Reptile & Amphibian Forums

I am responding to some recent discussions (several threads) in which CKing wrote, among other stuff:

“Ignoring taxonomic proposals is indeed an option. For example, Kluge proposed that the genus Chondropython should be synonymized with the genus Morelia because recognizing Chondropython renders Morelia paraphyletic. Both names are available under the rules of the ICZN, but I disagree with Kluge's proposal because it is based on an intolerance of paraphyletic taxa. Such disagreements are based on classificatory philosophy and are well known among taxonomists. Therefore it is unlikely that my disagreemnt with Kluge will be published in a journal. Nevertheless nobody is required by the ICZN to follow Kluge's taxonomic proposal nor is anyone required to publish a paper refuting Kluge. The customary practice is for each person to review the evidence presented by an author and to either accept or reject the proposal based on one's own evaluation of the evidence.”

and

”The rules cannot be used to settle taxonomic disputes between different schools of taxonomy, such as those between the Hennigians and the Darwinians. The rules are silent on the acceptability of paraphyletic taxa and will likely remain so forever. Therefore there is nothing in the rules that would require anyone to splinter paraphyletic taxa nor is there anything in the rules that requires anyone to follow taxonomic proposals that are meant to disqualify or destroy paraphyletic taxa. Conversely, no Hennigian is required by the rules to accept paraphyletic taxa. However, the stated purpose of the ICZN is to promote taxonomic stability. Hence one can very well argue that the stated purpose and the very existence of the ICZN are both contradictory to the taxonomic instability that is the direct result of the Hennigian's intolerance of paraphyletic taxa.”

Kluge’s (1993) classification of pythons is an interesting case in regard to the anti-Hennigian crusade CKing is all het up about. It presents two neat examples of nomenclatural changes made to produce a monophyletic classification (on the assumption the preferred cladogram accurately represents history, i.e. is a true phylogenetic tree): the sinking of Chondropython (into Morelia) and the splitting of Liasis (into Antaresia, Leiopython, Bothrochilus, Liasis and Apodora). To a non-dogmatist, each should be considered independently on its merits.

Speaking of dogma, this talk of ‘Hennigians vs. Darwinians’ reminds me of Lysenko, who also claimed to be a good Darwinian in his attacks on the Morganist-Weismannists (e.g. ‘The Situation in Biological Science’, 1948). If there really is a Darwinian school of taxonomy, it could be based on what Darwin actually had to say on the subject (‘On the Origin of Species’, Ch. XIV):
“No doubt organic beings, like all other objects, can be classed in many ways, either artificially by single characters, or more naturally by a number of characters… Some authors look at it merely as a scheme for arranging together those living objects which are most alike, and for separating those which are most unlike… I believe that community of descent – the one known cause of close similarity in organic beings – is the bond, which though obscured by various degrees of modification, is partly revealed by our classifications.”

Right on, Mr D! Let’s base our classification on our best hypothesis of ‘community of descent’ (i.e. monophyly, or holophyly if you insist on a distinction), as soon as we have enough evidence (‘number of characters’) to be reasonably sure of the actual patterns, and not be misled by the ‘various degrees of modification’ (i.e. symplesiomorphy in paraphyletic assemblages, and autapomorphy in certain specialized lineages).

Suppose we would like to know “What species is/are the closest genealogical relatives of Green pythons?” or “What other species is/are most similar to Green pythons in morphology, ecology and behaviour?” What use is a classification going to be in finding answers? It depends: if we’re working with the ‘traditional’ classification that CKing prefers, with a distinct genus Chondropython, the answer is “none at all”, whereas if the species is called Morelia viridis and Morelia is a monophyletic group, we will be able to infer that Green pythons are most closely related (had most recent common ancestry with), and most likely quite similar in biology to (one, some or all) other species of Morelia. In either case, it may be necessary to get behind the classification to the cladogram (or more complicated phylogeny) and raw data, but this is an easier process with one set of names than the other.

Keeping the name ‘Chondropython’ is a bit like Steven King going back to writing under pseudonyms: the product is simply WORTH MORE, these days, if the name on the cover tells you where it actually comes from. Morelia viridis (or the M. viridis complex) is in fact a highly derived member of the M. spilota group, not a creation ex nihilo.

For the same reasons, when I found that the Miocene python Montypythonoides riversleighensis was not as distinctive as originally supposed by Smith and Plane 1985 (see also Kluge 1993), I had no problem sinking the genus into Morelia (Scanlon 2001. Memoirs of the Association of Australasian Palaeontologists 25: 1-35). I can’t say for sure whether M. riversleighensis is closer to the amethistina group, the spilota group in general, or (what it’s most similar to) oenpelliensis, but calling it a Morelia covers all these bases. Montypythonoides is a cute name but was (it turned out) as useless as Chondropython.

