Thursday, April 20, 2006

The serpent beguiles?

More blasted intermediates to deny today. This hasn't been a good month to be a creationist. Najash rionegrina from the Late Cretaceous of Patagonia is no ordinary snake: it has legs.

This is not the first fossil snake with legs to be found. Previously, there had been Pachyrhachis, Haasiophis and Eupodophis. What's especially interesting about Najash is the presence of a sacrum: usually a modified rib that forms a strut between the vertebral column and the pelvis and helps support the body on land. This is a feature that is found even in the earliest tetrapods and tends only to be lost when limbs are lost or when tetrapods become more fully acquatic.

The authors analyzed the characters of various snakes (fossil and living) and generated the following tree:

The interesting thing to note is that Najash is the most primitive snake. However, the other three legged snakes find themselves deep inside snakes. In fact, they're what are known as the Macrostomata, the group that includes the things you most commonly refer to as modern snakes: boas, pythons, as well as the more derived vipers and the elapids (cobras and their kin).

Now, the last thing I want to get myself into is the sordid debate on snake origins. This new hypothesis will not go without contention since it requires several independent losses of limbs or even the independent re-development of limbs in order to account for the distribution. The debate on snake origins is a hot topic and the key players show no mercy with each other! So, I'll try to ride the fence here.

This phylogeny lies at the heart of the debate on snake origins. One issue here centers on whether or not the mosasauroids (large, extinct marine lizards) are the sister group of snakes and whether or not snakes have a marine origin. The mosasauroids (which includes the mosasaurs proper, as well as the smaller aigialosaurs) are not likely to include the ancestor of snakes, but they share a lot of features in common, such as body elongation and the way the teeth fit into sockets in the jaws -- a condition called "thecodonty" (tooth in hole). In other varanoid lizards (like the Komodo Dragon), the base of the tooth is sort of "squished" to the inside of the jaw, a condition called "pleurodonty" (see a diagram of these tooth types here).

Interestingly, however, a lower jaw referred to Najash apprears to present a pleurodont type condition, or something very similar. This means that the pleurodont condition might actually be ancestral for snakes, and therefore the condition seen in mosasauroids might have nothing to do with the condition seen in snakes.

A matter of outgroup:

The dichotomy that arises between a mosasauroid vs. 'other' sister-group hypothesis is also one between a terrestrial origin for snakes and a marine or aquatic origin. Najash is apparently a terrestrial snake and, in its position as the most 'primitive' snake of all would suggest that terrestriality is ancestral for snakes. But all of this is dependent on how we interpret their tree. This, of course, is further dependent on how the authors treat the data.

One of the potential problems with this new work is that it doesn't actually test the question of who is most closely related to the snakes. If you look at the tree they've obtained, one can see the most deeply stemming branch is labeled "Varanoid root". This is the outgroup, the taxon in the analysis against which allows the analysis to infer the "direction" of evolution inside the tree. The authors constructed a hypothetical varanoid from observations of different varanoids, including mosasauroids. However, since it's only a single terminal taxon, it is impossible for snakes as a total group to "move around" in the analysis. They're stuck there next to a hypothetical taxon. So, the question of mosasauroid relationships is unaddressed. Furthermore, the different outgroup will have different characters, resulting in

The other (and somewhat related) argument that may fall under contention is the inference of the ancestral condition for snakes as a whole based on the environment of Najash. This may be somewhat problematic, especially where they have the other snakes with legs higher up in the tree. If, instead, the other snakes with legs are more primitive (as one would initially think), and one were to test whether or not there is a marine outgroup, then terrestriality might be a specilaization of Najash rather than an ancestral feature. The placement of the legged snakes within the macrostomatans has been contested int he past and will continue to be, so watch for that!

The take-home message here is actually me 'busting down' the sacrosanctitiy of the concept of 'transitional forms'. Najash is almost certainly intermediate and, at this time, I hardly doubt that it is the most primitive snake. As such, it tells us about the early origins of snakes. But it is not an ancestor, and even as an approximation of an ancestor, it is still specialized in its own way. The same could be said for Tiktaalik which has its own specializations that are not necessarily intermediate between other fishes and tetrapods.

The other thing to bear in mind is that perhaps our problem with the re-elaboration of limbs is fraught with 'common sense' thinking -- often a danger in science. It is difficult to imagine how snakes could simple re-evolve their legs, but it would be unwise to use such an argument. The development of an embryo is a hierarchical process and all or most of the genes used in 'leg-making' in snakes probably never disappear (though perhaps somebody more up to speed on this could comment here). A few or a single 'up-stream' switch in the hierarchy of development could perhaps turn the whole leg on or off. Our personal preference or bias that snakes should not 'grow back' their legs must not be an argument. More robust phylogenetic arguments as well as developmental biology should be looked to for clues.

The goal of the study of phylogeny is to sort out between three different kinds of characters: those that are ancestral and therefore ancient and general (plesiomorphy), those that are specialized and unique to a taxon (autapomorphy), and those that are the characteristics unique to a group that shares a unique common ancestor (synapomorphy). When we have sorted out the best and most heavily tested solution to this trichotomy, we can then infer what is the ancestral state. The data on the fish-tetrapod transition are so unequivocal that it is pretty easy to sort this out for Tiktaalik. The same can't yet be said for the origin of snakes, and Najash is by no means the last word. We'll be hearing a lot more about this soon.

Other blogs:

ApesteguĂ­a, S. and Zaher, H. 2006. A Cretaceous terrestrial snake with robust hindlimbs and a sacrum. Nature 440: 1037-1040. <link>

No comments: