Daniel Atha, revised February 6, 2024
Introduction
This post is a preliminary summary of my findings and is written for the benefit of all interested in Smartweeds. The results of this research will be incorporated into a formal scientific publication. I am grateful for all of the observations that are helping clarify the taxonomy.
The issue is that the Water Smartweeds were long considered by most botanists to represent two or more distinct species with extremely variable morphologies depending on the habitat. The consensus abruptly shifted after 1968, when Richard Mitchell published his dissertation research (Mitchell, 1968) concluding that the presence of intermediates suggesting a continuum of variation proved that there was just one highly variable species called Persicaria amphibia.
The work presented here is not an in-depth analysis of Mitchell's findings. That will be presented later. What I wish to show is that based on the totality of evidence available today and sound taxonomic principles, the plants collectively known as Water Smartweeds are indeed at least two distinct species. Yes, there are plants that seem to be intermediate in one or more characters as Mitchell elegantly and conclusively demonstrated. But they are a small fraction of the total and their presence, however interpreted, is not proof of a single species.
I am inspired by the beauty and ecological importance of these plants and motivated by a desire to protect them from further degradation. That’s why I posted this extended explanation and why I include a link to the post along with my identifications. Thank you for taking the time to read it.
My interest in Persicaria began in 2008, when I found a Smartweed plant (Persicaria extremiorientalis) right outside my office door and all over New York City that had been in North America for fifty years, yet was unrecognized by all botanists, including Arthur Cronquist, one of the most famous and influential botanists of the twentieth century, who wrote the flora of the northeastern United States and who had an office in the same building (although not at the same time). Unfortunately he passed away in 1992, just months before I started my internship.
I've collected hundreds of Persicaria specimens and examined thousands more in major herbaria. I have studied the literature old and new, published papers, been a peer-reviewer for journals and corresponded with the handful of Polygonaceae specialists practicing today. I know the North American species pretty well and can recognize most from photographs. Ever so gradually, the plants are teaching me how to distinguish the species and since 2010, I've discovered three species new to North America (Persicaria hispida and Persicaria posumbu, in addition to Persicaria extremiorientalis).
The Water Smartweeds, Persicaria amphibia and Persicaria coccinea attracted my attention early on because they are such beautiful plants, yet no one seemed interested in them, perhaps because of their tortured taxonomic history. The more I learned about them and the sorry state of our professional conclusions, the more I wanted to “understand” them and reveal their unique qualities and relationships to each other and their surroundings.
Figure 1. Water Smartweed, Persicaria amphibia var. stipulacea (Photo 160158626, (c) Ian Guthrie, some rights reserved (CC BY-NC), uploaded by Ian Guthrie). Is it an exaggeration to say this is one of the most beautiful plants there is? The wonderful photo helps us see it.
Figure 2. Scarlet Smartweed aka Longroot Smartweed, Persicaria coccinea (Photo 298785464, (c) Ocean, some rights reserved (CC BY-NC), uploaded by Ocean). This is another spectacular photo that captures the essence of the plants with superb aesthetic qualities such as light, texture, color, form, movement, balance and composition.
Smartweeds are a genus of about 100 species primarily of the north temperate zone of both hemispheres. Most species are annuals with simple, alternate, entire, ovate or elliptic leaves and spicate inflorescences with small, five-parted flowers. Hybridization, introgression and polyploidy are especially common in the core “Eupersicaria” group of the genus. There are few autapomorphies (unique derived traits) that clearly distinguish one species from another. Often a suite of characters are necessary to define a species and distinguish it from others. In addition, the species can be quite variable and morphologically plastic in response to environmental conditions, especially periodic inundation (as are the two species discussed here). The Pale Smartweed Persicaria lapathifolia, Lady’s Thumb Persicaria maculosa and Persicaria densiflora can also form inflated, floating stems when flooded.
Persicaria amphibia var. stipulacea and Persicaria coccinea (Persicaria amphibia var. emersa) are perennial North American natives that inhabit high-quality, oligotrophic and mesotrophic wetlands, especially the OBL oligotrophic Persicaria amphibia var. stipulacea. They are both adapted to fluctuating water levels (hence the “amphibious” epithet). Persicaria amphibia (sensu stricto) is normally an aquatic with floating leaves, but when stranded on dry banks can grow aerial shoots (with often with flared ocreae). Persicaria coccinea is normally a palustrine species with aerial shoots, but can tolerate temporary flooding and may sometimes develop floating stems and leaves, but never flared ocreae.
The two species are treated as varieties in the current iNaturalist classification, strictly adhering to Plants of the World Online (POWO), which itself follows the influential opinion of Richard Mitchell (1968) who did his dissertation research on these plants. Recognition at the varietal level is far better than just one heterogeneous species (as Persicaria amphibia) advocated by Mitchell, but still short of the two species classification widely accepted before 1968 and suggested by the evidence presented here.