The Liasis situation is different in a few ways, and contradicts CKing’s claim that Kluge is (consistently) a ‘lumper’. Kluge’s departure from ‘tradition’ in this case is in splitting a group of more-or-less similar species (Liasis sensu lato) into numerous genera rather than sinking one useless name, making the binominal classification more complex and more redundant (adding several monotypic genera) rather than more streamlined and informative (information on suprageneric groups was instead added as a separate layer, using an indentation scheme). The most important difference, in my view, is that the Morelia spilota group was quite well resolved by Kluge’s character data (leaving no doubt that viridis belongs inside it), while the whole region of the tree between Morelia and Aspidites (i.e. Liasis s.l.) has a lot of short internal branches and conflicting characters, so the particular topology favoured by Kluge was quite weakly supported. I don’t blame him for choosing a method and running consistently with its implications, but I’ve continued to use Liasis in the broad and possibly paraphyletic sense, partly to justify slipping a fossil species into an existing genus when its exact phylogenetic position was uncertain (Scanlon and Mackness, 2002. A new giant python from the Pliocene Bluff Downs Local Fauna of northeastern Queensland. Alcheringa 25: 425-437). Liasis dubudingala (probably the largest snake in Australia’s history) is distinctive in a few ways, but to create a monotypic genus for it would have just created more clutter. There was not enough information to place it in any of Kluge’s genera, and in such a case of uncertain relationships and similar species, paraphyletic taxa are (I think, sometimes) better than some alternatives. (If unable to place it in a genus, no species name could be assigned under the Code, and anybody could have come along and named it later without).

But once the phylogeny is well resolved, paraphyletic taxa are just positively misleading: they group together species that are KNOWN not to be each others closest relatives, and I don’t think Mr Darwin would approve. Despite statements by Bob Carroll (1988 etc.) and some others, it is biologically meaningless to say that genera and families are objectively real even if paraphyletic.

And what about all those ‘contrived taxa that cannot be defined’? Are we talking definition or diagnosis? In either case, this phrase (much repeated, without a cited reference) sounds like empty rhetoric. It’s always possible to DEFINE a taxon either by listing its constituent species, or (as in PhyloCode) by various means such as ‘all organisms more closely related to X than to Y’ or ‘the most recent common ancestor of X and Y and all its descendants’. These definitions are unambiguous and logically guaranteed to be monophyletic, though assigning species to higher taxa, apart from the reference species actually named in the definitions, is still always a matter of evidence for a phylogenetic hypothesis.

And then there’s DIAGNOSIS. Theoretically, in the case of almost total evolutionary stasis in two or more successive outgroups to a clade, there may exist paraphyletic assemblages of species that truly cannot be split up and diagnosed using shared derived characters. To assert that this is the case in New World Elaphe or any other group would require thorough investigation of all morphological systems in the relevant species (Scott Thomson mentioned some turtle examples where a close look at morphology found diagnostic characters for groups discovered and defined using genetic data). And a cladist’s diagnosis (using only apomorphies) can be constructed for any tree, whether or not it is a true representation of phylogeny: compare those of Kluge (1993), and mine of Liasis s.l., based on Kluge’s data set, in Scanlon and Mackness (2002), which can’t both be true. Both represent perfectly adequate diagnoses, but I prefer the one that refrains from radical change to preceding systems because (in my opinion) such change is not justified by the state of knowledge of actual relationships. That’s the interim version of taxonomic stability; but once we know the true relationships, we can pick a monophyletic taxonomy that can stand for all time.

Well said John,

I personally think there are two waring issues in this argument that get intermingled. The need to be teased apart because they really are two separate issues independant of each other.

1. is the issue of paraphyly and I agree John that there is no room for paraphyly and that it goes against the pricipals of Evolution by decent as well as the the principals of systematics by Mayr and others.

2. The issues of nomenclatural stability, ie splitting versus lumping.

A third issue creeps in and that is the cladistic fear of monotypy.

There is no valid reason to keep a nomenclatural system that is wrong for the sake of not changing the current system. If it is shown to be wrong the names must change. However should you fix a paraphyly by lumping or splitting.

The ultra conservatives would argue you should lump. This is not necessarily a good idea. The genera created from this must stand up as a genus under some definition of the genus and not be a means to avoid the monotypic taxon.

The ultra radicals would argue that you should split, but again the genera created must stand up to a defined value as a genus.

