Taxonomic History
Below is a biased and narrowly focused excerpt of a document dealing almost exclusively with the taxonomic history of the Acaulescent Blue violets of Viola subsect. Boreali-Americanae. It is further biased by particularly summarizing the methodological approaches and taxonomic interpretations of the four Viola specialists who published a large body of work examining the Acaulescent Blue violets. This summary sets the stage, as it were, to introduce our recent integrative taxonomic reevaluation (see “Integrative taxonomy” for more information). For the moment, it will at least serve as a placeholder for a more comprehensive taxonomic history of all violets in the Great Plains and eastern North America. While obviously not complete, it does provide a broad background on the main players working with violets in the Great Plains and eastern North America. Many others have made valuable contributions to our knowledge of the violets at state and regional levels, and they are not mentioned here. A future iteration of this will be more thorough and will note their contributions as well.
Background
Several taxonomists in the last century have focused their investigations on North American violets (Viola) on a regional or continental level, and their treatments of taxa vary greatly in the number and circumscription, and to some extent the names used. Divergence in taxonomic treatment is particularly acute with the Acaulescent Blue violets of subsect. Boreali-Americanae (W.Becker) Gil-ad, by far the most diverse and perplexing lineage on the continent. While many taxonomists have contributed knowledge on Boreali-Americanae violets, I will restrict my discussion here to the major contributions of four specialists, out of chronological order. Two specialists, Ezra Brainerd and Nir Gil-ad, stand out in utilizing diverse empirical approaches to illuminate biological and evolutionary aspects in subsect. Boreali-Americanae. Their research has greatly advanced our knowledge of this taxonomically challenging group.
Ezra Brainerd
Ezra Brainerd was the first specialist to comprehensively study violets in North America, making several critical contributions to our knowledge of taxonomy and nomenclature, hybridization and other evolutionary processes. Brainerd published a series of papers and books from 1904 to 1924 (1904a, 1904b, 1905, 1906a, 1906b, 1906c, 1907a, 1907b, 1907c, 1908a, 1908b, 1909, 1910a, 1910b, 1911a, 1911b, 1911c, 1912, 1913a, 1913b, 1913c, 1913d, 1915, 1917, 1921a, 1921b, 1924), with the bulk of his active investigations occurring prior to 1915. Aside from his keen acumen in disentangling nomenclatural problems, his key taxonomic contributions relied on intensive field studies in New England and sporadically across the eastern U.S., and his extensive observations on the morphology and reproduction of plants over two decades in his garden (at one point, with more than 3,500 plants of diverse violet taxa including countless hybrid progeny). He grew plants of tentatively accepted species or infraspecific taxa of North American violets, including subsect. Boreali-Americanae, and putative hybrids involving them, and compared their macromorphology in all structures throughout the growing season, as well as their chasmogamous and cleistogamous fruiting behavior and mature seed output. For putative hybrids based on macromorphological traits, he also grew progeny from F1 through F3 or F4 hybrid generations from cleistogamous capsules (where viable seeds were produced) and studied their macromorphology and fertility. In many instances he maintained cultivation and observations through several years. Of Brainerd's many taxonomic and evolutionary results, some of his most profound and intriguing conclusions on the Boreali-Americanae violets are the following, based on observations of 80 intertaxon (mostly interspecific) hybrid combinations. Plants he accepted as members of a valid species reproduced through chasmogamous and cleistogamous flowers, displayed essentially complete fertility in mature seed output, produced completely uniform cleistogamous seeds, and generated progeny essentially identical in all traits to the maternal plant. Plants he accepted as hybrids were found in the vicinity of the hypothesized parental species, expressed intermediacy of individual traits and/or combinations of traits found otherwise uniquely in the parental species, failed to produce seeds from chasmogamous flowers, nearly always displayed compromised fertility in cleistogamous reproduction (typically less than 1/3 of ovules maturing to seeds, often 0-10%), produced heterogeneous cleistogamous seed morphologies (where seeds were produced at all), and generated macromorphologically heterogeneous progeny that spanned the gamut of parental