No, this is not another review of a show, not yet anyway. That will be on Tuesday. This is some speculation as to what selective pressures lead to the development of snake fangs. Researchers recently discovered how fangs develop and PZ Myers described the research in relatively easy-to-grasp terms. If you would like the original paper, the citation is at the bottom of PZ's review of it.
The paper addresses what we know about how snake fangs evolved, but it does not answer why snake fangs evolved. The selective pressures for snake fang development are fairly easy to identify in a stepwise manner.
1) no fangs, no venom
2) venom gland formation from modified secretory cells located in the mouth
I leave this one open for several purposes I shall get to later.
3) fang development (normal teeth being modified to form grooves)
4) in some species, the grooves were refined into hollow fangs, and in others, further evolution of the grooves occurred.
We do not, as yet, know how the venom glands specifically evolved, however, it is probable that some mutations prior to the separation of various lineages. Here's a phylogeny based upon mitochondrial DNA...below the fold.
If you look at the points of mutation among all genera containing venomous snakes (Viperidae, Elapidae, Colubridae, Atractaspis), all of them are related by a common ancestor dubbed Colubroidea. Dr. Brian Grieg Fry made a nice little breakdown of Colubroidea evolution.
It is likely that the evolution of venom glands directly lead to the development of fangs due to efficiency of delivery. The method of feeding for colubrid snakes is generally that of constriction with the venom serving the role of accelerating the death of the prey. It is also a misnomer that colubrid snakes are nonvenomous. Many colubrid snakes, in fact, have venom glands, but many are not harmful towards mice. Of the ~1400 species of colubrid, 700 have venom of some kind. This indicates that some species commonly classified as colubrids are actually closer related to other groups than previously thought. This is reflected in Dr. Fry's phylogeny.
It may also be noted that no snake species outside of Colubroidea has been observed to have venom indicating that venom production likely developed in one of the common ancestors of these organisms. I would also note that many of the proteins found in venom are made in many other parts of the body, so it could be a modification of the expression of these proteins which lead to the development of venom.
I would like to add that much of what I said is purely an educated guess as we do not have the data yet to prove it, but it does have evidence in the form of protein homology in a number of venom proteins for origins in other.
Now, for the glands. Because some species of lizards also have venom glands, it is probable that the development of a new oral gland much earlier than previously thought.
Since I find this so very interesting, I can go on at length about the the origin of the glands and how they form, blah, blah, blah. Instead, check out Dr. Fry's page, he's got lots of good information on there.