An “Emily Litella” Moment in Herpetological Classification

Emily Litella

“Emily Litella” was one of the many comedic personae created by the late, great comedienne Gilda Radner, an original cast member of the television show Saturday Night Live back in 1975.  As played by Radner, “Miss Litella” was an elderly woman with a hearing impairment that frequently caused her to misinterpret—and become outraged by—some of the news stories of the day in the show’s news parody segment, “Weekend Update.” As was the practice then, local television stations allowed citizens with views opposing those expressed in news stories to appear on the air with a brief editorial reply.  The “Miss Litella” character was a regular on the newscast parody, mistaking stories about the oppression of “Soviet Jewry” in the USSR with “Soviet jewelry,”  “sex and violence on TV” with “sax and violins on TV,” and “endangered species” with “endangered feces,” and so on.  After “Miss Litella” was allowed to rant for several moments, the anchor character, usually played by Chevy Chase or Jane Curtain, would politely interrupt her, explaining what the actual topic was that she was confused about.  At that point, “Miss Litella” became meek, stating softly, “Oh, that’s different,” then turning to the camera with an innocent smile, stating “Never mind!”

Taxonomy is regrettably tasked with two different functions that are often in conflict with each other.  The first function is simply to pigeonhole each and every organism with a unique label that serves as its identity and with which we may reference it.  This function is arguably the most important to the largest number of people.  As with most references, in order to be useful it must be somewhat conservative, adopting changes only when absolutely necessary, avoiding capricious and whimsical alterations that tend to render its reference function confusing or unreliable. 

The second and more difficult purpose of taxonomy is that it should reflect the phylogeny of the organisms it classifies.  At its most functional level, this typically boils down to ensuring that all species included within a genus are actually related to each other (homophyly); a genus containing species that are not all actually derived from a common ancestor is said to be “paraphyletic.”  Obviously, this is of most concern to specialists who comprise only a tiny fraction of the persons who use taxonomy.  Most people would not be bothered if a copperhead snake was named “Crotalus contortrix,” placing it in the genus containing rattlesnakes; its moniker would still be unique, distinguishing it from other snakes.  A specialist, however, would be apoplectic as such a combination would make the genus containing most of the rattlesnakes (Crotalus) paraphyletic, containing a species that is not descended from an ancestor shared with other rattlesnakes.  Since, even though copperheads are distantly related to rattlesnakes, the very closest relatives of copperheads are contained within another genus (Agkistrodon), that genus would be the appropriate place for them. 

Some of you may remember the wave of consternation that swept through the herpetological hobbyist community several years ago with the publication of a paper (Burbrink and Lawson 2007) contending that the North American Ratsnakes (genus Pantherophis) were so closely related to the Pine and Bull Snakes (genus Pituophis) that, to avoid paraphyly, they should all be included in a single genus. This conclusion was so counter-intuitive to those of us who cut our teeth on morphologically-based taxonomy that it caused us to shake our heads in disbelief, given the phenotypical differences between these two genera.  Nevertheless, many among the “genes never lie” contingent immediately accepted this proposal as another example of phylogenetics trumping morphology despite the “obvious” incongruities involved. After all, science has a substantial history of disproving intuition. However, the “genes never lie” aphorism should perhaps be slightly modified to include the phrase “when correctly interpreted.”

It should be pointed out that the justification offered for this proposed realignment was due to two factors:

1) the sequencing of four mitochondrial genes and one nuclear gene for many—but not all—taxa considered to belong to Old World and New World “Ratsnakes,” and

2) the resulting production and cladistic interpretation of computer-generated tree diagrams (cladograms) reflecting this incomplete genetic data. 

There was no new anatomical or other traditional evidence advanced to support the proposal that two groups so “obviously” distinct as Pantherophis and Pituophis should, in fact, be included in a single genus; a mere computer-generated cladogram showing Pituophis melanoleucus (the only species of the genus included in the trial) as a “sister species” to Pantherophis vulpinus was taken as sufficient to upset the taxonomic applecart [Figure 1].  This is especially remarkable when one considers that the statistical support for the particular node (presumed speciation event) leading to this proposal was not at all strong. 

