Reptile & Amphibian Forums

Just published online:

WÜSTER, W., A.J. DUMBRELL, C. HAY, C.E. POOK, D.J. WILLIAMS & B.G.FRY (2005, in press) Snakes across the Strait: Trans-Torresian phylogeographic relationships in three genera of Australasian snakes (Serpentes: Elapidae: Acanthophis, Oxyuranus and Pseudechis). Molecular Phylogenetics and Evolution 34(1): 1-14.

This can be downloaded as a PDF at the link below.

The paper also addresses some systematic issues concerning death adders and Pseudechis which have caused some discussion in the past...

Cheers,

WW
WW Publications

-----
WW Home

Thanx for sharing Wolfgang

It's a very interesting publication giving clarity to the many discussions been placed on this and other forums over the last year or two i think.

Is a revision of the Acanthophis species on its way?
-----
Regards

Jan Grathwohl

HERPBREEDER.com - The Herpetological database
HERPBREEDER.dk - My private collection

>>Thanx for sharing Wolfgang
>>
>>It's a very interesting publication giving clarity to the many discussions been placed on this and other forums over the last year or two i think.
>>
>>Is a revision of the Acanthophis species on its way?

Thanks for that A revision is in the works, in the medium term.

Cheers,

WW
-----
WW Home

Yes I agree, an excellent paper. Not just because of the phylogeny, but I am finding other aspects very interesting as well.

Cheers, Scott

Scott Thomson
Carettochelys.com

-----
Scott Thomson

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

I am not sure the latest paper by WW et al. really is "giving clarity to the many discussions been placed on this and other
forums over the last year or two" or not. I haven't read the paper. But judging from the title of the paper and where it has been submitted, it looks like molecular characters were used, most likely mtDNA. It does confirm my claim that systematists are increasingly turning to molecular characters (whenever possible) to investigate phylogenetic relationships.

Even though I haven't read the paper, I am pretty sure that if there are any paraphyletic taxa that have been uncovered by the data, these taxa will be disqualified from the classification, given WW's professed intolerance of paraphyletic taxa because of his adherence to cladistic dogma.

WW,

I am not trying to start a flame war.

I have read the paper and i have a few questions:

Why did'nt you sample O. microlepidotus?

You say that you analyzed the phylogeny of the genus Oxyuranus and yet your sample size for the Coastal Taipan sample was one from Irian Jaya, One from PNG and a single specimen from Cairns. What about the ones in W.A and in the N.T?

Any reason why you did'nt sample Cannia weigeli, I believe the holotype is in the W.A Museum?

Why not sample P. colletti, P. guttatus with locality data?

Why not use P. porphyriacus from its Type locality, or better still why not sample at least one of each of the four disjunct populations?

Why did'nt you sample any of the Acanthophis from the Northern Side of the Central Highlands and the Central highlands themselves as well as the insular populations to the north and west?

In Table one you refer to many Acanthophis sp., why did'nt you use the names that you believe are now correct for those species eg Acanthophis sp. Anthony lagoon, Barkely Tableland N.T....would therefore become either A. hawkei or being conservative A. rugosus?

Why not use the Name Pseudechis pailsei instead of "Pseudechis cf. australis" Mount Isa and Pseudechis rosignoilii for "Pseudechis cf. australis" Irian Jaya.

On the name Pseudechis itself why not make the change to Cannia?

In reference to Pseudechis porphyriacus... "the position of the latter sister taxon of all other Pseudechis is only weakly supported"

so therefore should they remain Pseudechis (Being the type species" and the others moved into Cannia.

Was ther any reason for not looking at specimens from the type locality for Cannia australis?

Regards,

Scott Eipper.

I would like to take the opportunity to make a point here on names, not on any of the other points raised.

The issues concerning the use or non-use of names has become a favorite among herpetologists in Australia. Starting obviously with the publications by Wells and Wellington, 1985.

In the paper by Wuster et al. that discussed amature contributions to taxonomy, although I do not like the term amature I do agree with one point that is made. Determining if a name should be used is a separate issue from wether the taxon it is applied to is valid.

A. hawkei and A. rugosa are both Wells and Wellington names. Now as to wether they should be used is not up to anyone unless they can demonstrate it to be unavailable. In other words refute or accept, liking the the name, or not liking it has nothing to do with it.

Therefore, I ask is there a published paper refuting the validity of these names? Has anyone examined the paper, and published in a reviewed journal that the names cannot be used, stating the reasons? If not then they have to be used. Or they have to be sunk in the paper in question. The same goes for the genus name Cannia (Gray, 1842), is there a publication that resurrects it. If not then the fact that it has been synonimised with Pseudechis stands.

Wuster claims that O.s.barringeri is a nomen nudem as an example, giving reasons, though short reasons. Hence the name has been dealt with. After the original author posted that his name was valid I had a look myself and agree the name is Nomen Nudem, it had no diagnosis that used taxonomic characters as defined by the ICZN. In other words you cannot define a taxon by its locality as the author did.

Iverson et al, went through all the names proposed for turtles by Wells and Wellington and determined their validity against the ICZN rules, this was published in a peer reviewed International Journal. This is the best and really only way to sort out contentious names issues. We examined the types, the type data, the descriptions. We did not set out to destroy every name, we did not take the politics into account at all. If a name was valid then it stood, if not we sank it. Hence we ended up accepting E. purvisi, E. s. worrelli and the genus Macrochelodina. These are now all valid names, and have to be used. Wether people like the idea of splitting Chelodina into two general is irrelevant. If they do not like it they have to prove Chelodina and Macrochelodina are conspecific, not that the name should not be used. Hence it has gone onto the next step. This is the way it should be done.