How we recognize these taxa is of critical conservation concern. Species of Persicaria are an important source of food for wildlife, especially waterfowl which share the same habitat and consume large quantities of the seeds. Persicaria amphibia and Persicaria coccinea often occur in extensive, dense populations and are primary producers, cycling nutrients and supplying food and habitat for macro- and micro- invertebrates as well as vertebrates (Partridge, 2001).
The vast majority of specimens and observations across North America will key out clearly with the key below. But there are populations that don’t, especially in the mountain west. These anomalies may be genetic mixtures from hybridization and introgression, both phenomena common and well-documented in the genus. Each species’ extreme anatomical plasticity and the existence of intermediate specimens has thrown botanists into taxonomic fits for 200+ years, lumping the entire range into one artificial super-species (e.g., R. Mitchell) or dividing every minor morphotype into a separate species (e.g., E.L. Greene).
Persicaria amphibia is a circum- boreal and -north temperate species occurring in America, Asia and Europe. It has long been recognized that the American plants are distinct at some level from the Eurasian as evidenced by the capacity to produce flared ocreae on aerial shoots when stranded. The Eurasian plants never do. Persicaria amphibia has over one hundred heterotypic synonyms just in North America! When recognized at the varietal level the correct name for the American plants is Persicaria amphibia (L.) Delarbre var. stipulacea (N. Coleman) H. Hara. At the subspecies level, the correct name is Persicaria amphibia subsp. laevimarginata (Hultén) Soják.
Persicaria coccinea, Scarlet Smartweed, is endemic to North America. It has almost as many synonyms. The correct name at the species level is Persicaria coccinea (Muhl. ex Willd.) Greene. As a variety of Persicaria amphibia the correct name is Persicaria amphibia var. emersa (Michx) J.C. Hickman.
Materials and Methods
Most taxonomic studies are conducted on a limited number of herbarium specimens, usually a handful and rarely numbering in the hundreds. Molecular studies today often use one sample per species. Herbarium specimens are very informative in ways photographs and illustrations will never be. They can be dissected, measured and their DNA analyzed. But ecological data is often poor or absent. The conclusions presented here are based on close examination of the nearly 10,000 observations of Water Smartweeds (Persicaria amphibia sensu lato) in iNaturalist, most of the extant herbarium specimens and observation in the field.
An often overlooked and rarely discussed issue with many taxonomic studies is the logical problem of sample selection, potentially leading to confirmation bias and circular reasoning. These can be avoided by random sampling. But that has its own set of biases and could exclude plants with the traits we wish to analyze.
This study is essentially a meta-random sample. All of the available specimens and observations are analyzed with the same criteria. The thousands of observations made by others across the entire range of the taxa are included. The iNaturalist observations were made by almost 6,000 observers for a range of purposes unrelated to this study. They are examined and analyzed by me randomly as they appear in the simple taxon search and none are excluded.
Each specimen and observation is assigned a name based a suite of morphological criteria traditionally used to distinguish the taxa. These include leaf shape (when floating and stranded); ocreae flared or not; and length and shape of inflorescence spike.
Key to the species of Water Smartweeds. (For an extended key see Reveal and Atha, 2012).
A. Plants palustrine, usually with emergent leafy stems; ocreae never with flared apices; leaves ovate or lanceolate (widest below the middle); inflorescence spikes usually 2 (unequal), cylindrical, usually > 4 cm long; flowers usually scarlet...… Persicaria coccinea (Persicaria amphibia var. emersa).
A. Plants aquatic, usually with floating stems and leaves; ocreae with flared apices (when stranded); leaves oblong or elliptic (widest at the middle); inflorescence spikes usually 1, ovoid, < 4 cm long; flowers usually pink...…Persicaria amphibia var. stipulacea.
Results
The hypothesis that there are two distinct species (Persicaria amphibia and Persicaria coccinea) and possibly one or more hybrids is strongly supported by the data.
When applied to the nearly 10,000 observations of Water Smartweed (Persicaria amphibia s.l.), the character states in the key above prove to be highly consistent and predictive. Each of the characters are significantly correlated, so that if you find a palustrine plant with ovate or lanceolate, aerial leaves it will have long, cylindrical inflorescence spikes (99% of the time). On the other hand, if you find an aquatic plant with oblong, floating leaves it will have short, ovoid spikes (99% of the time). The flared ocreae is significantly correlated with oblong leaves and short spikes. Richard Mitchell's studies found the same results, but he emphasized the significance the few apparent intermediates.
