A New Integrative Taxonomic Approach

The following explanation of the new integrative taxonomic approach we now utilize has been excerpted from a document detailing our investigations with the Acaulescent Blue violets, sect. Nosphinium, subsect. Boreali-Americanae (W.Becker) Gil-ad. This lineage has received the bulk of our efforts in recent years, but we have been employing the same approach and methods with other species complexes to reevaluate Viola taxa throughout the Great Plains and eastern North America. For previous taxonomic background on the Boreali-Americanae (another large excerpt), see “Taxonomic History”.

A different approach

My research group is applying the General Lineage or Unified Species Concept (USC) of Kevin de Queiroz (de Queiroz 2007) as a taxonomic filter, to interpret diverse lines of evidence we gather from field, herbarium and laboratory studies. We examine results from each data set for correspondence and corroboration to delineate taxa, infer their evolutionary status (e.g., species, de novo recent hybrid, geographic or populational variant), and determine appropriate taxonomic status and rank. Rather than choosing a criterion or objective that will immediately bias our acceptance or circumscription of taxa, we accept all known or suspected speciation models or pathways as possible for producing taxonomic diversity in Viola. We make no a priori judgements on hybrid derivatives. Taxa may in fact be products of more ancient hybrid speciation, but the evolutionary processes which may have produced distinct taxa do not in themselves constitute grounds for acceptance or rejection of the taxa unless the taxa under consideration do not meet all the criteria under the USC. Recently produced de novo hybrids are expected to follow Brainerdian behavior, involving limited single-site interactions with both species in the area and exhibiting compromised fertility or anomalous reproductive behavior relative to the parental species. Our ongoing studies with limited sampling have only confirmed Brainerd's early observations for all hybrid combinations we have encountered. Such cases do not adhere to the metapopulation dynamics invoked by the USC because individual plants do not perpetuate progeny with “uniform” phenotypes and genotypes (they recombine traits of the parental species), and so they are not accepted as distinct species. In the spirit of the USC, it is incumbent upon us to use as many lines of evidence as possible to detect and circumscribe species, and to distinguish species from de novo hybrids. In this regard we consider all previously and currently recognized taxa, as well as phenotypes we find that do not obviously match accepted taxa, as taxonomic hypotheses to be tested and either accepted or rejected. For methods of detection, we combine Brainerd's intensive field observations and herbarium studies, common garden cultivation of living plants from multiple populations per taxon or phenotype, and examinations of reproductive behavior and fertility; and Gil-ad's rigorous macromorphological and micromorphological evaluation of taxa, including scanning electron microscopic examinations of seeds and lateral petal trichomes. We have supplemented these approaches with our own comprehensive univariate and multivariate statistical analyses of data from different lines of evidence; examinations of soil factors (texture, pH, organic matter) to assess ecological niche; and, in a few cases, applications of appropriate genetic markers (until recently, microsatellites) to determine genetic distinctness. Thus far, combinations of these lines of evidence, with sufficient population and regional sampling, have provided us with abundant evidence to confidently and objectively delineate evolutionary species under the USC, at a level that has never before been achieved. Even in the absence of genetic information (which has been limited to the Subsinuata group), the macromorphological, micromorphological, reproductive and ecological data have unequivocally delineated substantially more distinct evolutionary species demanding recognition than previous treatments have indicated. This includes several highly distinct new species that we are in the process of describing. The majority of our efforts have been focused on the largest North American violet group and one in the most desperate need of comprehensive revision: the Acaulescent Blue violets, Viola subsect. Boreali-Americanae.

