3.1 (2.5-3.7) MYA.  Relationships within Hawaii, however, remain unresolved (Dupuis et al., in review).

Clermontia tuberculata at Waikamoi

Clermontia tuberculata at Waikamoi

Diversification in other Tropical Plant Groups – With support from the Gordon and Betty Moore Foundation, I recently led an effort to identify DNA barcoding genes from the nuclear genome that would be useful for recent plant radiations, targeting the species-rich groups Clermontia (Campanulaceae) and Cyrtandra (Gesneriaceae) in Hawaii.  In addition to identifying DNA barcodes from Roche 454 pyrosequencing of pooled-species runs (Pillon et al., 2013; Applications in Plant Sciences), we were able to reconstruct cytoplasmic-DNA-based and nuclear-DNA-based evolutionary relationships among populations and taxa within Hawaii to reveal significantly greater coalescence times in nuclear genes (Pillon et al. 2013; BMC Evolutionary Biology), significant phylogenetic discordance among nuclear genes (Pillon et al. 2013; Molecular Phylogenetics and Evolution), and high-resolution insights into the colonization and hybridization patterns of these groups on Hawaii Island (Johnson et al., in press).  Work on these two plant groups is ongoing; e.g., studies of reproductive isolating barriers in Cyrtandra (Johnson et al. 2015) and parallel niche-space evolution across islands (current).  I have also collaborated on a study of cryptic adaptive radiation in trees in New Caledonia (Pillon et al., 2014) and phylogenetic analyses and the documentation of new records within non-Hawaiian plant groups (Pillon et al., 2014; Hopkins et al., 2015; Pillon et al., 2018).

Kaala summit views