Cheers, Scott
-----
Scott Thomson

http://www.carettochelys.com
http://ittn.net

Hi Scott,

Just a few little corrections to your comments. Firstly, Acanthophis rugosus was not a Wells and Wellington name. It was actually first described by that true gentleman of herpetology the late Arthur Loveridge - as Acanthophis antarcicus rugosus - in the Memoirs of the Museum of Comparative Zoology (Harvard), 101: p. 392. We have however, always regarded it as a valid species in its own right from the island of New Guinea. I would also respectfully like to suggest that anyone who seriously entertains the idea that rugosus is conspecific with Acanthophis hawkei should apply for a white cane and a guide dog as soon as possible.
Secondly, the name Cannia was not described by Gray ! It was in fact described by that pair of imbeciles Wells and Wellington in 1984 (see the Synopsis). Not being picky...just keeping you on your toes Scott.

Best Regards from

Richard Wells

Sorry Richard,

I am not up on the snakes as well as can be said for the turtles.

I did however see a reference that stated that Cannia was described by Gray 1842, with a full citation so I assumed that was correct. However, if you described it and it is valid and no-one has sunk it, then it is valid and should be used.

As for A rugosus, if you resurrected it and no one has since refuted that, again it is valid.

ahh nomenclature.... makes you go grey early doesn't it...

Cheers, Scott
-----
Scott Thomson

http://www.carettochelys.com
http://ittn.net

>>I did however see a reference that stated that Cannia was described by Gray 1842, with a full citation so I assumed that was correct. However, if you described it and it is valid and no-one has sunk it, then it is valid and should be used.

Correction/addition: It should be used if there is reason to recognise as a separate genus the group of species for which it is the oldest available name. At present, that depends on whether Pseudechis sensu Cogger is monophyletic. Our data did not resolve the issue one way or another, others are working on this with more data, so we stuck with the status quo for the time being.

Cheers,

WW
-----
WW Home

WW wrote:
"It should be used if there is reason to recognise as a separate genus the group of species for which it is the oldest available name. At present, that depends on whether Pseudechis sensu Cogger is monophyletic."

Me:
Actually, what WW meant was that whether Cannia is recognizable as a valid genus will depend on whether Pseudechis is paraphyletic or not. If it is paraphyletic, then the species in Cannia would need to be transferred back to Pseudechis, or else Pseudechis would need to be splintered into as many pieces as neccessary to eliminate paraphyly. In the case of the green tree python, Kluge thought that it was necessary to include Chondropython viridis as a member of the genus Morelia so as not to [gasp] recognize a paraphyletic Morelia. In the case of the ratsnakes and their descendants, Utiger et al. took the opposite tact. Instead of invalidating Lampropeltis, Pituophis, Arizona, Bogertophis, Cemophora, Stilosoma and Rhinocheilus and transferring all the species within these genera back to Elaphe in order to eliminate paraphyly, they instead splintered Elaphe into a dozen or so genera (which are morphologically indistinguishable from one another) to eliminate paraphyly instead.

The cladists' ideological intolerance of paraphyletic taxa therefore would result in wholesale taxonomic changes no matter which extreme option they exercise. And the cladists are straightjacketed by their own ideology to either split excessively or lump excessively to eliminate paraphyletic taxa.

A sane alternative would be to recognize paraphyletic taxa as valid, as Charles Darwin and the Darwinians do. To an evolutionary or Darwinian systematist, it does not matter whether Pseudechis is paraphyletic or not. It only matters whether Cannia is monophyletic (or paraphyletic) and whether it is sufficiently different morphologically from Pseudechis to merit generic recognition. If Cannia is monophyletic or paraphyletic (meaning that all members share the same common ancestor even though not all descendants of the same ancestor are included) and if it is different enough from its ancestral genus, then it merits recognition as a distinct genus. The Darwinian alternative is sane and it is much less destructive taxonomically than the cladists insane intolerance of paraphyletic taxa.

Again, not having read WW's latest unpublished paper, I will not comment on the taxonomic issues therein. But Scott Thomson made an assertion which which I disagree.

Scott Thomson:
"Determining if a name should be used is a separate issue from wether the taxon it is applied to is valid."

Me:
I do agree with this statement. Whether a particular name is valid or available is a separate issue from whether a taxon to which the name may apply is valid or not.

Scott Thomson:
"Iverson et al. ... examined the types, the type data, the descriptions. We did not set out to destroy every name, we did not take the politics into account at all. If a name was valid then it stood, if not we sank it. Hence we ended up accepting E. purvisi, E. s. worrelli and the genus Macrochelodina. These are now all valid names, and have to be used. Wether people like the idea of splitting Chelodina into two general is irrelevant. If they do not like it they have to prove Chelodina and Macrochelodina are conspecific, not that the name should not be used. Hence it has gone onto the next step. This is the way it should be done."

Me:
Chelodina and "Macrochelodina" are higher taxa. Higher taxa are entirely categories of artificial human construct. There is no universally accepted definition for any higher taxa, although it has been customary to define animal phyla on the basis of their different body plans. Different biologists may recognize different numbers of kingdoms for example. Hence it is possible that some people will recognize two or more higher taxa (including genera) while others may only recognize one. Such disagreements are entirely philosophical and subjective. Hence there is no way to "prove" whether two species belong in the same higher taxon or not. Therefore there is nothing that could compel anyone to recognize "Macrochelodina." Further, even if the genus Chelodina were to be split, the name "Macrochelodina" may or may not be used. For example, if the type species of "Macrochelondina" should be considered a member of Chelodina, then the name becomes a junion synonym. In that case, another name may need to be applied to the other genus or genera resulting from the split. Whether "Macrochelodina" is a valid taxon or not is therefore, as Scott Thomson had stated earlier in his post, a different issue than whether the name is valid or not.

ME:
I was trying to not get carried away on the details...

Ok first up I agree with most of what you said and reading what I said I used an incorrect term in saying "conspecific" at one stage.

All we looked at was the availablity of the name, we did not analyse wether the group of species it refers to (also known as the Chelodina B group) warrants generic recognition. That said we took the view that the original description of Macrochelodina claimed that this was a genus in its own right. There has been no evidence presented that suggests otherwise though in saying that the issue was not directly tested. Hence if the name applied to a valid group that was unnamed and did not fail any other tests you mention then it was the oldest available name for that taxon.

