“Everybody knows” that species can lose features through evolution: snakes, whales and sea cows all lost hind limbs. But occasionally, they can “revert”, like the snake shown in the photo below, which has grown one limb.
However, many people are confused as to why this happens.
A popular misunderstanding of evolution is that traits come and go rather fluidly: if you don’t need eyes, for example, then you lose them, like cave dwelling fish do. Likewise, when snakes no longer needed their legs, because they were burrowers or undergrowth crawlers, they simply lost them. But this is not how evolution works.
Organisms are developmental systems. This means they have mechanisms in place that, when the time or conditions are right, will do things like grow limbs or eye. These systems are usually pretty deeply embedded in the overall lifecycle of the organism – for instance, limbs in tetrapods grow from early in fetal development, about week 4 in the case of humans.
So if evolution had a “need” to eliminate limbs, it would be easier simply to deactivate the mechanisms by overlaying a “confounder”, so to speak, than to remove the mechanisms themselves, largely because the genes and other cellular processes that cause limbs probably are closely involved in a host of other, equally significant, traits as well. Remove the leg-genes or limb-budding tissues and you may find that you no longer get, say, a head, or a liver. [Made up examples!]
So blocking, or “downregulating” genes in the tissues that cause limb growth is easier, and more effective. But this means that snakes are not, as you might expect, “legless”; they are instead four limbed in an unusual way. That is to say, they are tetrapods still, but they have legs in a manner that results in no legs being grown, unlike, say, a worm, which has no deactivated mechanisms for limbs at all (it is not a tetrapod).
Sometimes this is called “phylogenetic inertia“, but a much better term is Bill Wimsatt’s “generative entrenchment” – in snakes, limb mechanisms are entrenched, so turning them off means they remain there, but don’t do the limb growing they used to in snake ancestors. Such things are called, in phylogeny (the working out of the evolutionary relationships between taxa), plesiomorphies, which simply means the ancestral or initial states from which all subsequent states have derived.
There is nothing quite like physical inertia in evolution. Things are not retained because they have some impetus to persist over millions of generations; they persist because mutation cannot eliminate them, and the reason why this is, is because they are selectively maintained. When genes are deactivated and have no other function, they slowly become, in computer parlance, “corrupted”. But they aren’t immediately eliminated, because the mechanisms that produce them are important (the mechanisms that copy DNA during mitosis and meiosis).
The lesson is that while history destroys functional information, it rarely destroys all trace of the past. This is why we can, to a degree, work out relationships between species on the basis of their developmental, morphological and molecular traits. This is also why sometimes you get atavisms.
An atavism is a trait that recurs in a group of organisms that notionally have eliminated it. It used to be thought that this could happen if you allowed your sheep or goats to breed in sight of the trait (as in Genesis chapter 50); in Darwin’s day it was thought that you could activate atavisms through exposing the mother to the wrong environmental influences (something Darwin himself thought). Before the modern synthesis, it was thought that atavisms were a sign of some kind of retrograde evolution, or “ontogeny recapitulating phylogeny” in an imperfect manner.
But now we know that what regulates the expression of genes are often post-transcription, and that occasionally the post-regulation fails. Snakes occasionally grow limbs because the mechanisms for limb growth are still there, and the stop signs get knocked down. Bird occasionally grow claws on their wings (and the hoatzin has them normally as a chick) for the same reason.
When you read that something is a vestige of evolution, often it is meant that the organ or trait has some residual function, but not the function for which it evolved (things don’t literally evolve “for” purposes – in my view all purpose ascription in evolution is after the fact). Atavisms are a kind of vestige, where they do not appear usually because the “function for” is completely lost. But they never quite disappear…
Late note: Jerry Coyne has a similar, and more detailed, post on tooth enamel genes.