Reptile & Amphibian Forums

troy h wrote:
as Dr. Fry pointed out, the loss of venom (or toxins if you prefer) only occurs once among Colubrids - in the lampropeltines . . . and a reversal in one segment of this lineage (Coelognathus) is not a giant leap. It is far more parsimonious (therefore vastly more probable) that venoms evolved once (as is supported by shared chemical composition of many venoms) and that one group lost the venoms (once) and for it to reappear (once) in a subset of that group. Your hypothesis, that venoms evolved many times, is not supported by chemical analysis of venoms (as Dr. Fry pointed out) or by the principle of parsimony.

Further, the evidence that Colubrids lost constriction and that it has re-evolved in many colubrid groups, while not the most parsimonious explanation, is well supported by evidence from the widely varying patterns of constriction found in Colubrids - all primitive constrictors constrict in precisely the same pattern while advanced constrictors use a wide variety of constriction patterns.

The evidence is strong, follow the evidence. Or stick to dogma - your option.
=================================================================

Troy h later gave the following reference as evidence that colubrid snkes lost constriction and constricting behavior then re-evolved in many colurbid groups:
Greene, H.W. & Burghardt, G.M. (1978) Behavior and phylogeny: constriction in ancient and modern snakes. Science 200:74-77

Let’s examine the evidence. Greene and Burghardt state that “There are four possible origins of similar phenotypes in different taxa.” These 4 possibilities are (i) chance, (ii) similar experiences in the lives of individual animals, (iii) convergence, (iv) phylogenetic continuity. They believe that the first 3 can be falsified for the evolution of constricting behavior. That means they believe that if two snakes constrict in a similar fashion, then their phenotypic similarities are due to common ancestry. In footnote number 14, on page 76, they state: “Among colubrids, some coils of Arizona and Lampropeltis resembled those of boids.... A published drawing of Regina alleni (Colubridae) suggests that this species might also at least sometimes use the boid pattern.”

That means, according to Greene and Burghardt, the similarities between Regina, Arizona, Lampropeltis and the boids are due to “phylogenetic continuity” since they had “falsified” the other 3 possibilities for these similarities. If there is phylogenetic continuity in constriction behavior between the boids and colubrids, then needless to say there has not been any loss or discontinuity in this behavior in Regina, Arizona, Lampropeltis, in the common ancestor of these snakes, or for that matter in the common ancestor of the Colubroidea. Lampropeltis, Arizona and Regina therefore have retained the constricting behavior that has apparently evolved in the common ancestor of all snakes, which lived before the end of the Cretaceous.

I agree with troy h that the evidence is strong, not for the loss and re-acquisition of constricting behavior in colubrid snakes but for “phylogenetic continuity” in constricting behavior from primitive to adavanced snakes. If there is loss of constriction among colubrids, it almost certainly occurred independently multiple times, but not in the common ancestor of the colubrids.

If the common ancestor of the coluroid snakes never lost constriction, then, like Lampropelits and Arizona, this ancestor is probably non-venomous. Powerful constrictors such as boids and Lampropeltis simply do not need venom to be successful. Judging from the holarctic distribution of the ratsnakes and the distribution of the genus Lampropelis (from Canada to northern South America) and the distribution of boids, these snakes are indeed very successful.

If the colubrid ancestor did not lose the ability to constrict, then there is no adaptive reason for it to evolve venom and then lose it, as it has been claimed to have happened in the lampropeltine snakes. It is far more parsimonious for Lampropeltis to have never lost constriction and never evolved venom then for it to have evolved from an ancestor that had lost the ability to constrict and thence evolved venom, and then re-evolve constriction and finally lose the venom. Evolution is not parsimonious, but it is not ad hoc either.

Mayr and Ashlock write:
****************************************************************
Evaluation of Homoplasy

When there is discordance among characters because of homoplasy, characters should be weighted and ranked according to their quality. The following criteria are helpful in determining the character to which the highest weight should be assigned (Bock 1977: 890).