The original type species of Macrochelodina was M. oblonga. This caused serious issues for two reasons. First up because this species is a member of the Chelodina A, and second because the name C. oblonga has been incorrectly applied.

Thomson 2000, showed that the holotype of C. oblonga was in fact a C. rugosa and that the valid name for the Western Australian species is Chelodina colliei. However I took the stand in that paper to maintain the status quo awaiting a decision by the ICZN on the issue. We cannot right it off as Nomen oblitum or something because the name C. oblonga was correctly applied till 1974.

That is well and good for Chelodina oblonga (=colliei) but the effect on Macrochelodina is different. Types are conserved in nomenclature hence the type species of the genus is that taxon to which the holotype belongs, ie Chelodina rugosa. All these corrections were made in the paper.

So the conclusion was that the name Macrochelodina was the available name for the Chelodina "B" (or Chelodina expansa group) and was the oldest available name for the genus.

Wether it should have been recognised is another matter. There has been considerable evidence published that demonstrates a completely different bauplan for the Chelodina and the Macrochelodina. I use the name now bexause I feel it has been described and I use the view of "refute or accept", despite describing two species in this group and assigning another I have no evidence that can refute this as a genus. As the members of each genus have a different bauplan, they both form monophyletic groups (I know you hate that one) and there is a "decided gap" between them that is more extensiive than that between any short neck genus. Though I recoignise that genera do not have to be equal.

CKing:
"Chelodina and "Macrochelodina" are higher taxa. Higher taxa are entirely categories of artificial human construct. There is no universally accepted definition for any higher taxa, although it has been customary to define animal phyla on the basis of their different body plans. Different biologists may recognize different numbers of kingdoms for example. Hence it is possible that some people will recognize two or more higher taxa (including genera) while others may only recognize one. Such disagreements are entirely philosophical and subjective. Hence there is no way to "prove" whether two species belong in the same higher taxon or not. Therefore there is nothing that could compel anyone to recognize "Macrochelodina." Further, even if the genus Chelodina were to be split, the name "Macrochelodina" may or may not be used. For example, if the type species of "Macrochelondina" should be considered a member of Chelodina, then the name becomes a junion synonym. In that case, another name may need to be applied to the other genus or genera resulting from the split. Whether "Macrochelodina" is a valid taxon or not is therefore, as Scott Thomson had stated earlier in his post, a different issue than whether the name is valid or not."

Cheers, Scott
-----
Scott Thomson

http://www.carettochelys.com
http://ittn.net

Scott Eipper asked many questions of WW et al. as to why some localities were not sampled. Most likely it is because they either cannot obtain tissue samples from those localities or in their judgement the samples they have is adequate to address the question they are attempting to answer.

As to the following question, also asked by Scott Eipper:

"Why not use the Name Pseudechis pailsei instead of 'Pseudechis cf. australis' Mount Isa and Pseudechis rosignoilii for 'Pseudechis cf. australis' Irian Jaya."

This is an easy enough question to answer. WW et al. are adhering strictly to Hennigian classificatory convention of recognizing only "clades," or holophyletic groups, which means groups that consist of one single ancestor and all of the descendants of that single ancestor.

That means WW et al. will only recognize clades, but not paraphyletic groups. If the P1b clade is removed from the genus Pseudechis, the remaining species will not form a holophyletic group, since it will then consist of clades P3, P2, P1a and P1c, but not P1b. In the Hennigian's or cladist's jargon, Pseudechis minus P1b (or any other descendant group) would be paraphyletic. No self respecting cladist or Hennigian will ever do such an unthinkable thing as to recognize paraphyletic taxa, no matter how different the P1b clade is morphologically from the genus Pseudechis. WW et al. are no exception. This sort of behavior was predicted by me before I even read their paper. The cladists' dogmatic intolerance of paraphyletic taxa is perhaps one of most destructive forces facing taxonomy today.

Their decision not to use Hoser's specific names for the P1b clade is stated on page 8. Their analysis finds equivocal support for the "monophyly" (or holophyly) of the P1b clade. Hence to be safe, they have apparently decided to consider the P1a and P1b clades to be a single species, at least for now. Whether or not the P1b populations represent a different species from P. australis cannot be answered by the mtDNA data. Since 2 specimens of Mt Isa P. australis show a great deal of mtDNA divergence, it would be interesting to find out if these 2 specimens are from two geographically isolated populations or whether they are sympatric but fail to interbreed. Such data would be a lot better at answering the question of whether one or more species should be recognized in the P. australis complex than mtDNA data alone.

CKing said:
"The cladists' dogmatic intolerance of paraphyletic taxa is perhaps one of most destructive forces facing taxonomy today."

Richard Wells said:
Well, being one that can identify with "destructive forces", I can well see the point you are making..., a bit dogmatically, but with no small amount of justification.

I remember first reading the papers of Willie Hennig and his apparatchiks and thinking what the hell is going on here? I must confess that as the Cladist dogma spread, it turned me right off "Systematic Zoology", a journal that I just couldn't wait to read previously. I know it might seem a bit inappropriate, but I have always seen cladism also as a convenient way for creationists to get involved in evolutionary biology at a purportedly taxonomic level. Prayer, blind faith and slavish idolatry would be major prerequisites for a cladistic approach to taxonomy, so creationists would be preadapted to the method. It seems to me that some cladists believe that they have true power to comprehend it all, as though they are on a path that will lead to a holy grail of biology - a sort of Combined Field Theory of taxonomy. But if Einstein had been a cladist...who knows, maybe his work might not have become "one of most destructive forces facing [humanity] today", so maybe there are worse problems than cladists...Sorry, must go...time's getting on you know. I have to go and split-up a few more paraphyletic genera before bedtime!

Richard Wells

Richard Wells wrote:
"Prayer, blind faith and slavish idolatry would be major prerequisites for a cladistic approach to taxonomy, so creationists would be preadapted to the method."

