The Origin of Snake Feeding

More evidence in support of a mosasaur origin of snakes.

Lee, M.S.Y., G.L. Bell and M.W. Caldwell 1999. The Origin of Snake Feeding. Nature 400:655-659

Synopsis:

Snakes are renowned for their ability to engulf extremely large prey, and their highly flexible skulls and extremely wide gape are among the most striking adaptations found in vertebrates_l-5_. However, the evolutionary transition from the relatively inflexible lizard skull to the highly mobile snake skull remains poorly understood, as they appear to be fundamentally different and no obvious intermediate stages have been identified_4,5_. Here we present evidence that mosasaurs——large, extinct marine lizards related to snakes——represent a crucial intermediate stage. Mosasaurs, uniquely among lizards, possessed long, snake-like palatal teeth for holding prey. Also, although they retained the rigid upper jaws typical of lizards, they possessed highly flexible lower jaws that were not only morphologically similar to those of snakes, but also functionally similar. The highly flexible lower jaw is thus inferred to have evolved before the highly flexible upper jaw——in the macrophagous common ancestor of mosasaurs and snakes——for accommodating large prey. The mobile upper jaw evolved later——in snakes ——for dragging prey into the oesophagus. Snakes also have more rigid braincases than lizards, and the partially fused meso- and metakinetic joints of mosasaurs are transitional between the loose joints of lizards and the rigid joints of snakes. Thus, intermediate morphologies in snake skull evolution should perhaps be sought not in small burrowing lizards, as commonly assumed, but in large marine forms.

As far as I'm aware, this is the only evidence, and not especially persuasive at that. All this research does is push the question 'how did flexible skulls evolve from lizards' a little further back - mosasaurs were simply varanid lizards, so at some stage in their development lizards must have evolved flexible joints. Mosasaurs are one transitional stage, and an interesting avenue to explore, but as you are so fond of pointing out characteristics have been known to evolve convergently. Any lizard evolving from a rigid-jointed form into a flexible-jointed one would exhibit the same intermediate characteristics as the mosasaurs, and the fact that the earliest known snakes were terrestrial forms with strong hindlimbs casts doubt on the mosasaur hypothesis, presuming as it does that a marine lifestyle provided the impetus for snakes to lose their limbs. They may be derived from varanid lizards, and may even have diverged from the lineage which gave rise to the mosasaurs after the loose jaws had developed, but I'm sceptical about a marine origin for snakes.

Philip Bowles

This is not the only evidence.

As this article points out, the only lizards that have evolved a flexible jaw joint are the varanoid lizards.

The earliest known snakes, Pachyrhachis and Dinylysia, are marine. You are quite simply wrong.

Since you are wrong, the earliest known snakes support the mosasaur hypothesis, contra your unsupported assertion.

You are entitled to your skepticism, but a wide variety of evidence now supports the marine origin of snakes and at the same time contradicts the burrowing lizard origin of snakes. The blindsnakes may in fact be the most primitive snakes extant, but Pachyrhachis and Dinilysis are even more primitive. Pachyrhachis has hindlimbs and a large fish within its gut, showing that it is capable of engulfing large prey. It shows that the snakes's characteristic large gape had already evolved prior to the loss of the hindlimbs and it shows that the blindsnakes's relatively small gape evolved secondarily, as a byproduct for a burrowing existence, as Carroll (1988, Vertebrate Paleontology and Evolution) points out.

The idea that snakes are derived from varanid stock has support in the features of living monitors, and particularly the snake-like head of the earless monitor, and is not especially controversial. This in itself does not suggest that, if indeed varanids are ancestral to snakes, the particular varanids from which snakes evolved were mosasaurid.

Pachyrrhachis is no more primitive than Haasiophis, and similar to modern forms (Rieppel, Science 2000). The fact that these two species are marine no more implies marine ancestry than the existence of living sea snakes.

You are right that I misremembered the reason for Rieppel's dispute, which is in fact that marine snakes show no particular affinity with mosasaurs rather than that the early forms are terrestrial, but that does not invalidate the dispute itself. Pachyrrhachis is not an especially primitive snake; akin to boids as it is, it is apparently rather less primitive than the blindsnakes.