Geography is only sometimes considered a significant taxonomic character, probably because resolution of habitat and geographic ranges have traditionally been poor due to very limited samples available in herbaria and the fact that plants do in fact move over time through dispersal and habitat alteration. iNaturalist is changing that in a big way. Now with thousands or sometimes tens of thousands of samples available we are getting closer to true and accurate range maps for many of the charismatic species (which includes these two). It is now apparent that correlation between geography and morphology is very strong and stark in these two species. More broadly, iNaturalist has now incorporated geography into its AI algorithms with significant improvement in predictive values. Richard Mitchell sampled only three populations from the vicinity of San Francisco.
Persicaria amphibia var stipulacea (plants with oblong leaves, short spikes and flared ocreae) is not found south of the Laurentide Ice Sheet (and see here) except in the mountain west and Mexico. See map of this species here. Note the observation near St. Louis, tracking precisely the southernmost extent of glaciation.
Persicaria coccinea (plants with ovate leaves, long spikes and no flared ocreae) is found nearly throughout North America except the extreme southeastern coastal plain and the far north. It has not been observed north of Edmonton, Canada, unlike Persicaria amphibia which extends through Alaska across the Bering Sea and into Asia. See map of this species in North America here.
The results suggest that Persicaria amphibia is somewhat rare in North America and that the species should probably be tracked by Natural Heritage programs throughout much of its range, especially along the southern boundary in New York, Pennsylvania, West Virginia, Kentucky, southern Ohio, Indiana, Illinois and Missouri. The true abundance at a local level is impossible to ascertain today because so many datasets do not distinguish the two taxa, even as varieties. But as the species are segregated in more data sets and floristic accounts it will become clearer how rare true Persicaria amphibia is today. Failure to recognize Persicaria amphibia as distinct from Persicaria coccinea could result in a cryptic decline in the former-- if it has not happened already. At the very least, the recognition of varieties is critical to conserve the gene-pools of both taxa. The most prudent course is to conserve all taxa wherever and whenever possible and that requires accurate identification of the taxa involved.
Results Summary
-
Distinct morphologies. The vast majority of plants clearly exhibit a number of quantitative, discontinuous and correlated character states consistent with one species or the other.
-
There are rare exceptions that appear to combine character states of the two species: See for example this observation. It has two, unequal elongate spikes like Persicaria coccinea, but has pale flowers, flared ocreae and oblong leaves like Persicaria amphibia. In this case, the plant overall is more like Persicaria amphibia, just with two, somewhat elongate spikes. And here is an example of a terrestrial plant with aerial shoots, large, ovate leaves and flared ocreae. But these leaves are glabrous, unlike the usually strigose palustrine form of Persicaria coccinea. These are two 1,000 that seem to combine character states of both species. Before we jump to the conclusion that these two plants and perhaps others like them "prove" they are one species, we have to eliminate all other possible explanations, like random mutation, hybridization, introgression, misinterpretation of the evidence or even one or more additional species.
-
Distinct geographic ranges. This is becoming more and more clear as the observations accumulate. There are now almost 10,000 observations of the two species in North America and they clearly have different distributions. There are no plants with typical Persicaria coccinea character states in the far north. And there are no plants with typical Persicaria amphibia character states south of the Laurentide ice sheet in the eastern US. Very few plants better illustrate the correlation between past glaciation and present range than Persicaria amphibia.
-
No single plant has ever been found to possess the traits of Persicaria coccinea at one end of the long rhizome and Persicaria amphibia at the other, even though there is ample opportunity for them to do so based on the position of the rhizome and level of inundation. For an example of this, see the observation here. I have seen most of the herbarium specimens of both species in North America and all the iNaturalist observations and I have never seen a plant with erect shoots and long inflorescences on the stem portion out of water and oblong floating leaves and short inflorescences on the stem portion in the water. But there are many examples of Persicaria amphibia with floating leaves at one end of the rhizome and erect shoots at the other.
-
If they are one species with blended genetics, how can it be that no Persicaria coccinea like plant has ever been found with flared ocreae? That character is found exclusively in plants with the character states of Persicaria amphibia var. stipulacea.
-
The most parsimonious explanation for plants with some intermediate character states is that the they are hybrids. To consider the entire range as one species based on these very rare specimens is the least good explanation. R.S. Mitchell did not consider this possibility (as the null hypothesis) when he lumped them for his PhD thesis in 1968.
-
Taken all together these data are consistent with the consensus definition of a species in botany.
-
If we reject morphological boundaries in Persicaria as currently defined based on the tiny fraction of apparent intermediates, we would lose half the species in Eupersicaria.
Discussion
To consider the null hypothesis (for this study) and treat them as a single species (even with varieties or subspecies), as many did after 1968 (and some still do), requires ignoring the many morphological, genetic and geographic discontinuities between the taxa as proved by the evidence. These discontinuities are so distinct and so consistent, in most groups there would be no question they are two distinct species. The presence of a tiny fraction of individuals (ca. 0.01%) that appear to be intermediate in some characters does not prove they are a single species.