Me:
I think you have a good observation of the behaviors of some cladists, because--as P.C.H. Pritchard pointed out--sometimes cladism does take on the overtones of religious dogma and that some cladistic texts read more like the Koran than a scientific writing. Many cladists' adherence to Hennig's principle of holophyly (which states that a taxon must consist of one ancestor and all of its descendants) and Hennig's suggestion that only branching order should be taken into account when ranking organisms can only be described as being slavish.

That said, the cladists' approach to taxonomy is not really in any way similar to the creationists'. After all, ever since Charles Darwin, biologists have adhered to Darwin's suggestion that "the arrangement of the groups within each class, in due subordination and relation to other groups, must be strictly genealogical in order to be natural." The classifications of the Darwinians and cladists, and of course those of the pheneticists as well, are based on the idea (though not often realized in practice because of a lack of knowledge) that each taxon must share a common ancestor in order to be scientifically acceptable. Known unnatural polyphyletic groups that are based on convergent similarities are routinely rejected by all schools of biological classification.

If taxonomists of all schools reject polyphyletic groups, then what is the real difference between the various schools? For a detailed description of the many differences between the evolutionary systematists and the cladists, please refer to chapter 10 of Mayr and Ashlock's excellent book "Principles of Systematic Zoology, 2nd ed." The main difference between the cladists and all other schools of taxonomy is that cladist are unique in their adherence to Hennig's principle of holophyly. Cladists reject taxa that do not include some but not all descendant groups of a common ancestor. For example, referring to WW et al.'s paper, WW et al. would not recognize "Pailsus pailsei" because if the genus "Pailsus" is recognized, then Pseudechis becomes "paraphyletic," the same way that Reptilia is paraphyletic because two of the descendant groups of the common ancestor of Reptilia, namely birds and mammals, are excluded from Reptilia and classified in 2 other separate taxa, Aves and Mammalia. WW et al.'s rejection of "Pailsus" therefore has nothing to do with whether "Pailsus" is morphologically divergent from Pseudechis or not. To give an extreme example, even if "Pailsus" were to evolve a shell and a beak (thus making these snakes look more like turtles than snakes), the principle of holophyly would still forbid the cladists from recognizing it as a separate genus.

There is however, one way that a shelled and beaked "Pailsus" can be recognized that would nevertheless be acceptable to the cladists. That alternative would be to splinter the paraphyletic genus Pseudechis into many different but morphologically indefinable and indistinguishable genera. That means P. porphyriacus and P. guttatus would need to be classified in two different genera in order for "Pailsus" to be recognized, according to cladistic dogma. This second approach, while cladistically correct, is nonsensical since it substitutes one set of problems (way too many genera that are morphologically indistinguishable from one another) for another problem (a single morphological disparate genus). It is this second approach that Utiger et al. took to splinter the paraphyletic genus Elaphe.

Hopefully it is clear why evolutionary taxonomists have so stronly objected to many of the cladists' radical reclassification of familiar taxa. Fortunately, there is hope. As Ernst Mayr pointed out, "There is nothing in any theory of classification that would require one to rely on [Hennig's] principle of holophyly." Indeed, even systematists who are not cladists can rely on Hennig's principle when classifying taxa. Hedges, for example, relied on Hennig's principle of holophyly when he reclassified Hyla regilla as "Pseudacris regilla," even though the type of data (distance data) he used is routinely rejected by cladists on philosophical grounds.

Cladistic analysis, therefore, is not to blame for much of the taxonomic chaos being generated, because after all, a phylogenetic analysis is little more than a hypothesis of phylogeny. It is most cladists' slavish reliance on Hennig's untenable principle of holophyly which is the main source of taxonomic chaos being generated.

Scott,

The answer to all your questions is: because, as the title indicates, it is in the first instance a biogeography paper, not a detailed taxonomy paper. Therefore, we only included samples that were relevant for the biogeographical question at hand, concerning biogeographical relationships between Aussie and NG representatives of these genera. Yes, we did draw the taxonomic inferences our data allowed, but also indicated the many issues that remain unresolved.

Cheers,

WW
-----
WW Home

Scott,

Although my previous reply addresses most issues you raise, some aspects of nomenclature require further elaboration:

>>In Table one you refer to many Acanthophis sp., why did'nt you use the names that you believe are now correct for those species eg Acanthophis sp. Anthony lagoon, Barkely Tableland N.T....would therefore become either A. hawkei or being conservative A. rugosus?

Because mitochondrial DNA data alone cannot resolve species status. As we wrote on page 11:

"Our data reveal strongly supported, hitherto unsuspected patterns of genetic diversity within the death adders (Figs. 2 and 3). These patterns differ radically from all previous interpretations of species limits within this medically important genus. This study thus provides a first phylogenetic framework for a thorough and much-needed revision of the systematics of the genus, although a phylogeographic study based on mitochondrial DNA may not be able to resolve species limits without additional evidence, especially where the suspected species are parapatric or sympatric (Puorto et al., 2001). "

and:

"However, we refrain from attempting to diagnose species limits on the basis of mtDNA phylogeography alone (Puorto et al., 2001). More detailed and rigorous studies combining an analysis of morphology (particularly multivariate morphometrics—e.g., Wuster
and Broadley, 2003; Wuster et al., 2001b) and nuclear markers are required to resolve the population systematics of these snakes. The Merauke death adders were described as Acanthophis antarcticus
rugosus by Loveridge (1948). We suggest referring to the rough-scaled Irian Jaya death adder and the Northern Territory and northwestern Queensland death adders of clade A1 as the A. rugosus complex, pending further revisions."

>>Why not use the Name Pseudechis pailsei instead of "Pseudechis cf. australis" Mount Isa and Pseudechis rosignoilii for "Pseudechis cf. australis" Irian Jaya.

Because using the names as such would have implied acceptance of these forms as separate species. There was no basis for doing so in the intro and materials etc of the paper, since there was little or no published evidence supporting their species status.