species traits. Regarding the last phenomenon, Brainerd published several papers detailing "Mendelian" segregation of parent species' traits in the F2 through F4 progeny in numerous hybrid combinations. Brainerd was the first taxonomist to point out the tremendous taxonomic value of traits in cleistogamous flowers, capsules and seeds and used these traits to delineate species, demonstrate hybrid status, and organize species into groups. The most pertinent impact here is the set of useful criteria Brainerd has provided us for objective species delineation and detection of earlier-generation ("de novo") hybrids without resort to genetic data, namely precise characterization of macromorphological traits across the plant body and throughout the growing season, assessment of chasmogamous and cleistogamous reproduction, and determination of phenotypic uniformity in cleistogamous seed size, shape and color pattern (where these traits differ among potentially hybridizing species) and in cleistogamous progeny. Two other completely overlooked implications can be gleaned from Brainerd's prolific studies. One is the failure of hybrids to reproduce by chasmogamous flowers, which demands rejection of introgression, in its traditional sense as an active backcrossing process, to account for the later-generation hybrid derivatives; his evidence suggests a second, presently uncharacterized mechanism of interspecific gene flow. The other implication is that extensive segregation of traits in progeny from F1 and later hybrid generations can only result from recombination across the genome during formation of gametes in the maternal hybrid plant (see discussion below on the newly revealed high-allopolyploid nature of the Boreali-Americanae lineage). Brainerd's classification of the genus Viola in North America was accepted by his contemporaries and stood until the 1950s. It is striking and peculiar that although every taxonomic publication notes Brainerd's documentation of numerous interspecific hybrids in Viola and especially in the Boreali-Americanae, not one mentions his remarkable findings on the reproductive behavior of species and hybrids. Perhaps it is not surprising, then, that taxonomic studies and treatments since the early 1950s have circumscribed many species in subsect. Boreali-Americanae very broadly, paying no attention to Brainerd's circumscriptions and subsuming many previously recognized taxa into synonymy or dismissing them as hybrids.
Norman Russell
Norman Russell conducted a series of investigations focusing on species delimitation and hybridization in acaulescent blue violets (Russell and Cooperrider 1955, Russell and Risser 1960, Russell 1952, Russell 1953, Russell 1955, Russell 1956a, Russell 1956b, Russell 1958a, Russell 1958b, Russell 1959, Russell 1960, Russell 1963), using almost exclusively leaf blade length and width and vestiture as traits, ignoring details of chasmogamous flowers and cleistogamous fruits. His taxon delimitations were considerably broader than those of Brainerd and earlier taxonomists, and he synonymized some of Brainerd's recognized taxa and dismissed others as probable hybrids (not always on the basis of solid evidence). Nevertheless, his seminal "survey" of eastern North American Viola (Russell 1965) was the first to provide order to our cumulative understanding of eastern North American violets and present general biological, ecological and geographic information on them. It continues to serve as a valuable resource for some revisionary studies.
Landon McKinney
Landon McKinney (McKinney and Blum 1978, McKinney and Blum 1981) later made considerably broader taxon delimitations than those of Russell without substantial analysis, using few foliage traits and ignoring reproductive traits as evidence. He was prone to reducing taxa to infraspecific rank on the basis of overall resemblance, initially utilizing varietal rank and later adopting subspecies rank. Although he acknowledged the contribution of hybridization as an element contributing to taxonomic confusion, he considered "phenotypic plasticity" as the basis of variability in much of genus Viola, using this as the basis for quite massive synonymization and reduction of violet diversity to a small number of species and several taxa. His broader species concepts are often so heterogeneous that they are difficult to reconcile with plants "on the ground". McKinney summarized his efforts in a comprehensive treatment of the acaulescent blue violets (McKinney, 1992). McKinney incorporated his taxonomic concepts in a regional treatment of violets for the southeastern U.S. with Russell (McKinney and Russell, 2002), and he made critical contributions to the treatment of violets for the “Flora of North America” with R. John Little (Little and McKinney, 2015).