Hence I feel that the mistake in all this is the view of many cladists to split as a general rule but only to do so if it does not produce a monotypic taxon. Their error is one of circularity. They are avoiding the production on monotypy for reasons of methodology. Hence their nomenclature is influenced by methodology. As this is not permissable under nomenclature their names are invalid.

Cheers, Scott
Carettochelys.com

-----
Scott Thomson

If you believe you can or you can't you are always right.

Let me start by saying that I am not on an anti-Hennigian crusade. The Hennigians are entitled to their believes of course, as are the pheneticists and the Darwinians. It is in fact the Hennigians who are imposing their views upon the traditionalists in their war against traditional taxonomy. A prime example of this imposition is their hijacking of traditional terms for their own use. The term monophyletic has never meant groups that consist of a single ancestor and all of its descendants. But the cladists insist that all other biologists must now adopt their new meaning. Is that not evidence that the cladists are on a crusade? To turn around and then accuse some Darwinians of being on a crusade because they do not accept the Hennigians’ imposition is indeed quite unjust.

Yes there is a Darwinian school of taxonomy, because Darwin is unambiguous about how organisms ought to be classified in a "natural system." Darwin has sometimes been misinterpreted as being a cladist in his classificatory philosophy. Nothing can be further from the truth. Darwin wrote:

“But I must explain my meaning more fully. I believe that the arrangement of the groups within each class, in due subordination and relation to each other, must be strictly genealogical in order to be natural; but that the amount of difference in the several branches or groups, though allied in the same degree in blood to their common progenitor, may differ greatly, being due to the different degrees of modification which they have undergone; and this is expressed by the forms being ranked under different genera, families, sections, or orders. the reader will best understand what is meant, if he will take the trouble to refer to the diagram in the fourth chapter.”

As one can see, Darwin is advocating the ranking of taxa based on “degrees of modification” or morphological disparity, which is of course a criterion that is rejected by the cladists, who only want to classify organisms according to branching order and nothing else. In fact, Darwin is advocating the recognition of paraphyletic taxa.

Darwin, referring to the only diagram in his book, continues:

“The forms descended from A, now broken up into two or three families, constitute a distinct order from those descended from I, also broken up into two families. Nor can the existing species, descended from A, be ranked in the same genus with the parent A; or those from I, with the parent I. But the existing genus F14 may be supposed to have been but slightly modified; and it will then rank with the parent-genus F; just as some few still living organic beings belong to Silurian genera.”

By putting A and its descendants in different genera, Darwin has in effect created a paraphyletic genus which contains A but not all of its descendants. He did the same for I. Darwin therefore clearly shows that he has no philosophical aversion to paraphyletic taxa, unlike the Hennigians. In fact, he makes it quite clear that paraphyletic taxa are unavoidable if biologists are to classify organisms according to “degrees of modification” by giving us an exercise on how to create a Darwinian or "natural" classification.

johnscanlon:
“Suppose we would like to know “What species is/are the closest genealogical relatives of Green pythons?” or “What other species is/are most similar to Green pythons in morphology, ecology and behaviour?” What use is a classification going to be in finding answers? It depends: if we’re working with the ‘traditional’ classification that CKing prefers, with a distinct genus Chondropython, the answer is “none at all”, whereas if the species is called Morelia viridis and Morelia is a monophyletic group, we will be able to infer that Green pythons are most closely related (had most recent common ancestry with), and most likely quite similar in biology to (one, some or all) other
species of Morelia. In either case, it may be necessary to get behind the classification to the cladogram (or more complicated phylogeny) and raw data, but this is an easier process with one set of names than the other.”

Me:
If you want to answer the first question, it can be answered with a tree. Darwin drew trees. Darwinians draw trees. Cladists draw trees. Trees are useful in depicting relationships. Use them. If one wants to answer the second question, then one way that has proven useful is to classify it in a different higher taxon.

johnscanlon:
"Keeping the name ‘Chondropython’ is a bit like Steven King going back to writing under pseudonyms: the product is simply WORTH MORE, these days, if the name on the cover tells you where it actually comes from. Morelia viridis (or the M. viridis complex) is in fact a highly derived member of the M. spilota group, not a creation ex nihilo."

Me:
If Chondropython is so highly derived, then it should be removed from the same genus as its ancestor, as Darwin did by removing A’s descendants from the genus which includes A. By doing so, we are showing that Chondropython is quite distinct morphologically, behaviorally and ecologically than Morelia. Of course, removing highly derived members from a group is strictly prohibited by those who adhere religiously to Hennig's principle of holophyly. You asked, "When, if ever, are paraphyletic taxa useful?" The answer is that they are useful when one wishes to show that one or more members of a holophyletic taxon is highly derived.

