Reptile & Amphibian Forums

Feldman and Parham (2002) are proposing to split the small North American turtle genus Clemmys because "The paraphyly of the genus Clemmys Ritgen 1828 requires a taxonomic revision of the nonhinged emydines."

The word "required" is dubious. There is no requirement, for example, under the rules of the ICZN to revise a taxon because it is paraphyletic. Nor does the Darwinian or evolutionary school of systematics require the revision of a taxon because it is paraphyletic. Only the Hennigians who straightjacket themselves with Hennig's arbitrary principle of holophyly are "required" by their methodology to destroy paraphyletic taxa.

In fact, R. L. Carroll (1988:13) wrote: "The existence of paraphyletic groups is an inevitable result of the process of evolution. Their existence requires that we define them in terms of the absence of the characters of their descendants as well as the presence of characters absent in their ancestors."

Hence all that a paraphyletic taxon requires is a definition. There is no requirement that paraphyletic taxa be revised, disqualified or destroyed.

Mayr and Ashlock (1991) went further than Carroll. They wrote: "Hennig developed the cladistic method on the basis of the assumption that new species originate by the splitting of a stem species into two daughter species. hence on the principle of dichotomy: The parental species disappears with the birth of the two daughter species. He saw the origin of new higher taxa in an equivalent manner as a dichotomous process. Much research of the last 50 years has indicated, however, that budding is a far more frequent way of originating new taxa than is splitting. When a new species originates by means of peripatric speciation, this has no effect on the parental species from which the neospecies has budded off. However, by Hennig's criteria, a species which has given rise to a new species by budding thereby becomes paraphyletic and has to be removed from the classification, even though it has not been affected by the budding event. The same alignment pertains to higher taxa, most of which evidently originated by the budding off of an enterprising new species that was successful in a new niche or adaptive zone. The parental taxon continued to flourish unchanged in its traditional niche, but it has become paraphyletic by cladistic definition and must be excluded from the classification. It is now fully evident that the proposal to disqualify paraphyletic groups from recognition in classifications is not only impractical and destructive but scientifically untenable."

If Hennig's classificatory philosophy is based on an erroneous assumption, then why do many systematists follow this scientifically untenable practice of destroying paraphyletic taxa?

Regardless of the reason for the disqualification of paraphyletic taxa, the result is often the splintering of a morphologically homogeneous group that is also monophyletic to most scientists who do not subscribe to Hennig's redefinition of the term monophyletic. In the case of Clemmys, Feldman and Parham are proposing to splinter the 4 species within this genus into 3 different genera: Emys, Calemys and Clemmys. Just like Utiger et al., Feldman and Parham do not even attempt to define the genera they resurrect. That is not a mystery because as Lazell (1989) pointed out, contrived genera cannot be consistently defined because the species in which these genera are placed are similar to one another and to their common ancestor because they have not changed for long periods of time. So, the cladists are replacing morphologically homogeneous paraphyletic taxa that can be defined, as Carroll pointed out, "in terms of the absence of the characters of their descendants as well as the presence of characters absent in their ancestors" with contrived taxa that cannot be consistently defined because they are not distinguishable from one another. This is scientific progress? Or is the cladists' destruction of paraphyletic taxa, as Mayr and Ashlock pointed out, scientifically untenable?

Reference

Carroll, R.L. 1988. Vertebrate Paleontology and Evolution. W.H. Freeman and Company. New York

Feldman, C. R. and J. F. Parham 2002. Molecular Phylogenetics of Emydine Turtles: Taxonomic Revision and the Evolution of Shell Kinesis. Molecular Phylogenetics and Evolution 22(3): 388–398

Lazell, J.D., Jr. 1989. Phylogenetic systematics of iguanine lizards (Review). Copeia 1989:807-809

Mayr, Ernst and P. Ashlock 1991. Principles of Systematic Zoology, 2nd. ed. McGraw Hill

CK:
I do not frequent this forum often because most of the threads are well be my understanding of the concepts mentioned, the terminology all due to my lack of background in these areas.

, However,I believe I understand the major thrust of your thread. The example of the Rubber Boa (Charina bottae) may represent a situation similar to your position dealing with the concept of 'budding'. If I understand correctly, this is where a parental type continues to persist but has given rise to a different form that adapted to a set of environmental conditions somewhat different from where the parental stock exists or in which the parental stock could not successfully compete with the new form.