1 It is a complex feature rather than a simple one because the probability of evolving independently two or more times is considerably lower for a complex than it is for a simple feature.

2 It is a new feature rather than the reduction or loss of a feature because the probability of a new feature evolving independently two or more times is low while reduction or loss of a feature may occur several times independently.
***************************************************************

It is more likely for constriction to have been lost independently multiple times in many colubrid lineages than for some (such as Lampropeltis) to have lost it, evolve venom, reevolve constriction and then lost venom without a trace.
Venom evolved once?

Retry. Link not working in last post
Venom evolved once?

Now that you have a preview of my response, you should know how ad hoc it is to argue that the common ancestor of the Colubroidea evolved venom.

As I said, powerful constrictors do not need venom. Boids are powerful constrictors, and they never evolved venom. Their near cosmopolitan distribution is testimony to their success. In troy h's scenario, the common ancestor of the Colubroidea evolved venom, then it lost the ability to constrict, then in the lineage leading to Lampropeltis and Elaphe radiata, constriction reappeared, causing the common ancestor of Elaphe radiata and Lampropeltis to abandon venom without a trace. Subsequently, E. radiata re-evolved venom from a non-venomous ancestor.

Such an implausible scenario with multiple reversals of the same character is needed if one were to account for the non-venomous nature of the ratsnakes and their descendants such as Lampropeltis. A more reasonable explanation is that the ancestor of the Colubroidea is a powerful constrictor that is not venomous. It evolved the Duvernoy's gland as a gigantic salivary gland. The postorbital position of this gland is perfect for lubrication of large prey for easy passage to the gut but not so perfect for venom delivery. If the Duvernoy's gland evolved originally as a venom gland, there should be no mucous cells in them, only serous cells. And it should have been located anteriorly so that toxins can be delivered to the prey at the earliest opportunity. Thus both the posterior position of this gland, and the anterior positioning of the fangs in viperids and elapids argue against the Duvernoy's gland having evolved originally as a venom gland. As for the biochemical similarities between the venoms of various colubroid snakes, biochemical homoplasy can explain their similarities. If the evolution of the eye in animals is any guide, pre-existing chemicals can be recruited independently in different snake lineages to serve as venom. Since the chemicals (whether they are enzymes, proteins, or products of enzyme assisted biochemical reactions) that are recruited are similar in related snake lineages, the chemical structures of the end products are also similar. If there is a way to make poison out of a particular chemical that already exists within a snake, then nature will likely make use of the same recipe again and again in different but related lineages of snakes through convergent evolution. Multiple independent evolution of venom in colubroid snakes is therefore not only possible, it is probable given the ad hoc hypotheses one would have to invent for Elaphe ["Coelognathus"] radiata to re-evolve venom from a secondarily non-venomous ancestor.

The fatal flaw with your scenario is that it assumes that Coelognathus post-dates the evolution of the American Lampropeltis/Patherophis/Pituophis lineage when in fact all the American colubrines are descended from Asian invaders. Thus, mapping over the taxonomical tree shows that venom evolved a single time, was lost in the American lineage and then thats it. Not too terribly difficult to follow unless you are deliberately (for reasons that only you know) putting on the blinkers.

As for homoplasy, this is disproven by the phylogenetic analysis unless you want to argue against phylogenics as a whole. In which case I would like to see how you would explain PLA2 toxins obviously having been involved twice in snake venoms, once in the Viperidae (through recruitment of a synovial (type II) phopholipase gene for use in the venom and another time by the common ancestor of elapids 'colubrids' through the recruitment of a pancreatic (type I) phospholipase gene into the chemical arsenal. When analysed by bayesian, maximum parsimony and neighbor-joining, the two toxin types for two separate monophyletic groups, with an abundance of non-venom proteins in between. In contrast, the 3FTx, CRISP, cystatin, nerve growth factors, M12B peptidases, ANP/BNP-toxins form monophyletic groups, obviously indicative of a shared history and consequently recruitment of into the venom right at the base of the Colubroidea tree. This shows not only that venom evolved a single time but also which toxins were part of this first recruitment. Further evidence against homoplasy is the CNP-toxins represent a second recruitment of natriuretic pepetides into the venom and form a monophyletic group distinct from the ANP/BNP toxins. This demonstrates quite nicely that natriuretic peptides were recruited twice into the venom. Similarly, the lectin toxins form two monophyletic groups, once as an ancestral type and a second when recruited again into the viper venoms.

