I arrived at the Missouri Botanical Garden’s herbarium in the middle of a summer afternoon in 2014. I had spent the past two days driving and looking for Viburnum in the farm country of northern Illinois where occasional natural populations can be found in wooded roadside ditches. I was a second-year graduate student at Yale University, and it was my second collecting trip. This herbarium was an important destination because it is the second largest in the United States, and it has excellent collections of North American plants. I was planning to photograph all the viburnum specimens. I especially wanted to learn how to identify the seven viburnum species that taxonomists have traditionally grouped in the Lentago clade. A curator met me and took me to the viburnum section, where he showed me five full floor-to-ceiling cabinets. As I began to sort through the specimens that afternoon, I was totally overwhelmed: there were far too many to photograph, and I couldn’t tell the species apart.
The Viburnum Lentago clade is a small lineage of wide-ranging shrubs and small trees that are common in woodland edges of North America. Two members of this clade, nannyberry (V. lentago) and witherod viburnum (V. cassinoides), occur in New England and the upper Midwest. The lineage also includes the blackhaw (V. prunifolium), rusty blackhaw (V. rufidulum), and possumhaw viburnum (V. nudum), which have more southern distributions. Walter’s viburnum (V. obovatum) occurs primarily in Florida, and V. elatum is restricted to Mexico. The species in this lineage all flower in the early spring and produce large blue or black bird-dispersed fruits.
One of the most notable features of the group is that its members look very similar to one another, and nearly all the traits that distinguish the species are subtle. After my first day at the herbarium, I reread the two best treatments of the Lentago clade: Thomas Jones’s 1983 graduate thesis and Waldo McAtee’s A Review of the Nearctic Viburnum. Both taxonomic keys are full of equivocal phrases like “leaves usually short-pointed or rounded apically,” “petioles more or less crinkly-margined,” and “veins less scurfy, usually glabrous.” Resigned to the fact that I might not be able to tell the species apart, I decided to photograph as many specimens as I could with the hope that I would be able to figure them out once I was back in New Haven.
Over the next couple of years, I spent many weeks collecting viburnums from wild populations in the eastern United States, collecting 330 individuals. I spent many more weeks looking at herbarium specimens. I did eventually learn to tell the species of the Lentago clade apart. Many botanists know the feeling: familiar species become easy to identify, but it is almost impossible to describe what is distinctive about them. Once I knew the species, I understood the reason for ambiguous descriptions in the taxonomic keys. My own ways of identifying the species are just as difficult to describe—like picking out friends in a crowd of people.
This detailed work on a small plant lineage led to surprising results and new insights into the biology of the clade. We even rediscovered a species, Viburnum nitidum, in the southeastern United States, which has been ignored for most of the past two centuries. It turned out that a lot can be learned from a group of plants that was known even to Linneaus, a lineage collected hundreds of times and studied by generations of taxonomists.
History of the Clade
Although most viburnums today occur in temperate or boreal environments, phylogenetic evidence suggests that the ancestors of Viburnum, and of the Lentago lineage specifically, lived in the warm temperate forests of Asia. Then, around thirty to thirty-five million years ago, Lentago moved into North America (Landis et al., in prep). The only close relatives of the Lentago clade are V. punctatum and V. lepidotulum, which both live in warm forests in Southeast Asia, many thousands of miles of away. Because no closely related species occur in Europe or Northern Asia, it is impossible to be sure how the Lentago clade’s ancient migration occurred. Researchers, however, have found a fossilized viburnum pollen grain in Iceland, which has distinctive characteristics typical of species in the Lentago clade. This suggests that Lentago was likely present in Iceland about fifteen million years ago and may have migrated to North America through Europe, over the North Atlantic Land Bridge (Landis et al., in prep).
