A raised-letter certificate of appreciation, with wax seal and red ribbon, is hereby awarded our good friend Tom Coote for his important contribution to the general understanding of mtDNA sequence divergence in the freshwater gastropod fauna of North America. The paper he based on his 2011 dissertation , published late last year in Northeastern Naturalist , offers us the best survey of intraspecific mtDNA sequence variation among hydrobioid populations as yet available, anywhere . His results will surprise many and enlighten us all.
Populations of the (relatively large-bodied) Marstonia lustrica can reach locally high densities on rocks and macrophytes in big lakes and other cold, clear bodies of water above the glacial maximum. In recent years they have disappeared from many habitats where they were once common [eg Oneida Lake, 6] and hence have been the object of some conservation concern. Our buddy Tom obtained COI and NDI sequences from 20 M. lustrica populations scattered across six northern states and one Canadian province. He recorded at least one COI sequence from 18 of those 20 populations, for a total of 54 sequences. And he recorded at least one NDI sequence in 14 of the 20, including the two populations that he missed with the COI, for a total of 41 sequences.
Let us first look at Tom’s COI results. That big black triangle at the top of his maximum likelihood COI tree below (labelled “modal”) represents the most common sequence, and variants of no greater than 0.5%, which he recorded in 39 of his 54 individuals. Tom documented at least one copy of the modal sequence in 15 of the 18 populations for which he had any COI data.
|Sample sites, from Coote |
Now look immediately below the black triangle. The three sequences marked “HLMN,” from Harriet Lake, Minnesota, differed from the modal sequence by 1.4%. And the sequences marked KBMI (B and C, from Keeweenaw Bay, in Lake Superior) differed from the mode by 0.5%. And the sequences marked LLMN, from Limestone Lake, MN, differed from the mode by 2.6%, and so forth. The maximum divergence from the mode was 4.5%, recorded by two sequences collected at Grand Island, New York (GINY-A and GINY-D) and the one collected in Indiana’s Tippicanoe River (TRIN). The maximum COI divergence anywhere in Tom’s dataset was 5.7%, between GINY-A and IRMI-A, from Isle Royal in Lake Superior.
Does this mean that Tom Coote has discovered a big swarm of new Marstonia species? Does his TRIN/GINY sequence for example, 4.5% different from modal M. lustrica, sitting on its own branch of the tree, way outside three other good Marstonia species from Alabama and Texas, reveal to us the existence of an undescribed species of hydrobiid lurking in the Niagara River of upstate New York, cryptic under Marstonia lustrica? No, of course not.
Because Tom obtained four COI sequences from the GINY population, and two of them (GINY-B and GINY-C) were modal. The Marstonia lustrica population on the Grand Island shores of the Niagara River is heterogeneous for (at least) two COI sequences – the common one that is found everywhere across six northern states and a rare one, 4.5% different, that matches – if you can believe it – a population in Indiana.
|COI maximum likelihood tree, from Coote |
Well, then. How about that sequence Tom found at Moira Lake way up in Ontario (MLCA-B), the one that seems to match a Marstonia hershleri haplotype collected down in the Heart of Dixie? Maybe Moira Lake is inhabited by an Alabama species? No, of course not. Because again, Tom obtained three COI sequences at Moira Lake, and two of them (MLCA-A and MLCA-C) were modal. The Moira Lake population is heterogeneous for two COI haplotypes differing by 2.4%.
What could that mean? Maybe Tom’s NDI results will cast some light on his peculiar COI data. He did not publish his NDI results separately, preferring to offer us a concatenated COI+NDI network in his Northeastern Naturalist paper. But the figure down below shows the NDI (only) neighbor-joining tree from his dissertation .
Good grief! Sticking way out on a distant limb of Tom’s NDI tree we find the same three sequences that were sticking way out on that distant limb of his COI tree: GINY-A, GINY-D, and TRIN-A. And again, GINY-B and GINY-C are lumped into the mode. And there’s MLCA-B sticking way out as well, again with MLCA-A and MLCA-C in the mode. Could the striking similarity of Tom’s COI and NDI trees be a coincidence?
No, of course not. The two M. lustrica data sets we are comparing here correspond precisely as the COI and 16S data sets that Nathan Whelan and Ellen Strong developed for pleurocerids in Alabama  and published in 2016. And indeed, Dillon & Frankis documented an identical correspondence between COI and 16S haplotypes in Pleurocera proxima of the Southern Appalachians way back in 2004 .
At this point I would invite you to open a new browser window and re-read my blog post of 15Mar16, entitled “Mitochondrial superheterogeneity: What we know.” In that 2016 essay I reviewed the striking patterns of mtDNA sequence divergence reported by Whelan & Strong in North Alabama pleurocerids and compared them to a set of 18 other such studies published for a variety of other gastropod populations both terrestrial and freshwater. Of the six bullet-points that Whelan & Strong confirmed for us (or taught us!) about mitochondrial superheterogeneity (mtSH), let me call your attention to #2, “peculiar patterns of interpopulation haplotype sharing” and #3, “tight coupling in the evolution and distribution of separate mitochondrial genes.” Do those phenomena look familiar?
|NDI Neighbor-joining tree, from Coote |
The patterns in mtDNA sequence divergence in Tom Coote’s M. lustrica populations correspond strikingly to the patterns that Whelan & Strong documented in the Leptoxis populations of North Alabama. The only difference is one of scale: Marstonia lustrica haplotypes only vary to a maximum of 4.5% within populations, while intrapopulation sequence heterogeneity in Alabama pleurocerids ranges up to 20%, and elsewhere in the southern Appalachians up to 21.9% .