Philip Bowles

It will be better if you read the article I cited by Lee et al., or at least the synopsis I quoted in an earlier post, because you seem to be totally oblivious to the evidence cited therein for a mosasaur-snake link.

The following passage is from the synopsis that you deleted and apparently never read:

'Here we present evidence that mosasaurs——large, extinct marine lizards related to snakes——represent a crucial intermediate stage. Mosasaurs, uniquely among lizards, possessed long, snake-like palatal teeth for holding prey. Also, although they retained the rigid upper jaws typical of lizards, they possessed highly flexible lower jaws that were not only morphologically similar to those of snakes, but also functionally similar. The highly flexible lower jaw is thus inferred to have evolved before the highly flexible upper jaw——in the macrophagous common ancestor of mosasaurs and snakes——for accommodating large prey. The mobile upper jaw evolved later——in snakes ——for dragging prey into the oesophagus.'

Yes they are similar to modern forms in their jaw morphology, but they have hindlimbs. They are indeed more primitive than the most primitive living snakes.

The fact that the oldest known snakes are marine and the fact that they have lost their front limbs, have an elongate trunk and still retain their hindlimbs means that they have not quite evolved the complete suite of characters of modern snakes, none of which have retained as much of the hindlimbs as Pachyrhacis or Haasiophis. These snakes demonstrate that the origin of snakes occurred in a marine environment, as much as Longisquama demonstrates that the evolution of birds occurred in an arboreal setting.

'Rieppel (1980, 1983) questions varanoid affinities and suggests that the pattern of kinesis of both the skull roof and the lower jaws differ from that of modern lizard groups and probably evolved separately from the eolacertilian level.' (Carroll 1988:236)

Rieppel has a long history of disbelieving the varanoid affinity of snakes. So his latest denial of the varanoid affinities of Pachyrhachis and Dinilysis is no surprise. But Rieppel's objections have been rebutted by Lee et al., who demonstrates unique similarities between the lower jaws of snakes and mosasaurs in the cited article, similarities that are not found in any other group of lizards.

Rieppel's objections to a mosasaur-snake link, based largely on the morphology of the lower jaw, has been refuted by Lee et al., who demonstrates unique similarities between the lower jaws of the snakes and mosasaurs.

The largely disproven hypothesis that snakes evolved from a burrowing, visually challenged ancestor demonstrates the danger of inferring evolutionary history entirely from an assessment of the living species. Even though the blindsnakes are fossorial, and even though they are the most primitive of living snakes, that does not necessarily mean that they are morphologically most similar to the snake ancestor. The small gape of the blindsnakes is probably a function of their secondary fossorial adaptations, according to Carroll (1988)'s analysis, and therefore not an indication of their primitiveness. Their skulls may in fact be highly derived because of fossorial adaptations and do not necessarily represent the primitive condition. Besides, the characteristic large gape of snakes is already present to a limited extent in mosasaurs, which had narrow snouts because of the constraints of hydrodynamics, but which preyed upon large animals such as sea turtles (Lee et al. 1999).

When one does take into account the fossil record, a different picture emerges. The fossil record shows that all of the earliest known fossil snakes (e.g. Dinilysia and Pachyophis), not just Pachyrhachis and Hassiophis, are marine species. Pachyrhachis and Hassiophis also retain hindlimbs. They are definitely more primitive than the blindsnakes because of hindlimb retention. The contention that Pachyrhachis re-evolved hindlimbs from a limbless ancestor is simply implausible. Pachyrhachis shows that the characteristic large gape of snakes had already evolved before the snake lineage had lost the hindlimbs. All of the available evidence (stratigraphic, comparative anatomical, etc.) therefore strongly suggest a mosasaur origin of snakes.

In fact, if there is a 'mistaken extinction', it is not the dinosaurs, which have indeed become extinct without leaving any living descendants. The mosasaurs, OTOH, did not really become extinct. They have evolved into snakes and their descendants are thriving.