Now, from our data it certainly looks like they are separate species, but, as noted above, mtDNA alone cannot determine this. Using the term "Pseudechis cf. australis" simply indicates that we have reserved judgement on the issue.

>>On the name Pseudechis itself why not make the change to Cannia?
>>
>>In reference to Pseudechis porphyriacus... "the position of the latter sister taxon of all other Pseudechis is only weakly supported"
>>
>>so therefore should they remain Pseudechis (Being the type species" and the others moved into Cannia.

See my response to Richard Wells. In the absence of clear evidence that Pseudechis is non-monophyletic, we decided not to fix it.

Cheers,

Wolfgang

>>
>>Was ther any reason for not looking at specimens from the type locality for Cannia australis?
>>
>>Regards,
>>
>>Scott Eipper.
-----
WW Home

Mmmmmm. Molecules and Invasions...interesting - though a bit confusing to one that is not part of the molecular illuminati, but definitely interesting. I am a little surprised at the apparent relatively recent separation of the two Oxyuranus scutellatus members...I think you should go over your data again or apply some other bit of molecular gymnastics because I have seen both scutellatus and canni - and I considered them to be parapatric species - and if my suspicion is correct, then I think they may have differentiated more than your present data indicates. You alluded to the genus Oxyuranus also including microlepidotus as most others seem to believe, but I feel this should be revisited to - Parademansia has two quite different populations of microlepidotus within it, one in south-western New South Wales (Yes, the species is NOT EXTINCT within this State as is commonly believed) and the other in south-western Queensland and north-eastern South Australia. Sure, they have some similarities, but I still believe that their similarity to Pseudonaja may indicate more than just convergence. But on scutellatus I agree with your rejection of barringeri - although I still remain convinced that scutellatus will be found to be a polytypic species within Australia. Similarly, I suspect that variation within canni itself may be more significant than hitherto realized. As for your treatment of Acanthophis, well you guys are pretty much on the right track I think, and much more needs to be done of course, for when you look at more material it will become clearer I feel. As for "Pseudechis" well as I have long maintained, the traditional use of this genus for that assemblage of species makes no ecological or morphological sense at all. I still hold the view that Pseudechis contains only the porphyriacus populations. The australis members (except the weigelli complex) should be placed within Cannia. The weigelli complex are so distinct that Ross Wellington and I almost placed them in a separate genus back when we described weigelli in 1987, but decided not to as "Cannia" seemed like an easier pill for others to swallow at the time (yes, we were being conservative!). Hoser's "Pailsus" may in fact be an available generic name for the weigelli complex, but this will have to be for others to determine with more certainty than I can.
Anyway, I enjoyed your paper (with the exception of the boring stuff about molecules), as I have always had a keen interest in biogeography. Keep it up. 7/10 for effort.

Richard Wells

>>Mmmmmm. Molecules and Invasions...interesting - though a bit confusing to one that is not part of the molecular illuminati, but definitely interesting. I am a little surprised at the apparent relatively recent separation of the two Oxyuranus scutellatus members...

Join the club - my first thought was that this was a sampling mix-up, so everything was redone - but the results stayed the same, also with additional samples.

> I think you should go over your data again or apply some other bit of molecular gymnastics because I have seen both scutellatus and canni - and I considered them to be parapatric species - and if my suspicion is correct, then I think they may have differentiated more than your present data indicates.

Quite possibly. Before we got those sequences back, I had rather assumed I would end writing about evidence in favour of canni being a separate species.

All the mtDNA sequences really tell us is that there has clearly been recent and extensive gene flow between the two. It is of course possible that the Aussie and NG populations have been in situ much longer, but that introgressive hybridization swamped out the mtDNA haplotypes of one of the two populations, while other alleles were retained - we can't tell, at least not from our current sampling, which is also why we did not sink canni formally. Other molecular methods (e.g., AFLPs) would certainly be of interest also.

> You alluded to the genus Oxyuranus also including microlepidotus as most others seem to believe, but I feel this should be revisited to - Parademansia has two quite different populations of microlepidotus within it, one in south-western New South Wales (Yes, the species is NOT EXTINCT within this State as is commonly believed) and the other in south-western Queensland and north-eastern South Australia.

I look forward to your package with samples from these localities

> Sure, they have some similarities, but I still believe that their similarity to Pseudonaja may indicate more than just convergence.

Basically, O. microlepidotus retains a more ancestral morphology, which would have been more similar to Pseudonaja, whereas O. scutelatus evolved its present much more derived morphology later. We do have sequence data for O. microlepidotus, and they show it to be the closest relative of O. scutellatus - in fact that relationship is one of the most strongly corroborated phylogenetic relationships among Aussie elapids, from a numebr of different marker systems. Obviously, whether one chooses to recognise Parademansia as a separate genus or not based on morphological differentiation is, at the end of the day, a purely personal and arbitrary decision.

> As for "Pseudechis" well as I have long maintained, the traditional use of this genus for that assemblage of species makes no ecological or morphological sense at all. I still hold the view that Pseudechis contains only the porphyriacus populations. The australis members (except the weigelli complex) should be placed within Cannia. The weigelli complex are so distinct that Ross Wellington and I almost placed them in a separate genus back when we described weigelli in 1987, but decided not to as "Cannia" seemed like an easier pill for others to swallow at the time (yes, we were being conservative!). Hoser's "Pailsus" may in fact be an available generic name for the weigelli complex, but this will have to be for others to determine with more certainty than I can.