In the case of C. bottae, there are two rather distinct size morphs with a dwarf form found in the extreme southern part of the species' range and a larger morph that occupies the remainder of the species' range in N. Am. Since 1995 when I confirmed the existence of the dwarf form in the San Bernardino Mts., questions naturally arose as to the origins of both forms C. bottae. There of course exists a variety of possibilities. For a variety of considerations, it seemed more reasonable that the large morph arose from a dwarf form of C. bottae. Just recently and independent of my assessment, an individual that goes by R.S. Newton that occasionally frequents these forums came to the same position using a somewhat different set of reasoning.

With C. bottae, I do not believe we have reached the point in which two distinct species (more likely subspecies) are present but nevertheless, there is some chance that the C. bottae example would conform to the 'budding' concept rather than the two forms arising from a common ancestor that is now extinct. Of course, this is all speculation and the latter scenario is still an option to be considered. But I tend to agree with your thesis in that evolutionary processes have undoubtedly used every conceivable manner in which to produce new forms and rejecting or the excluding of any of these pathways would seem unwise.

Richard F. Hoyer

Hi, you have a very good grasp of the concept of budding. However, budding usually refers to the origin of new species or new higher taxa, not the origin of new geographic races. Therefore the rubber boa may not be considered an example of budding unless a new species is recognized.

Indeed, the common ancestor of both the large and small morphs of the rubber boa can either be a large morph, a small morph or something totally different. However, given the fact that both the large morphs and the small morph are morphologically very similar, it is extremely unlikely that their common ancestor would be very different morphologically from either of these.

Therefore the only question that remains is whether this ancestor is large or small morph. A possible answer to this question comes from the fact that umbratica is the first lineage to branch off from this ancestor. Since it branched off first, this lineage is most likely to retain the ancestral characters found only in the common ancestor of all rubber boas, but it will not have any of the characters of the next lineage to branch off. A lineage which branches off later may contain features found in both the ancestral and the new lineage. For example, in the evolution of the Chordata, the primitive chordates retain the ancestral character of the notochord but lack the more recently evolved characters of lineages that diverged later such as limbs. The lineage leading to birds, for example, diverged much later, and so it has both limbs and the notochord (at least in some part of the life cycle).

We see that umbratica (the first lineage to branch off) consists only of small morphs and the Sierra Nevada subclade (which branches off later) consists of both large and small morphs. Applying the general principle discussed above to the rubber boas, it is most probable that the small morph is the ancestral morphotype since it is found in umbratica, the oldest lineage and that the large morph evolved later since large morph snakes are found only in the newer lineages (Northwestern and Sierra Nevada subclades).

If umbratica is considered a different species, then bottae indeed budded off of umbratica since umbratica continues to exist and is now a paraphyletic species. The problem with recognizing umbratica as a new species is that some of the small morph populations, which is morphologically indistinguishable from umbratica even on the basis of adult size would be classified as a different species: bottae.

CK:
I knew your post dealt with speciation but simply used the C. bottae examples as being somewhat analogous.

Depite the paper by Javier Rodriguez-Robles and others elevating umbratica to the level of species, I believe it this position does not have sufficient support. The morphological features upon which the authors relied are weak at best and may be invalid due to some problems of which they were not aware.

That being said, I am now of the position that the dwarf form (regardless of the mtDNA results), possesses a set of morpholgical features to warrant subspecific distiction. The characteristics of maximum mid dorsal scale count and ventral count appear to be consistent with the factor of lengths. Lower values for these two scalation features are concordant with the dwarf form whereas higher scalation values are associated with the large form immediately north from central Tulare County northward to about south Plumas County and west to include the boas in the Bay Area of central Calif. The feature of dorsal scale count is muddled to the north with another cline of lower maximum mid dorsal counts occurring in specimens from Plumas Co., Calif. northward and eastward.

My reasons for considering that a dwarf form was the original ancestor is based on biological factors. I am not certain as to how you arrived at your interpretation that the dwarf form was likely to be the ancestor to current populations.

You mentioned: "A possible answer to this question comes from the fact that umbratica is the first lineage to branch off from this ancestor."

Further on: "We see that umbratica (the first lineage to branch off) consists only of small morphs and the Sierra Nevada subclade (which branches off later) consists of both large and small morphs. Applying the general principle discussed above to the rubber boas, it is most probable that the small morph is the ancestral morphotype since it is found in umbratica, the oldest lineage and that the large morph evolved later since large morph snakes are found only in the newer lineages (Northwestern and Sierra Nevada subclades)."

I am not certain how you determined that umbratica branched off first and thus is the oldest lineage.