I read and responded to it; indeed the very comment of mine quoted below that contains this response. All this suggests is that the mosasaurs and snakes had similar dietary habits - as indeed this quote claims by describing the jaw arrangement of the two groups as functionally similar. If this is so, then in the absence of further compelling evidence convergence is as likely as ancestry. As the remaining evidence does not seem compelling, I would not be inclined to favour one hypothesis over the other.

Rieppel suggests that this is a secondary adaption in the boid family which developed for grasping mates as spurs do today. Whether this is true or whether they simply hadn't lost their original hindlimbs by this stage, their modern affinities are clearly with the boids and this argues in favour of Rieppel's hypothesis that these animals are too far advanced in terms of snake development to reveal very much about the group's origins.

Given the uncertain affiliations of Longisquama this is perhaps not the definitive analogy you were hoping for. The major argument for an arboreal origin of birds lies in the fact that both Longisquama and Microraptor adopted this lifestyle, so whatever birds were derived from it probably lived in a tree. Applying the same logic to the snakes is probably not possible as these animals, while not completely similar to contemporary forms, are too different from any known ancestral candidate for this sort of judgement to be made.

Nor is Feduccia's continuing denial of evidence presented in favour of a dinosaurian origin of birds, yet this is not sufficient argument against his position. In showing that the comparisons recent research has drawn are suspect, Rieppel's strategy is essentially identical to Feduccia's tendency to cast doubt on the reliability of what he now calls 'dino-fuzz' as evidence for bird ancestry.

But Rieppel's objections have been rebutted by

An abstract I read from the ABC news archives contained the suggestion that lizards may be the wrong place to look for a snake ancestor altogether - where do amphisbaenids feature in the current argument? Or extinct forms of sphenodont?

For someone who favours convergence in other areas you seem remarkably unwilling to entertain the possibility here - all the similarities here show are that snakes and mosasaurs developed the same methods of consuming and digesting food, which is interesting but not necessarily important in determining ancestry. It's a specialised adaption, and the more specialised it is the more similar convergent animals will have to become in order to produce the same behaviour.

No, but the fact that they appear to be more primitive than the living animals which most closely resemble Pachyrrhachis and its kin is an observation which deserves further investigation, as it suggests that the boids had already diverged from blindsnake-like animals 97 million years ago.

The small gape of

Which would necessarily have to suppose that the fossorial adaptions are secondary. If one does not make the initial presumption that snakes are not derived from burrowing animals, then one would not reach this conclusion.

>When one does take into account the fossil record, a different picture

Yet they are uniformly no more than about 97 million years old, and if Rieppel's recent analysis is correct then none represent forms ancestral to the lineage. These are, as you say, marine animals, and they are large-bodied animals, making them more likely to be preserved than a small fossorial desert-dweller. The fact that so many large marine snakes have been preserved yet no land snakes of the same period is very likely an artefact of the fossil record.

They are definitely more primitive than the blindsnakes because

Assuming first that Rieppel's analysis of a secondary adaption is incorrect, and second that the earliest blindsnakes lacked hindlimbs. All hindlimb 'retention' shows us is that these animals were more primitive than modern boids.

The contention that Pachyrhachis re-evolved

Care to explain in more detail?

Philip Bowles

-cq.aol.com>:

That is like arguing that because the wings of pigeons have a similar function as the wings of ravens, therefore their wings are likely to be convergent. In order to show that two structures are convergent, you cannot merely say that they have a similar function. You have to show evidence that would support convergence; you have not done so. Yet you think that because there is a possibility that two structures are convergent, and therefore they must be convergent. That is a gross error in logic.

You should definitely read the article I cited. They proffer evidence to support homology between the jaws of mosasaurs and snakes, unlike you, since you proffer zero evidence in support of convergence.

Since you are unaware of the evidence, you are in no position to judge whether they are compelling or not.

This suggestion has been refuted. Cohn and Tickle (1999) have shown that pythons do have true hindlimbs. These are vestigial hindlimbs, not secondary structures.

It is simply not true.

Pachyrhachis and Haasiophis et al. are not by any stretch of the imagination or logic 'too far advanced' to reveal very much about snake origins. They are the oldest snake fossils known, and they retain the ancestral featurs of hindlimbs. Since the pythons retain less of the hindlimbs than these early snakes do, it shows that pythons are much more derived than Pachyrhachis, Haaiophis et al.