Yes it is indeed possible to define taxa as the descendants of some ancient species, such as A in Darwin’s diagram. Such definitions, however, tell us nothing about the morphology, ecology and behavior of this group, which is what you said you are interested in. Therefore, an adherence to Hennigian dictates actually makes it less possible to answer some of the interesting biological questions that you posed.

Even if there is a good degree of certainty that a particular genealogy is correct, as is the case in Darwin’s hypothetical tree, he still advocates taxa be ranked on the basis of not just genealogical relationships but degrees of modification as well. Doing so of course meant the recognition of paraphyletic taxa. Not only are paraphyletic taxa the inevitable result of the process of evolution, as Carroll pointed out, but it is also the inevitable result of classifying organisms according to both genealogy and the degrees of modification, as Darwin demonstrated with his diagram.

Below is a very well drawn version of Darwin's only diagram in Origin:

A few remarks on CKing’s latest…

First the appeal to inclusivity and tolerance, including the remarkable statement that pheneticists are entitled to their beliefs. Anybody who still thinks that grouping by similarity is the best method of inferring phylogeny (tree-building) is simply mistaken in their belief. Tolerance can be taken too far.

Hijacking of terms: I have come across the term ‘holophyletic’ and the sense of ‘monophyletic’ that includes paraphyly in one or two obscure places, but NEVER, that I recall, in an actual phylogenetic or taxonomic work from the last 30 years. I really don’t know the tradition they belong to; while it’s good to be able to read the dead classical languages, they are no longer a suitable medium for scientific discourse. Whether a hijacking was involved or not, the process of language evolution can hardly be reversed. As for “The term monophyletic has never meant groups that consist of a single ancestor and all of its descendants” – actually, that is exactly what it has meant since I first heard it (about 1980). HOW OLD ARE YOU, that the last few decades have passed without notice?

”Darwin has sometimes been misinterpreted as being a cladist in his classificatory philosophy. Nothing can be further from the truth.” That last bit is an exaggeration, you must admit; many things further from cladism can be imagined (e.g. cheerful acceptance of polyphyly, defining groups by their ‘destiny’ rather than their content or characters, creationism). Cladistic classification is ‘pure’ grouping by genealogy, whereas Darwin believed (as you quote) that “the arrangement of the groups within each class, in due subordination and relation to each other, must be strictly genealogical in order to be natural” (sounds very Hennigian so far), but also that different degrees of modification could usefully be expressed by (to simplify his circumlocution) sometimes using different Linnean ranks for sister taxa, or assigning ancestors and descendants to separate groups of the same rank.

Well, Darwin was a great biologist – a revolutionary, saint and prophet of science indeed – but only came close, without ever explicitly stating or applying the principles of homology and synapomorphy that turned phylogenetics into a science. That all happened after his time (though it was happening well before Hennig got hold of the idea). Linneus’ great contribution had been the hierarchical system of classification (as opposed to the eclectic lists or weird numerological/astrological arrangements that went before); Darwin’s advance in systematics (i.e. apart from Natural Selection) was the principle that a natural, genealogical classification is possible and desirable, and would be strictly hierarchical as a direct consequence of descent with modification (so the manifest success and ‘naturalness’ of Linneus’ system was confirmation for Darwin’s theory). Hennig explicitly proposed that a natural classification should use only monophyletic taxa (i.e. only clades should be named). Traditionalists like CKing find this objectionable, but that does not give them exclusive rights to be called Darwinians. Most of the people who make and use classifications these days do prefer (or expect) taxa to be monophyletic, but that does not make us ‘Hennigians’ in any useful sense because most of his concepts, methods and terms have been superseded; modern systematists are much more likely to read, cite and learn from Darwin’s timeless classics than from Hennig.

In Darwin’s day, importing arbitrary judgements about what degree of divergence justifies a certain ‘rank’ was unavoidable. Hennig kept ranks, and thought they should correspond to the age of clades (genera are so many millions of years old, families somewhat older, etc.). These days we have vastly more detailed phylogenetic information and are increasingly tossing out the idea that taxa of the same Linnean rank are equivalent in any meaningful way, so that while the standard endings of ‘family’, ‘order’ and other taxon names can still be useful for organizing lists and memory, many of us have decided to exclude such ranks from formal classification completely. When we mention Mammalia or Reptilia, we mean clades, not ‘Classes’; Caenophidia, Colubroidea, Elapidae and Hydrophiinae are a sequence of nested clades, and it means nothing real to (formally) recognize one of them as equivalent in rank to another taxon apart from its immediate sister group. Linnean binomina are an extremely handy mnemonic device that makes it possible for ordinary mortals to organize information on thousands of species in their heads, so I can’t see any net advantage in abandoning genera; but I like them monophyletic, and this is always possible except in the case of species that are actual ancestors of two or more genera that it would be impractical to lump. Since actual ancestors are unlikely to be observed (even as fossils) but may be indistinguishable from some of their descendants (like F – F14 in Darwin’s diagram), nothing is lost by naming a monotypic genus for the ‘ancestral’ species in such a case.

