Dr. Rob Dillon, Coordinator

Friday, October 14, 2016

The Fat simplex of Maryville matches type

Editor’s Note.  This is the second installment of a three-part series on the discovery of a cryptic pleurocerid species in East Tennessee.  You might want to back up and read my essay of 13Sept16 if you haven’t already.  The present essay was published in the FMCS Newsletter Ellipsaria 18(2): 16 - 18 [pdf] if you’re looking for something citable.

So when last we left our intrepid malacologist, puzzling at his lab bench in the summer of 2008, he had come around to the realization that the population of Pleurocera simplex inhabiting Pistol Creek at the Courthouse Park in Maryville, Tennessee, was actually two reproductively isolated populations, fats (S6f) and skinnies (S6s).  Which population might be the bona fide P. simplex of Thomas Say (1825)?  And what might be the identity of the other?

I actually started out with a pretty good hunch.  In 2007 John Robinson and I ran allozyme gels on five populations of P. simplex from up in Virginia, including a nice sample from Say’s type locality in Saltville [1].  The Saltville population (S5) was fixed for the Oldh100 allele that turned out to be diagnostic for the Maryville S6f fats.  And interestingly enough, the simplex population John and I sampled from Indian Creek at Kesterson Mill way down at the southwest tip of Virginia (S1) was nearly fixed for that Oldh104 allele diagnostic for the Maryville S6s skinnies.  I hadn’t noticed any shell morphological differences between the Saltville simplex and the Kesterson Mill population at the time, however.

So in August of 2008 [2] I drove up to Virginia for additional samples from Saltville and from Kesterson Mill, which I returned to Charleston to run alongside a fresh batch of Maryville samples, to make sure all the bands matched up. Sure enough, the Maryville S6f fats were indeed genetically similar to P. simplex collected from its type locality at Saltville S5, not just at the Oldh locus but over all ten of the allozyme loci I have found informative for studies of this sort.  And the Maryville S6s skinnies were similar to Kesterson Mill S1.

Figure 1.  See footnote [3] for methodological details.

I also took the calipers to a bunch of the shells from Saltville and Kesterson Mill.  And sure enough, exactly the same pattern reappeared that we first noticed at Maryville.  Dividing the total shell length into body whorl height (B) and apex height (everything else, A), the (N = 37) shells I sampled from Kesterson Mill showed a significantly greater ratio of A to B than the (N = 40) shells from Saltville.  Figure 2 shows that the simple regression of apex height on body whorl height for the Saltville type population S5 was A = 0.157B + 2.46 (r = 0.36), very significantly different (ANCOVA t = -9.52, p < 0.0001) from the regression for the Kesterson Mill population S1, A = 0.556B - 0.09 (r = 0.84).  Saltville snails are fat, and Kesterson Mill snails are skinny.

Figure 2.

I have picked two shells directly off the regression lines to illustrate the difference between the two species.  So since the Maryville S6f fats match that lower-sloping S5 sample of N = 40 from Saltville both genetically and morphologically, and since Saltville is the type locality of Pleurocera simplex (Say 1825), clearly the Maryville S6f fats must be the bona fide simplex [4].

Then what might be the identity of the Maryville S6s skinnies, matching the higher-sloping S1 population from Kesterson Mill both genetically and morphologically? Tune in next time!


[1] Dillon, R. T., Jr. and J. D. Robinson. 2007. The Goniobasis ("Elimia") of southwest Virginia,  I. Population genetic survey.   Report to the Virginia Division of Game and Inland Fisheries.  25 pp. [pdf]

[2]  Yes, this was the same trip I featured last month, in which I resampled Maryville.  I’ve been sandbagging my story a little bit, to make it unfold in a more linear fashion.  To tell you the truth, I got the “hunch” mentioned in this month’s essay pretty much immediately upon reading my first Maryville simplex gel, and last month’s story unfolded pretty much simultaneously with this month’s story.

[3] Subsamples of 14 individuals from population S5 and 41 from population S1 were analyzed electrophoretically together with the 20f + 51s individuals from Pistol Creek I featured in last month’s blog post.  These data were combined with the data sets of N = 34 from population S5 and N = 37 from population S1 previously published by Dillon & Robinson (2007), and with the N = 17f + 13s data I had previously accumulated from Maryville. Then BIOSYS version 1.7 (Swofford & Selander, 1981) was used to calculate matrices of Nei's (1978) unbiased genetic identity (below the diagonal in Fig 1) and Cavalli-Sforza and Edwards (1967) chord distances between all pairs of control populations S1 and S5 and the two cryptic S6 populations co-occurring in Maryville. Chord distances were used as the basis for the neighbor-joining tree shown above the diagonal in Figure 1, as calculated using PHYLIP v3.65 program NEIGHBOR (Felsenstein, 2004).

[4]  Ultimately I only used that sample of 20 + 17 = 37 fats as my S6 Maryville control in Dillon 2011.  The existence of a second Pleurocera population at Maryville cryptic under simplex was not mentioned in my 2011 paper at all:
  • Dillon, R. T., Jr.  2011. Robust shell phenotype is a local response to stream size in the genus Pleurocera (Rafinesque 1818).   Malacologia 53:265-277. [pdf]

Tuesday, September 13, 2016

The Cryptic Pleurocera of Maryville

Editor's Note.  This is the first installment of a three-part series on the discovery of a cryptic pleurocerid species in East Tennessee.  The present essay was published in the FMCS Newsletter Ellipsaria 18(2): 15 - 16 [pdf] if you're looking for something citable.
"endless forms most beautiful and most wonderful have been, and are being, evolved."
I like to imagine that, as Charles Darwin wrote the poetic final clause of his masterwork, he may have been gazing over some well-curated systematic collection of molluscan shells [1].  Shells are captivating things, aren’t they?

This month’s installment in my long-running fascination with The Shell began in 2007, when I first glimpsed the phenomenon we ultimately christened “cryptic phenotypic plasticity" [2].  Using a survey of gene frequencies at a pair of highly polymorphic allozyme-encoding loci, John Robinson and I had just discovered that the pleurocerid populations we were identifying as Goniobasis acutocarinata, Goniobasis clavaeformis, and Pleurocera unciale on the basis of their strikingly different shell morphologies apparently constituted a single biological species [3].  Our samples had come from Indian Creek, a tributary of the Powell River on the Virginia/Tennessee line.  And I was curious to see if these intriguing results might extend generally through East Tennessee and into North Georgia, implying that the shell morphological distinction historically used to distinguish the pleurocerid genera Goniobasis/Elimia from Pleurocera might have no heritable basis.

So I picked three fresh study areas – the Little River drainage in the vicinity of Maryville (TN), the Conasauga drainage east of Etowah (TN), and the Coahulla near Dalton (GA).  And from each of these three regions I sampled three populations from the acutocarinata-clavaeformis-unciale-carinifera continuum to analyze together with my original data set from the Powell drainage.  And from each of these (now four) regions I also needed a control – a population of some common, widespread pleurocerid that everybody recognizes, the specific status of which nobody doubts.

Populations of Pleurocera simplex are very nearly omnipresent in small streams throughout Southwest Virginia, East Tennessee, and North Georgia.  They are also rather distinctive, with their dark, smooth, teardrop-shaped shells and strikingly black bodies.  Thomas Say described “Melania” simplex in 1825 from “a stream running from Abingdon to the Salt Works, and from the stream on which General Preston’s grist-mill is situated, as well as in a brook running through the salt water valley and discharging into the Holstein River.”  The eighteenth-century salt mines to which Say must have been referring are still identifiable in modern day Saltville (VA), and (indeed) snail populations matching Say’s small figure and description still inhabit streams in the vicinity.  And Say’s nomen “simplex” is among the oldest names available for any American pleurocerid [4], so is in no danger of being synonymized under anything else.

I included two populations of Pleurocera simplex (as “Goniobasis simplex”) in the first allozyme study I ever published, way back in 1980 [5], and added five fresh simplex populations to control that 2007 gray-literature report with John Robinson I referenced to open the present essay, extending west from the Saltville type locality across the Holston, Clinch and Powell drainages of SW Virginia [6].  Clearly Pleurocera simplex would be the perfect control for my follow-up study of the acutocarinata-clavaeformis-unciale-carinifera continuum now extending south, through East Tennessee and into North Georgia.  Am I right?