Once in North America, the Lentago clade split into two lineages: the core Lentago clade and the Viburnum nudum species complex. The core Lentago includes five species: V. lentago, V. prunifolium, V. rufidulum, V. obovatum, and V. elatum. The V. nudum species complex is a small, variable lineage, which I’m calling a “species complex” because when I started working on it, there was a lot of uncertainty about how many species were included and how they were different from one another. Like most of Viburnum, the species in this complex bear fruit on stalked inflorescences (umbel-like compound corymbs), but in the core Lentago clade, the inflorescences are sessile or unstalked, which means that leaves are produced immediately under the inflorescence branches. Each of these two lineages has radiated into habitats that today span the range of eastern North America from central Florida to Nova Scotia.
One of the most enigmatic species of the Lentago clade is the blackhaw, Viburnum prunifolium. This species occurs in a geographic area that overlaps with both V. lentago and V. rufidulum, and it is intermediate between them in many traits. Joe Brumbaugh and Arthur Guard, who observed the overlapping ranges of all three species in Indiana, concluded, in 1956, that repeated backcrossing between these species—a process known as introgression—could be a significant cause of taxonomic confusion. Botanist Linda Rader went further and argued, in 1976, that V. prunifolium might actually be a hybrid species formed by an ancient hybridization event between V. lentago and V. rufidulum parents. Although V. lentago and V. rufidulum do not come in contact today, it is reasonable to imagine that they might have shared a geographic range in the past and might have had opportunities to hybridize.
Each of these theories about hybridization is supported by the fact that hybrids among the species are possible. The cross between Viburnum lentago and V. prunifolium is known as V. × jackii. Alfred Rehder, a taxonomist who worked at the Arnold Arboretum for much of his career, proposed the name in honor of his colleague John George Jack, who, in 1908, noticed plants in the Arboretum that appeared intermediate between V. lentago and V. prunifolium and assumed that they were spontaneous hybrids. Although those individuals are no longer at the Arboretum, a specimen of V. × jackii growing at the Morton Arboretum was obtained from the Arnold Arboretum and is likely to be from the same original lineage.
The only documented set of controlled crosses between Viburnum lentago and V. prunifolium was carried out by Donald Egolf, a graduate student at Cornell University who would become a leading research horticulturist at the United States National Arboretum. His findings, in 1956, found that a cross between V. lentago and V. prunifolium yielded twenty-eight seeds and twenty-six plants, roughly the same number as crosses between only V. lentago. These numbers are surprising because they suggest that intrinsic barriers that could prevent hybridization are very low for this species pair: if an insect transported pollen from V. lentago to a V. prunifolium flower, the V. prunifolium would likely produce fertile seeds. All of this indicates that hybridization is possible and could even be common in natural populations where both species occur.
In order to test for hybridization and get a better understanding of these species, I went on a series of road trips to collect leaf samples from across the range of Viburnum lentago, V. prunifolium, and V. rufidulum, including areas where multiple species occur. At each natural population, I collected leaves to use for morphological measurements and leaves to use for DNA extraction and sequencing. Across all of these areas, I found no evidence of hybrid zones or of widespread gene flow among species. While the species sometimes appear morphologically similar to one another, it turned out that this variation is not related to genetic structure or hybridization. That is, in cases where V. rufidulum looked somewhat like V. prunifolium, genetic sequencing showed that it was still 100 percent V. rufidulum. Overall, our sequencing found that V. prunifolium did not originate through hybridization. Instead, V. prunifolium is sister to the southern V. rufidulum, and the northern V. lentago is a more distant relative (Eaton et al., 2017; Spriggs et al., 2019).
Out of all the 180 individuals sequenced, only two appear to be admixed (have genes from two different species). These two individuals are from a small population in northern Kentucky where Viburnum prunifolium and V. rufidulum occur together in a rocky roadside woodland, five miles east of the Kentucky River. These individuals were morphologically similar to V. rufidulum (I labeled them as V. rufidulum when I collected them), but when sequenced, they appeared to be half V. prunifolium and half V. rufidulum.