Picture1
Figure 1. Relevant portion of cladogram presented by Burbrink and Lawson (2007) purporting to show Pantherophis vulpinus and Pituophis melanoleucus as sister species (closest relatives). The numerical value at each node indicates the Bayesian posterior probability (PP) for each node as determined by a computer program, serving as an estimate of the statistical confidence at each node. (Click on image to enlarge)

Under cladistic interpretation of such data, since Pituophis melanoleucus (Eastern Pine Snake) and Pantherophis vulpinus (Eastern Fox Snake) had resolved out as each other’s closest relatives and as having both apparently descended from the same hypothetical immediate ancestor, to continue to include them in two separate genera would violate the rules of cladistic analysis (i.e., render Pantherophis “paraphyletic”).  Cladistic classification strongly discourages the recognition of paraphyletic genera and, consequently, all the species in the Pantherophis “clade” (i.e., branch), at least as diagramed by the authors, would have to assume the same generic name.

Burbrink and Lawson's problem, then, seems to have arisen from having tissue samples from only a single species of the entire genus Pituophis (P. melanoleucus), leading them to their outwardly erroneous conclusion that Pituophis melanoleucus and Pantherophis vulpinus were each other's closest living relatives and, therefore, according to cladist convention, should reside within the same genus.  The weakness in Burbrink and Lawson's study apparently derives from this failure to sample all—or at least a majority—of the species within the genus Pituophis

In retrospect, it should have been apparent to the researchers that if one is using a computer program that displays inferred relationships among species as paired branches on a cladogram, and if only one species of a polytypic genus is being tested, then the sister species on the two-branched clade would have to be a representative from another genus, thus falsely implying that the species from another genus is the closest relative only because the truly closest-related species were not being tested.  The computer program functioned as it was designed, rendering an incorrect result only because it was provided with inadequate data.  The program’s considering P. vulpinus to be the closest relative of P. melanoleucus was correct, given the limited data it had to work with. 

A similar lack of thoroughness possibly accounts for a disparity in a different study by Utiger et al. (2005) resulting in their earlier but similar conclusion (again solely for cladistic reasons) to lump the colubrid genus Masticophis into Coluber: only three species (total) from these two polytypic genera were included in the study (although in fairness it should be pointed out that earlier authors using differing methods have suggested the same arrangement [e.g., Auffenberg 1955; Dowling et al. 1983]).  In both of the present cases, however, the authors were performing studies of exceptionally broad scope involving many genera from widely dispersed areas, so in a sense under-representative sampling of larger genera may have been difficult to avoid.  Nevertheless, the possibility of distorting experimental results by allowing an entire polytypic genus to be represented by a single species should not be discounted. 

At this point, we encounter another arbitrary but arguably more necessary rule, that of taxonomic priority:  the older genus name must be used.  Pituophis was coined by Holbrook in 1842, predating Fitzinger's (1843) Pantherophis by a year.  Thus, Burbrink and Lawson (op. cit) nonchalantly suggested that all species of Pantherophis be lumped into Pituophis as the most convenient method of satisfying both cladistic doctrine and established rules of priority (regardless of the obvious anatomical differences between them).

Enter Messrs. Collins and Taggart (2008, Figure 2), for whom it appeared more generic names were preferable to fewer.  By coining a new genus (Mintonius) for the Fox Snakes, coupled with the resurrection of Scotophis Baird and Girard for the obsoletus complex within Pantherophis, Collins and Taggart cleverly managed to placate most of the cladistic objections to the phylogenies produced by Burbrink and Lawson (op. cit).  The description and diagnosis of Mintonius was ostensibly accomplished via a data matrix comprised of an eclectic set of characters, which Collins and Taggart assured us would serve to differentiate their new genus from other members of the problematic Pantherophis-Scotophis-Pituophis clade as diagrammed by Burbrink and Lawson (op. cit.).

Picture2
Figure 2. Collins and Taggart’s (2008) revised phylogeny for the ratsnakes and bull-pine snakes, using data from other authors to fill in those species missing from Burbrink and Lawson’s analysis.  The genus Scotophis was revived to accommodate the distinctive obsoletus group and the genus Mintonius is coined for the equally distinctive vulpinus group. (Click on image to enlarge)

It should be noted that cladistic analysis is not restricted to the investigation of only molecular data and, in fact, such data were unavailable when this analytical method became popular in zoological systematics.  Consequently, Collins and Taggart were on firm cladistic ground when they assembled their diverse character matrix to support their proposals.  Ideally, it would seem that reference to multiple sources and types of data would be preferable to only very limited molecular information, especially when it is painfully evident that many kinks remain to be worked out before phylogenetics alone should be trusted to be the singular authoritative source on questions of classification.