And so it was that on 7May08, I came to stand on the banks of Pistol Creek in Courthouse Park, Maryville.  What a lovely little city is Maryville, Tennessee!  Spreading like a cool oasis between the county courthouse and historic Maryville College I found a shady park full of happy picnickers and laughing children.  And Pistol Creek is chock full of Pleurocera clavaeformis acutocarinata, which was the focus of my attention on that sunny spring afternoon, mixed with a population of good old familiar Pleurocera simplex, the perfect control.  I was probably there no more than 30 minutes.  In my field notes I wrote, “Very pretty spot!  Friendly girls drove me off.”

The genetic analysis would, of course, take somewhat longer.  I started running the gels for the study that was ultimately published as “Robust shell phenotype is a local response to stream size in the genus Pleurocera” [7] later that month, extending through the summer.  I pulled the first sample of 7 individuals from the bag labelled “Maryville simplex S6” on July 9, with another 15 individuals on July 11.  The PowerPoint slide below [8] shows photos of two of the nine gels I ran on 7/11/08 – one stained for octanol dehydrogenase (Octl), the other stained for octopine dehydrogenase (Octp).  The 15 Maryville simplex (labeled “S6”) were interleaved that day with ten simplex individuals from someplace else and six P. clavaeformis – don’t worry about those other 16 samples.

Much to my wondering eyes, the 15 “simplex” from Maryville S6 proved to comprise 9 homozygotes for Octl104 and Octp98, and 6 homozygotes for Octl100 and Octp96, with no heterozygotes in evidence at either locus.  There were also striking differences between the set of 9 and the set of 6 at the PGM locus.  I wrote in my lab notebook, “News Flash!!  There are TWO species inside Maryville S6.  Criminy!” 

Over the course of the next several runs, the sample of 30 snails I had collected from Maryville on the afternoon of 7May08 ultimately proved to include 17 of the one species and 13 of the other.  And almost immediately an important follow-up question began to nag me.  Might there have been some subtle morphological difference between these two sets of snails I had initially lumped together as Pleurocera simplex?  What about the shells?  Were they absolutely indistinguishable?  Alas, I had cracked the shells and thrown them all away when I froze my “Maryville simplex S6” sample back in May, following my standard practice.  

So in August of 2008 I returned to Maryville Courthouse Park for a second sample.  And on this second visit, I examined the shells much more critically.  Standing ankle-deep in Pistol Creek, I developed the impression that significant variation might exist in the simple shell proportions of the snails I had previously lumped together as P. simplex, particularly with respect to the relative heights of their body whorls.  Although the differences were slight – indeed negligible in juveniles and subadults – it seemed possible to me that the creek might be inhabited by an admixture of two snail populations with dark, smooth, teardrop-shaped shells and strikingly black bodies, one with “fat” shells and the other with “skinny.”

So, for the 71 Maryville snails I carried back to Charleston the next day, I measured the maximum shell dimension (or "shell height"), and body whorl height (B), defined as the length of the final 360⁰ of whorl, along the axis of coiling. I then defined apex height (A) as shell height minus body whorl height, and analyzed the relationship between body whorl height and apex height by analysis of covariance using the separate slopes model (JMP version 7).  I then classified my fresh sample of 71 individuals by their phenotype at 10 allozyme-encoding loci using standard methods.

A total of 20 snails proved homozygous for Oldh100, while 51 were homozygous for Oldh104, with no putative heterozygotes again in evidence. Differences were also very marked at the Opdh and PGM loci, although a few heterozygotes were observed in both groups.  The combined sample of 20 August snails plus 17 snails from the May sample showed Opdh96 = 0.946 and Pgm96 = 0.946, and the combined sample of 51 + 13 showed Opdh98 = 0.953 and Pgm102 = 0.852.  Again no variation was detected at the seven additional genetic loci examined.

The figure above compares the regressions of (A) on (B) for the two subsamples, the N = 20 fixed for Oldh100 and the N = 51 fixed for Oldh104.  Sure enough, the regressions are quite significantly different [9].  I provisionally designated the (N = 20) snails fixed for Oldh100 as population S6f, with f appended for “fat,”, and the (N = 51) snails fixed for Oldh104 as population S6s, for “skinny.”

One of these populations must surely match Thomas Say’s bona fide simplex from that “brook running through the salt water valley and discharging into the Holstein River,” I should hope!  But which one?  And what might be the identity of the other population?  Tune in next time!


[1] Charles Darwin, Freshwater Malacologist [25Feb09]

[2] I originally christened the phenomenon “Goodrichian Taxon Shift” in February of 2007, focusing entirely upon freshwater snails.  Stephen Jacquemin, Mark Pyron, and I broadened the concept into “cryptic phenotypic plasticity” in a (2013) paper we published in Hydrobiologia [PDF].  Most of the (now 17) blog posts filed under “Phenotypic Plasticity” in the blog index at right above touch upon the pervasive phenomenon of CPP in freshwater gastropod shells.  But see particularly:
  • Goodrichian Taxon Shift [20Feb07]
  • Pleurocera acuta is Pleurocera canaliculata [3June13]
[3] Dillon, R. T. & J. D. Robinson (2007b)  The Goniobasis ("Elimia") of southwest Virginia, II.  Shell morphological variation in G. clavaeformis.  Report to the Virginia Division of Game and Inland Fisheries.  12 pp.  [PDF]

[4] Thomas Say published four pleurocerid names in 1825: simplex, proxima, subglobosa, and fluvialis.  There are eight older pleurocerid names in the literature: virginica (Gmelin 1791), carinata (Brug. 1792), verrucosa (Say 1820), armigera (Say 1821), canaliculata (Say 1821), praerosa (Say 1821), catenaria (Say 1822), and carinifera (Lam. 1822).

[5] Dillon, R.T. and G.M. Davis (1980) The Goniobasis of southern Virginia and northwestern North Carolina: Genetic and shell morphometric relationships. Malacologia 20: 83-98. [PDF]

[6] Dillon, R. T. & J. D. Robinson (2007a)  The Goniobasis ("Elimia") of southwest Virginia, I.  Population genetic survey.  Report to the Virginia Division of Game and Inland Fisheries.  25 pp.  [PDF]

[7] Dillon, R. T. (2011)  Robust shell phenotype is a local response to stream size in the genus Pleurocera (Rafinesque 1818). Malacologia 53: 265-277. [PDF]  I featured these results in two blog posts:
  • Mobile Basin III: Pleurocera Puzzles  [12Oct09]
  • Goodbye Goniobasis, Farewell Elimia [23Mar11]
[8]  This slide comes from a presentation I made at the 2011 AMS meeting in Pittsburgh:
Dillon & Robinson (2011) When Cometh Our Reformation?  Molecular typology meets population genetics in the pleurocerid gastropod fauna of east Tennessee.

[9] The regressions of A = 0.34B + 1.96 (r = 0.69) for the Oldh100 group and A = 0.57B + 1.02 (r = 0.74) for the Oldh104 group differed significantly in their slopes, although not in their intercept. Separate-slopes ANCOVA returned a value of t = -7.14 (p < 0.0001) testing for a difference between the two groups, the group of 20 Oldh100 snails showing a very significantly lower apex (holding body whorl constant) than the group of 51 Oldh104 snails.

Thursday, August 18, 2016

The Classification of the Hydrobioids

I confess that simply composing a header for the present blog post was a minor challenge for me.  How should I entitle an essay about a great big group of little tiny snails that really isn’t even a group?  The noun “hydrobioid” (without taxonomic status) has found widespread application in the literature [1], but if “tiny little odd-lot leftovers” could be Latinized into something that sounded fancy enough, that’s the word I’m fishing for.

The non-group of snails that will be the subject of this month’s blog post are the freshwater representatives of the Superfamily Rissooidea, vanilla gastropods bearing cusps around the base of their median radular tooth.  They give the impression of being smaller-bodied than is typical for the Class Gastropoda, although I’m not sure that’s true, and seem to include an unusually high proportion of shallow water, intertidal, or amphibious taxa.  Sexes are separate, as is true for most gastropods, the penis arising from behind the head.

And I should hasten to add the disclaimer that hydrobioids are not “leftovers” in the American West.  Throughout the Great Basin and Pacific drainages, hydrobioids can be the dominant (sometimes only) element of the freshwater gastropod fauna.  Their endemicity is legendary, and conservation concerns legitimate [2].  But here in The East, we naturally tend to focus on the pulmonates and the pleurocerids and the viviparids, and maybe, at the end of the afternoon, we might find a couple sullen little hydrobioids sucked onto a stick.