“If Chondropython is so highly derived, then it should be removed from the same genus as its ancestor, as Darwin did by removing A’s descendants from the genus which includes A. By doing so, we are showing that Chondropython is quite distinct morphologically, behaviorally and ecologically than Morelia.” What such a split actually does is to obscure the fundamental similarity (homology) of viridis with other members of the Morelia spilota group. It has diverged in some ways, but on average (looking beyond superficial appearance, and using the tree and character data that we now have) far less than M. spilota has from M. amethistina. It is false to say that clades “tell us nothing about the morphology, ecology and behavior”: as Darwin said, “community of descent [is] the one known cause of close similarity in organic beings”, and it is therefore the best predictor of similarity in characters that have yet to be studied.

With no formal ranks above the genus, and no taxa at any level (above species) that are provably non-monophyletic, we can have a workable but streamlined system. Some may prefer classifications with a lovely baroque feel - taxonomic equivalents of festoons, curlicues and little plaster cherubs - but I can’t see the point.
-----
John D. Scanlon
Riversleigh Fossil Centre
Outback at Isa
Mount Isa, Queensland, Australia

The term holophyletic has remained obscure perhaps because most cladists find it unnecessary, since they have already altered the meaning of the term monophyletic to fit their need. The traditionalists, on the other hand, simply continue to use the term monophyletic as they have always done so, and find no need to use the term holophyletic since they do not make a distinction between paraphyletic and holophyletic taxa.

“Well, Darwin was a great biologist – a revolutionary, saint and prophet of science indeed – but only came close, without ever explicitly stating or applying the principles of homology and synapomorphy that turned phylogenetics into a science. That all happened after his time....”

Darwin in reality pays close attention to homology and synapomorphy. He points out that taxonomically informative characters can often be found in the embryos and among atrophied or rudimentary organs, and that both of these types of characters are often of higher taxonomic value than vital organs. He therefore realized the possibility of convergent evolution producing superficial similarities and warned his readers about relying on these characters to group taxa. In fact, he and his contemporary systematists display far more insight than many modern systematists, such as the pheneticists and the cladists, who treat all characters equally. Darwin knows better; the Darwinians know better. Now it appears that some cladists have independently rediscovered the wisdom contained in Darwin’s book Origin. Some cladists have reinvented character weighting on their own! Darwin opens his chapter with the following statement:

“The existence of groups would have been of simple signification, if one group had been exclusively fitted to inhabit the land, and another the water; one to feed on flesh, another on vegetable matter, and so on; but the case is widely different in nature; for it is notorious how commonly members of even the same subgroup have different habits.”

Thus Darwin realizes that similarities in some characters (such as diet or habitat preference) may be uninformative of phylogenetic relationship because they are found in different groups. Contrast this insight with recent attempts by some cladists to use food habits as a taxonomic character (de Queiroz’s use of herbivory in analyzing the iguanian lizards and Rodriquez-Robles’ use of dietary preference in analyzing the lampropeltine snakes) and one should realize how unwise it is to ignore the wisdom of past great minds, as cladists have often been accused of doing.

Ernst Mayr writes: “Monophyletic: This term was coined by Haeckel in 1866 in support of Darwin’s theory of common descent. He was unalterably opposed to then current theories of Darwin’s opponents, who espoused numerous independent origins of living things. Haeckel termed such multiple origin theories polyphyletic.”

Johnscanlon:
“Hennig explicitly proposed that a natural classification should use only monophyletic taxa.”

Ever since Darwin’s theory of common descent became widely accepted, all of the then currently recognized taxa would have been subjected to the test of monophyly. By the time Hennig was practicing systematics, practically all known taxa were already considered monophyletic according to the best available data at the time. Hence Hennig did not advocate that only monophyletic taxa be recognized, he came along and mandated that only holophyletic taxa be recognized. If anyone should insist that taxa must be holophyletic to be acceptable, he/she is thus following the dictates of Hennig. He/she is thus a Hennigian in his/her classificatory philosophy.

Hennig, as Mayr points out, has considerable difficulty coming up with a workable ranking system. Mayr writes:
“His followers have abandoned both of Hennig's basic criteria (equal rank of sister groups and geological age) not only because they led to too great an imbalance between explosively evolving and slow groups (including living fossils) but also because strict coordination of sister groups led to a veritable explosion of categorical levels.”

