I keep saying this to everyone: if you want to understand the origin of novel morphological features in multicellular organisms, you have to look at their development. "Everything is the way it is because of how it got that way," as D'Arcy Thompson said, so comprehending the ontogeny of form is absolutely critical to understanding what processes were sculpted by evolution. Now here's a lovely piece of work that uses snake embryology to come to some interesting conclusions about how venomous fangs evolved.
Basal snakes, animals like boas, lack venom and specialized fangs altogether; they have relatively simple rows of small sharp teeth. Elapid snakes, like cobras and mambas and coral snakes, are at the other extreme, with prominent fangs at the front of their jaws that act like injection needles to deliver poisons. Then there are the Viperidae, rattlesnakes and pit vipers and copperheads, that also have front fangs, but phylogenetically belong to a distinct lineage from the elapids. And finally there are other snakes like the grass snake that have enlarged fangs at the back of their jaws. It's a bit confusing: did all of these lineages independently evolve fangs and venom glands, or are there common underpinnings to all of these arrangements?
Let's start by looking at the phylogenetic tree and different fang arrangements. As you can see, those snakes with fangs at the back of their jaws (Natrix natrix, second from top) are interposed between one group of snakes with front fangs (the elapids, top), and another group with fangs in front (the vipers, third from top). We can imagine all kinds of scenarios that would produce that condition — front fangs are primitive, and Natrix secondarily lost them, or the fangs of all three are of independent origin and this is an example of convergence — but to resolve the question we need to look at some evidence. We need to examine embryos.
This is where we begin to see some underlying unity. Vonk and others used sonic hedgehog staining to visualize the dental primordia in snake embryos (O Sonic Hedgehog, is there no process in which you are not involved, nothing in which your expression is not enlightening?) and mapped out the pattern of tooth generation. They identify an odontogenic band, a thin strip of tissue that gives rise to teeth, and note an interesting peculiarity: there are subdivisions into independent anterior and posterior dental lamina, and ablating the anterior lamina does not perturb the development of the posterior lamina. In essence, the snakes simply have a couple of separate tooth-generating zones in their embryonic jaws.
The cool observation is that even in front-fanged snakes, it is the posterior zone that generates the fangs. It is also this same primordium that buds off a tube and a sac that will make the post-orbital venom gland — even in the front-fanged snakes, they have a gland located way back behind the eye to produce venom.
These observations are diagramed below. The unspecialized dental lamina, the part that sprouts the generic small pointy teeth, is in green; the specialized posterior dental lamina, which makes fangs and the venom gland, is in orange. In all the venomous snakes, the venom gland is a tube that first extends forward, and then curls back to make the bulk of the gland even more posteriorly. The important point is that all of these snakes use the same small posterior scrap of embryonic odontogenic tissue to make fangs and glands — we can make a pretty solid argument that these structures are all homologous.
Now hang on, you may be thinking, if all the fangs develop at the back of the jaw, how do they end up out front in the front-fanged snakes? We can find the answer in development, too. Remember that the two tooth generating zones, front and back, are independent, and the front one can be repressed without disturbing the development of the posterior zone. In the front-fanged snakes, the anterior part of the upper jaw lacks sonic hedgehog expression, and the posterior teeth move forward naturally as part of the allometric expansion of the jaw in embryonic growth. This is sweet: not only does development reveal a homology, it also exposes the process that led to a morphological difference.
Here is the authors' summary:
Our results suggest a new model for the evolution of snake fangs. A posterior subregion of the ancestral tooth-forming epithelium became developmentally uncoupled from the remaining dentition, resulting in posterior and anterior dental laminae that are developmentally independent. This condition is retained in the non-front-fanged snakes, such as the grass and rat snake. This model would imply that the front-fanged elapids and viperids have independently lost the anterior dental lamina, which is supported by the lack of shh expression anterior in their upper jaws.
The developmental uncoupling of the posterior from the anterior tooth region could have allowed the posterior teeth to evolve independently and in close association with the venom gland. Subsequently, the posterior teeth and venom gland could have become modified and formed the fang-gland complex—an event that underlies the massive radiation of advanced snakes during the Cenozoic era.
The key innovation in snake evolution was a subtle one, an uncoupling of two tooth-generating regions that opened the door to more flexibility in the modification of the jaws. The fang/venom gland complex probably evolved once in the common ancestor of these groups, but the elapids and vipers independently stumbled on a secondary change, the suppression of the anterior region, that allowed the posterior fangs to move forward to make a more effective poison delivery system.
Vonk FJ, Admiraal JF, Jackson K, Reshef R, de Bakker MAG, Vanderschoot K, van den Berge I, van Atten M, Burgerhout E, Beck A, Mirtschin PJ, Kochva E, Witte F, Fry BG, Woods AE, Richardson MK (2008) volutionary origin and development of snake fangs. Nature 454:630-633.