Renewed studies

Since 2012, my students and I have employed intensive population- and site-level field studies of each phenotype (whether formally recognized taxon or informal “variant”) over specific geographic areas, guided by consultation with local and regional floristicians to visit a wide array of habitats in protected landscapes. We have coupled these intensive field studies with scrutiny of local and regional herbarium collections to ensure we are sampling the breadth of diversity in each geographic area (although in some instances, our new discoveries are the first, or among the first, to provide vouchers of new taxa to herbaria). To date, we have examined specimens comprehensively for all eastern North American Boreali-Americanae taxa at A/GH/NEBC, BHO, MO, NY, PENN/PH, US, VPI and VT; and specimens of select species complexes at AUA, BRIT/SMU/VDB, CM, DUKE, EKY, F, FLAS, GA, MICH, MIN, MSC, NCU, ND/NDG, TENN, USCH, WILLI, WIS, WMU and WVA, as well as numerous specimen images online for dozens of herbaria not listed here. After identifying a population of a recognizably different taxon at a site, and any putative hybrids, we collect approx. 5 representative plants of each taxon, give each plant a unique tag, put them in ziplock bags with the native soil, and bring them back to grow for multiple years at the new violet garden at Ohio University. At the collection site, for each narrowly uniform taxon, we collect vouchers as well as chasmogamous flowers (either pressing these or preserving them in SafeFix), and we preserve leaf tissue in silica gel from 10 uniform individuals of each taxon for subsequent genetic studies (thus far limited to microsatellite genotyping of taxa in the Subsinuata group). Back at OU's violet garden, we study and photograph structures and record traits across the entire plant body—foliage, chasmogamous flowers and fruits and cleistogamous flowers and fruits—on each plant throughout the growing season. We bag maturing cleistogamous capsules on each plant to harvest and study mature seeds (as well as any aborted ovules) following dehiscence. After retrieving bags of dehisced cleistogamous capsules, we count the number of seeds and aborted ovules per capsule, and we image capsules and batches of seeds, recording traits on these reproductive structures. We conduct scanning electron microscope examinations of the cleistogamous seeds and of the lateral petal trichomes collected earlier from chasmogamous flowers on the same plant. We have isolated hundreds of microsatellite loci from Viola sororia, and we have begun applying 11 of them in multiplexed reactions with various species groups as funding becomes available. We are also presently pursuing our first double-digest Genotyping-by-Sequencing run with numerous Viola taxa, to evaluate the efficacy of those markers for our population- and species-level studies in the immediate future. We are beginning to examine potential bioclimatic variables responsible for determining the broad geographic ranges of distinct taxa using niche modeling, placing our site-based evidence for substrate and site environmental variables in a broader context.

These combined morphological, ecological, reproductive and genetic approaches are now being utilized in integrative taxonomic investigations of morphological species groups in the Boreali-Americanae, focusing heavily on taxa in the Appalachian Mountains, Piedmont and Atlantic Coastal Plain. Bethany Zumwalde recently completed Master's thesis research on the systematics of the homophyllous cut-leaved violets of the Viola pedatifida and Viola subsinuata species groups (Zumwalde, 2015), providing unequivocal evidence to support all taxa under investigation as distinct species under the USC. She demonstrated that the so-called disjunct populations of "Viola pedatifida" in Virginia's shale barrens region were in fact an undescribed, highly distinctive county-level endemic species (V. tenuisecta ined.). Jennifer Hastings reevaluated the Viola subsinuata species group in Virginia in greater detail (Hastings, 2018), revealing the existence of a new Blue Ridge Mountain endemic (V. monacanora ined.) and adding yet more mysterious phenotypes that do not match known taxa and demand future study. Remington Burwell has recently initiated thesis research with me on the Viola affinis and Viola edulis species group, reassessing several of Brainerd’s species and investigating a couple of new species (V. “Gulf Coastal Plain” and V. impostor ined.) we detected in early 2018 along the Atlantic and Gulf Coastal Plains.

I must point out that some of the key methodologies employed by us are those used successfully by Brainerd and Gil-ad—with a more local and regional focus than accomplished previously, supplemented by our own innovations of ecological niche analysis, statistical analyses and genetic diversity investigation. We are building on the amazing work of all the specialists before us, and we realize that any new insights we may gain are far more due to the collective work of others than our own. As the taxa below are investigated, accepted and circumscribed, reexamination of many herbarium collections and further fieldwork will be needed to delineate the morphological range of variation and comprehensive geographic distribution of each of the accepted taxa. This comprehensive research will obviously take us several to many years to complete.

Current understanding of the Boreali-Americanae

Through recent integrative taxonomic studies and more thorough scrutiny of herbarium collections since 2012, our understanding of evolutionary and taxonomic diversity in eastern North American Viola has undergone a "revolution" of sorts. We have sought to detect and narrowly delineate uniform sets of populations meeting the criteria of an evolutionary species under the USC. We have examined every taxon accepted by Brainerd and his peers early in the last century, at a minimal level using comprehensive morphological studies as well as preliminary ecological and geographic investigation. In many instances, we have also had the opportunity to conduct field studies of local niche specificity, common garden observations, document micromorphological traits of cleistogamous seeds, and examine reproductive behavior. In a few instances we have microsatellite data as well. This multifaceted approach has provided convincing evidence not only to support every Brainerdian taxon but provides compelling evidence of multiple additional evolutionary species hidden in nearly every apparently polymorphic species. The most extreme examples of "cryptic" (more accurately, overlooked) evolutionary diversity have been found in Viola emarginata sensu lato (which itself has been virtually disregarded as synonymous with V. sagittata), Viola subsinuata sensu lato, Viola palmata sensu lato, and Viola sororia sensu lato.