Through most of the 20th century, classifications of the hydrobioid taxa have typically recognized no more than five families worldwide.  Essentially all workers have always recognized the (huge, diverse) Hydrobiidae of Troschel (1857), to which Thiele (1929) added the Old World Micromelaniidae and the marine Rissoidae, for example.  Wenz (1939) also recognized three families: the Hydrobiidae, the Micromelaniidae, and the Truncatellidae, including Pomatiopsis.  Any of my readership with a taste for taxonomic arcana are referred to the 1993 monograph of Kabat and Hershler [3] for a tabulation of 16 different, and often strikingly discordant, 19th and 20th century hydrobioid classification systems.
Table 1 of Kabat & Hershler (1993)
The FWGNA project, at our birth in 1999, adopted the broad-sense definition of the Hydrobiidae that arose from Kabat & Hershler’s scholarly review.  So for the last 18 years we have recognized here in North American freshwaters the Bithyniidae (Gray 1857) with just one species, the Pomatiopsidae (Stimpson 1865) with just a couple species, and a huge, diverse Hydrobiidae with a zillion tiny little odd-lot leftovers.

But the dawn of molecular phylogeny was rendering Kabat & Hershler’s broad-sense concept of the Hydrobiidae obsolete even as we were adopting it.  In 1998 an international team of researchers headed up by Tom Wilke and George Davis began publishing the first gene trees with hydrobioid taxa at their tips [4].  The analysis of the Wilke team suggested that Troschel’s old family Hydrobiidae was polyphyletic, implying that many of the taxa previously considered subfamilies underneath the Hydrobiidae deserved promotion to the full family rank.

In retrospect, we might have updated our hydrobioid classification back there as early as 2001, and probably at least once or twice again in the mid-2000s as well.  I confess that I’ve just been letting the pot simmer.  But at some point one must serve.

So by 2013, the Wilke team, now including our good friends Bob Hershler, H-P Liu, and Winston Ponder among others, had “pushed their short DNA fragments to the limit” [5].  Using a concatenation of two mitochondrial genes and one nuclear gene (CO1, 16s, and 18s) Wilke and colleagues classified individual representatives from 90 mostly [6] genus-level rissooidean taxa into approximately 21 family-level groups [7].  The most important result, from our standpoint in the FWGNA kitchen, is that the old Troschel concept of a vast, inclusive Hydrobiidae has been boiled away.

From this day henceforth the old subfamily Amnicolinae (Amnicola, Lyogyrus) stands promoted to the Amnicolidae, the old Lithoglyphinae (Gillia, Somatogyrus, Holsingeria) is the Lithoglyphidae, and the old Cochliopinae (Littoridinops, Pyrgophorus) is the Cochliopidae. Our unwelcome guest from New Zealand, Potamopyrgus, is henceforth segregated into the separate-but-equal Tateidae.  Left behind in the Hydrobiidae pot (s.s.) is the subfamily Nymphophilinae, which includes Marstonia, Floridobia, Notogillia and Spilochlamys.  An updated FWGNA website went online this morning to reflect all this taxonomic churn.

The subfamily Fontigentinae has also been left behind in the old Hydrobiidae pot, but this ingredient should be considered especially volatile.  The single Fontigentine species analyzed by Wilke and his colleagues, a Fontigens nickliniana sample from Michigan, actually clustered more closely with the European Bithyinellidae and Emmericiidae than the Hydrobiidae (ss).  But I get the impression that the Wilke team really didn’t feel as though they had enough information to deal with little, local exceptions such as Fontigens seems to be, and ran short of patience.  So the pot probably isn’t quite off the stove just yet.

I shouldn’t fail to mention that, although the Wilke classification system was based entirely upon molecular evidence, their 2013 paper did feature an extensive Appendix B tabulating 50 morphological and anatomical characters for 13 “selected” rissooidean families.  All of the freshwater families mentioned above are included in this useful resource – the old Pomatiopsidae and Bithyniidae as well as the new Amnicolidae, Lithoglyphidae, Cochliopidae and Tateidae, and the condensed Hydrobiidae (ss).  It will be helpful to have the 1998 review work by Hershler & Ponder [8] open on your desk if you want to dig through Wilke’s Appendix B.

As is universally expected for studies of this sort, Wilke and colleagues concluded their 2013 paper by calling for “additional analyses based on more and/or longer gene fragments.”  Additional samples of tiny little odd-lot leftovers are not (apparently) wanted.


[1] I originally thought that the term “hydrobioid” was first proposed in 1979 by my mentor, G. M. Davis, in the same monograph in which he proposed elevating the subfamily Pomatiopsinae (Stimpson 1865) to the family level. But my good buddy Gary Rosenburg has more recently called my attention to several much earlier uses of the term, including Pilsbry, H. A. (1896) Notes on new species of Amnicolidae collected by Dr. Rush in Uruguay.  Nautilus 10: 86 - 89.

[2] I have previously published several posts on the conservation of western hydrobioids, most recently:
  • Megapetitions of the Old West [14July09]
I also did a 2005-06 series on Pyrgulopsis robusta, culminating with:
  • FWS Finding on the Idaho Springsnail [4Oct06]
[3] Kabat, A. R. & R. Hershler (1993) The prosobranch snail family Hydrobiidae (Gastropoda: Rissooidea): Review of classification and supraspecific taxa.  Smithsonian Contributions to Zoology 547: 1 – 94.

[4]  Davis, G.M., Wilke, T., Spolsky, C., Zhang, Y., Xia, M.Y., Rosenberg, G. (1998)  Cytochrome oxidase I-based phylogenetic relationships among the Hydrobiidae, Pomatiopsidae, Rissoidae, and Truncatellidae (Gastropoda: Prosobranchia: Rissoacea). Malacologia 40, 251–266.  Wilke, T., Davis, G.M., Gong, X., Liu, H.-X. (2000)  Erhaia (Gastropoda: Rissooidea): phylogenetic relationships and the question of Paragonimus coevolution in Asia. Am. J. Trop. Med. Hyg. 62, 453–459.  Wilke, T., Davis, G.M., Falniowski, A., Giusti, F., Bodon, M., Szarowska, M. (2001)  Molecular systematics of Hydrobiidae (Mollusca: Gastropoda: Rissooidea): testing monophyly and phylogenetic relationships. Proc. Acad. Natl. Sci. Phila. 151, 1–21.

[5] Wilke T., Haase M., Hershler R., Liu H-P., Misof B., Ponder W. (2013)  Pushing short DNA fragments to the limit: Phylogenetic relationships of “hydrobioid” gastropods (Caenogastropoda: Rissooidea).  Molecular Phylogenetics and Evolution 66: 715 – 736.

[6] Our buddy Tom and his extensive network of colleagues apparently all subscribe to the U1S2NMT3 rule – each genus usually being represented by one species, sometimes two, never more than three.

[7] The Wilke team listed 21 rissooidean nomina in the column labelled “classification” of their Appendix 1.  Of these, 18 ended with the “idae” suffix suggesting that the team endorses family-level status, two nomina ended in “inae,” and one was simply called a “group.”  The Wilke team also excluded several obscure family-level rissooidean taxa from their analysis entirely, for various reasons.

[8]  Hershler, R. & W. F. Ponder (1998)  A review of morphological characters of hydrobioid snails.  Smithsonian Contributions to Zoology 600: 1 – 55.

Tuesday, July 12, 2016

Pleurocera clavaeformis in the Mobile Basin?

This is the final installment of my five-part series [1] reviewing an excellent recent paper by Nathan Whelan and Ellen Strong [2] on DNA sequence variation within and among an assortment of pleurocerid populations sampled from North Alabama.  The most important contributions made by Whelan & Strong will turn out to be their large and fine data set on mitochondrial superheterogeneity, and their demonstration that no morphological characters seem to correlate with it, putting to rest the hypothesis of cryptic speciation.  Their paper also sheds new light on evolutionary relationships among the pleurocerid populations of the Mobile and Tennessee drainages, if one is able to contextualize the information they have developed.  Such will be our business in the essay that follows.