And I agree when you stated the following: "The problem with recognizing umbratica as a new species is that some of the small morph populations, which is morphologically indistinguishable from umbratica even on the basis of adult size would be classified as a different species: bottae."

Richard F. Hoyer

It would help greatly if you refer to the diagram below:

I drew this diagram as a simplified version of Rodriguez-Robles's tree presented in their paper. If you look at the bottom of the diagram, you will see an arrow with the word "Time" underneath it. Above this arrow is a diagram depicting the relationship among the 3 different lineages of rubber boas based on the mtDNA data. Most such diagrams do not show the arrow and the word "Time" because it is assumed that the reader already knows that this is meant in such a diagram. However, in a cladogram, the time dimension is not drawn proportionally. That means as we go from left to right, it is understood that events that have occurred to the far left are the most ancient and that events that are depicted on the farthest right are the most recent, usually the present. However, there is no attempt to depict time in such a way that two different lines of the same length represent the same amount of time. This is not the case, however, in Rodriguez-Robles’ fig. 4, in which there is an attempt to depict time proportionally.

Again, referring to this diagram, one sees a dot with two lines diverging from the dot. The dot represents the common ancestor of all living rubber boas at a given time. We do not know when that time might have been. The two lines branching off this first dot represents the two descendant lineages of this common ancestor. In a budding event, only one of these two lineages are different from the common ancestor. In a splitting event, both lineages are different. That means at that particular time represented by the dot, two rubber boa mtDNA lineages evolved. The lineage eventually known as umbratica evolved at that time, since the line that leads from this first dot ultimately to a group of populations known collectively as umbratica is a straight line, without further diverging lines. On the other hand, the other line (closer to the bottom than the line that led to umbratica) from this first dot further diverges into two different lineages at a later time. One is the Sierra Nevada subclade and the other is the Northwestern subclade. Since the appearance of the umbratica lineage predates the appearance of the other two lineages, systematists consider umbratica to be the first lineage to have evolved. The other two lineages of rubber boas evolved later. However, the mtDNA data tells us nothing about the morphology of any of these lineages. They only tell us that their mtDNA are different because of substitutions in their nucleotide sequences.

Without any knowledge of the rubber boas at the time the umbratica lineage evolved, since such information can only come from the incomplete fossil record, systematists can only infer the morphology of the lineages at the particular periods of time depicted in the cladogram. Since all of the populations of the umbratica lineage are dwarf forms, there is also a good chance that the dwarf form is the ancestral morphotype of the umbratica lineage. However, we still do not know whether the common ancestor of all living boas is a dwarf form or a large morph. We can infer that by examining the morphology of the other two lineages. One of these, the Northwestern subclade, is a large morph, whereas the other lineage, the Sierra Nevada subclade, is a mix of small and large forms. Therefore one of two things could have happened: 1) the common ancestor of all living rubber boas is a large morph, and dwarfism evolved in the umbratica lineage and then dwarfism evolved once again in the Sierra Nevada lineage independently or 2) the common ancestor of all living rubber boas is the dwarf form, and that the large morph evolved independently in the Sierra Nevada and Northwestern subclades. As I discussed in my previous post, I believe that possibility number 2 is more likely since dwarfism is found in the oldest lineage-umbratica and one of the descendant lineages, whereas the large morph is unknown in umbratica.

I hope I have adequately explained myself. If not, feel free to pose further questions.

CK:
Thanks for the added explanations. We are in the moving process so I have not been able to completely absorbed all you mentioned at the time I reviewed your post along with Javier's fig. #4. But it does make better sense. Will get back to it later near the end of the week.

Richard F. Hoyer

The ICZN rules deal with Nomenclature not systematics. Therefore wether there is a requirement under the ICZN to deal with paraphyly or not is irrelevant.

What is relevant is the paradigm in which the researchers find themselves. If they are under a cladistic paradigm then the science of that will require a taxonomic solution to a paraphyletic taxon.

I am neither a splitter or a lumper but will do what I consider necessary under the paradigm I am working from. I happen to agree that morphologically homologous taxa are sometimes separated. I have also found that when these taxa are re-examined and looked at more closely, particularly internally, they are not the same.

I am not personally in favour of the use of DNA sequence data in systematics, as I am a taxonomist/ palaeontologist I see it to be useless as the majority of the testible taxa return a non-result. Hence I think DNA has its place in population studies and other below species work.

Cheers, Scott
Carettochelys.com

"The ICZN rules deal with Nomenclature not systematics. Therefore wether there is a requirement under the ICZN to deal with paraphyly or not is irrelevant."

