Phlox is a very taxonomically complex genus, with several large species complexes, especially in the mat-forming species of the plains and mountains of western North America. However, among the causlescent species of eastern North America there are three species complexes P. drummondii (annuals in Texas), P. pilosa (widespread) and P. carolina (widespread). The P. carolina complex (PCSC) comprises two species, P. carolina and P. glaberrima and at least seven subspecies. The lines between these species are very unclear, as evidenced by the fact that subspecies have switched classification between species through time.
I have recently been working on trying to identify purported members of the PCSC and have become increasingly frustrated and confused by what appears to be a continuum of variation in both floral and vegetative traits. So, what are the supposed differences between members of the PCSC? In his monograph of Phlox Wherry (1955) relies on calyx characters to separate the two species: Calyx shape, calyx lobe width, and calyx texture are all mentioned as distinguishing features. From, here, the subspecies of P. carolina are differentiated based on wishy-washy vegetative characters, which I won’t show. For P. glaberrima, both (poor) vegetive as well as calyx characters are emphasized, specifically calyx length:
Calyces in the PCSC are indeed variable in shape, size, and prominence of the midvein, as illustrated in a few exemplary specimens below (scale bars =1mm).But many of these characters seem to intergrade making identification difficult as noted on the specimen label below, annotated as P. glaberrima subsp. galaberrima by Wherry (circa 1942) and P. carolina subsp. angusta by Phlox expert Donald Levin. Apparently Wherry later changed his mind about the morphological limits of P. glaberrima subsp. glaberrima as by 1955 he only reports P. carolina subsp. angusta from this part of Mississippi. To facilitate identification of members of the PCSC I have two questions 1. do any calyx differences hold up to delimit species and subspecies in the PCSC?, 2. are any calyx characters more useful than others?. Another, peripherally, related question is 3. how do any groupings compare to evolutionary relationships between members of the PCSC?
To do this I’m using herbarium specimens annotated by either Wherry or Levin (ideally both) and therefore “expertly” identified. I’ll be looking at calyx tube shape, calyx lobe width, and calyx length as flagged as important by Wherry, but will also look at a few other calyx characters. Calyx texture is a little hard to quantify, especially without calyx cross sections, so I’ll ignore it for now. Complicating the picture of the relationships in the PCSC is the possibility for hybridization with species outside the complex, namely P. maculata as hypothesized by Wherry (1955) and demonstrated by Levin (1963), so I’ve included this species in these analyses as well. Another possibility is introgression from P. pilosa (Levin and Schall 1970), but this will be ignored for now.
To look at calyx shape, we can either get a rough estimate (elongate or round) by simply looking at the length to width ratio. But a better method is using Elliptic Fourier Analysis (EFA) by fitting trigonometric functions to outlines of a particular structure. In this case, we have the following 51 outlines color-coded by subspecies: Even from this panel, we can see that there is a lot of overlap in shape between so-called taxa. Looking at the distribution of the shapes we see the following:This method does a pretty good job of explaining the variation in shape and there appears to be some separation between groups but also a lot of overlap. A test for differences between groups suggests that at least one group is different (presumably glaberrima interior).
When we look at shape as well as some linear measures as box plots we can see some more separation, although there is overlap in almost all characters:This separation is clearest for two subspecies of P. glaberrima; P. g. interior (campanulate shape and small length) and P. g. triflora (long and broad calyces with broad lobes). All other taxa are morphologically indistinguishable based on calyx characters, except possibly P. carolina alta (narrow, tubular calyces).
Looking at these measures in a multivariate way, we do see that P. glaberrima interior and P. glaberrima triflora are distinct based on calyx morphology, but not very different when compared to other members of the PCSC:
So….it looks like we may be able to distinguish some subspecific taxa based on calyx morphology but not species. Additionally, it looks like P. carolina carolina is a catch-all taxon (if we assume that calyx differences are all that important).
Below is a phylogenetic tree of members of the PCSC from DNA sequence data. These sequences stem mainly from the study of Ferguson et al. (1999). I’ve colored members of the PCSC. Those in red are “expertly” identified and those in green are dubious identifications.The molecular data suggests the PCSC is even more of a mess than the morphological data suggests, as multiple individuals of the same species do not form clade. To what extent this is due to reticulate evolution in Phlox remains to be seen. However, the molecular data seem to suggest that messy calyx morphology in P. carolina might be due in part to multiple species currently being lumped into a single species. In addition, the distinct calyx morphologies of P. glaberrima interior and P. glaberrima triflora seem to reflect evolutionary divergence. As can be seen from the few results above, Phlox is in desperate need of continued morphological and molecular studies at the genus, species, and population level!
Ferguson, C., F. Krämer, and R. Jansen. 1999. Relationships of eastern North American Phlox (Polemoniaceae) based on ITS sequence data. Systematic Botany 24: 616-631.
Levin, D. 1963. Natural hybridization between Phlox maculata and P. glaberrima and its evolutionary significance. American Journal of Botany. 50:724-729.
Levin, D. and B. Schaal. 1970. Reticulate evolution in Phlox as seen through protein electrophoresis. American Journal of Botany. 57:977-987.
Wherry, E. 1955. The genus Phlox. Morris Arboretum Monographs III. Morris Arboretum of the University of Pennsylvania, Philadelphia. 174 pp.