How exactly Longisquama is related to other reptiles may be unknown, but the fact that it is arboreal and the fact that it is the closest known relative of birds is quite well established.

You are arguing that the earliest snakes do not resemble mosasaurs very much I presume. That is easily refuted. Their lower jaws are quite similar. As the article I cited point out, they both share unique features such as snake-like long palatal teeth. Is this feature also convergent between mosasaurs and snakes too? If so, what is your evidence for convergence, if any?

You have a bias against Feduccia I can tell. Feduccia REFUTED the evidence in support of the dinosaurian origin of birds. You, OTOH, has DENIED the existence of evidence in support of a mosasaur origin of snakes. There is a big difference between refutation and denial.

Rieppel is not by any stretch of the imagination or logic similar to Feduccia. Rieppel has been criticized by Dodson (2000 Amer. Zool.) for ignoring stratigraphic evidence. Rieppel bases his theory primarily if not exclusively on the basis of morphology. Feduccia definitely takes into account stratigraphy in challenging the dinosaurian origin of birds. In fact Feduccia (1996) maintains that the temporal paradox is one of the strongest pieces of evidence against the dromaeosaurid origin of birds. Besides morphology, Feduccia (1996) also considers biogeography, biophysics, developmental biology, physiology, ecology, and of course stratigrahy, in formulating his theory, again unlike Rieppel, who relies largely if not exclusively on morphology.

You are relying on the wrong sources for scientific information. Try reading scientific journals instead of the popular press. Amphisbaenians are burrowing reptiles of uncertain affinity. They are not closely related to the squamates. Snakes, OTOH, are squamates, and one of the uniquely shared characters between snakes and lizards is the paired hemipenes. I guess you are going to suggest that this is also a convergence between snakes and lizards.

I favor convergence if there is evidence for it. I don't favor it blindly, unlike you. I don't favor it because it fits my preconceptions, unlike you and many cladistic paleontologists.

I will 'entertain' it if you can show me some evidence for convergence between mosasaurs and snakes. I will even 'entertain' the possibility that snakes and lizards evolved paired hemipenes convergently if you can show me supporting evidence.

If we had only fossils to go on, and a raven was as different anatomically from a pigeon as a snake from a mosasaur, it would be certainly be an avenue worth pursuing even though it would eventually be found to be incorrect.

In order to show that two structures are convergent, you

Nor have I attempted to show that any convergence exists - I raise the possibility as just that, a possibility. My position is simply that there is not enough evidence about snake ancestry to take a dogmatic position either way; nowhere have I stated that mosasaurs are definitely not snake ancestors, or even that they are unlikely to be, simply that there are reasons for supposing we lack the full story, and that for that reason I am sceptical about claims that Lee et. al.'s work is conclusive.

Yet you

I have made no comment about what 'must' or 'must not' be the case; I noticed in earlier discussions about the dinosaur/bird debate you chose to caricature my position as one of cladistic dogma, but if you take that approach you are merely attacking a straw man.

I was referring to the evidence you've presented to date; I may well look into the issue further but I make no statements about other work at present.

Until I look into the issue in more detail, I'll take your word for it, but as Rieppel has repeated his hypothesis in his 2000 work I would presume that he has reasons for disputing Cohn and Tickle.

More derived, but not necessarily 'much more' - the modern pythons seem to have changed little except to reduce these hindlimbs to spurs, and the loss of a single unnecessary adaptation is not tremendously significant.