According to your evidence-free theory, if there were multiple recruitements then all the toxin should form monophyletic groups due to homoplasy. We have demonstrated that this simply is not the case.

I fail to understand why you are having such a hard time grasping this but really don't care actually. However, it does provide a perfect backdrop for our results since you are a lovely example of the wrong interpretation.

Thanks mate, couldn't have done it without ya

Ciao
B
-----
Dr. Bryan Grieg Fry
Deputy Director
Australian Venom Research Unit
University of Melbourne

www.venomdoc.com

BGF wroet:
“The fatal flaw with your scenario is that it assumes that Coelognathus post-dates the evolution of the American Lampropeltis/Patherophis/Pituophis lineage when in fact all the American colubrines are descended from Asian invaders. Thus, mapping over the taxonomical tree shows that venom evolved a single time, was lost in the American lineage and then thats it. Not too terribly difficult to follow unless you are deliberately (for reasons that only you know) putting on the blinkers.“

You are laboring under a misunderstanding of ratsnake evolution. Let me briefly summarize the available evidence on ratsnake evolution. The racers (e.g. Coluber, Ptyas, Gonoysoma, Masticophis) are either the paraphyletic parental group which gave rise to the ratsnakes (e.g. Old and New World Elaphe, Lampropeltis, Arizona, Bogertophis, Pituophis) or these two groups are monophyletic (holophyletic) sister groups that share a common ancestor. Unfortunately the systematic position of Elaphe [“Coelognathus”] radiata remains unclear because Utiger et al. excluded all species of “Coelognathus” from their otherwise rather comprehensive analysis. Therefore there are two scenarios:

“Coelognathus” is a racer. In that case, either the common ancestor of racers and ratsnakes is non-venomous, which disproves your theory that the colubroid ancestor is venomous, or the common ancestor of the racers and ratsnakes is venomous and the genus Elaphe has since lost its venomous ancestry completely without a trace.

The second possibility is that “Coelognathus” is a ratsnake. In that case it would mean that either venom evolved independently multiple times in Colubroidea, which of course falsifies your hypothesis, or else “Coelognathus” underwent the following evolutionary sequence:

nonvenomous (booid ancestor) -> venomous (colubroid ancestor) -> nonvenomous (ratsnake ancestor) -> venomous (Elaphe radiata).

You seem to favor the hypothesis that “Coelognathus” is a racer, but that would contradict your earlier claim that it is a ratsnake and that the ratsnakes lost their venom. If “Colognathus” is a racer, then there is not a single species of ratsnake that is venomous. It also means that there is no reason to suppose that either their ancestor or the colubroid ancestor is venomous. There is no “fatal flaw” in my argument. You are simply mistaken about ratsnake and racer relationships.

“As for homoplasy, this is disproven by the phylogenetic analysis unless you want to argue against phylogenics as a whole. “

As Feduccia points out, convergent evolution is an insidious trap. It appears that you have fallen into that trap.

“In which case I would like to see how you would explain PLA2 toxins obviously having been involved twice in snake venoms, once in the Viperidae (through recruitment of a synovial (type II) phopholipase gene for use in the venom and another time by the common ancestor of elapids 'colubrids' through the recruitment of a pancreatic (type I) phospholipase gene into the chemical arsenal. “

Did I not already point out to you that biochemical homoplasy is rampant in eye evolution? Why would you still insist that venom evolution must follow the most parsimonious route according to these chemicals? Besides, the “Coelognathus” problem I discussed above shows that it is simply not parsimonious for venom to have evolved once, since it would need to evolve twice if “Coelognathus” is a ratsnake, and if it is racer, then the common ancestor of the ratsnakes almost certainly is nonvenomous, meaning that the colubroid ancestor is also likely to be nonvenomous.