At this time, in our investigations of the Boreali-Americanae lineage specifically, we have found compelling evidence to recognize 39 distinct evolutionary species and one variety under the USC by employing our integrative taxonomic approach. We currently have some evidence of an additional 17 anomalous phenotypes requiring further study. Each accepted taxon possesses a unique suite of macromorphological features across the plant body (spring and summer foliage, chasmogamous flowers, chasmogamous and cleistogamous capsules, seeds), usually distinctive micromorphological traits of lateral petal trichomes and/or cleistogamous seed coats, unique modal ecological niche (microhabitat), full chasmogamous and cleistogamous seed fertility, capacity to hybridize with other local species (often with taxa under which they were previously synonymized) to yield usually impaired cleistogamous seed fertility or even complete sterility, and often possess a distinctive geographic or elevational distribution (although it may overlap heavily with others regionally). Under the USC, only one infraspecific taxon is recognized (V. palmata var. triloba). In studying the Boreali-Americanae taxa, I recently created nine morphologically and ecologically defined species groups, quite different from Brainerd's three based only on cleistogamous fruit morphology, to which I assigned taxa. I measured approximately 50 quantitative and qualitative macromorphological traits on individual taxa and analyzed the nine groups statistically, finding that the groups are largely or completely distinct. They serve mainly as a convenient “container” for grouping similar taxa for our systematic reevaluation of sets of the most similar entities. (See “Genera and Groups” for more information.)

The narrower taxon delineations here are, obviously, a different perspective from treatments in the last half-century. Many of the taxa accepted here are circumscribed similarly or exactly to taxa presented by Fernald (1950) and Alexander (1963), that are themselves derived from the research of Brainerd and his contemporaries. The evolving taxon delimitations are importantly aimed at detecting and formally recognizing evolutionary species rather than utilitarian taxonomic species, although they are hopefully one and the same in most instances. The use of the USC species concept necessarily minimizes infraspecific taxa, reducing their default as provisional dumping grounds as has often been done in the past. A broad comparison with the Boreali-Americanae taxa recognized by previous specialists and our current treatment is presented below, in Table 1. The situation with the Boreali-Americanae is the most dramatic departure from previous treatments, but other groups of Viola in our region are proving to yield similar results, mainly with taxa deserving species rank.

Table 1. Taxa recognized by four previous specialists in the Boreali-Americanae group, and in the ongoing investigations by Ballard and students. *a few infraspecific taxa were reinstated by Little and McKinney (2015)

Other Viola taxa

As mentioned, we have applied the integrative taxonomic approach and methods described above with a number of other species complexes in other Viola lineages. These include the caulescent yellow-flowered Viola tripartita complex in sect. Chamaemelanium, a project pursued by Nick Chilson, in which we have obtained compelling morphometric evidence to distinguish three species (V. tripartita Elliott, V. glaberrima (Ging.) House, and V. tenuipes Pollard). We have examined the Viola blanda complex in the acaulescent white-flowered subsect. Stolonosae and have accepted V. blanda and V. incognita as separate species, attributing some of the apparent intergradation in the Appalachian Mountain region of sympatry to quite extensive hybridization and the rest to substantial within-population variation in these allo-octoploids. We have been studying other Stolonosae violets and found sufficient evidence (including additional traits) to distinguish V. lanceolata and V. vittata at the species level; we propose that not uncommon (but universally overlooked) sterile hybrids between V. primulifolia and V. vittata should be checked as the probable source for sporadic and suspicious reports of V. lanceolata deep in the range of V. vittata.

PLEASE NOTE that this is very much a work in progress. Taxon recognition, delineation, and inference of evolutionary status will proceed as our knowledge increases—limited by time, personnel and available funds. The taxonomic situation should be expected to change for several years, as we garner additional evidence on each taxon to make more confident decisions and interpretations.

Literature Cited

Alexander, E. J. 1963. Violaceae. Britton, N. L. and A. Brown, The new Britton and Brown illustrated flora of the northeastern United States and adjacent Canada, pp. 552–567. Hafner Publishing Co., New York, NY.
De Queiroz, K. 2007. Species concepts and species delimitation. Systematic Biology 56: 879–886.
Fernald, M. L. 1950. Violaceae. Gray’s Manual of Botany, pp. 1022–1042. American Book Company, New York , NY.
Hastings, J. L. 2018. Systematic and ecological studies of the Viola subsinuata species complex. M.S. thesis. Ohio University, Athens, OH.
Little, R. J., and L. E. McKinney. 2015. Violaceae. Flora of North America: Cucurbitaceae to Droseraceae. Oxford University Press, New York, NY.
Zumwalde, B. A. 2015. A systematic revision of the Viola pedatifida group and evidence for the recognition of Viola virginiana, a new narrow endemic of the Virginia shale barrens. M.S. thesis. Ohio University, Athens, OH.

Authored by Harvey Ballard on 21 March, 2020; updated on 19 April, 2020.