Marstonia populations sampled above the glacial maximum are younger than the pleurocerid populations of the American South. I feel sure that M. lustrica, the original population of that species, diverged from whatever its ancestor might have been prior to the Pleistocene, probably up north somewhere, and that it spread, and that isolated populations diverged in a fashion not unlike the pleurocerids. Glacial advance may well have extinguished the species through most of its range, to survive only in isolated refugia. Populations have then subsequently re-colonized the northern latitudes, but not in a smooth wave rolling from south to north. Rather, individual colonists have been airlifted north from genetically divergent refugia in a chaotic and unpredictable fashion, some lakes skipped and other lakes colonized twice.
I would now invite you to open another browser window and re-read my essay of 6Apr16, “Mitochondrial superheterogeneity: What it means.” My “wildebison model” will be easy to recognize in the paragraph above.
And finally, I would invite you to review my essay of 3May16, “Mitochondrial superheterogeneity and speciation.” Under the wildebison model of mtDNA genome evolution, we now understand that results such as Tom Coote has reported for COI and NDI do not call into question the specific identity of scattered Marstonia lustrica populations. Just the opposite. Strikingly divergent haplotypes such as he has labelled GINY-A and TRIN on the distant branches of his gene trees constitute direct, positive evidence that the Marstonia populations inhabiting the Grand Island shores of the Niagara River in upstate New York and the Tippencanoe River of Indiana are conspecific.
One last point, and then a summary, begging your indulgence. All of my 2016 essays were directed toward the phenomenon of mitochondrial superheterogeneity, which I defined on 15Mar16 as follows: “A population can be said to demonstrate “mitochondrial superheterogeneity” when two or more of its members demonstrate 10% sequence divergence or greater for any single-copy mtDNA gene, not sex-linked.” Well, Tom Coote’s Marstonia populations do not display mtSH by that definition. Their maximum intrapopulation divergence (GINY-A and GINY-B) is just 4.5%. So, let us call intrapopulation sequence divergence greater than 2% but less than 10% “simple mitochondrial heterogeneity” (mtH) nothing super about it, shall we? I think future research will find mtH to be the rule in freshwater gastropod populations, not the exception.
Gene trees are dependent variables, not independent. Examined in isolation, they can lead the naïve researcher to terrible misunderstandings about the evolution of the group which he studies. But bringing a robust understanding of the evolution of the group to his study, a wise researcher can peer through the window of a gene tree to find evolutionary processes of great generality and importance.
 Coote, T. W. (2011) The phylogeography of Marstonia lustrica: Understanding the relationship between glaciation and the evolution and distribution of a rare snail. Ph.D. dissertation, University of Massachusetts, Amherst. Open Access Dissertations 399. [html]
 Coote, T. W. (2019) A phylogeny of Marstonia lustrica (Pilsbry 1890) (Gastropoda: Hydrobiidae) across its range. Northeastern Naturalist 26: 672 – 683.
 Well, maybe I should back up and qualify my opening paragraph just a little bit. Tom’s research involved more populations of any single hydrobioid species than have ever been heretofore surveyed, although not more individuals. Tom Wilke and colleagues did publish, back in 2000, a survey of COI sequence variation across ten populations of the medically-important pomatiopsid Oncomelania hupensis sampled from SE Asia, ten individuals per population . The maximum sequence divergence recorded in the Wilke study was a not-insubstantial 2.2%. Here in North America, the largest study of intraspecific mtDNA sequence variation for any hydrobioid published to date was that of Hershler and colleagues , involving 12 Idaho populations of Taylorconcha serpenticola, 5 individuals per population.
 Wilke, T., G.M. Davis, C-E. Chen, X-N. Zhou, X.P. Zhang, Y. Zhang & C.M. Spolsky (2000) Oncomelania hupensis (Gastropoda: Rissooidea) in eastern China: molecular phylogeny, population structure, and ecology. Acta Tropica 77: 215-227.
 Hershler, R., H-P. Liu, T.J. Frest, E.J. Johannes & W.H. Clark (2006) Genetic structure of the Western North American aquatic gastropod genus Taylorconcha and description of a second species. Journal of Molluscan Studies 72: 167-177.
 For more about the lost malacofauna of Oneida Lake, see:
- Harman, W. N. & J. L. Forney (1970) Fifty years of change in the molluscan fauna of Oneida Lake, New York. Limnology and Oceanography 15: 454 – 460.
- Dillon, R. T., Jr. (1981) Patterns in the morphology and distribution of gastropods in Oneida Lake, NY, detected using computer-generated null hypotheses. American Naturalist 118: 83 – 101.
 Whelan, N.V. & E. E. Strong (2016) Morphology, molecules and taxonomy: extreme incongruence in pleurocerids (Gastropoda, Cerithiodea, Pleuroceridae). Zoologica Scripta 45: 62 – 87. With subscription: [html]
 Dillon, R. T., Jr. & R. C. Frankis. (2004) High levels of DNA sequence divergence in isolated populations of the freshwater snail, Goniobasis. American Malacological Bulletin 19: 69 - 77. [PDF].
 Dillon, R. T., Jr. & 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]. For a personal perspective, see:
- The snails the dinosaurs saw [16Mar09]