Whenever hybridization is possible but rare, it suggests that something is acting to prevent or eliminate hybrids, in other words, mechanisms of reproductive isolation. Hybridization in natural populations has several possible outcomes. One possibility is that hybridization between two species will be so common that the species will eventually merge to become a single species. At the other extreme, if hybrids between two species are unfit, natural selection can cause species to evolve ways of avoiding one another. In some plant lineages, closely related species have different numbers of chromosomes, often leading to inviable offspring. Another common way plants avoid interbreeding is with flowers that evolve to attract different pollinators. In the Lentago clade, however, the species have the same number of chromosomes (Egolf, 1956), and their flowers are extremely similar (Donoghue, 1980).
How, then, do these species in the core Lentago group maintain their separation? The answer seems to lie in subtle habitat differentiation and phenological timing. Brumbaugh and Guard, in 1957, described habitat distinctions for the three overlapping species based on their work in Indiana: “V. lentago occupies poorly drained areas; V. prunifolium, moist borders of woods, and V. rufidulum, dry rocky slopes.” This characterization matches my own experience in the field, although I have also found that it is not difficult to find areas with two or more of the species occurring together. This kind of habitat differentiation could decrease how often the species flower in close proximity to one another. If the species are also adapted to slightly different habitats, hybridization might be disadvantageous because it could separate beneficial traits or cause beneficial traits to be lost altogether.
Even more intriguing, several authors who have observed these species in the field or at arboreta sometimes mention offset flowering times. Viburnums typically flower once a year for only ten to fourteen days, and flowering time is often very synchronized within species (Donoghue, 1983). Rehder, in 1920, wrote that Viburnum prunifolium flowers about a week before V. lentago, and Rader, in 1976, noted that V. prunifolium flowered about two weeks before V. rufidulum. My doctoral advisor Michael Donoghue observed this same pattern, where V. prunifolium flowers about a week ahead of the other two species, while he was conducting his own dissertation research at the Arnold Arboretum (Donoghue, 1980). To investigate whether flowering time varies between the species in wild populations, I drove through Ohio, Indiana, and Kentucky in the early spring of 2015. In this region, V. prunifolium was very common and V. rufidulum less so, but over the course of a week, I found multiple populations of each species and consistently found V. prunifolium in full flower while V. rufidulum had green buds.
Anecdotal evidence from single locations (even wild populations) is helpful but not all that convincing because each of these species occurs across a large geographic area, and each year populations experience a variety of climate conditions. To try to understand patterns of flowering time, I used herbarium specimens, which were excellent because the species are relatively common and have been collected hundreds of times from across their ranges. Even more importantly, most specimens are reproductive, meaning they have flowers or fruits, so we found plenty of suitable material, representing a large number of years. With the help of Caroline Schlutius, an undergraduate researcher, I found 1,379 flowering specimens that spanned the range of each of the species.
By examining these specimens, we found clear geographic patterns in flowering time and consistent differences in flowering between species. First, we found that within each species, southern populations in warmer climates flowered earlier than northern populations in colder climates. More surprisingly, we found small but consistent differences in flowering time that remained remarkably constant across the regions where species co-occurred. In any given location, a nine- to ten-day difference in flowering time occurs between species, with a sequence that matches previous observations—Viburnum prunifolium flowers before V. lentago and V. rufidulum. Because viburnums only flower for ten to fourteen days total, this small offset can dramatically decrease the opportunities for pollen transfer among plants. To be clear, these findings do not suggest that all of the individuals will flower in sequence every year in every location, only that in any given place for a particular year, the majority of individuals of one species will flower ahead of the majority of individuals of the other.
Hidden in Plain Sight
A second major focus of my work on the Lentago clade was the Viburnum nudum species complex. For the V. nudum complex, I wanted to sort out how many species there are, where they occur, and whether any traits consistently differentiate the species. In my first year collecting viburnums, I started in Florida and drove north. I was specifically targeting the V. nudum species complex, but it was very hard to find. I was surprised because viburnums are easy to find in New England, and I had expected to encounter populations driving down sideroads or in the state parks where I had permits to collect. This was not the case, and for future trips, I researched locations extensively using herbarium specimens and talking with local botanists.