Some (but not all) cladists who have not abandoned Hennig’s criteria are advocating the abandonment of the Linnaean ranks, since it cannot accommodate the explosion of categorical levels. Nevertheless, Linnaean ranks are useful for most biologists. Therefore most biologists will continue to use them.

johnscanlon:
"I can’t see any net advantage in abandoning genera; but I like them monophyletic, and this is always possible except in the case of species that are actual ancestors of two or more genera that it would be impractical to lump. Since actual ancestors are unlikely to be observed (even as fossils) but may be indistinguishable from some of their descendants (like F – F14 in Darwin’s diagram), nothing is lost by naming a monotypic genus for the ‘ancestral’ species in such a case."

I agree, there is no reason to abandon the rank of genus. Neither is there any reason to abandon the other Linnaean ranks. Forcing species into strictly holophyletic groups would simply lead to either excessive splitting (e.g. Utiger et al.’s split of Elaphe) or excessive lumping (Kluge’s lumping of Morelia and Chondropython) or both (Frost and Etheridge’s splitting of the Iguanidae and their lumping of the Chameleonidae and Agamidae). Not surprisingly these taxonomic proposals have been controversial. More than a decade later, Frost and Etheridge’s classification has been reversed and Kluge’s proposal has largely been ignored. Utiger et al.’s proposal will likely be ignored by most veteran herpetologists, since none of the genera they resurrect and erect is morphologically distinguishable from Elaphe, unlike Pituophis, Arizona, Lampropeltis, Cemophora, Stilosoma, Rhinocheilus and Bogertophis. There is no reason to splinter a slowly evolving genus such as Elaphe simply because one member migrated to North America and evolved rapidly to give rise to a number of divergent species, classifiable in different genera, thus rendering Elaphe paraphyletic. The parental genus Elaphe remains unaffected by the evolution of new taxa in the New World and it should not be sacrificed at the altar of the Hennigian's ideological distaste for paraphyletic taxa.

To amplify a point from earlier, where I said “Darwin’s advance in systematics (i.e. apart from Natural Selection) was the principle that a natural, genealogical classification is possible and desirable, and would be strictly hierarchical as a direct consequence of descent with modification (so the manifest success and ‘naturalness’ of Linneus’ system was confirmation for Darwin’s theory).”

… with another succinct statement by Darwin (Origin, Ch. XIV):

“In considering this view of classification, it should be borne in mind that the element of descent has been universally used in ranking together the sexes, ages, dimorphic forms, and acknowledged varieties of the same species, however much they may differ from each other in structure. If we extend the use of this element of descent – the one certainly known cause of similarity in organic beings, - we shall understand what is meant by the Natural System: it is genealogical in its attempted arrangement, with the grades of acquired difference marked by the terms, varieties, species, genera, families, orders, and classes.”

It is ambiguous in Darwin’s text whether ‘grades of acquired difference’ correspond in his thought to what we call ‘grades’ (typically paraphyletic if not polyphyletic in 20th century practice), or if he considered as distinct (and rejected?!) the idea of naming clades instead. I suspect it was just too soon for him to realize there was a difference; the concept of a clade and its diagnosis could be stated – e.g. “on the principle of inheritance, all the forms descended, for instance, from A, would have something in common” – but nobody had yet though of coining a rank-free term for “all the forms descended from A”.

Linnean classifications are hierarchical, consisting of ‘groups under groups’ as Mr D so often expressed it, but the traditional taxonomic hierarchy is typically not the same as the actual genealogical (phylogenetic) structure (which we can increasingly often determine exactly enough, hence the death of the ‘no-hope’ school of phenetics). Darwin belaboured the fact that there’s no criterion to regard hierarchy as real or natural except as a result of branching common descent, so why would he plump for a hierarchy that’s known to be contradictory to nature’s? I don’t think he ever meant to do so, though anti-cladists do it consciously every day.

Since Darwin’s day, users of classifications (for example, those designing comparisons and experiments to test evolutionary theory and particular patterns and processes) expect members of the same taxon to be ‘related’ to each other more closely than to those of other taxa at the same rank. Much of the time, accepted classification has failed to do what was expected, and much experimental effort has been wasted as a result. An example, probably typical: which species of Python should be compared serologically with Australian and New Guinea pythonines to test the systematic ideas of McDowell 1975? Schwaner and Dessauer 1981 picked the wrong one (P. regius) to test, precisely because McDowell had a paraphyletic concept of the genus but they assumed he meant it to be a clade.

CKing: “I agree, there is no reason to abandon the rank of genus. Neither is there any reason to abandon the other Linnaean ranks.”