Nir Gil-ad
Nir Gil-ad (1995, 1997, 1998) studied Viola subsect. Boreali-Americanae during the 1990s for his dissertation research. His initial premise was that every plant could be a hybrid derivative unless evidence was found to the contrary, and his primary goal was to distinguish and delineate "orthospecies" from potential hybrid derivatives. He used a modified Minimum Diagnosability criterion of the Phylogenetic Species Concept, postulating that an accepted species must possess at least one unique macromorphological trait in a reproductive structure and also must express unique micromorphological traits that cannot be construed as the product of hybridization between accepted orthospecies. This unfortunately prevented him from accepting taxa which might exhibit a distinctive suite of traits but lacked a single unique reproductive feature. He was the first to employ scanning electron microscopy as a proxy for genetic studies, to document orthospecies-specific micromorphological traits on cleistogamous seed coats and lateral petal trichomes. He characterized the traits of cleistogamous capsules and seeds more rigorously and precisely than any previous specialist, and he grew plants of many taxa in the greenhouse in order to obtain mature cleistogamous seeds for investigation. He was the first to provide truly comprehensive and precise morphological descriptions of orthospecies and infraspecific taxa he recognized in the subsection. He was methodical in his typification studies and cleared up a number of nomenclatural and typification issues. Like others, Gil-ad made no reference to Brainerd's findings on the reproductive behavior of interspecific hybrids, held no discussion of hybrid speciation as an accepted mechanism of diversification in subsect. Boreali-Americanae, and he made no attempt to distinguish between recent de novo hybrids from descendants of ancient hybridization. He dismissed plants or taxa inferred as hybrid derivatives from further consideration if any evidence failed to demonstrate unique macromorphological or micromorphological structures. In many cases his sampling was sporadic and limited, owing to the enormity of his task—reevaluating the subsect. Boreali-Americanae as a whole. One of his most valuable conclusions regarding de novo hybrids is that they potentially represent a sizeable fraction of herbarium and field specimens examined. He demonstrated that flowering specimens alone do not provide the majority of diagnostic traits to distinguish orthospecies from themselves or from hybrids, as well as the extreme value of fruiting collections that include mature seeds harvested from naturally dehisced capsules. Gil-ad's exacting and comprehensive descriptions of accepted orthospecies, complete literature review, and his presentation of the taxonomic history of many taxa, are invaluable beginning points for understanding the morphology and taxonomy of Boreali-Americanae violets.
These four specialists have produced wildly divergent treatments of Boreali-Americanae taxa, both in terms of delimitation of recognized taxa and the ranks at which the taxa were assigned. They provide somewhat different regional perspectives, with Brainerd's main work focusing on northeastern violets, McKinney's and Russell's on the Midwest, and Gil-ad's taking in eastern North America. A few well-defined species, e.g., Viola cucullata Aiton and Viola nephrophylla Greene, have remained essentially unchanged in their circumscription and taxonomic rank, but many others have received very different treatment, depending on the extent of morphological traits utilized to delineate taxa, use of any additional biological phenomena (such as reproductive biology, in the case of Brainerd's studies), and the opinion of individual specialists on the cumulative differences required to recognize distinct species versus infraspecific taxa.