The first year, I collected a couple of individuals near Gainesville and then didn’t find V. nudum again for a full week. The populations I had collected in Florida were in sandy soil along shallow streams, and I was looking for similar habitats as I made my way up the East Coast. I was driving on a small road along the edge of a black-water swamp in coastal South Carolina when I found the next population. A small group of viburnums was strung out along the edge of the road, several inches deep in muddy water, near a swamp with bald cypress (Taxodium distichum) and water tupelo (Nyssa aquatica). Not only was this a totally different environment than I was expecting but the plants looked different. The Gainesville plants had seemed delicate: they had small, narrow leaves, and the inflorescences bore bright pink and more mature black fruits simultaneously. This South Carolina population had thick leaves, larger than my hand, and pale green fruits. After several more collecting trips over the next several years, allowing us to sequence individuals from many populations, we discovered that these habitats are both typical for the V. nudum complex, but they contain totally separate genetic lineages, each adapted to its own environment.
These results were surprising because they are at odds with the generally accepted taxonomy of Viburnum nudum species complex, which dates to the eighteenth century. Linnaeus described V. nudum in the first edition of Species Plantarum, published in 1753, and added V. cassinoides in the second edition, in 1782, distinguishing V. cassinoides by its leaf shape. Then, in 1789, William Aiton, the first director of the Royal Botanic Gardens, Kew, proposed two additional species, V. nitidum and V. laevigatum. Since then, more than eight other names have been proposed for segregates within the complex (McAtee, 1956), but none of these have been widely recognized. In recent floras, only two species or subspecies are recognized: V. cassinoides in the North and V. nudum in the South (Jones, 1983; Small, 1933; Gleason and Cronquist, 1991; Ferguson, 1966; Radford et al., 1968; Strausbaugh and Core, 1978; Weakley, 2012).
When we sequenced DNA of eighty individuals, we found three different lineages in the Viburnum nudum species complex, all of which seem to be evolving independently. Most importantly, we found that the way V. nudum is typically described makes it paraphyletic, meaning that the name refers to a partial evolutionary lineage. The oldest genetic split in the V. nudum complex lineage is not between the northern and southern populations (traditionally V. cassinoides and V. nudum). Instead it is between the large-leaved populations that occur in swamps and the rest of the complex (including sandy stream populations and northern populations). Our genetic data show that each of these three lineages is distinct and evolving independently, and therefore all three deserve to be recognized as species (Spriggs et al., 2018). Linnaeus’s V. nudum matches the large-leaved species that occurs in swamps; V. cassinoides corresponds to the northern species in our analyses; and the sandy-soil species seems to match Aiton’s V. nitidum.
After surveying herbarium specimens from across the Southeast, we determined that Viburnum nitidum is mostly restricted to the coastal plain. From Florida it extends up to North Carolina, along the coast and in the Sandhills region, then west into the eastern edge of Texas. V. nudum is more widespread and occurs throughout the coastal plain and the Piedmont, from Delaware to Arkansas. The habitats of these two species are interdigitated across the Southeast, and the species occur in close proximity to one another frequently yet remain distinct and do not hybridize.
Our findings support recent arguments that the flora of the North American Coastal Plain is under-described, meaning it is more diverse than the current taxonomies suggest (Sorrie and Weakley, 2001; Noss et al., 2015). It seems that this disregard was not always the case. Over the past century, field botanists, particularly southern botanists like William Ashe, of North Carolina, or Alvan Chapman, who spent most of his career in Georgia, recognized subtle variation in the habitats of the coastal plain and described many species. These proposed species have been systematically ignored or lumped into larger widespread “species” that may turn out to be paraphyletic. Genetic sequencing may vindicate at least some of these descriptions, as was the case for Aiton’s V. nitidum.