Whelan and Strong sampled five populations of Leptoxis picta (which they identified as L. “ampla”) from tributaries of the Alabama River draining south into the Mobile Basin and one population of the widespread and well-studied Leptoxis praerosa sampled from a tributary of the Tennessee River, draining west.  Despite the strikingly high incidence of mitochondrial superheterogeneity demonstrated by the five picta populations, the praerosa population remained genetically distinctive, demonstrating about 13% CO1 sequence divergence from modal picta.  This observation dovetailed neatly with the (nuclear) histone H3 sequence data also developed by Whelan & Strong, and with a much larger (9 locus, 11 population) allozyme data set published by Chuck Lydeard and myself in 1998 [3], all of which combined to strongly reinforce the validity of the specific distinction between L. picta populations inhabiting the Mobile Basin and L. praerosa inhabiting the Tennessee.

But such did not seem to be the case with the five populations Whelan & Strong identified as “Pleurocera prasinata” from tributaries of the Mobile Basin, when compared with the population they identified as “Pleurocera pyrenella” from a tributary of the Tennessee.  As we brought our essay of 3May to a close, we noticed that the histone H3 data did not return any evidence of a genetic distinction between any of the six Pleurocera populations whatsoever.

And indeed the larger dataset developed by Whelan & Strong on CO1 sequence divergence did not demonstrate a distinction between any of the six North Alabama Pleurocera populations either.  The gene tree above shows that the modal CO1 haplotype sequenced from the Tennessee “pyrenella” population (“clade 6”) was just 2% different from both the modal prasinata clade 8 and the first-runner-up prasinata clade 7 [4].  But the gene tree below shows that CO1 sequence divergence within the Tennessee “pyrenella” population ranged up to 8% [5].

Nor indeed is there any significant morphological distinction among Whelan & Strong’s six Pleurocera populations.  Last month we reviewed the general topic of ecophenotypic plasticity in shell morphology, as it applies to the Pleurocera populations of North Alabama.  It was our strong impression that the Tennessee drainage Pleurocera population sampled by Whelan & Strong, referred by them to “Pleurocera pyrenella,” was misidentified.  In fact, Whelan & Strong’s Figure 7, reproduced in last month’s post, suggested that all six North Alabama Pleurocera populations are morphologically indistinguishable from Pleurocera clavaeformis.

Now here is the second-most amazing revelation to come from the Whelan & Strong’s remarkable data set [6].  In order to explain the amazingness of this revelation, however, I’ll need to digress a bit into the basic mechanics of the NCBI GenBank.  And backtrack to 2011.  So bear with me.

Among the most useful and powerful tools on the GenBank menu is “BLAST,” the “Basic Local Alignment Search Tool.”  A user can call up any sequence from the database, say for example KT164088, the first CO1 sequence Whelan & Strong listed in their Table 1 for modal CO1 clade 8 of “Pleurocera prasinata.”  And then simply click the BLAST button (optimized for more dissimilar sequences), and wait a minute or two.  And the BLAST tool will return the (default) 100 sequences in GenBank most similar to KT164088.  Amazing.  I never thought I would live to see the day.

So W&S uploaded all 5 x 20 = 100 of their “prasinata” CO1 sequences to GenBank, plus all 20 of their “pyrenella” CO1 sequences, for a total of 120 CO1 sequences from North Alabama Pleurocera [7].  You might think that if you simply picked a representative sequence from the modal “prasinata” CO1 cluster and hit the BLAST button, the search tool would return 100 other North Alabama Pleurocera sequences, wouldn’t you?  But this is not the case.  Down toward the bottom of the big list you will get if you perform the experiment I have outlined above, at about 95% similarity with KT164088, you will begin to see a sprinkling of P. clavaeformis from East Tennessee.

At this point in the history of the FWGNA Blog, I have referred to my 2011 paper in Malacologia – the one that synonymized Goniobasis & Elimia under Pleurocera [8] – so often that I imagine you all are sick of it [9].  That was an allozyme paper, involving 9 populations of P. clavaeformis from East Tennessee, 4 control populations of P. simplex, and a pair of carinifera/vestita populations from North Georgia, more about which later.  The summer that paper was published, John Robinson and I sequenced single individual CO1 haplotypes from each of those same 15 populations, in connection with an oral presentation we were planning for the 2011 AMS meeting in Pittsburgh.  We uploaded those 15 sequences to GenBank promptly, although our results were not ultimately published until quite recently [10].  The gene tree looks like this: 
The point that John and I were making in our little note had to do with the hazards of DNA barcoding.  A researcher who naively sets 5% sequence similarity as a cut point for species recognition, for example, will imagine that our 15 populations comprised 10 species, lumping only the 6 P. clavaeformis bearing the modal CO1 haplotype at the top of the figure.  And even at 10% sequence similarity, a naïve researcher would recognize 8 species, where (at most) only 4 exist.

Now here’s a mental exercise for you.  Imagine Figure 3 above rotated 180 degrees.  It will actually fit almost like a lock-and-key into the top half of Whelan & Strong’s Figure 2, like this: 

The modal CO1 sequence that Dillon & Robinson obtained from 6 of our 9 P. clavaeformis populations in East Tennessee is approximately 95% similar to the modal “prasinata/pyrenellum” sequence that W&S obtained from North Alabama.  And in fact, one of our outlying East Tennessee sequences (JF837315) is approximately 95% similar to a bunch of the outlying North Alabama sequences.  The bottom line is that all six North Alabama Pleurocera populations are genetically indistinguishable from East Tennessee P. clavaeformis.

I first preached a series of sermons on this topic back in 2009 [11], but The Spirit has moved me into the pulpit again.  The evolutionary relationships between the malacofaunas of the Tennessee River and the Mobile Basin are much closer than anyone has ever realized.  Nineteenth-century taxonomic tradition has always held that the two pleurocerid faunas are completely disjoint, sharing no species whatsoever.  But the shell morphological intergradation we see in East Tennessee from acutocarinata to clavaeformis to unciale to curta is strikingly parallel to the intergradation from carinifiera to vestita to praesinata to foremani observed in the Mobile Basin.  That was my rationale for analyzing a carinifera/vestita set from North Georgia together with the three sets of acutocarinata/clavaeformis/unciale I sampled from East Tennessee in 2011.  It seemed possible to me that all 11 populations might prove conspecific.

On the basis of the ten polymorphic allozyme-encoding loci I analyzed in 2011, the genetic relationship between the two North Georgia populations and the nine East Tennessee populations was ambiguous.  As were the CO1 results I subsequently obtained with John Robinson, reproduced above.  Ultimately I decided to refer to the entire basket-full of Tennessee/Mobile Basin populations sharing this similar, slippery shell morphology as the “carinifera group,” and set the matter aside [12, 13].

The results of Whelan & Strong have reawakened the issue, and brought it roaring back to the fore.  In North Alabama, apparently Pleurocera “prasinata” of the Mobile Basin is genetically and morphologically indistinguishable from Pleurocera clavaeformis of the Tennessee.  This is big news [14].

Science is the construction of testable hypotheses about the natural world.  So I will concluded my long, rambling series of essays on the North Alabama Pleuroceridae with some science.  Pleurocera prasinata populations are moderately common in the main Coosa River and in scattered larger tributaries downstream from the North Georgia populations of P. carinifera and P. vestita I sampled for my 2011 paper [15].  I hypothesize that Coosa River populations of P. prasinata are more genetically similar to North Georgia carinifera and vestita than they are to the Cahaba populations of P. prasinata sampled by Whelan & Strong.  I challenge any of my colleagues with research interests in the Alabama Pleuroceridae to test my hypothesis, using any genetic tools at their disposal.  We’ll be standing by.


[1] Previous installments in this series:
  • Mitochondrial Superheterogeneity: What we know [15Mar16]
  • Mitochondrial Superheterogeneity: What it means [6Apr16]
  • Mitochondrial Superheterogeneity and Speciation [3May16]
  • The Shape-shifting Pleurocera of North Alabama [2June16]
[2] Whelan, N.V. & E. E. Strong (2016)  Morphology, molecules and taxonomy: extreme incongruence in pleurocerids (Gastropoda, Cerithiodea, Pleuroceridae). Zoologica Scripta 45: 62 – 87.  Open Access: [html]

[3] Dillon, R.T., and C. Lydeard (1998) Divergence among Mobile Basin populations of the pleurocerid snail genus, Leptoxis, estimated by allozyme electrophoresis.  Malacologia. 39: 111-119. [PDF]

[4] The estimated 2% sequence divergence values shown in Figure 1 were obtained by blasting a randomly chosen modal (clade 6) pyrenella CO1 sequence, KT164127, against randomly-chosen Clade 7 sequence KT164125 and Clade 8 sequence KT164088.

