A recent study (Ruane et al 2014) attempts to clarify the admittedly messy phylogeny of the Western Hemisphere's colubrid snake species known as Milk Snakes (Lampropeltis triangulum). The previous monograph on this complex (Williams 1978, revised 1988) was completed in the era before the pervasiveness of phylogenetics (as well as "subspecies denial") and was based entirely on morphological characteristics. Williams envisioned a single species, widely distributed from southern Canada to northern South America, composed of 25 subspecies, all of which, except for the northeastern-most two, sport a similar ringed, tricolored pattern (ostensibly a Batesian mimicry of coral snakes). This monograph was a distillation of Williams' doctoral dissertation and suffered from a number of problems, for which it was appropriately taken to task by Greene (1979). Nevertheless, it has served as the working paradigm for this complex for the past 35 years or so, despite the number and quality of subspecies that it espoused.
Having ruminated upon both the recent paper and Dr. Ruane's dissertation, from which it is derived (available ] HERE, thanks to Bob Hansen), I have reached the conclusion that it's far from the final word on the subject. Admittedly, it appears to offer reasonable solutions to some of the problems within the complex, especially so with respect to the relationships of the tropical forms. However, regarding the Nearctic forms, it is essentially replacing one unsatisfactory paradigm (Williams 1978, 1988) with another. Restricting my comments to those forms found along and to the north of the U.S.-Mexican border, since those are the only ones I am familiar with, I'll enumerate my reservations below. These, of course, are merely my opinions - feel free to enlighten me!
1) Poorly defined taxon boundaries - Since it is obvious that most of us would have to resort to the current paper's range maps to identify which "species" of milk snake we were dealing with, it is unfortunate that the authors chose to propose altering the taxonomy of the group without better establishing the geographic boundaries of each taxon. For example: in south Texas and adjacent Mexico the Rio Grande is mapped as the boundary between the putative "Lampropeltis annulata" and the adjacent "L. gentilis" even though they actually had no specimens/tissue samples from that area. Considering that the type locality for Kennicott's L. annulata is Matamoros, Tamaulipas (i.e., across that little stream from Brownsville, Texas), plus the fact that the natural environments on either side of the river are virtually identical for many miles in any direction (i.e., the Tamualipan Biotic Province [Dice 1943]), it is highly improbable that snakes from the respective vicinities of Brownsville and Matamoros belong to different, discrete species or that they belong to a stable hybrid population. Additionally, despite a strongly recurrent tendency for proposals emanating from certain phylogenetics labs to postulate rivers as barriers to various species' distributions, the Rio Grande does not appear to be - or ever have been - much of a barrier to most species of herps.
2) Lack of sympatry among alleged species
My second reservation with this new milk snake revision is that the geographic distribution of these purported "species" is troubling. One of the "acid tests" of whether closely related forms are discrete species or not has always been if they are found in sympatry with each other - if they are sympatric then they must have developed some form of reproductive isolation and are therefore distinct species. As mapped in this study, the ranges of the three U.S. milk snake "species" are contiguous, fitting together like tiles (parapatry*), with little or no overlap.
Most fully discrete, broadly distributed, species, however, have ranges that overlap widely and even occur in broad sympatry in many areas. In much of the country it's possible to find three species of Lampropeltis occurring together; none of these three milk snake "species," however, exhibit this behavior amongst themselves (except for a small area in the southern Appalachians where the Eastern Milk Snake ["Lampropeltis triangulum"]and the Scarlet Kingsnake ["L. elapsoides"] apparently occur sympatrically - but not syntopically). Elsewhere they commonly "hybridize" (intergrade?) only at the contact points on the periphery of their respective ranges, which also means that they are routinely exchanging genes within these peripheral zones; according to several species concepts, that would make them the same species. Hybridization between actual full species occurring sympatrically is a relatively rare phenomenon (as it must be if they are to retain their discreteness).