Significance of this Clade
It is reasonable to question whether it is important to know how many species are in the Lentago clade or to know how exactly they are related to one another. If Viburnum nitidum is very similar to V. nudum, does it really matter that it has a separate name? Is it useful to know whether V. prunifolium originated as a hybrid species? Does knowing the recent evolutionary history of these species have any broader implications? I believe that all this lineage-specific knowledge about species limits, occurrences, and history is important for conservation and also provides insight into the ecology and evolution of North American forests.
For one thing, species identities are fundamentally important because they affect how species are recognized and valued, and whether they are conserved. Some might view the long list of proposed (and subsequently ignored) botanical names for segregates of the Viburnum nudum complex as wasted effort, but I view these names and descriptions as essential contributions, stepping-stones leading along a path to accurately characterize plant diversity. The numerous common names for V. nudum, including witherod viburnum, possumhaw, wild raisin, and Appalachian tea, suggest that even nontaxonomists knew these plants and attempted to differentiate them. With genetic sequencing, we have new opportunities to get this right, to rigorously test species limits so that the current names and species descriptions are accurate. This is especially important because relatively few plant lineages in North America (or elsewhere) have been studied with fine-scale genetic sequencing, and there is much to learn.
Accurate species descriptions are also an essential starting point for ecological studies that seek to understand species distributions or interactions. If Viburnum nitidum and V. nudum are considered as a single entity, their distribution might be confusing or bimodal because it would include two very different habitat types. If, instead, V. nitidum is considered alone, its habitat preferences would likely be much more specific, and it might even be a useful indicator species that could serve as a quick way to identify a particular habitat or plant community. Similarly, species often differ for traits that are not readily apparent to human observers, and these can be critical in shaping interactions with insects or other species. Imagine, for instance, that there was reason to suspect that V. nitidum was a host species for a rare caterpillar, but in an assessment of this relationship, a researcher sampled V. nudum thinking it was the same as V. nitidum. This kind of mistake could lead to inconsistent or misleading results. Species are an essential unit for many ecological studies, and when species descriptions are inaccurate, there can be significant consequences.
Finally, understanding the evolution of this small lineage can provide important insights into the evolution of the North American flora more generally. Each of the two major lineages of the Lentago clade diversified in eastern North America over many millions of years, and each lineage has differentiated into a series of morphologically similar but ecologically distinct species. Similar patterns of slow diversification into subtly different species are also apparent in other North American lineages like maples (Acer), dogwoods (Cornus), or ash (Fraxinus). Our work showed how the species of the Lentago clade have persisted through climate fluctuations associated with glaciation and provided evidence on one of the mechanisms (flowering time) that these species use to maintain their separation. Tensions around species identifications have hindered our understanding of the Lentago clade for over a century, but after years of observation in the field and extensive genetic sequencing, the species that had seemed incomprehensible to me at the Missouri Botanical Garden herbarium became clear. Many other plant lineages that form the foundation of the flora of eastern North America likely have similar histories of subtle differentiation and persistence, promising countless stories waiting to be revealed.
Brumbaugh, J. H. and A. T. Guard. 1956. A study of evidence of introgression among Viburnum lentago, V. prunifolium, and V. rufidulum based on leaf characteristics. Proceedings of the Indiana Academy of Sciences, 66: 300.
Clement, W. L., and M. J. Donoghue. 2011. Dissolution of Viburnum section Meaglotinus (Adoxaceae) of Southeast Asia and its implications for morphological evolution and biogeography. International Journal of Plant Sciences, 172: 559–573.
Clement, W. L., M. Arakaki, P. W. Sweeney, E. J. Edwards, and M. J. Donoghue. 2014. A chloroplast tree for Viburnum (Adoxaceae) and its implications for phylogenetic classification and character evolution. American Journal of Botany, 101: 1029–1049.
Donoghue, M. J. 1980. Flowering times in Viburnum. Arnoldia, 40(1): 2–22.
Donoghue, M. J., B. G. Baldwin, J. Li, and R. C. Winkworth. 2004. Viburnum phylogeny based on chloroplast trnK intron and nuclear ribosomal ITS DNA sequences. Systematic Botany, 29: 188–198.