For Pseudechis, we went with nomenclatural stability. I agree, P. porphyriacus is very distinct indeed genetically from the other species of the genus, but since our data did not provide clear evidence that it is NOT monophyletic, since there is some morphological evidence indicating monophyly of all Pseudechis, and since others are working on this, we left it in the traditional way. Nomenclature is, in the first instance, an informaion retrieval system, and nomenclatural changes are disruptive, if not for us taxonomists then for everyone else (just like I get p*ssed off when my local supermarket changes all the shelves around and it taes me twice as long to do my shopping). Consequently, when changes are proposed, I regard the onus of proof as resting on those proposing the change. Our data did not clearly show Pseudechis to be broken, so we didn't fix it. If later work does show Pseudechis to be non-monophyletc, ten I will be happy to cherr Cannia along for the other species of the genus. Heck, I could even countenance recognising Panacedechis for guttatus, colletti and papuanus if Pseudechis is going to be split at all. However, I don't see any reason to place the various weigeli, pailsi, rossignolii etc. into a separate genus from Cannia - they are similar enough to australis (which is why we regard them as different species), and the evidence for the monophyly of "Pailsus" is practically non-existent in any case.

Cheers,

Wolfgang
-----
WW Home

WW:
All the mtDNA sequences really tell us is that there has clearly been recent and extensive gene flow between the two. It is of course possible that the Aussie and NG populations have been in situ much longer, but that introgressive hybridization swamped out the mtDNA haplotypes of one of the two populations, while other alleles were retained - we can't tell, at least not from our current sampling, which is also why we did not sink canni formally. Other molecular methods (e.g., AFLPs) would certainly be of interest also.

Me:
I wouldn't say that. If one of these two localities was not occupied by snakes within this complex, then all that the mtDNA data tells us is that there had been a recent migration from one locality to another. Migration into a previously unoccupied area should not be called gene flow between two populations. If a single gravid female or only a few individuals migrated to a new area and never interbred with any individuals of the ancestral population, then there is technically no "gene flow" between the new population and the ancestral population. Further, if only a few individuals migrated, then founder effect divergence may result in rapid morphological divergence, as can adaptation to a new habitat.

WW:
Obviously, whether one chooses to recognise Parademansia as a separate genus or not based on morphological differentiation is, at the end of the day, a purely personal and arbitrary decision.

Me:
The recognition of a new genus on the basis of mophological disparity is indeed subjective. However, the grouping of taxa on the basis of holophyletic groups is even more subjective. As Kluge pointed out in his classification of "Charina," he could have recognized either one, two or three genera for the 3 species he was classifying. His decision to recognize only one genus for these 3 species is purely personal and arbitrary even by his own reckoning. His decision is based on "taxonomic efficiency," which is of course subjective and arbitrary.

Richard Wells:
> As for "Pseudechis" well as I have long maintained, the traditional use of this genus for that assemblage of species makes no ecological or morphological sense at all. I still hold the view that Pseudechis contains only the porphyriacus populations. The australis members (except the weigelli
complex) should be placed within Cannia.

WW:
For Pseudechis, we went with nomenclatural stability. I agree, P. porphyriacus is very distinct indeed genetically from the other species of the genus, but since our data did not provide clear evidence that it is NOT monophyletic, since there is some morphological evidence indicating monophyly of all Pseudechis, and since others are working on this, we left it in the traditional way.

Me:
Taxonomic stability is not always on the top of the list of WW's reasons for supporting a taxonomic proposal or not. For example, Utiger et al.'s proposed splintering of the paraphyletic Elaphe was fully supported by WW, even though it will create taxonomic chaos by the truckload. Similarly, Kluge's lumping of Morelia and Chondropython into a single genus and Kluge's lumping of Charina, Calabaria and Lichanura into another genus were both fully supported by WW. As long as a taxonomic proposal conforms to the cladistic intolerance of paraphyly, it appears that WW will tend to support it, regardless of whether the proposal will result in taxonomic stability and regardless of whether the proposal represents excessive lumping or excessive splitting. However, if someone, anyone, were to propose a taxonomic rearrangement that will result in the recognition of paraphyletic taxa, it is almost a sure bet that WW will oppose it.

WW:
Heck, I could even countenance recognising Panacedechis for
guttatus, colletti and papuanus if Pseudechis is going to be split at all. However, I don't see anyreason to place the various weigeli, pailsi, rossignolii etc. into a separate genus from Cannia - they are similar enough to australis (which is why we regard them as different species), and the evidence for the monophyly of "Pailsus" is practically non-existent in any case.

Me:
Of course WW will support recognizing guttatus colletti and papuanus as a genus since these three species form a neat little holophyletic group in his cladogram. He cannot, however, support the recognition of "Pailsus" because doing so will render Pseudechis paraphyletic. I am not arguing that Paisus should be recognized or not. That decision should be left to those who are experts in this taxon. If someone presents good evidence to show that it is evolutionarily divergent from Pseudechis, then I would definitely support its recognition, even if Pseudechis will be rendered paraphyletic. On the other hand, as I pointed out elsewhere, even if "Pailsus" were to evolve a shell and a beak, the cladists and WW would oppose recognizing "Pailsus" as a separate genus from Pseudechis, unless Pseudechis is also broken up into as many holophyletic groups as necessary so that no paraphyletic group will result from the recognition of "Pailsus." Therein lies the difference between WW and the cladists on the one hand and the Darwinians and I on the other hand. There is no middle ground between these two incompatible classificatory philosophies. That is why WW and I will continue to disagree on most taxonomic matters for the foreseeable future.

Richard Wells wrote:
"Mmmmmm. Molecules and Invasions...interesting - though a bit confusing to one that is not part of the molecular illuminati, but definitely interesting. I am a little surprised at the apparent relatively recent separation of the two Oxyuranus scutellatus members...I think you should go over your data again or apply some other bit of molecular gymnastics because I have seen both scutellatus and canni - and I considered them to be parapatric species - and if my suspicion is correct, then I think they may have differentiated more than your present data indicates."