The lower jaw, out of an entire animal. One thing which seems not to have been addressed as far as I've seen is the change in the overall body shape of the animals. For all their adaptations, mosasaurs were simply sea-going monitor lizards, not snake-like. The only examples of a lizard body form developing into a snake-like form are burrowing lizards. No habitually marine or even aquatic reptile has developed a similar body form as a result of its habitat, and it must be asked why some mosasaurids would have been an exception when the mosasaur body itself proved adequate for their marine existence throughout their evolution. Isolating characteristics to discuss is only relevant in the context of the entire animal - why lose the limbs when turtles and early marine crocodiles developed flippers to assist in their underwater locomotion? Why go to the trouble of elongating the body shape, losing the extra lung and rearranging the internal organs when crocodiles, lizards and turtles adapt perfectly well to aquatic and even marine life with squatter forms? There is no obvious evolutionary or ecological reason why a mosasaur would have turned into a snake, yet we see a number of snake-like adaptions in burrowing lizards - either their body form is convergent with the snakes, which implies similar requirements imposed by a common habitat, or the mosasaur jaws are the convergent feature, and suggest a similar method of consuming food. Granted I don't know of any particular evidence to support convergence between mosasaurs and snakes, but nor do I know of any particular evidence to support convergence between legless lizards and snakes. What I do know is that the legless lizards, amphisbaenids and caecilians which are convergent with one another all share a burrowing, and usually terrestrial, lifestyle, and that animals which do not share this lifestyle have not been known to develop serpentine bodies.

As the article I cited point out, they both share unique features

It may be; in order to make a definitive statement on the matter I would need more information regarding mosasaur feeding behaviour. Convergence is a consequence of functional adaptations to meet similar ecological requirements and can't be inferred from purely skeletal similarities.

Not at all - being willing to concede that he is wrong is hardly a statement of bias. I did not say he refuted the evidence because he has not done so - a refutation would imply that he has categorically shown the opposing interpretation of the evidence to be incorrect, which is not the case on an unbiased appraisal of the situation, whether or not one personally believes that his assessment is correct. He has presented evidence for his position and carried out valuable examinations of specimens in support of that position, but this has not yet been sufficient to prove that interpretations of apparently avian characteristics in dinosaurs are incorrect and may never be.

Feduccia REFUTED the evidence

I have denied nothing; I have made mention of Rieppel's position. I do not have detailed information about how rigorous or otherwise Rieppel has been in providing evidence to dispute the arguments of his critics, but that was not the point of my comparison.

-ft.aol.com>:

I am sure that no competent paleontologist would consider a pigeon to be as different from a raven as a mosasaur is from a snake.

Correct, you never made any attempt to show that.

No, don't take my word for it. My words are not evidence. Look instead at the evidence presented in Cohn and Tickle's paper. That is evidence. Don't be afraid to look at the evidence. It won't bite.

They are much more derived than these primitive snakes, unless of course you believe that blindsnakes are more advanced snakes than pythons.

Wrong. It has been addressed. Please read Cohn and Tickle. They addressed that issue.

Again, if you are capable of understanding Cohn and Tickle, they provide a developmental mechanism for the evolution of not only limblessness, but the elongation of the trunk, and how such an evolutionary change can easily take place. In fact, as Lee et al. point out, more than 50 lineages of reptiles and amphibians have undergone a similar change independently. It is by no means rare for vertebrates to lose limbs and evolve elongate bodies.

That is your misunderstanding. At this point, you should familiarize yourself with the literature on the prevalence of limb reduction and body elongation among reptiles instead of assuming that what little you know is fact.

Again, this statement stems from your ignorance, not from scientific fact. If you wish to disagree, please cite some evidence to support the two statements you just made.

Why don't you ask the Amphiuma why it lost its limbs? It may have a good answer for you.

That is because you know very little about herpetology.

That is bias. You are claiming that he is wrong without any evidence. Very strong bias. In fact, your assertion that he is wrong is about as well supported as your assertion that mosasaurs are 'possibly' convergent upon the snakes.

No, you did not say that. You said he denied the evidence. That is of course wrong and strong evidence of bias. He did not deny the evidence, unlike you.

You deny there is evidence in support of the mosasaur-snake link, calling their similarities 'possibly' convergent but without proffering a single shred of evidence in support of convergence. IOW, you have ignored the evidence without giving any reason as to why the evidence is unreliable.

'Uncertain' means no one really knows which group it is closest to.

I guess it is pointless arguing with you further since have not presented any evidence to refute the mosasaur origin of snakes.

You are showing your bias. You characterize the evidence as an 'article of faith' even though you admit that you are incapable of judging the

I didn't say they wouldn't - I said that they would have to for your analogy to be relevant to this situation.