[5] The estimate of 8% mtDNA sequence shown in Fig 2 above was obtained by blasting modal (clade 6) CO1 sequence KT164127 against clade 3 CO1 sequence KT164138.

[6] The first-most amazing thing is CO1 sequence KT163940, that single blue dot showing at Red-star-14 in the final figure of my May post.

[7] Well, to be precise, one individual was excluded by misidentification, and no CO1 sequence was obtained for 6 others.  So 113 North Alabama Pleurocera CO1 sequences, to be precise.

[8] Dillon, R. T. (2011)  Robust shell phenotype is a local response to stream size in the genus Pleurocera (Rafinesque 1818).  Malacologia 53: 265-277. [PDF]

[9] Goodbye Goniobasis, Farewell Elimia [23Mar11]

[10] Dillon, R. T. Jr, and J. D. Robinson (2016)  The hazards of DNA barcoding, as illustrated by the pleurocerid gastropods of East Tennessee.  Ellipsaria 18(1): 22-24. [PDF]

[11] My 2009 series on genetic relationships in the Mobile Basin Pleuroceridae:
  • Mobile Basin I: Two pleurocerids proposed for listing [24Aug09]
  • Mobile Basin II: Leptoxis lessons [15Sept09]
  • Mobile Basin III: Pleurocera puzzles [12Oct09]
  • Mobile Basin IV: Goniobasis WTFs [13Nov09]
[12] As I pointed out in both my essay of 12Oct09 and in my 2011 paper, the oldest name in the basket seems to be Lamarck’s (1822) Melania carinifera.  Lamarck gave his locality as “pays des Chérokées, dansun ruisseau qui se jette dans la rivière d'Estan-Alley,” but I don’t think any English speaker in 200 years has ever had a clue where “la rivière d'Estan-Alley” might be.  Binney (1864) simply abbreviated Lamarck’s locality as “Cherokee Country,” which Tryon (1873) understood to mean Cherokee County, Georgia.  Cherokee county lies in North Georgia’s Etowah River valley, draining SW toward the Alabama/Coosa and the Mobile Basin.  So both Goodrich (1941) and Thompson (2000) followed Tryon in restricting carinifera to tributaries of the Mobile Basin, and I don’t see any reason to question that call here in 2016.  I don’t think any pleurocerid populations looking like Lamarck’s carinifera actually inhabit Cherokee County today, but that whole region has been impacted by the Atlanta sprawl.  Pleurocera populations matching Lamarck’s carinifera do indeed inhabit Etowah tributaries in other North Georgia counties nearby, such as Whitfield, from whence my 2011 population was sampled.

[13] So following the logic above, I seriously considered entitling the present essay “Pleurocera carinifera in the Tennessee Basin,” not “Pleurocera clavaeformis in the Mobile Basin.”  But ultimately I decided that this entire topic is already sufficiently complex and obscure.  There aren’t five people in the world who know what I mean when I say “Pleurocera clavaeformis,” and if I swapped over to Pleurocera carinifera, I’d confuse even those.

[14] Here’s another mental exercise.  In what sense of the adjective “big” is this news big?

[15] Here I’m looking at a 1993 gray literature report submitted to the Alabama Natural History Program by Art Bogan and Malcolm Pierson entitled, “Survey of the Aquatic Gastropods of the Coosa River Basin, Alabama: 1992.”

Thursday, June 2, 2016

The Shape-shifting Pleurocera of North Alabama

The Tennessee River south of Chattanooga reminds me of Danica Patrick on three tires.  Swerving wildly out of the fertile Valley & Ridge Province of East Tennessee, she slices through the Cumberland Plateau at Walden Ridge, bounces off the NW Georgia guard rail, and plunges into the Interior Plains of North Alabama.  The character of the freshwater gastropod fauna changes slightly but perceptibly – not so much in the main river as in the tributaries.  The little black Pleurocera simplex populations, so common in springs and small streams of East Tennessee, disappear west of Chattanooga, to be replaced by populations of the larger and more impressively-shelled P. laqueata.

The most conspicuous element of the Tennessee River macrobenthos through this entire region is Pleurocera canaliculata (Say, 1821) bearing shells of the typical, robust form.  The smaller tributaries are inhabited by a more slender and lightly-shelled ecophenotypic variant, originally described by T. A. Conrad in 1834 as “Melania pyrenellum.”  In 2013 Stephen Jacquemin, Mark Pyron and I presented evidence that the distinction between Conrad’s pyrenellum in small tributaries and Say’s canaliculata in the main Tennessee River is “cryptic phenotypic plasticity,” intrapopulation morphological variance so extreme as to prompt an (erroneous) hypothesis of speciation [1]

Fig 1. Typical Pleurocera canaliculata from the Tennessee R at Site #1.
The sample of typical P. canaliculata canaliculata depicted in Figure 1 above was collected from the Tennessee River at Decatur, Alabama – marked as site #1 on the map below [2].  Stephen, Mark, and I showed that this population is conspecific with the population of P.canaliculata pyrenellum depicted in Figure 2 below, collected from Limestone Creek at Site #2.  I reviewed the evolutionary biology of P. canaliculata throughout eastern North America generally in a pair of related essays posted on this blog back in June of 2013 [3].

Fig 2.  Pleurocera canaliculata pyrenellum from Limestone Creek at Site #2.
The situation is similar in the case of P. clavaeformis, another notorious shape-shifter.  Pleurocera clavaeformis is the most common freshwater gastropod in East Tennessee, ranging from small creeks into large rivers, although not inhabiting the main Tennessee River itself.  Populations of clavaeformis bearing more robust shells in the larger rivers were for many years identified (erroneously) as the distinct species Pleurocera unciale and P. curtum [4].  I first documented cryptic phenotypic plasticity in East Tennessee populations of P. clavaeformis in February of 2007, returning to the subject again in October 2009 and March 2011 [5].

Fig 3.  The Tennessee River west of Knoxville.
Populations of P. clavaeformis are common in several of the Tennessee tributaries of North Alabama, including the Elk, Flint, and Paint Rock Rivers.  These populations generally seem to bear robust shells, however, and have historically been identified as “Pleurocera curtum.”  Goodrich, in fact, stated that Pleurocera (Goniobasis) clavaeformis “does not make the westward turn around Walden Ridge” because he was unaware of any North Alabama populations bearing shells of the lighter, more typical form inhabiting streams of the Cumberland Plateau or Interior Plains [6].

I myself cannot boast of extensive field experience in the state of Alabama.  But I collected the mixed sample of P. clavaeformis and P. canaliculata shown in Figure 4 below from the Paint Rock River SE of Huntsville in March of 1988 (marked as Site #4 above).  The shell variation was rather dramatic in the P. clavaeformis population (top row), ranging from medium-typical to robust-curtum.  The P. canaliculata population (bottom row) was more uniformly heavy-shelled, ranging only from skinnyish-typical to plain-vanilla-typical.

Fig 4.  Pleurocera clavaeformis and P. canaliculata from the Paint Rock at Site 4.
Most significantly, notice that my 1988 sample of Pleurocera from the Paint Rock River did not include any shells demonstrating the pyrenellum morphology, as depicted in Figure 2.  Phenotypic plasticity manifests itself most cryptically where small streams like Limestone Creek empty directly into large rivers, like the Tennessee.  In such situations, the population bearing the more gracile shell morphology will differ strikingly from the population bearing the robust shell, sometimes even mixing unconformably at the mouth of the creek, looking like reproductively-isolated biological species.  But in rivers that grow larger gradually, such as the Paint Rock, the shell morphology of the pleurocerid population changes gradually, and taxonomists are usually not fooled.

It is perhaps for this reason that Goodrich [6] did not report Pleurocera of the pyrenellum form from the Paint Rock River drainage, or indeed from any of the larger tributaries of the Tennessee in North Alabama, such as the Flint or the Elk.  Goodrich gave the distribution of pyrenellum as “tributaries of the Tennessee River in Morgan and Limestone Counties” downstream from the Paint Rock, “and Walker County, Georgia” upstream from the Paint Rock, but not in Madison or Jackson Counties, through which the Paint Rock River runs.  The canaliculata population of the Paint Rock subdrainage just looks like typical canaliculata.  It doesn’t fool anybody.