Even with the Scarlet Kingsnake ("Lampropeltis elapsoides"), which I consider the most likely candidate for full species status among the three US "species", I have to wonder why the other "full species," of milk snake, "L. triangulum" (sensu stricto) from the north and "L. gentilis" from the west, have not occupied more of the range of "L. elapsoides." If they are, in fact, fully discrete and independently evolving lineages, why don't they occur more broadly sympatric with the Scarlet Kingsnake ("L. elapsoides"), much as Prairie/Mole Kingsnakes (L. calligaster) and Common Kingsnakes (L. getula [sensu lato]) do? Of course, in order to do so they would have to have had developed some sort of reproductive isolation mechanism to maintain their distinctness.
*Parapatric speciation - Theoretically speciation can occur in populations that are distributed parapatrically, at least according to Coyne and Orr in their book Speciation (2004), but in order for this to happen the amount of genetic exchange between adjacent populations must be extremely limited, not to the pervasive extent one commonly sees in intergrade zones of bordering subspecies. In fact, Coyne and Orr were hard-pressed to find unequivocal examples where parapatric speciation had definitely occurred in any form of animal.
3) Discounting local diversity (subspecies)
It is interesting that the authors claim that their proposal will increase species diversity within the milk snake complex by six species while at the same time they are diminishing another measure of diversity (subspecies) by 18 taxa. Of course, if a study employs a species concept that fails to accept the reality or relevance of subspecies, that criterion will not appear to matter. By choosing to sequence only "anonymous" and other apparently neutral genes, which theoretically vary only by the random, non-selective process of drift, they have eliminated from consideration the very genes that result in localized phenotypic variations, some of which have been described as subspecies. Subspecies are characterized by traits determined by nuclear genes that are likely being actively selected, and which are almost certainly reflected in their morphology; natural selection acts most directly upon phenotypes, not genotypes. While conceding that there are probably poorly described and diagnosed subspecies of milk snake both north and south of the Mexican border, it seems that the present study has proposed to dispose of them in a rather doctrinaire manner.
Of course the classic defense of phylogenetic studies such as this one will state that since subspecies are currently defined almost entirely on the basis of morphology, localized morphological variants are simply beyond the scope of purely genetic studies. If, however, such phylogenetic efforts propose to dismiss, or "sink", recognized local taxonomic variants (subspecies), then they are clearly opening themselves to criticism on this point. If the methods employed in a study are inadequate to detect the genetic causes for localized morphological variation, it follows that the authors should refrain from arbitrarily dismissing taxonomic entities that have been defined using morphological criteria. Proposing to sink morphologically-defined subspecies using genetic techniques that are not yet sufficiently fine-tuned to detect the gene(s) responsible for those diagnostic trait(s) would be comparable to rejecting the existence of viruses in the era before the development of electron microscopy because they were not demonstrable using light microscopes.
Conversely, dismissing localized morphological variants - that must ultimately be genetically based, regardless of whether a particular study can detect the relationship or not - may rest on a philosophical bias, which has no place in a scientific investigation.
When Rodriguez-Robles et al. (1999) proposed to sink the various subspecies of Lampropeltis zonata, they did so by first re-examining the morphological diagnostic characters of the various races. Using more specimens from many more localities, they convincingly demonstrated that the diagnostic characters overlapped to such an extent that they were useless in distinguishing the previously recognized subspecies. Their accompanying mitochondrial genetic data was effectively irrelevant to that determination. Thus it would seem that subspecies defined on the basis of morphological characters should be re-evaluated on the basis of whether the same morphological characters can withstand additional scrutiny. Such morphologically-defined subspecies should not necessarily be contradicted by the failure of a handful of neutral genes to provide evidence of their existence.
I am also troubled at the casual dismissal of the problematic so-called Coastal Plains Milk Snake (L. triangulum "temporalis"), which morphologically appears to represent an intermediate form between "L. triangulum" and "L. elapsoides," by summarily sinking it into "L. triangulum." I doubt that the puzzle of U.S. milk snakes can be satisfactorily resolved without also solving the "temporalis" problem.