“When analysed by bayesian, maximum parsimony and neighbor-joining, the two toxin types for two separate monophyletic groups, with an abundance of non-venom proteins in between. In contrast, the 3FTx, CRISP, cystatin, nerve growth factors, M12B peptidases, ANP/BNP-toxins form monophyletic groups, obviously indicative of a shared history and consequently recruitment of into the venom right at the base of the Colubroidea tree. This shows not only that venom evolved a single time but also which toxins were part of this first recruitment. Further evidence against homoplasy is the CNP-toxins represent a second recruitment of natriuretic pepetides into the venom and form a monophyletic group distinct from the ANP/BNP toxins. This demonstrates quite nicely that natriuretic peptides were recruited twice into the venom. Similarly, the lectin toxins form two monophyletic groups, once as an ancestral type and a second when recruited again into the viper venoms. “

See the “Coelognathus” problem above to see why venom most likely did not evolve in the colubroid ancestor.

“According to your evidence-free theory, if there were multiple recruitements then all the toxin should form monophyletic groups due to homoplasy. We have demonstrated that this simply is not the case. “

Far from being evidence free, the multiple recruitment theory is supported by the evolution of other body parts, including the eye. I am not sure how multiple recruitments can result in monophyletic groups for the toxins due to homoplasy. That is not a logical extension of the multiple independent evolution of venom. Therefore you are only erecting a strawman. I point out that convergence can result in similar chemical structures among related taxa.

“I fail to understand why you are having such a hard time grasping this but really don't care actually. However, it does provide a perfect backdrop for our results since you are a lovely example of the wrong interpretation.

Thanks mate, couldn't have done it without ya “

I have no difficulty understanding the basis of your hypothesis. You are fooled by the chemical similarities of the venoms of different lineages, not realizing that similar chemical structure can result even when they are products of convergent evolution. I showed you how it happened in the evolution of the eye, but you simply ignores it and continue to have blind faith in these similarities. We all have eyes, but only some of us will use them to see evidence of convergent evolution.

So, you are subscribing to the hypothesis that "the common ancestor of the racers and ratsnakes is venomous and the genus Elaphe has since lost its venomous ancestry completely without a trace"

There are several problems with that scenario.

1. Fry et al. present no data to support any member of the genus Elaphe has any trace of a venomous ancestry. So they are guessing.

2. Fry et al. can provide no plausible explanation why the lampropeltines should lose venom.

BGF claims that a shift in diet from ectotherms to rodents is responsible for the loss of venom but I pointed out that Ptyas eats rodents, but it has not lost its venom. Neither has the viperids. Rattlesnakes feed largely on rodents, and they are more venomous than the copperhead, which feed largely on ectotherms. So a predominantly rodent diet should at least favor the retention of venom rather than losing it. You can map your character over a tree, but that does not explain seemingly maladaptive changes. Maladaptive changes are possible on a tree, but they are very difficult to achieve in nature.

"And I've pointed out that its not homoplasy but rather useful ancient tools being misunderstood by recent useless tools."

Yes you made that claim. But like your other claims, it makes no sense. And it is contradicted by scientific evidence. The Pax-6 gene is found in the development of all sorts of things, not just the eye, and the only plausible explanation is that it has been co-opted many times independently and convergently by different groups of animals to regulate eye development.

"Then we are in fundamental agreement since some toxin lineages form monophyletic groups (and thus are ancestral) while others (e.g the PLA2 toxins) form two separate monophyletic groups (and thus represent independent recruitement events)."

Unfortunately we are not in any agreement. You still believe that similarity means homology. That is simply borne of ignorance. Experienced systematists would not be as easily fooled by the insidious trap of convergent evolution as Fry et al. were.