So for the last three months, we’ve been studying (in excruciating detail) the recent paper by Nathan Whelan and Ellen Strong on mtDNA sequence divergence among the pleurocerid populations of North Alabama [7].  Previously our attention has focused on populations of Leptoxis [8].  But Whelan & Strong also sampled five populations they identified as “Pleurocera prasinata” from tributaries of the Cahaba River draining south through Alabama into the Mobile Basin, and one population they identified as “Pleurocera pyrenellum” from a tributary of the Paint Rock River in Jackson County [9].  That Paint Rock tributary site is marked as Site A on the map above.

Whelan & Strong's Fig 7.  Row C shows putative "Pleurocera pyrenella."  
Whelan & Strong’s Figure 7 is reproduced above, showing example shells from their sample of “Pleurocera pyrenellum” on Row C, at the bottom.  These specimens do not look like the pyrenellum form of P. canaliculata to me.  The shells borne by pyrenellum are flat-sided, almost trapezoidal, essentially demonstrating no suture or indeed any whorl, as shown in Figure 2.  North Alabama pyrenellum populations also typically bear shells marked with spiral cords, especially around the aperture.  But the shells depicted in Row C of Whelan & Strong’s figure seem to lack spiral cords, and demonstrate the slightly rounded whorls and impressed sutures typical of Pleurocera clavaeformis

I hasten to stipulate that I have no personal observations from the Paint Rock subdrainage as far upstream as the Whelan & Strong sample.  And of course, the entire theme of the present essay has been one of caution regarding the use of shell morphology to distinguish species of pleurocerids.  So what might the mtDNA sequence data tell us about the genetic relationships between Whelan & Strong’s Pleurocera populations (all six of them) and Pleurocera populations sampled from elsewhere in the drainage of the Tennessee?  Tune in next time…


[1] Dillon, R. T., S. J. Jacquemin & M. Pyron (2013)  Cryptic phenotypic plasticity in populations of the freshwater prosobranch snail, Pleurocera canaliculata.  Hydrobiologia 709: 117-127.  [PDF]

[2] Here are the complete locality data for all four sites mentioned in this month’s blog:
  • Site #1 – Tennessee River at Decatur, Morgan County, AL.  (34.6279N; -86.9564W).  This was site “CT” of Dillon et al. (2013).
  • Site #2 – Limestone Creek at Nick Davis Road, 2 km N of Capshaw, Limestone County, AL. 
  • (34.8027N; -86.8163W).  This was site “PT” of Dillon et al. (2013).
  • Site #4 - Paint Rock River at US 431 bridge, 25 km SE of Huntsville, AL.  (34.4992, -86.3905).
  • Site A – Estill Fork at Co. 140 bridge, Jackson Co, AL.  (34.9653, -86.1537).
[3]  Cryptic phenotypic plasticity in P. canaliculata:
  • Pleurocera acuta is Pleurocera canaliculata [3June13]
  • Pleurocera canaliculata and the process of scientific discovery [18June13]
[4] Dillon, R. T. (2011)  Robust shell phenotype is a local response to stream size in the genus Pleurocera (Rafinesque 1818). Malacologia 53: 265-277. [PDF]

[5] Cryptic phenotypic plasticity in P. clavaeformis:
  • Goodrichian Taxon Shift [20Feb07]
  • Mobile Basin III: Pleurocera Puzzles [12Oct09]
  • Goodbye Goniobasis, Farewell Elimia [23Mar11]
[6] Goodrich, C. (1940) The Pleuroceridae of the Ohio River system. Occas. Pprs. Mus. Zool. Univ. Mich., 417, 1-21.

[7] Whelan, N.V. & E. E. Strong (2016)  Morphology, molecules and taxonomy: extreme incongruence in pleurocerids (Gastropoda, Cerithiodea, Pleuroceridae). Zoologica Scripta 45: 62 – 87.  Open Access: [html]

[8]  I have featured the Leptoxis mtDNA data set of Whelan & Strong in three previous essays:
  • Mitochondrial Superheterogeneity: What we know [15Mar16]
  • Mitochondrial Superheterogeneity: What it means [6Apr16]
  • Mitochondrial Superheterogeneity and speciation [3May16]
[9] Our good friends Nathan Whelan and Ellen Strong seemed unaware that Conrad’s pyrenellum is an ecophenotypic variant of Say’s canaliculata, and indeed entirely neglected to cite the (2013) paper by Jacquemin, Pyron, and myself.  I’m sure this was just a simple oversight.

Tuesday, May 3, 2016

Mitochondrial Superheterogeneity and Speciation

Editor’s Note – This is the third essay in my series on the phenomenon of double digit intrapopulation mtDNA sequence variation, which I have begun calling “mitochondrial superheterogeneity.”  The points I plan to advance here will depend upon my crazy “wildebison” model of 6Apr16, which (in turn) depended on the field observations I reviewed in my essay of 15Mar16.

Understood in their proper context, typical mtDNA sequence data sets yield weak, null models of population relationship, not directly relevant to the question of speciation [1], but certainly correlative.  A mitochondrial gene sequence is a perfectly fine single character.  Given a decent sample size per population, typical mtDNA sequence data can reasonably be used to estimate interpopulation genetic divergence, which (especially in combination with data from other sources) is generally correlated with the likelihood of speciation [2].

So in the last couple months we have focused on the mtDNA sequence data collected by Whelan & Strong from five populations of Leptoxisampla” inhabiting tributaries of the Cahaba River, draining south through Alabama into the Mobile Bay [3].  Whelan & Strong also sampled n = 20 Leptoxis praerosa from a tributary of the Tennessee River, draining west across the top of Alabama toward the Mississippi River.  Their data set included many striking examples of mitochondrial superheterogeneity.  So while intriguing as an evolutionary phenomenon, doesn’t the occurrence of up to 20% sequence divergence within conspecific populations of “ampla” compromise the utility of mtDNA as a tool for distinguishing “ampla” from praerosa?

No, perhaps just the opposite.  But before I develop that surprising point, let’s back up a few years and review what we actually know about the evolutionary biology of Leptoxis populations in Alabama.

The Leptoxis fauna of the Mobile Basin was last comprehensively monographed by Calvin Goodrich in 1922 [4], prior to his 1934 - 1941 awakening to the Modern Synthesis [5].  The entire pleurocerid fauna of Alabama was then decimated by a massive program of impoundment and channelization, leaving behind what may be the single biggest taxonomic mess anywhere in North American malacology [6].

In 1998, Chuck Lydeard and I published a survey of genetic variation at 9 allozyme-encoding loci in eight populations representing all four nominal species of Leptoxis known (at that time) to persist in the Mobile Basin of Alabama [7], calibrated with three populations of the widespread and well-studied Leptoxis praerosa of Tennessee.  Our Alabama populations included three identified as Leptoxis ampla from the Cahaba drainage [8], two populations identified as L. taeniata from tributaries of the Coosa River, and one population of Leptoxis picta from the main Alabama River way downstream in Monroe County [9].  The allozyme variation among those six populations was comparable to that observed among our three control populations of Leptoxis praerosa, and negligible compared to the genetic divergence between praerosa and the ampla+taeniata+picta group combined.  Thus our (9 gene, 11 population, 333 individual) results suggested that Leptoxis ampla (Anthony 1855) is a junior synonym of L. picta (Conrad 1834) [10].

So returning to the mtDNA sequence data of Whelan & Strong.  It is reassuring to note that, despite the high incidence of superheterogeneity W&S observed scattered among the five Cahaba drainage pleurocerid populations of (what should better be referred to as) Leptoxis picta, their 16S+CO1 gene tree nevertheless depicted the single population of Leptoxis praerosa they sampled from that Tennessee tributary as genetically distinct.

The bottom half [11] of Whelan & Strong’s Figure 2 (reproduced above) shows negligible sequence divergence within their 20-individual sample of L. praerosa, but approximately 13% divergence between modal ("clade 9") L. praerosa and modal (“clade 15”) L. picta [12].  Granted, the sequence variance within populations of L. picta (“ampla”) does indeed range well above 13%.  But a mode-to-mode difference of such magnitude is certainly consistent with an hypothesis of speciation.