4) Species trees vs. Gene trees
The use of species trees, based upon multiple nuclear gene sequences, in this investigation is recognized as a major step forward. Previously, gene trees, based upon the sequencing of a single mitochondrial gene, had been the norm. But, as this and as several other recent papers have acknowledged, phylogenetic trees based on mitochondrial genes have been shown to be occasionally misleading, especially when there have been past hybridization events (such as apparently happened between Lampropeltis alterna and L. triangulum [sensu lato]) that have resulted in the introgression of mitochondrial genes of one species into the mitochondrial genome of another (e.g., Bryson et al. 2007).
Nevertheless it would appear that the application of nuclear gene sequencing in phylogenetics is yet in its infancy and perhaps not quite ready for prime time, despite its improvement over previous technologies. Indeed, it has always seemed a bit hubristic to this humble observer to presume to determine the phylogeny of a complex organism from the variation in the base sequences of a single gene or even a handful of them. Differences that develop between the base sequences of two isolated populations would appear to be merely incidental in the case of randomly evolving genes, as opposed to the directed changes accumulating in genes under the influence of natural selection. The histories of individual genes can vary; the greater the number of genes in an analysis, the better the resolution of localized populations. The recent sequencing of the human nuclear genome, for example, has revealed the existence of "genetic clusters," more or less corresponding to traditional, continental racial categories, that had not been detectable when only a single or small number of individual genes were sequenced previously. Thus, ultimately, accurate solutions to phylogenetic problems in herpetology may have to await the availability of more reasonably priced complete genome sequencing.
On a more positive note, this paper at last presents a phylogeny that depicts the Scarlet Kingsnake (L. elapsoides) as being more closely related to other members of the milk snake complex, rather than to the various "mountain kingsnakes" (L. pyromelana, L. zonata) as some previous phylogenetic works have proposed (e.g., Pyron and Burbrink 2009). Ironically, this paper's proposed phylogeny, besides being more geographically probable, stands in agreement with more traditional, morphology-based phylogenies. At times it seems that phylogeneticists fail to appreciate the fundamental relationship between genes and morphology: for a morphological trait to exist it must also be coded for in the genes.
Bryson, R. W., Jr.; J. Pastorini, F. T. Burbrink, and M. R. J. Forstner. 2007. A phylogeny of the Lampropeltis mexicana complex (Serpentes: Colubridae) based on mitochondrial DNA sequences suggests evidence for species-level polyphyly within Lampropeltis. Molecular Phylogenetics and Evolution 43:674-684.
Coyne, J.A. and H.A. Orr. 2004. Speciation. Sunderland, MA: Sinauer and Associates.
Dice, L.R. 1943. The Biotic Provinces of North America. Ann Arbor: University of Michigan Press.
Greene, H.W. 1979. Review of: Systematics and natural history of the American Milk Snake, Lampropeltis triangulum. Copeia 1979(1): 180-181.
Pyron, R. A., and F. T. Burbrink. 2009. Neogene diversification and taxonomic stability in the snake tribe Lampropeltini (Serpentes: Colubridae). Molecular Phylogenetics and Evolution 52:524–529.
Rodríguez-Robles, J. A., D. F. Denardo, and R. E. Staub. 1999. Phylogeography of the California mountain kingsnake, Lampropeltis zonata (Colubridae). Molecular Ecology 8:1923–1934.
Ruane, S. 2013. Phylogenetics, phylogeography, historical demography, and morphology of milksnakes (genus Lampropeltis). PhD dissertation, City University of New York. 165 pp.
Ruane, S., R. W. Bryson, Jr., R. A. Pyron, and F. T. Burbrink. 2014. Coalescent species delimitation in milksnakes (genus Lampropeltis) and impacts on phylogenetic comparative analyses. Systematic Biology 63(2):231–250.
Williams, K. L. 1978. Systematics and Natural History of the American Milk Snake, Lampropeltis triangulum. Milwaukee Public Museum, Publications in Biology and Geology, Number 2. 258 pp.
Williams, K. L. 1988. Systematics and Natural History of the American Milk Snake, Lampropeltis triangulum. 2nd edition, revised. Milwaukee Public Museum, Milwaukee, Wisconsin. 176 pp.