Eaton, D. A. R., E. L. Spriggs, B. Park, and M. J. Donoghue. 2017. Misconceptions on missing data in RAD-seq phylogenetics with a deep-scale example from flowering plants. Systematic Biology, 66(3): 399–412.
Egolf, D. R. 1962. A cytological study of the genus Viburnum. Journal of the Arnold Arboretum, 43(2): 132–172.
Egolf, D. R. 1956. Cytological and interspecific hybridization studies in the genus Viburnum (Unpublished doctoral dissertation). Cornell University, Ithaca, NY.
Ferguson, I. K. 1966. The genera of Caprifoliaceae in the southeastern United States. Journal of the Arnold Arboretum, 47(1): 33–59.
Gleason, H. A. and A. Cronquist. 1991. Manual of vascular plants of Northeastern United States and adjacent Canada (2nd ed.). Bronx, NY: New York Botanical Garden.
Jones, T. H. 1983. A revision of the genus Viburnum section Lentago (Caprifoliaceae) (Unpublished doctoral dissertation). North Carolina State University, Raleigh, NC.
Landis, M. J., D. A. R. Eaton, W. L. Clement, B. Park, E. L. Spriggs, P. W. Sweeney, E. J. Edwards, and M. J. Donoghue. In prep. Joint estimation of geographic movements and biome shifts during the global diversification of Viburnum.
McAtee, W. L. 1956. A review of the Nearctic Viburnum. Chapel Hill, NC: Author.
Noss, R. F., W. J. Platt, B. A. Sorrie, A. S. Weakley, D. B. Means, J. Costanza, and R. K. Peet. 2015. How global diversity hotspots may go unrecognized: lessons from the North American Coastal Plain. Diversity and Distribution, 21: 236–259.
Rader, L. L. 1976. A biosystematics study of Viburnum prunifolium and Viburnum rufidulum (Caprifoliaceae) (Unpublished M. S. dissertation). University of Tennessee, Knoxville, TN.
Radford, A. E., H. E. Ahles, and C. Bell. 1968. Manual of the vascular flora of the Carolinas. Chapel Hill, NC: The University of North Carolina Press.
Rehder, A. 1920. New species, varieties and combinations from the herbarium and the collections of the Arnold Arboretum. The Journal of the Arnold Arboretum, 2(2): 121–128.
Small, J. K. 1933. Manual of the Southeastern Flora. New York: Author.
Sorrie, B. A. and A. S. Weakley. 2001. Coastal plain vascular plant endemics: phytogeographic patterns. Castanea, 66: 50–82.
Spriggs, E. L., C. Schlutius, D. A. R. Eaton, B. Park, P. W. Sweeney, E. J. Edwards, and M. J. Donoghue. 2019. Differences in flowering time maintain species boundaries in a continental radiation of Viburnum. American Journal of Botany, 106(6): 833–949.
Spriggs, E. L., D. A. R. Eaton, P. W. Sweeney, C. Schlutius, E. J. Edwards, and M. J. Donoghue. 2018. Restriction-site-associated DNA sequencing reveals a cryptic Viburnum species on the North American Coastal Plain. Systematic Biology, 68(2): 187–203.
Spriggs E. L., W. L. Clement, P. W. Sweeney, S. Madriñán, E. J. Edwards, and M. J. Donoghue. 2015. Temperate radiations and dying embers of a tropical past: the diversification of Viburnum. New Phytologist, 207(2): 340–354.
Weakley, A. S. 2012. Flora of the Southern and Mid-Atlantic States. Retrieved from http://www.herbarium.unc.edu/flora.htm
The map in this article was created using Esri, USGS, USFS, NGA, NASA, CGIAR, N Robinson, NCEAS, NLS, OS, NMA, Geodatastyrelsen, Rijkswaterstaat, GSA, Geoland, FEMA, lntermap and the GIS user community.
Elizabeth Spriggs is a Putnam Fellow at the Arnold Arboretum.