Me:
I think you are unfamiliar with the scientific fact that molecular evolution occurs independently of morphological evolution. Basically, many molecules that the molecular systematists study are selectively neutral, or nearly so. Mutations in these molecules occur rather randomly, so that the chance that a particular mutation will be fixated in a population is virtually entirely due to chance. Therefore, like the atomic clock, the "molecular clock" really measures (albeit much less accurately than the atomic clock) roughly how long ago two species may have shared a common ancestor. mtDNA is very useful for this purpose. In fact, mtDNA from a Neandertal man suggests that Neandertals last shared a common ancestor with modern humans about 500,000 years ago, whereas all living human populations last shared a common ancestor about 150,000 years ago. However, that does not necessary mean that Neandertals are a different species than modern humans. To determine that, one has to look for evidence to see if Neandertals can interbreed successfully with modern humans. Since Neandertals are extinct, that determination is quite difficult and the taxonomic status of Neandertals therefore remains controversial as well as unresolved.

Because molecules evolve independently of morphology, sometimes a species can diverge rather quickly from its closest relative in a relatively short period of time. Conversely, two or more populations of a slowly evolving species may display a great deal of mtDNA divergence (or a great deal of nuclear DNA divergence) and yet be morphologically similar. A real world example is presented below. In an earlier post,

http://forums.kingsnake.com/view.php?id=609826,622134

I reproduced part of an actual branching diagram (based on mtDNA) from a published paper but I removed all of the names. I then challenge anyone to tell me how many species are represented in that diagram. Not surprisingly there are no takers.

In the diagram below I will review the answer. The names I use differ from those in the original paper, as I personally do not agree with the split of the genus Scaphiopus into the genera Spea and Scaphiopus. As one can see from the diagram, the species Scaphiopus bombifrons and S. intermontanus have only recently diverged from one another, and yet they are recognizable as distinct species. On the other hand, the populations of S. hammondii in San Diego and Alameda Counties, California show much greater mtDNA divergence than that between S. bombifrons and S. intermontanus. This is illustration that a little mtDNA divergence does not necessarily mean that there cannot be a disproportionately large amount of morphological divergence. WW et al.'s data may accurately reflect the timing of the split between the taxa they studied, but their data cannot tell us how different two taxa may be, simply because mtDNA data is not a measure of morphological evolution.

I have also chosen to classify S. multiplicatus as being conspecific with S. hammondii. This is done to highlight the fact that despite the considerable mtDNA distances between hammondii and multiplicatus, these two taxa are still recognizable as close relatives and they were for a long period of time regarded as different, allopatric populations of the same species. In sum, it is not possible to learn from mtDNA data whether two populations are the same species and how different they may be morphologically. These are questions that must be answered by other sorts of data. What mtDNA is very good at is to reveal past migratory patterns and which populations of a species may be the oldest, as well as revealing which particular species may be basal or even ancestral to another, provided of course that the researchers choose their outgroups carefully.

CKing: "Therefore, like the atomic clock, the "molecular clock" really measures (albeit much less accurately than the atomic clock) roughly how long ago two species may have shared a common ancestor."

Richard Wells: Yeah, so I have been told...and this, like much of molecular biology I pretty well have to take the word of others better informed. I claim no skill or special knowledge in this obviously important area at all, yet this inadequacy has been no impediment at all to me in the identification of biological species. And indeed, I can confidently challenge any molecular biologist professing expertise in vertebrate taxonomy to join me in field work and we shall see who is the superior in identification. But in my blissful ignorance, I do chuckle a little when I see "facts" emerge from bases like the concept of "DNA molecular clocks" that are at best theoretical postulates. I was under the impression that the rate of nucleotide substitution is presumed to be a constant across the entire genome of a species, and the degree of divergence in these nucleotide sequences between two different species enables an estimation of the time of separation (ie the point when they branch from their common ancestor). Exciting stuff to be sure, but how reliable is the basis for this postulate, when the significance of the divergence it is presumably based on the delta T50H? I mean to say, this seems pretty hairy stuff to me - that the genetic relationship between the nucleotide sequences of two or more species can be determined by the difference between the temperatures at which DNA homoduplexes and DNA heteroduplexes undergo that magic 50% dissociation. And what about the accuracy of identifying repetitive sequences for their removal by interpretation of the reassociation kinetic data? Also, what if there is no independant dating reference point from the fossil record, thus obstructing the conversion of the delta T50H value to an absolute time interval? Don't get me wrong - I'm not bagging the molecular clock hypothesis as erroneous - I just see fertile ground for unwarranted extrapolation. To me the concept is more a "work in progress", than a complete understanding, so I am a little uneasy at the "absolutist" interpretations that are made on the basis of this tool at times. So please forgive me for not immediately jumping up and down in fits of joyous commendation everytime a molecular theory is transformed into a molecular fact by the slick application of some sort of statistical gymnastics or another. My scepticism may be more a product of ignorance than understanding...but I often wonder how much the molecular maestros REALLY understand about the game they are playing as well?

Richard Wells

Richard Wells wrote:
I do chuckle a little when I see "facts" emergec from bases like the concept of "DNA molecular clocks" that are at best theoretical postulates.

Me:
Like you, I was skeptical when I was first introduced to the idea of the molecular clock. However, I was won over because, like evolution, the molecular clock is not "just a theory" (as opponents of evolution suggest) but the only reasonable explanation for the set of accumulated facts in support of it, for most people who are rational. S.J. Gould writes in the essay "Our Greatest Evolutionary Step" in The Panda's Thumb:

'During the past fifteen years, students of molecular evolution have accumulated a storehouse of data on the amino acid sequences of similar enzymes and proteins in a wide variety of organisms. This information has generated a surprising result. If we take pairs of species with securely dated times of divergence from a common ancestor in the fossil record, we find that the number of amino acid differences correlates remarkably well with time since the split—the longer that two lineages have been separate, the more the molecular difference. This regularity has led to the establishment of a molecular clock to predict times of divergence for pairs of species without good fossil evidence of ancestry. To be sure, the clock does not beat with the regularity of an expensive watch—it has been called a "sloppy clock" by one of its leading supporters—but it has rarely gone completely haywire.'