Once the artificial and misleading features of paraphyletic taxa are recognized and we adopt a monophyletic arrangement instead, actually practicing systematics on real organisms (in contrast to bombinating in a vacuum [Julian Huxley’s turn of phrase, slightly out of context]) will show the impracticality of ranks. For example, take Squamata (phylogenetic understanding of which has got quite detailed but not changed in broad outlines since Camp 1923):

All extant snakes belong to a taxon named Serpentes, traditionally called a suborder (sometimes an order). (There are also fossil snakes which – according to most of the evidence, and it should come as no surprise - lie outside the clade comprising all living snakes; clearly, whether to include them in Serpentes is a matter of arbitrary boundary fixing.) But snakes are (again, according to most of the evidence) derived from among the varanoid lizards; Varanoidea is traditionally called a superfamily within the infraorder Anguimorpha. And all anguimorphs are scleroglossans (… but I don’t know or care what the rank of that is supposed to be called…), and scleroglossans (together with Gekkota) are autarchoglossans, which along with Iguania make up Squamata (order, or subclass or whatever). All squamates that aren’t snakes would traditionally have been called ‘Lacertilia’ or ‘Sauria’ (oh no, that’s not right; I was forgetting mosasauroids and amphisbaenians, often given orders or suborders of their own…).

The actual (or likely) relationships can easily be expressed by a tree (cladogram), and having once seen such a tree (like the one in Camp’s 1923 monograph, or those of Estes et al. 1988), why would anyone go back to a confusing mess of redundant and meaningless terms (‘rank’ titles for higher taxa) and taxonomic concepts that falsify what we actually know? If we know (as we think we do) that snakes are varanoid, anguimorph, scleroglossan squamates, and only one of numerous independently evolved groups of limbless lizards, why ‘promote’ them to the same rank as ‘other lizards’? There is no imaginable justification apart from claiming ignorance of actual relationships – an interim solution at best, now requiring heroic effort to ignore the accumulated evidence.

Give up rank, or give up monophyly; the choice is easy. Thanks for the directions, Linneus, but we can see where we’re going now.
-----
John D. Scanlon
Riversleigh Fossil Centre
Outback at Isa
Mount Isa, Queensland, Australia

“In considering this view of classification, it should be borne in mind that the element of descent has been universally used in ranking together the sexes, ages, dimorphic forms, and acknowledged varieties of the same species, however much they may differ from each other in structure. If we extend the use of this element of descent – the one certainly known cause of similarity in organic beings, - we shall understand what is meant by the Natural System: it is genealogical in its attempted arrangement, with the grades of acquired difference marked by the terms, varieties, species, genera, families, orders, and classes.”

Very interesting quote from Darwin. He realizes that the species category is not typological. Species concepts that rely on typology, for example, the phylogenetic species concept on which many cladists rely, are thus scientifically untenable. Darwin advocates that classification be "genealogical." He also advocates that paraphyletic taxa be recognized by ranking an ancestor in a different genus as its descendants if the amount of morphological disparity between ancestor and descendant justifies it. Therefore paraphyletic taxa is not anathema to his genealogy-based classification. Conversely, strictly holophyletic taxa, such as those advocated by Hennig and his followers, do contradict Darwin's natural system, since ranking different genera according to their degrees of modification is not possible if only holophyletic taxa are recognized.

As I said, ever since Haeckel coined the term monophyletic in 1866, and ever since Darwin's theory gained widespread acceptance, all taxa that are recognized have been tested for monophyly. Hence the decades old 'grade vs. clade' debate is not a 'polyphyletic vs. monophyletic' debate. It is instead a monophyletic vs. holophyletic debate. To recognize grades in classification, it is therefore necessary to recognize paraphyletic groups. Hence it would be inaccurate to suggest that some biologists (including those who are called 'gradists') are advocating the recognition of polyphyletic taxa. It is true that some taxa recognized on the basis of morphological similarities are polyphyletic but then again many cladistically delimited taxa have also been shown to be polyphyletic. For example, some cladists consider the hooked fifth metatarsal to be a synapomorphy and they claim that turtles are diapsids. Other evidence suggest instead that turtles are anaspids. Hence, if the hooked fifth metatarsal is indeed a homoplasy, as Michael Lee has advocated, the group that consists of turtles diapsid would be polyphyletic, even if it is a cladistically delimited group.

“Since Darwin’s day, users of classifications (for example, those designing comparisons and experiments to test evolutionary theory and particular patterns and processes) expect members of the same taxon to be ‘related’ to each other more closely than to those of other taxa at the same rank. Much of the time, accepted classification has failed to do what was expected, and much experimental effort has been wasted as a result. An example, probably typical: which species of Python should be compared serologically with Australian and New Guinea pythonines to test the systematic ideas of McDowell 1975? Schwaner and Dessauer 1981 picked the wrong one (P. regius) to test, recisely because McDowell had a paraphyletic concept of the genus but they assumed he meant it to be a clade.”

