Dr. Rob Dillon, Coordinator

Thursday, April 12, 2018

REVIEW: Thorp & Covich Fourth Edition

Thorp, J. H., and D. C. Rogers (2016) Keys to Nearctic Fauna.  Thorp and Covich’s Freshwater Invertebrates, Fourth Edition.  Volume II.

“What a marvelous contribution!  I never thought I would live to see the day.”

Both these sentiments oscillated through my mind in rapid succession as I flipped through the pages of Christopher Rogers’ “Class Gastropoda” key in the new edition of Thorp & Covich displayed on the auction table in Raleigh [1] last June.  The taxonomy is complete, thorough, well-researched and modern.  Just three genera of living pleurocerids, check.  Hubendick’s single, broadly-inclusive genus of lymnaeids, check.  Just two genera of physids, check!  Bellamya yes, Planorbella no.  Onward as the pages turned.  What a marvelous contribution!  I never thought I would live to see the day.

This was not the first time such thoughts had flickered through my mind, however.  It had happened twice before.

I imagine that most of my readership will be familiar with the Thorp & Covich tradition.  The first edition of Ecology and Classification of North American Freshwater Invertebrates, as the collaboration was originally called, was published in 1991, with a second edition in 2001 and a third in 2010.  Over this 20-year span Thorp & Covich has become the primary bench reference for biologists working with the (non-insect) freshwater macrobenthic fauna of the United States [2].

Ken Brown was the solo author of the Gastropoda chapter through its first two editions, Chuck Lydeard joining him in 2010 [3].  Ken has an excellent background in freshwater gastropod ecology, which together with anatomy, physiology and (later) gene trees comprised the first 70% of the chapter.  The remaining 30% was allocated to a dichotomous key following the taxonomy of Burch [4] rather closely [5], but I am not criticizing.  That’s all there was.

So the first time that “marvelous-contribution-but-it’ll-never-happen” thought flickered through my brain was in March of 2011, when I opened an email from Jim Thorp.  Jim intimated to me that, even though his third edition was only a single year off the press, he and Christopher were already making plans for a major revision.  He projected that his fourth edition would expand into a series of volumes, the first of which would be dedicated to more ecological themes, the second of which would present dichotomous keys to the Nearctic fauna, the third to the Palaearctic, and so forth [6].  The fourth-edition keys would be much larger and more detailed than those of the third edition, extended down to the species level wherever possible.  And he wanted me to lead the Nearctic Gastropoda team.

I was flattered, of course.  But the series of negotiations into which we entered did not ultimately yield fruit.  I am a big fan of review and synthesis, and I greatly appreciate the value of reference works like Thorp & Covich for biologists in the field today.  But I climb onto the shoulders of giants with a pair of binoculars looking up, not down.  An historian I am not.

So for example, of the roughly 150 nominal species of pleurocerid snails Calvin Goodrich recognized in North American fresh waters [7], forwarded to the present day by Burch, I should estimate that no more than about 30 – 40 are biologically valid.  How am I to construct a dichotomous key to such a biota, I asked Jim, if most of the taxa to be keyed are not legitimately distinguishable?  Could I synonymize?

Jim (quite understandably) demurred.  Christopher suggested that, in cases such as the Pleuroceridae, I simply key down to the lowest firm taxon and drop it, with some language to the effect that “This genus is in need of revision and therefore identifications are left at genus level.”  Well, I thought to myself, all the North American freshwater gastropod genera are in need of revision.  Most have never had a vision in the first place.  Ultimately, I found it impossible to rationalize the expenditure of my own time and effort unless I could move the ball forward [8].

So Volume 1 of the Fourth Edition appeared in 2015, our good friend Mark Pyron’s name appearing above that of Ken Brown as author of Chapter 18, “Introduction to Mollusca and the Class Gastropoda.”  I was pleased to see that the Pyron & Brown contribution extended to 41 pages, up from the 30 pages of the 1991 original.  And sure enough, those 41 pages were focused entirely upon ecology and evolution, no dichotomous key anywhere in evidence.