And we have thus far given short shrift to the histone H3 sequence data adduced by Whelan & Strong, which is especially interesting because the gene is nuclear, and tends to evolve quite slowly.  Figure 3 of W&S (reproduced below) shows essentially zero sequence divergence within or among any of the five L. picta (“ampla”) populations, but 2.0% divergence between picta and praerosa [13].  

Levels of genetic divergence in allopatry, no matter how striking, do not directly address the question reproductive isolation, and hence can only constitute indirect evidence of speciation.  But the nine allozyme-encoding genes of Dillon & Lydeard, plus the 13% mode-to-mode sequence divergence W&S report for the CO1 gene, plus the 2.0% sequence difference at histone H3, taken together with whatever morphological evidence may have accumulated [14], certainly combine to suggest that the specific distinction between L. picta (aka “ampla”) and L. praerosa is indeed a valid one.

Now begging your indulgence, let me return to Figure 4 of Whelan & Strong – the third time I have reproduced the figure below in three months.  And let’s focus once again on the situation at red-star-14 in the lower left corner, showing that the Shades Creek Leptoxis sample included 16 copies of the modal haplotype and 1 copy of a 12% divergent haplotype, modal in the Cahaba River at US52.  (The red-star numbers in Fig 4 below correspond to the clade numbers in Fig 2 above.)  Last month I suggested that cases of mitochondrial superheterogeneity such as this are the signatures of very rare dispersal events, as though a wildebeest were airlifted into a herd of bison, and successfully interbred.  Seen in this light, the results depicted at red-star-14 constitute direct, positive evidence that the Leptoxis population of Shades Creek is conspecific with the Leptoxis population of Cahaba-US52.

And similarly at red-star-12.  Here we see that the Leptoxis populations of Cahaba-Bibb/Shelby and Cahaba-US82 share a unique, rare CO1 haplotype that Whelan & Strong didn’t discover modally anywhere.  This would seem to constitute direct, positive evidence that the Bibb/Shelby and the US82 Leptoxis populations are conspecific both with some third (unsampled) population where that particular haplotype is much more common, and with each other, as well.

So more broadly, I would suggest that rare, superdivergent mtDNA haplotypes be understood as markers of genetic compatibility with second populations some distance removed.  And at red stars 10, 11, and 13, we see three additional markers of genetic compatibility with three other populations of Leptoxis that Whelan & Strong did not sample.  All of these populations must be conspecific – the populations of Leptoxis picta in which W&S found the markers, and the four other populations elsewhere.  In an ideal world, given complete sampling, it seems possible to me that the phenomenon of mitochondrial superheterogeneity might afford a direct, positive tool to reconstruct specific relationships among extensive sets of highly allopatric populations in their entirety.

Well, perhaps you’d like to join us back here in the real world, Dillon?  The pleurocerid fauna of real Alabama is just a ghost of its former self.  And the Leptoxis populations where the sequences shown at red stars 10, 11, 12 and 13 are modal may well have been extinguished 80 years ago.  Or longer?  And those mtDNA haplotypes might be genetic fossils, of populations long gone to glory?  Interesting to think about.

But turning the page.  As I am sure most of my readership will have by now noticed, the mtDNA survey of Whelan & Strong extended to cover six populations of Pleurocera, sampled alongside the six populations of Leptoxis upon which we have focused these last three months.  Their Pleurocera results offer some striking contrasts with their Leptoxis results, one of which is illustrated in the copy of their Figure 3 I have reproduced above.  Why doesn’t the population W&S identified as “Pleurocera pyrenella” from the Tennessee drainage appear genetically distinct from the five populations they identified as “Pleurocera prasinata” from the Mobile Basin?  Tune in next time!


[1]  Gene trees and species trees are not the same thing:
[2] Examples of the correlative relationship between mtDNA sequence divergence and speciation:
  • Rhodacmea Ridotto [8Aug11]
  • The Lymnaeidae 2012: Stagnalis yardstick [4June12]
[3] Whelan, N.V. & E. E. Strong (2016)  Morphology, molecules and taxonomy: extreme incongruence in pleurocerids (Gastropoda, Cerithiodea, Pleuroceridae). Zoologica Scripta 45: 62 – 87.  Open Access: [html]

[4] Goodrich, C. (1922) The Anculosae of the Alabama River Drainage. Misc. Publ. Univ. Mich Mus. Zool. 7: 1-57.

[5] Goodrich’s (1934 – 1941) “Studies of the Gastropod Family Pleuroceridae” series may represent the earliest application of the modern evolutionary synthesis to malacology – I don’t know of any earlier.  For more, see my essays:
  • The Legacy of Calvin Goodrich [23Jan07]
  • Mobile Basin II: Leptoxis lessons [15Sept09]
[6]  But we can’t blame the entire mess on Alabama Power Company and the Corps of Engineers.  As was true throughout The South, the state of Alabama was blanketed with cotton and other row crop agriculture through the nineteenth century and well into the twentieth.  Bank-to-bank farming dumped tremendous loads of silt into the smaller rivers and streams, which still chokes most of the pleurocerid habitat, even to this day.

[7] Dillon, R.T., and C. Lydeard (1998) Divergence among Mobile Basin populations of the pleurocerid snail genus, Leptoxis, estimated by allozyme electrophoresis. Malacologia 39: 111-119. [PDF]

[8] One of the three Leptoxis ampla sites of Dillon & Lydeard matched one of the five Whelan & Strong sites precisely – the Cahaba River at HWY 52 bridge.

[9] We also included two populations of Leptoxis plicata from the Black Warrior drainage, not relevant to this particular essay.  We did not include the subsequently rediscovered Leptoxis “foremani/downiei” or Leptoxis “compacta.”

[10]  Our good friends Nathan Whelan and Ellen Strong neglected to cite Dillon & Lydeard (1998) in their more recent work on the Alabama Leptoxis under review here.  I’m sure this was just a simple oversight.

[11] The upper half of Whelan & Strong’s Figure 2 depicted the 16S+CO1 genetic relationships among their six populations of Pleurocera.  We’ll return to those data next month.

[12] For this 13% estimate I blasted randomly-chosen L. praerosa CO1 sequence KT164014 against randomly-chosen clade 15 L. picta sequence KT163938.

[13] To obtain this 2.0% estimate, I blasted randomly-chosen L. praerosa H3 sequence KT164460 against randomly-chosen L. picta (“ampla”) H3 sequence KT164387, just as in note [12] above.

[14]  Whelan & Strong did not offer any morphological observations on L. praerosa.  But it is my subjective impression that there may be subtle differences in the whorl shape or spire height between praerosa and picta (“ampla”) – perhaps noticeable only in juveniles, or in adults inhabiting calmer, more protected waters [15].

[15] And it is a crying shame that the captive-rearing study published by Whelan and colleagues (2015) in J. Moll. Stud. 81:85-95 was not controlled for current speed.

Wednesday, April 6, 2016

Mitochondrial Superheterogeneity: What it means

Last month [1] we reviewed, or at least touched upon, 15 separate reports of double-digit mtDNA heterogeneity within gastropod populations published over the last 20 years, the paper by Whelan & Strong [2] among the best and most recent.  Three additional reports from pulmonate snail populations have subsequently come to my attention [3].  The discussion sections of almost all of these works have featured lists of possible explanations for the phenomenon, including population fragmentation, great antiquity, introgression from other species and cryptic speciation, many of which do not seem to fit the data as it has now accumulated, none of which is complete. 

So Rob Dillon’s model to explain mitochondrial superheterogeneity begins with an observation (not an assumption, in the case of pleurocerid snails) of large numbers of super-isolated populations.  Then I add two more super-assumptions: super-long time and super-rare dispersal events.  And let me begin with an extended analogy.

The even-toed ungulates first evolved in the Eocene Epoch, about 50 mybp.  So today, the mtCOI sequence divergence between American bison (JF443195) and the wildebeest of the Serengeti (JX436977) is approximately 15%.  I suggest that when we stand on the bank of the Green River in Polk County, NC, and look down at the population of Pleurocera proxima grazing over the rocks, it is as though we were Lewis and Clark, looking over a vast herd of American bison [4].  And should we drive 80 km west to Buncombe Co, NC, and look down on the Pleurocera proxima population inhabiting Bent Creek, it is as though we were Frederick Russell Burnham standing before the wildebeests. 