Whether paraphyletic taxa are recognized or not is a taxonomic (nomenclatural) issue, not a systematics issue. A Darwinian (traditionalist) and a Hennigian (cladist) looking at the same tree may classify the animals quite differently depending on whether they tolerate paraphyletic groups or not and whether they take into account morphological disparity in their classification or not. Since the ICZN is silent on the issue of paraphyletic taxa, there is nothing in the rules that requires the destruction of paraphyletic taxa. The decision to splinter paraphyletic taxa is entirely that of the individual taxonomist. Therefore whether one accepts a proposal to reclassify animals on the basis of an intolerance of paraphyletic taxa is entirely up to the individual as well.

"I am not personally in favour of the use of DNA sequence data in systematics, as I am a taxonomist/ palaeontologist I see it to be useless as the majority of the testible taxa return a
non-result. Hence I think DNA has its place in population studies and other below species work."

DNA data may be useless to most paleontologists in most cases, but they have proven useful in some exceptional cases. For example, a recently published paper (Burger et al. 2003) dealt with the relationship of extinct cave lions by examining their mtDNA. Much older (20 million years) DNA preserved in the form of fossil leaves have provided invaluable insight for both paleontologists and evolutionary biologists. S. J. Gould, for example, dealt with the discovery of ancient DNA in his essay "Magnolias from Moscow" in the book "Dinosaur in a Haystack." It would be a mistake to dismiss DNA data simply because it is not a geologically stable molecule and hence is absent from most fossils, as big a mistake as it is to ignore the details of soft tissues that are occasionally preserved in exquisite detail in some fossils.

In many cases, the classification of animals is difficult because there are few morphological similarities and the high probability of convergence of morphological characters. The relationships among Burgess Shale organisms, if they had living descendants, would have been much clearer since their DNA can then be sequenced. DNA sequencing successfully resolved the relationship between the king crab and a single genus of hermit crabs. Morphological data alone would not have been able to do the same. In fact, the close relationship between whales and the artiodactyls was discovered long before the leg bones that confirm this relationship was finally found. As S.J. Gould points out, DNA sequences provide "an enormous increase in the number of useful characters for classification." I know of no scientist who is not elated by an abundance of data since they are most often frustrated by a lack of same. There can never be too much data if one is a scientist.

It would not be an exaggeration to say that DNA data revolutionized systematics. Cladistics, however, has done the opposite, as it has brought systematics back to the days of pre-Linnaean facile diagnosis. Cladistics was embraced by the morphologists because they desperately needed something to counter the ongoing revolution brought on by the increasing use of biochemical data, which threatened the morphological systematists with extinction. Sadly, one well known and highly respected herpetologist even committed suicide because he thought that there will be no place for the morphological systematist in the future. Paradoxically, many molecular systematists have unwisely allied themselves with the Hennigians (cladists), even though molecular systematics owes nothing to the Hennigians because it was developed independently of cladistics. Consequently, many molecular systematists are unwisely following the dictates of Hennig and are on a crusade to destroy paraphyletic taxa and replace them with contrived taxa that cannot be consistently defined.

Systematics has taken a giant leap forward in the use of DNA sequences but sadly taxonomy has taken a giant leap backwards at the same time because many taxonomists have become Hennigians.

Reference

Burger, J. et al. 2003. Molecular phylogeny of the extinct cave lion Panthera leo spelaea. Molecular Phylogenetics and Evolution (in press)

"Whether paraphyletic taxa are recognized or not is a taxonomic (nomenclatural) issue, not a systematics issue. A Darwinian (traditionalist) and a Hennigian (cladist) looking at the same tree may classify the animals quite differently depending on whether they tolerate paraphyletic groups or not and whether they take into account morphological disparity in their classification or not. Since the ICZN is silent on the issue of paraphyletic taxa, there is nothing in the rules that requires the destruction of paraphyletic taxa."

The system of nomenclature is independant of methodology. The methods employed by the research and the paradigm they work from is the science. Nomenclature is not a science. It is the application of a set of rules to a situation. It is about the validity of the publication of names and how they should be applied. It has nothing at all to do with how the relatedness of the taxa involved is determined.

As such what you are referring to is methodology of tree building and interpretation. This is systematics.

I can see your argument, you basically are from a different paradigm as the vast majority of taxonomists recently and are building arguments to justify your stance. Thats fine, all scientists do this. I am not saying you are wrong. I happen to also think that cladistics has more a kin to a religous fervour than to science. But I live with it.