And the second time that “marvelous-contribution-but-it’ll-never-happen” thought flickered through my brain was later that same year, when Christopher emailed me to request that I review the dichotomous key he himself had prepared for Volume 2.  His draft looked excellent – I really didn’t have a whole lot of suggestions to offer.  In retrospect, I think that from among the entire 3 x 10^8 population of the United States of America, I might have been the worst possible choice to review a freshwater gastropod key.  Because I understood what Christopher was trying to say, and I often skipped ahead, knowing where he was trying to go.

Regarding the problematic groups, I was gratified to see that Christopher had taken the approach he had first advocated during our 2011 negotiations.  Here’s a verbatim quote from his section entitled Limitations: 
“Particularly vexing is the strange confusion in the taxonomic literature, with many taxa accepted or rejected without explanation, and good quantitative taxonomic revisions ignored without explicit justifications.  Because of this issue, the keys are taxonomically conservative, often terminating with species groups rather than species.”
 So Christopher judged such terminations to be necessary for 13 “species groups” in the Nearctic freshwater gastropod fauna all told: the hydrobioid genera Amnicola, Lyogyrus, Clappia and Fluminicola, the hydrobioid families Cochliopidae and Hydrobiidae (ss), assorted prosobranch genera Valvata, Campeloma, Pleurocera/Lithasia, and Leptoxis, and the planorbid genera Helisoma, Carinifex, and Menetus.  Fine.  It’s hard to know which of those thirteen groups is the biggest mess.

The rest of the freshwater gastropods of North America Christopher did, very bravely, key all the way down to the species level.  Among the pulmonates he recognized 10 species of Lymnaea, 9 species of Physa, and 24 species of (non-Helisoma, non-Carinifex, non-Menetus) planorbids [9].  All good estimates – probably pretty close.  From among the entire 3 x 10^8 population of the United States of America, I may well appreciate the effort that Christopher must have put into this contribution the most.

So buy this volume.  Actually, you should back up and buy Volume 1 if you haven’t already, as well as Volume 2.  Do it for three reasons.  First, the gastropod taxonomy is correct, complete, and current.  Second, the new fourth-edition keys complement, but in many cases are strikingly different from, the old third-edition keys.  We’ll develop that theme next time.  And finally, assuage my guilt.  In retrospect, I may have been more a hindrance than a help to Christopher as he labored over this marvelous contribution.  I never thought I would live to see the day.


[1] SFS Raleigh  [4Apr17]

[2] Pennak held that position when I was coming up through the ranks, and is still an excellent reference.  Nothing in the essay above is intended to take anything away from Doug Smith’s (2001) Fourth Edition of “Pennak’s Freshwater Invertebrates of the United States” at all.

[3] Alan Covich and Ken Brown asked me to help with the 2010 edition.  I was flattered, but ultimately declined, for reasons similar to the ones I gave Jim Thorp in 2011.  I can be a difficult person to deal with.

[4] This is a difficult work to cite.  J. B. Burch's North American Freshwater Snails was published in three different ways.  It was initially commissioned as an identification manual by the US EPA and published by the agency in 1982.  It was also serially published in the journal Walkerana (1980, 1982, 1988) and finally as stand-alone volume in 1989 (Malacological Publications, Hamburg, MI). 

[5] The hydrobioid subfamilies were raised to the full family level in the third (2010) edition.

[6] Twelve volumes are now on the drawing board, according to a more recent communication I enjoyed with Jim.  Volume III is now Neotropical Hexapoda, and Volume IV is the Palaearctic Fauna, which was slightly delayed.  But both are currently in press. 

[7] The Legacy of Calvin Goodrich [23Jan07]

[8] My exact words to Jim and Christopher were, “It’s the fourth quarter of my career.  I won’t take a knee on the ten-yard line.”

[9] Ancylid limpets integrated with the Planorbidae.  OK, fine.


  1. Counterpoint: Don't buy this volume if you're interested in pleurocerids. The family needs revision, but lumping Lithasia/Pleurocera/Elimia is ridiculous and based on extremely questionable conclusions using outdated tools and poor taxon sampling.

  2. How else would you explain the observation that four populations of Pleurocera were each more genetically similar to upstream Goniobasis/Elimia than to other populations of Pleurocera?