Now suppose that every million generations or so, a wading bird with sticky feet transports one snail from Buncombe County to Polk County, as though a modern zookeeper were to airlift a wildebeest to Wyoming.   I don’t know, but it seems likely to me that our jet lagged wildebeest might find itself perfectly well-adapted to graze on the plains of Wyoming in all respects, morphologically and physiologically. 

But it does not seem likely to me that a single wildebeest could reproduce in a herd of American bison.  I imagine a variety of prezygotic reproductive isolating mechanisms have evolved between wildebeest and bison – appearances, smells, behaviors and so forth.  And even if a mating should occur, I feel certain all sorts of postzygotic incompatibilities have evolved - the chromosomes of the two species probably don’t match, a million screw ups occur in development, and nothing comes of it.

So (I admit) the most implausible element of my model is that, in the case of pleurocerid snails, a super-rare immigrant arriving via sticky-bird express must successfully mate within a host population from which it has been isolated for millions of generations.  There are several factors that make this phenomenon marginally less-implausible for pleurocerids, specifically.  Pleurocerid snails are aphallic, and no obvious mating behavior has been reported.  Apparently the male simply crawls over the female and deposits a spermatophore in her gonoduct, so unceremoniously that human observers typically don’t even notice it.  There’s very little light down where the snails live, so appearances don’t matter.  And pleurocerid populations typically inhabit waters with significant flow rates, so it seems unlikely that chemical mating cues could evolve.

Now why postzygotic reproductive incompatibilities have not evolved over millions of generations among such isolated gastropod populations, I do not know.  North American pleurocerids seem to demonstrate negligible anatomical diversity [5], and negligible chromosomal diversity to match [6].  I was tempted to add “super-stable environment” to my list of assumptions in paragraph two at the top of this essay, because it seems possible to me that stabilizing selection may have worked to prevent reproductive isolating mechanisms from evolving.

But I should hasten to add that very little of the reproductive biology I have reviewed directly above applies to pulmonate land snails, and that it was in land snail populations mitochondrial superheterogeneity was first discovered [7].  The bottom line seems to be that, over a wide range of mating systems, populations as isolated as bison and wildebeest have not speciated in gastropod world.  The plains of snail-America and snail-Africa are both grazed by reproductively compatible “wildebison.” 

Rob Dillon’s patented “wildebison model” fits all six of the points I reviewed last month regarding mitochondrial superheterogeneity as a general phenomenon [1].  So let us return to the Leptoxis population inhabiting Shades Creek in northern Alabama, marked in dark blue above.  Whelan & Strong reported 16 copies of the modal CO1 sequence (the “bison” sequence, at red-star-15) and one copy of a 12% divergent sequence, which we’ll call “wildebeest.”  The wildebeest sequence almost exactly matched the modal sequence in the Cahaba River at the US52 bridge, shown in black above (at red-star-14).  And not only did the rare CO1 sequence of Shades Creek match the modal CO1 sequence of the Cahaba River, the rare 16S sequence of Shades Creek matched the modal 16S sequence of the Cahaba River.  The conclusion seems almost inescapable that at some point in the (probably distant) past, a Cahaba wildebeest has been airlifted into the Shades Creek bison population.

The first-most amazing thing about this data set is that an individual Cahaba wildebeest was apparently able to mate into the Shades Creek bison herd, despite millions of generations of isolation.  And the second-most amazing thing is that Whelan & Strong happened to sample both Shades Creek and the Cahaba River, to show us what happened.  In most data sets of this sort, we simply find no clue. 

So for example, in addition to the 17 copies of the modal CO1 sequence Whelan & Strong reported from the Cahaba River at US52, they also reported two copies of a 20% divergent sequence that matches nothing else in their database, way off at red-star-11 [8, 9].  Did that rare mitochondrial genome come from a third population of “wildecamels,” currently grazing on some other continent that W&S did not sample?  Or might that genome have originated in an extinct population of “wilde-Irish-elk,” now gone?   A fascinating question.

I will close with two final notes.  First, I should emphasize that I have not evoked selection at any point during the essay above.  I very nearly used the S-word when I was waving my hands about the apparent absence of reproductive isolation among highly isolated gastropod populations with a surprising range of evolutionary backgrounds, but (ultimately) held off.  The wildebison model is grudgingly neutral.

And second, my good friend John Robinson and I are currently simulating the evolution of mitochondrial superheterogeneity using the “ms” program of R. R. Hudson [10]. Our preliminary results have not yielded the phenomenon in single-population models given any reasonable values of mutation rate (u) and population size (Ne).  Our two-population models do, however, yield mitochondrial superheterogeneity within reasonable values of u and (perhaps-surprisingly low values of) Ne, given (super-rare) migration rates around Nem = 0.0001 genomes per generation.

I find all this quite interesting as a study of evolutionary science, pure as the driven snow, and (I feel sure) a large fraction of my readership does as well.  But in situations such as the pleurocerid fauna of Alabama, research efforts have been at least partly motivated by conservation concerns regarding the distribution of potentially endangered species.  What might mtDNA sequence data sets such as developed by Whelan & Strong, riddled (as they are) with superheterogeneity, tell us about the species relationships of the populations sampled?  Tune in next time!


[1] Mitochondrial Superheterogeneity: What we know [15Mar16]

[2] Whelan, N.V. & E. E. Strong (2016)  Morphology, molecules and taxonomy: extreme incongruence in pleurocerids (Gastropoda, Cerithiodea, Pleuroceridae). Zoologica Scripta 45: 62 – 87.  Open Access: [html]

[3] Pinceel, J, K. Jordaens & T. Backeljau (2005)  Extreme mtDNA divergences in a terrestrial slug (Gastropoda, Pulmonata, Arionidae): Accelerated evolution, allopatric divergence and secondary contact.  J. Evol. Biol. 18: 1264 – 1280.  Parmakelis, A., P. Kotsakoizi & D. Rand (2013)  Animal mitochondria, positive selection and cyto-nuclear coevolution: Insights from Pulmonates.  PLOS ONE 8(4): e61970.  Nolan, J. R., U. Bergthorsson & C. M. Adema (2014)  Physella acuta: atypical mitochondrial gene order among panpulmonates (Gastropoda).  J. Moll. Stud. 80: 388 – 399. 

[4] The populations of Pleurocera I’m using as examples here were sampled in Dillon, R T. and J. D. Robinson (2009)  The snails the dinosaurs saw: Are the pleurocerid populations of the Older Appalachians a relict of the Paleozoic Era?  Journal of the North American Benthological Society 28: 1 - 11. [PDF].  I reviewed the Dillon & Robinson results in my blog post of March, 2009.  See:
  • The Snails The Dinosaurs Saw [16Mar09]
[5] Dazo B (1965) The morphology and natural history of Pleurocera acuta and Goniobasis livescens (Gastropoda: Cerithiacea: Pleuroceridae). Malacologia 3:1–80.   Strong, E., 2005. A morphological reanalysis of Pleurocera acuta Rafinesque, 1831, and Elimia livescens (Menke, 1830) (Gastropoda: Cerithioidea: Pleuroceridae). Nautilus 119: 119–132

[6] Dillon, R.T. (1989) Karyotypic evolution in pleurocerid snails. I. Genomic DNA estimated by flow cytometry. Malacologia 31: 197-203. [PDF]   Dillon, R.T. (1991) Karyotypic evolution in pleurocerid snails. II. Pleurocera, Goniobasis, and Juga. Malacologia 33: 339-344.  [PDF

[7]  I almost cut the two paragraphs about the absence reproductive isolating mechanisms in pleurocerid snails entirely out of this essay.  But then I figured, hell, if the land snail people don’t like it, they can start their own blog.

[8]  For this calculation I used KT164003 as an example as the modal CO1 sequence from the Cahaba River at US52, and KT164004 as an example of the rare, divergent sequence, and compared the two using blastn.

[9] And somewhat amazingly, Whelan & Strong also found one mitochondrial haplotype in the Cahaba River at US52 that matches the Shades Creek mode.  So their Figure 4 (reproduced above) should have shown one little black circle near their big blue circle, as well as that single little blue circle near their big black.  I have added one little black circle at red-star-15 above.  More about this next month.

[10] Hudson, R. R. (2002)  Generating samples under a Wright-Fisher neutral model of genetic variation.  Bioinformatics 18: 337-338.