This looks like a classic case of convergence to me.
Lumping them together as a single species has very serious conservation implications. Conservation plans should conserve distinctive genetic lineages and conflating the two species could lead to the extinction of one or the other in the false belief that the "species" is preserved by the presence of at least some Persicaria amphibia s.l. We all know that most people tend not to use trinomials (myself included) and even heritage botanists and environmental surveyors will use the species name for convenience or uncertainty. U.S. Army Corps of Engineers data forms often omit subspecies or varieties, compromising the integrity and usefulness of EIS surveys that might include “Persicaria amphibia”
Acknowledgements
Thanks to iNaturalist, every observation used in this study is available to anyone with an internet connection. My materials and methods can easily be repeated exactly by anyone willing to spend the time. No complicated equipment, chemicals, algorithms, programs, institutional permission, journal subscription or specialized knowledge is required. That's what I love about natural history study and iNaturalist.
References
Atha, D. E. 2004. Polygonaceae. Pp. 308–310 in N. Smith, S. A. Mori, A. Henderson, D. Stevenson and S. Heald (eds), Flowering Plants of the Neotropics. Princeton University Press, Princeton.
Atha, D. E. & W. Carr. 2010. First Report of Persicaria hispida (Polygonaceae) from North America north of Mexico. J. Bot. Res. Inst. Texas 4: 561–564.
Atha, D. E., M. H. Nee & R. F. C. Naczi. 2010. Persicaria extremiorientalis (Polygonaceae) is established in the flora of the eastern United States of America. The Journal of the Torrey Botanical Society 137: 333–338.
Atha, D. and S. Rall. 2020. First report of Persicaria posumbu (Polygonaceae) for North America. Phytoneuron 2020-86: 1–7.
Burke, J. M. 2011. Revised subfamily classification for Polygonaceae, with tribal classification for Eriogonoideae. Brittonia. 63: 510–520.
Galasso, G., E. Banfi, F de Mattia, F. Grassi, S. Sgorbati & M. Labra. 2009. Molecular phylogeny of Polygonum L. s.l. (Polygonoideae, Polygonaceae), focusing on European taxa: preliminary results and systematic consideration based on rbcL platidial sequence data. Atti Soc. it. Sci. nat. Museo civ. Stor. Nat. Milano, 150(1): 113–148.
Kim, S. T. & M. J. Donoghue. 2008a. Incongruence between cpDNA and nrITS trees indicates extensive hybridization within Eupersicaria (Polygonaceae). American Journal of Botany. 95: 1122–1135.
Kim, S. T. & M. J. Donoghue. 2008b. Molecular phylogeny of Persicaria (Persicarieae, Polygonaceae). Systematic Botany. 33: 77–86.
Kim, S. T., S. E. Sultan & M. J. Donoghue. 2008b. Allopolyploid speciation in Persicaria (Polygonaceae): insights from a low-copy nuclear region. Proceedings of the National Academy of Sciences. 105: 12370–12375.
Massart, J. 1902. L'Accommodation individuelle chez Polygonum amphibium. Bull. Jard. Bot., Brux. [write out in full if that is the format required] 1(2): 73–95.
Mitchell, R. S. 1968. Variation in the Polygonum amphibium complex and its taxonomic significance. Univ. Calif. Pub. In Botany 45: 1–65.
Partridge, J. W. 2001. Persicaria amphibia (L.) Gray (Polygonum amphibium L), Biological Flora of the British Isles. Journal of Ecology 89: 487–501.
Reveal, J. L. & D. E. Atha. 2010. New combinations and typifications in Bistorta, Persicaria, Polygonum and Rumex (Polygonaceae). Brittonia 62: 243–263.
Reveal, J. L. & D. E. Atha. 2012. 8. Persicaria (L.) Mill. Smartweed, pp 236–250. in Cronquist et al. (eds), Intermountain Flora. The New York Botanical Garden Press, Bronx, NY.
Ronse Decraene L. P. & J. R. Akeroyd. 1988. Generic limits in Polygonum and related genera (Polygonaceae) on the basis of floral characters. Bot. Journ. Linn. Soc., London 98: 321–371.
Sanchez, A., T. M. Schuster, J. M. Burke & K. A. Kron. 2011. Taxonomy of Polygonoideae (Polygonaceae): a new tribal classification. Taxon. 60: 151–160.
Schuster, T. M., J. L. Reveal, K. A. Kron. 2011. Phylogeny of Polygoneae (Polygonaceae: Polygonoideae). Taxon. 60: 1653–1666.
Timson, J. 1965. A study of hybridization in Polygonum section Persicaria. Journal of the Linnean Society of London, Botany 59: 155–161.
Asian Jumpseed, Persicaria filiformis 
American Jumpseed, Persicaria virginiana