We have recently found substantially more evolutionary diversity in many taxa in the field in and herbarium collections (including local and regional herbaria) than is accommodated or described in the literature—certainly this is so in subsect. Boreali-Americanae but also to a lesser extent with other groups. While major strides have been made in addressing this evolutionary diversity, as noted above, narrowly uniform and local or regional sets of populations can readily be found that are clearly dissimilar in foliage, chasmogamous flowers, cleistogamous fruits and seeds from other known taxa. A few of these are found in the Midwest and Northeast, more in the mid-Atlantic states (including the Appalachian Mountains down the Piedmont to the Atlantic Coastal Plain), and still more in the Southeast. While broad species delimitations creating heterogeneous taxa provide the easiest option for naming all specimens, scrupulous investigation of macromorphological and micromorphological traits over the entire growing season, and observation of microhabitat and reproductive behavior, readily reveal that such broad and shallow delimitations ignore much of the actual evolutionary diversity. Most taxonomic and floristic treatments of Viola in North America have largely neglected consideration of taxa as evolutionary products, although Brainerd's use of reproductive behavior and Gil-ad's attempt at interpreting the consequences of hybridization came closest. One can legitimately argue that any taxonomic classification ignoring evolutionary processes or delimiting taxa is bound to be inadequate, and the Boreali-Americanae and other violet groups have deserved reevaluation especially including genetic and other evidence. Much effort in plant taxonomy with some vascular plant lineages has often been conducted in a heuristic scientific vacuum, yet taxonomic research in principle can and should regard taxa as hypotheses to be tested using more and better data than applied previously (or not applied at all). Taxonomic "opinion" is not legitimately scientific, yet it has held sway to date in interpretation of the vast majority of taxa in North American Viola. Fine-scale investigations of species and species complexes, involving extensive herbarium and field studies as well as comprehensive analysis using evidence from diverse methods, are nearly non-existent. However, such studies are now required in order to reassess evolutionary diversity and provide rigorous and appropriate taxonomic decisions on North American violets. See “Integrative taxonomy” for further information.
Literature
Brainerd, E. 1909. Another hybrid between a white and a blue violet. Rhodora 11: 115–116.
Brainerd, E. 1910a. Five new species of Viola from the south. Bulletin of the Torrey Botanical Club 37: 523–528, Plates 34, 35.
Brainerd, E. 1913a. Four hybrids of Viola pedatifida. Bulletin of the Torrey Botanical Club 40: 249–260.
Brainerd, E. 1911a. Further notes on the stemless violets of the south. Bulletin of the Torrey Botanical Club 38: 1–9.
Brainerd, E. 1906a. How stems and stolons disappear in Viola. Bulletin of the Vermont Botanical Club 1: 11.
Brainerd, E. 1904a. Hybridism in the genus Viola. Rhodora 6: 213–223.
Brainerd, E. 1906b. Hybridism in the genus Viola,-II. Rhodora 8: 6–10.
Brainerd, E. 1906c. Hybridism in the genus Viola,-III. Rhodora 8: 49–61, Plates 66-70.
Brainerd, E. 1913b. Is Viola arenaria indigenous to North America? Rhodora 15: 106–111, plate 104.
Brainerd, E. 1907a. Mendel’s law of dominance in the hybrids of Viola. Rhodora 9: 211–216.
Brainerd, E. 1904b. Notes on New England violets. Rhodora 6: 8–17.
Brainerd, E. 1905. Notes on New England violets,-II. Rhodora 7: 1–8, Plate 50.
Brainerd, E. 1913c. Notes on new or rare violets of northeastern America. Rhodora 15: 112–115.
Brainerd, E. 1924. Some natural violet hybrids of North America. Vermont Agricultural Experiment Station Bulletin 239: 1–205.
Brainerd, E. 1907b. The behavior of seedlings of certain violet hybrids. Science 25: 940–944.
Brainerd, E. 1911b. The caulescent violets of the southeastern United States. Bulletin of the Torrey Botanical Club 38: 191–198.
Brainerd, E. 1910b. The evolution of new forms in Viola through hybridism. American Naturalist 44: 229–236.
Brainerd, E. 1907c. The older types of North American violets,-I. Rhodora 9: 93–98.
Brainerd, E. 1908a. Viola. Fernald, M. L., Gray’s manual of botany, 579–587.
Brainerd, E. 1913d. Viola. Britton, N. L. and A. Brown, An illustrated flora of the northern United States, Canada and the British possessions, 545–563.