My initial skepticism is shared by many Darwinians because of the way we were taught about evolution in schools. S.J. Gould explains:

'Darwinians were generally surprised by the clock’s regularity because natural selection should work at markedly varying rates in different lineages at different times: very rapidly in complex forms adapting to rapidly changing environments, very slowly in stable, well-adapted populations. If natural selection is the primary cause of evolution in populations, then we should not expect a good correlation between genetic change and time unless rates of selection remain fairly constant—as they should not by the argument stated above. Darwinians have escaped this anomaly by arguing that irregularities in the rate of selection smooth out over long periods of time. Selection might be intense for a few generations and virtually absent for a time thereafter, but the net change averaged over long periods could still be regular. But Darwinians have also been forced to face the possibility that regularity of the molecular clock reflects an evolutionary process not mediated by natural selection, the random fixation of neutral mutations."

To summarize, there are some molecules (such as serum albumin) that have no known function, and are therefore selectively neutral. A mistake in the replication of the gene coding for neutral molecules like serum albumin therefore would not be detrimental to the survival of an organism and are therefore not weeded out. In time, these mistakes accumulate in separate lineages, allowing us to trace the common ancestry of two organisms back in time. It is therefore not surprising that the molecular clock was formulated by Allan Wilson and his colleagues using a selectively neutral molecule such as serum albumin.

Richard Wells:
I was under the impression that the rate of nucleotide substitution is presumed to be a constant across the entire genome of a species, and the degree of divergence in these nucleotide sequences between two different species enables an estimation of the time of separation (ie the point when they branch from their common ancestor).

Me:
This is emphatically not the case. Different molecules do have different rates of mutation. That is because some molecules are constrained by natural selection while others are not. Those molecules that are necessary for a vital function tend to evolve very slowly whereas those that are selectively neutral can evolve quite rapidly. The same is true of morphological features as well. For example, in the article “Magnolias from Moscow” in Gould’s book Dinosaurs in a Haystack, Gould reported that DNA was successfully extracted from chloroplasts of 20million year old fossil leaves. Comparing the DNA to a closely related living species, only 17 mutations were found in a comparison of 820 base pairs, and 13 of these were found in the third position of the nucleotide sequence, resulting in no changes in the amino acid sequence. There were only 4 nucleotide substitutions that resulted in amino acid changes. In comparison, it has been estimated that one amino acid substitution occurs in serum albumin approximately every one million years. Chloroplast protein therefore evolves at a rate (4 amino acid substitutions over 20 million years or 1 per 5 million years) that is 1/5 the rate of serum albumin evolution. This is not surprising because a mutation that changes the function of chloroplast will be detrimental to the plant’s survival and thus these mutations will be quickly weeded out by evolution.

Richard Wells:
I'm not bagging the molecular clock hypothesis as erroneous - I just see fertile ground for unwarranted extrapolation.

Me:
Yes indeed. DNA data can be misinterpreted; mistakes can be made in obtaining data also. However, techniques are improving all the time. I have also pointed out some silly mistakes made by molecular systematists, who sometimes choose a member of the ingroup as an outgroup. Also, in the essay “Magnolias from Moscow”, Gould did report anomalous results from another study which reported a different rate of mutation for the same molecule. As he pointed out, there may be other factors involved. It is a long story. DNA data is far from bullet proof reliable, but the mistakes are being discovered and dealt with. Better molecular characters, such as SIN, are being found. More and more systematists, including those who used to rely on morphology alone, are turning to molecular characters to study phylogeny. However, phylogeny alone is only part of the picture. The other part of the picture is evolutionary divergence. At present, as Mayr and Ashlock pointed out, there is no way to classify organisms using molecules that are responsible for evolutionary divergence. The molecular systematists who try to classify organisms on the basis of neutral genetic changes such as serum albumin and mtDNA are giving us an incomplete and often misleading picture of evolutionary change. Those who ignore or dogmatically oppose molecular data, however, run the risk of behaving in a way that is similar to the creationists, who continue to maintain that evolution is not fact, just theory. The molecular clock, like evolution, is not merely a theory, but the most reasonable interpretation of the available data, at least according to most rational people. The molecular clock, like evolution, is a fact, not a theory.

Regards

Thank you CKing for your most enlightening response. As usual your point of view is laced with the kind of juicy references that I look forward to reading. Without dismissing its objective truths, I still see molecular systematics as running the risk of becoming one almightly card-house skyscraper unless the kind of rigor that you have repeatedly proposed is implimented. It seems to me that sooner or later the whole construct will just tumble down in a confusing miasma of misinterpretation. Its collapse is probably already underway by the way mtDNA data has been used, but a few good kicks at the bottoms cards might not go astray. I would love to lay the boot in, but generally I tend not to pick fights with gangs of thugs or with those bigger than me. If they pick on me...well, that's another matter...

Richard Wells

I am sorry to hear that you still think that molecular systematics is unreliable. I do not share your view that it is a house of cards and I see no reason why it should tumble down like one any time soon. I am not denying that morphology has a role in systematics. In fact, taxonomy is impossible without knowledge of evolutionary divergence, and such data can at present only be obtained through a study of morphological characters. Further, morphological characters can often serve as corrobative evidence for molecular systematics.

Over the past 2 decades or so, molecular systematics have answered many different questions in evolution. For example, the whales have evolved into such completely different animals that it was next to impossible to infer their ancestry using their morphology. With the help of molecular systematics, however, we now know that artiodactyls like the hippo, the pig, and cattle are among the closest living relatives of the whales. Recently discovered SINE characters (another type of molecular characters) confirm this relationship. Most excitingly, paleontologists then discovered that some fossil whales with legs have the signature leg bone structures that are found only in artiodactyls. Thus a mystery that has been around ever since Darwin claimed that whales evolved from land mammals has been solved. Molecular systematics is science, because predictions can be made and tested. In the case of whales, the prediction by molecular systematists that whales are descedants of an artiodactyl has been verified by fossil evidence.

As long as molecular systematics can produce results that are scientifically verifiable, it will live on as a science.

Regards