Since most taxa have traditionally been classified phenetically, it is thus no surprise that sometimes these taxa fail to inform us of branching order. Picking the incorrect outgroup is unfortunately a fairly common error in systematic analysis. In one recent study, for example, Lampropeltis (which is descended from a species of Elaphe which migrated to the Old World from the New) was picked as an “outgroup” in a mtDNA analysis of European species of Elaphe. That is analogous to picking Homo sapiens as the outgroup when analyzing the relationship among Gorilla, Pan and Pongo! One certainly cannot blame the classification for that mistake. The researcher is to blame in these cases. Even if were to adopt a holophyletic arrangement, the same mistake can still be made. Recognizing the dozen or so contrived taxa proposed by Utiger et al. to accommodate the pieces of the genus Elaphe, for example, will not prevent someone from mistakenly using Lampropeltis as an outgroup in an analysis of relationships among the European ratsnakes.

“Once the artificial and misleading features of paraphyletic taxa are recognized and we adopt a monophyletic arrangement instead, actually practicing systematics on real organisms (in contrast to bombinating in a vacuum [Julian Huxley’s turn of phrase, slightly out of context]) will show the impracticality of ranks.”

Actually it is well known for decades to taxonomists that holophyletic groups are impractical, hence the grade vs. clade debate. In order to recognize Lampropeltis and other derived genera, the basal genus Elaphe, which is paraphyletic, had to be splintered into a dozen or so contrived genera that cannot be defined or diagnosed under Utiger et al.’s proposal. The alternative is excessive lumping. Lampropeltis, Pituophis, Arizona, Bogertophis, Rhinocheilus, Stilosoma, and Cemophora have to be transferred back to Elaphe. Neither alternative is tenable as both require wholesale taxonomic changes. If Utiger et al.’s data is contradicted by a newer study, more wholesale reshuffling is necessary. Scientific communications may grind to a screeching halt, if the cladists’ ideological distaste for paraphyletic taxa must be accommodated. The better alternative is for the cladists to lose their religion, as some cladists have done, and accept paraphyletic taxa. Recently even Frost and Etheridge have not objected to the recognition of the paraphyletic Agamidae as their proposal to lump Chameleonidae and Agamidae was never fully accepted and it has been reversed by fellow cladists.

“If we know (as we think we do) that snakes are varanoid, anguimorph, scleroglossan squamates, and only one of numerous independently evolved groups of limbless lizards, why ‘promote’ them to the same rank as ‘other lizards’? There is no imaginable justification apart from claiming ignorance of actual relationships – an interim solution at best, now requiring heroic effort to ignore the accumulated evidence.”

Although I agree with you that snakes are most likely the descendants of a derived, varanoid lizard, some scientists nevertheless still believe that snakes and lizards are possibly sister taxa. Whether snakes should continue to be classified as a suborder under Squamata would require a certain amount of scientific discussion. The current arrangement of Squamata is certainly a good idea, vs. the older arrangement of separate orders for snakes and lizards. Order Squamata is certainly a genalogical arrangement. It is thus a Darwinian taxon that should be retained.

>> In one recent study, for example, Lampropeltis (which is descended from a species of Elaphe which migrated to the Old World from the New) was picked as an “outgroup” in a mtDNA analysis of European species of Elaphe. That is analogous to picking Homo sapiens as the outgroup when analyzing the relationship among Gorilla, Pan and Pongo! One certainly cannot blame the classification for that mistake.

Oh yes one can.

Actually, I can't think of a better example to show why the recognition of paraphyletic taxa is grossly misleading than the one you have just provided. Thank you for making the case for monophyletic taxa only for us.

Cheers,

WW
-----
WW Home

Firstly, I misspoke in my original message. I meant to say that Lampropeltis is descended from a species of Elaphe which migrated from the Old World to the New. This fact has been known for decades among herpetologists who are aware of Dowling et al.'s immunological data published in 1983 in the Journal of Zoology. Even without knowing Dowling et al.'s paper, it is still common knowledge among herpetologists that Lampropeltis is derived from Elaphe on the basis of morphology. Therefore there is no excuse for choosing Lampropeltis as the outgroup in analyzing relationships among European species of Elaphe. The same mistake, as I said in my original post, can still be made if one adopts Utiger et al.'s proposal to split Elaphe into a dozen or so different contrived genera, since Lampropeltis would still be a valid genus under Utiger et al.'s proposed classification.