Brainerd, E. 1917. Viola. Rydberg, P. A., Flora of the Rocky Mountains and adjacent plains, 565–569.
Brainerd, E. 1908b. Viola chinensis in the eastern United States. Rhodora 10: 38–40.
Brainerd, E. 1911c. Viola palmata and its allies. Bulletin of the Torrey Botanical Club 37: 581–590, plate 36.
Brainerd, E. 1915. Viola septentrionalis in British Columbia. Rhodora 17: 70–71.
Brainerd, E. 1912. Violet hybrids between species of the palmata group. Bulletin of the Torrey Botanical Club 39: 85–97.
Brainerd, E. 1921a. Violets collected at Tyson. Joint Bulletin of the Vermont Botanical & Bird Clubs 7: 18–19.
Brainerd, E. 1921b. Violets of North America. Vermont Agricultural Experiment Station Bulletin 224: 1–172.
Gil-ad, N. L. 1995. Systematics and evolution of Viola L. subsection Boreali-Americanae (W. Becker) Brizicky. Ph.D. dissertation. University of Michigan, Ann Arbor, MI.
Gil-ad, N. L. 1997. Systematics of Viola subsection Boreali-Americanae. Boissiera 53: 1–130.
Gil-ad, N. L. 1998. The micromorphologies of seed coats and petal trichomes of the taxa of Viola subsect. Boreali-Americanae (Violaceae) and their utility in discerning orthospecies from hybrids. Brittonia 50: 91–121.
Little, R. J., and L. E. McKinney. 2015. Violaceae. Flora of North America: Cucurbitaceae to Droseraceae, 106. Oxford University Press, New York , NY.
McKinney, L. E. 1992. A taxonomic revision of the acaulescent blue violets (Viola) of North America. Sida, Botanical Miscellany 7: 1–60.
McKinney, L. E., and K. E. Blum. 1981. A preliminary study of a polyspecies complex in Viola. Castanea 46: 281–290.
McKinney, L. E., and K. E. Blum. 1978. Leaf variation in five species of acaulescent blue violets (Viola). I. Castanea 43: 95–107.
McKinney, L. E., and N. H. Russell. 2002. Violaceae of the southeastern United States. Castanea 67: 369–379.
Russell, N. H. 1953. A resurvey of the violets of Iowa. Proceedings of the Iowa Academy of Science 60: 217–227.
Russell, N. H. 1952. An example of introgressive hybridization between Viola papilionacea Pursh and V. nephrophylla. Proceedings of the Iowa Academy of Science 59: 134–140.
Russell, N. H. 1956a. Hybridization between Viola pedatifida and its relatives. Castanea 21: 133–140.
Russell, N. H. 1955. Local introgression between Viola cucullata Ait. and V. septentrionalis Greene. Evolution 9: 436–440.
Russell, N. H. 1960. Studies in the photoperiodic responses of violets (Viola). Southwestern Naturalist 5: 177–186.
Russell, N. H. 1956b. Techniques for species comparisons in violets. Proceedings of the Iowa Academy of Science 63: 157–160.
Russell, N. H. 1958a. The nature of Viola missouriensis. Proceedings of the Iowa Academy of Science 65: 96–101.
Russell, N. H. 1959. The violets of Minnesota. Proceedings of the Minnesota Academy of Science 25/26: 126–191.
Russell, N. H. 1958b. The violets of Tennessee 1. Keys to the species and distribution maps. Castanea 23: 63–76.
Russell, N. H. 1965. Violets (Viola) of the central and eastern United States: An introductory survey. Sida 1: 1–113.
Russell, N. H., and M. Cooperrider. 1955. Prediction of introgression in Viola. American Midland Naturalist 54: 42–51.
Russell, N. H., and F. S. Crosswhite. 1963. An analysis of variation in Viola nephrophylla. Madroño 17: 56–65.
Russell, N. H., and A. C. Risser Jr. 1960. The hybrid nature of Viola emarginata. Brittonia 12: 298–305.