"DNA data may be useless to most paleontologists in most cases, but they have proven useful in some exceptional cases."

Exceptional cases while interesting do not help the majority of science.

"It would not be an exaggeration to say that DNA data revolutionized systematics."

Of course it isn't, I never denied it has been used, and as I have used allozyme electrophoresis and DNA sequencing in my own papers I would be a fool to say it could not be used. I just wonder at its validity in Systematics and wether we are making a mistake to utilise it so much. I have argued that considering every single locus as a character is a mistake because of the inherit protection systems in genes where limited point mutations can be present without effecting the gene. I also worry about the vast regions of "unknown", presumably non-coding areas that are used. But these are all parallel arguments.

Also one of the biggest difficulties with cladistics is weighting. If you sequence a gene and come up with several hundred characters and then combine it with a morphological annalysis with say 50-100 characters you have a weighted and biased dataset, biased to DNA as it swamps the dataset. Geneticists argue that the morphology should be excluded altogether and later mapped on to the gene sequence trees and that fossil evidence should be excluded altogether as well. Personally I call this manipulation of data.

"Systematics has taken a giant leap forward in the use of DNA sequences but sadly taxonomy has taken a giant leap backwards at the same time because many taxonomists have become Hennigians."

Well I think it has taken a leap sideways myself and I think it needs another 20 years or so before it gets understood what is happening and reigned in a bit. As for a backwards step in cladistics, again I think this depends how it is applied. Numerical Cladistics I would certainly agree with you. The deliberate attempt to turn systematics into a pure popporian science has done a lot of damage. However, there are alternatives and some are being developed.

Cheers, Scott

"The system of nomenclature is independant of methodology."

Quite correct. A cladist can, for example, recognize paraphyletic taxa, as a few of them have done. But these cladists are viewed with suspicion, as Mayr and Ashlock point out, by other cladists. Conversely, a systematist may adhere to Hennig's classificatory practice by recognizing only taxa consisting of one ancestor and all of the descendants of this ancestor even though his/her tree is based on distance data, which is categorically rejected by the cladists.

I am not referring to tree building or interpretation. I am referring to the fact that different taxonomists can classify animals quite differently even when both are basing their classification on the same tree or set of phyletic relationships. A taxonomist who tolerates paraphyletic taxa, for example, may recognize a paraphyletic Clemmys for the 4 North American species of emydid turtles, whereas a Hennigian may splinter the 4 species of Clemmys into two or more genera. This is the case because taxonomy is independent of systematic methodology.

"Exceptional cases while interesting do not help the majority of science."

Science is based on the accumulation of facts. Every little bit helps.

"Also one of the biggest difficulties with cladistics is weighting. If you sequence a gene and come up with several hundred characters and then combine it with a morphological annalysis with say 50-100 characters you have a weighted and biased dataset, biased to DNA as it swamps the dataset. Geneticists argue that the morphology should be excluded altogether and later mapped on to the gene sequence trees and that fossil evidence should be excluded altogether as well.
Personally I call this manipulation of data."

Character weighting is a problem for systematists in general, not just cladistics. Darwinians argue that the cladists who weight all characters equally are just as subjective as one who gives each character a different weight. Clearly, some characters are more informative than others. For example, the way a turtle retracts its neck into the shell cannot possibly be given the same weight as, say, whether its plastron is hinged or not. Character weighting is therefore necessary, but it is not easy.

I agree that there is no justification for ignoring scientific data. Molecular trees can be corroborated with morphological characters and vice versa. I am not sure the practice of dumping all the molecular and morphological characters into the same dataset and finding the shortest tree will give a better result. It all boils down to the goodness of the characters, not whether they are molecules or morphological characters nor how many characters one uses. Given good characters, just a few of them will suffice. For example, the SINE characters that systematists used to link artiodactyls with whales, though few in number, have been very informative of relationships.

"A taxonomist who tolerates paraphyletic taxa, for example, may recognize a paraphyletic Clemmys for the 4 North American species of emydid turtles, whereas a Hennigian may splinter the 4 species of Clemmys into two or more genera. This is the case because taxonomy is independent of systematic methodology."

If splitting them is problematic because you cannot tell them apart why not lump them? It is a valid means of dealing with paraphyly and should be looked at as an equal option as splitting. Then the most appropriate course taken.

As we did for the E. macquarii group, we got rid of E. signata, E. krefftii and a host of subspecies. Put them all in one taxa with just four subspecies.

Cheers, Scott