On the contrary; we have a living primitive snake and a fossil of an apparently comparatively advanced snake, which implies that we lack fossils of the early snakes, just as the discovery of Archaeopteryx, a fully-fledged bird, suggested that earlier avian forms existed.

However, in order to understand the reasons for this one needs to look for the similarities in behaviour and ecological conditions which have prompted this development. I've looked for information on the amphibian you mentioned, and I am willing to concede that amphibians with vestigial limbs do not fit the reptilian pattern of burrowing animals; as I know less about amphibian development in general than reptiles, I won't speculate as to the reasons none of these has developed flippers. However, since we are talking about reptiles, and since there are no marine amphibians, it is worth noting that no example of a reptile developing into a limbless form has been derived from a habitually aquatic animal which needs to propel itself against prevailing currents (would I be right in stating that Amphiuma and similar forms lead fairly sedentary lives in slow-flowing or stagnant bodies of water?), and indeed when snakes took to (or possibly returned) to the sea the basic form was unsuitable until they developed paddle tails unknown on the oldest fossil snakes. Nor, as I mentioned, did the mosasaur form seem to change substantially once it evolved - it didn't need to diverge into a limbless form to pursue a marine lifestyle.

You've provided no evidence yet to contradict it - the one animal you cited is a salamander with vestigial hindlimbs, not a lizard of any kind, and while I admit I was incorrect to apply the same blanket statement to amphibians this is not what you are responding to here.

>>No habitually marine or even aquatic reptile has developed a

How can I provide evidence of an absence? I could cite all the pygopodids, sandskinks, amphisbaenids, glass lizards, slow-worms etc. etc. by name but this would not in itself be evidence that aquatic legless lizards might not exist as I may have missed some, or there may be species yet to be discovered of this type. Give me a counter-example among reptiles and I shall see if I can provide evidence for a fossorial origin.

It might - it might tell me that salamanders naturally have comparatively weak limbs which could not be turned into muscular flippers, or that it has no particular need to move very much, or that a salamander has a more elongate body than most lizards to begin with, or any of a variety of other things, but it is not an aquatic reptile let alone a marine one.

Care to enlighten me rather than making baseless personal attacks? The salamander you pointed to belongs to none of the groups I mentioned, and seems of very little relevance to the mosasaur situation. Ignorance of a trio of North American salamander species hardly equates with knowing very little about herpetology.

I am not claiming any such thing - I said I am willing to concede that he is wrong, a hypothetical condition rather than a definitive judgement.

And every bit as hypothetical - I didn't get involved in this thread to present an argument, I became involved because I took issue with a dogmatic position based on a scarcity of hard evidence - in fact, the very same sort of argument you have criticised others for when they point out that X bird looks like Y dinosaur and so there must be a direct line of descent.

See the rest of the comment you responded to here - in the eyes of the paleontological community at large, the evidence stands despite Feduccia's attempts to refute it. Therefore the evidence has not been refuted.

Not at all. I have stated that the evidence is not compelling in itself; this is a far cry from saying there is no evidence to support it. The only thing I have denied is that there is enough evidence to make the sort of categorical claims in favour of one position or the other that you have been making. I wouldn't be happy claiming that two developed animals in different groups are obviously related on the basis of no more than theoretical argument, whether the link is between mosasaur/snake, bird/dinosaur, bird/Longisquamid or whatever. Of course paleontology is often an inexact science precisely because the fossil record is not as obliging as would be desired in terms of tracing evolutionary connections, which is why arguments rage for so long, but for that very reason it should have no room for dogmatism.

I am well aware of that, but you asserted rather more than that. The claim that 'They are not closely related to the squamates' is, in regard to current taxonomic practice both systematic and cladistic, incorrect - they have not been placed in their own order as have the sphenodonts, as would have been the case if they were not deemed to be close relatives.

It probably is; I didn't come to this

Any on-line resources about mosasaurs? In addition, as the the mosasaurs primarily aquatic, with well defined dorsal spines... bad analogy, more 'eel-like' than 'worm-like'? Are there land-based fossil records?

Jim M.