However, before the figurative ink was dry on the brainy Collins and Taggart workaround of the Burbrink and Lawson paper, the Burbrink lab was back with a fresh and more thorough study based on three nuclear genes, six mitochondrial genes, and perhaps most importantly, including representatives of all 31 traditionally recognized species of the tribe Lampropeltini (Pyron and Burbrink 2009, Figure 3).  The revised phylogeny resulting from this new study now clearly supported well-differentiated and distinctive clades for both Pituophis and Pantherophis, discounting the earlier assertion of Burbrink and Lawson (op. cit.) that the latter should be lumped into the genus Pituophis to avoid problems of apparent paraphyly.  The “Emily Litella” moment came when the authors furtively admitted the mistake in the previous paper: “With regard to the clade consisting of the rat and pine snakes and relatives, the taxonomic conclusions of Burbrink and Lawson (2007) and Collins and Taggart (2008) are shown to be inaccurate”—the taxonomic equivalent of Miss Litella's trademark, good-natured “Oh! That's different. . . . Never mind,” without actually addressing the cause of the error nor the cavalier willingness to upset taxonomic stability in perpetrating it.

Picture3

Figure 3. The relevant portion of Pyron and Burbrink’s (2009) revised phylogeny, which sampled all of the members of the tribe Lampropeltini,  clearly implying distinctive clades for the bull-pine snakes, corn snakes, fox snakes, and the obsoletus group, supporting Collins and Taggart’s (2008) alternative phylogeny.  Statistical support for this phylogeny is also much higher than for the previous Burbrink and Lawson (2007) study; the numbers above each node represent bootstrap values, those below are Bayesian posterior probabilities. (Click on image to enlarge)

The problem with this admission of error on the part of the Burbrink lab was that it also labeled Collins and Taggart’s logical solution to the quandary as being equally as mistaken as their own, which it clearly was not.  This slight did not fail to draw the notice of Collins:

“In summary, both the arrangement by Pyron and Bur­brink (2009) and that of Collins and Taggart (2008) are cor­rect and consistent with all available evidence . . . , but I feel that four distinct and easily recognized genera will always be a more infor­mative taxonomy than two genera, one of which contains a wide variety of morphologically dissimilar species.” (Collins 2010)

In short, the Collins and Taggart solution to this problem seems more intuitively satisfying than the Pyron and Burbrink one precisely because it does indeed recognize four genetically and morphologically distinct genera within the North American ratsnake complex, each of which is composed of species that are more closely related to each other than to other members of the larger complex.  Perhaps the best criticism of the Collins and Taggart proposal would be that it is the less parsimonious of the two, however in this case it would seem that the more informative view would trump the less informative one. 

Literature Cited

Auffenberg, W. 1955.  “A reconsideration of the racer, Coluber constrictor, in the eastern United States.”  Tul. Stud. Zool. 2(6): 87-155.

Burbrink, Frank T. and Robin Lawson. 2007. “How and when did Old World Rat Snakes disperse into the New World?”  Mol. Phylogenet. Evol. 43: 173-189.

Collins, Joseph T. 2010.  “An accurate accounting of an inaccurate statement about snake taxonomy.”  Jour. Kans. Herpetol. 36: 14-15.

Collins, Joseph T. and Travis W. Taggart. 2008. “An alternative classification of the New World rat snakes (genus Pantherophis [Reptilia: Squamata: Colubridae]).”  Jour. Kan. Herpetol.26: 16-18.

Dowling, H.G., Highton, R., Maha, G.C., and Maxson, L.R. 1983. “Biochemical evaluation of colubrid snake phylogeny.”  J. Zool. Lond. 201(3): 309-329. 

Fitzinger, Leopold. 1843. Systema reptilium. Fasciculus primus. Amblyglossae. Vindobonae. 106 + vi pp.

Holbrook, John E. 1842. North American herpetology. Second edition. Vol. 4. J. Dobson, Philadelphia. vi + 118 pp.

 Pyron, R. Alexander and Frank T. Burbrink. 2009. “Neogene diversification and taxonomic stability in the snake tribe Lampropeltini (Serpentes: Colubridae).” Mol. Phylogenet. Evol. 52: 524–529

Utiger U., B. Schätti & N. Helfenberger. 2005. “The oriental colubrine genus Coelognathus Fitzinger, 1843 and classification of Old and New World racers and rat snakes (Reptilia, Squamata, Colubridae, Colubrinae).” Russ. J. Herpetol. 12 (1): 39–60.

[This article originally appeared in the Winter 2022 issue of the SWCHR Bulletin]


08 May 2022

07 July 2021

07 April 2021

03 January 2021

07 October 2020

08 May 2020

17 April 2020

22 March 2020

03 October 2019

January 2023

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31