    1. Homoplasy, which is why allozymes are problematic, and/or outdated and inappropriate analyses of the data (e.g., use of chord distances).

      Also, possible misidentifications.

      But your question reveals the biggest flaw to the argument. You argued for synonymizing Pleurocera and Elimia without sampling even 1/4 of each genus's species, and only across a small part of their range. Thus, you have no evidence that other Elimia and Pleurocera species form a clade.

    2. Hmmm. I am not sure I understand how “homoplasy” could explain my 2011 results. Let us be clear. I sampled four different sites. At site #1, the population of genus A was most genetically similar to the population of genus B, and at site #2 population A2 was most similar to population B2, and A3 was most similar to B3, and A4 was most similar to B4. By “homoplasy” do you mean to imply that genus B evolved from genus A four separate times?

    3. No, I mean that since allozymes are prone to homoplasy they can give an incorrect signal, potentially like what you're describing. You would see the signal you describe, but it would not be explained by shared ancestry, but rather homoplasy/ convergence at allozyme loci.

      But it seems like you're unclear as to what homoplasy is so that explains how you came to the exceptionally weakly supported conclusions you made about Pleuroceridae relationships/taxonomy.

      But it could have also been a taxon sampling problem. With such poor overall taxon sampling of pleuroceridae it would be impossible to make a good assessment of genus taxonomy, regardless of what molecular tool you used.

    4. Here is the way science works. If you think that the phenomenon I documented in 2011 is due to convergence at ten different loci, in 30 individuals per population, in three populations per test, in four independent tests, you are expected to design an experiment to support your hypothesis, and refute mine. Until you do, my hypothesis prevails.

    5. Pot calling the kettle black here, don't you think?

      This blog is riddled with your dataless rants about why other people's hypotheses are wrong

      And honestly, in 2011 you should have been expected to use a modern tool to test any hypothesis, rather than allozymes.

      Until you can demonstrate that you understand what homoplasy is or why allozymes are useless for measuring/testing relationships above the population level, my hypothesis that you don't know what you're talking about still stands.

  3. Here's a pop quiz on "homoplasy." Which is more likely to demonstrate homoplasy - gene frequencies at ten loci or the robustness of a shell?

    1. Both could show homoplasy.

      Your question is like asking "Which is safer, riding in a car without a seatbelt or riding a motorcycle without a helmet?" Both are terrible options if you're trying to travel somewhere safely.

      Not all genetic markers are equally appropriate for testing different hypotheses. Allozymes were never an appropriate marker for above population-level analyses, and by the time you were doing the experiments for Dillon 2011 they were already outdated tools and should not have been used to test any hypothesis.

    2. Can you name a method of estimating phylogenetic relationships that does not also involve a risk of homoplasy?

    3. There's always some risk of homoplasy with molecular markers, but allozymes are much much much more prone to being homoplasious than gene sequences, single-nucleotide polymorphisms, or even microsatellites (but microsats would also have been a sub-optimal tool for genus-level systematics).

      It's obvious what you're trying to do here. You're trying to obscure the problem with allozymes by pointing out the all genetic markers COULD be homoplasious. However, the issue at hand is whether allozymes are an appropriate marker for above population genetics analyses. They are not, and one of the primary reasons is because they are expected to be much more homoplasious than modern tools, especially at the genus level.

      Moreover, distance methods are much more error-prone that likelihood methods for inferring relationships. This is a widely accepted fact.

      Dillon (2011) analyzed allozymes (an inappropriate and outdated genetic marker) with distance methods, coupled with insufficient taxon sampling, and then claimed that all Elimia and all Pleurocera species were more closely related to each other than any other pleurocerid (hence the synonymization). So yeah, I think the synonymization of Elimia into Pleurocera was without evidence and of questionable scholarship.

      How Dillon (2011) ever got through peer-review will probably always be a mystery.

    4. My goodness, I'm impressed! You seem to know exactly what is wrong with Dillon (2011), and have a clear experimental design in mind to fix it. Why don't you do so?

    5. Don't worry, people already are. It just takes a lot longer to do good science than bad science.

    6. Looking forward to it. Keep us posted on the progress...