"Parallel Evolution" is not really a kind of evolution, but rather, an observation we make about the pattern of evolution in particular cases. Many species have a "woodpecker adaptation" by which a hard sharp thing is used to get at grubs and other insect (or non insect) meat hidden beneath bark. Some of the sharp things are beaks, one bird uses a cactus spine, and there is a primate with a special elongated finger for doing this, and most or all "woodpeckers" (bird like or otherwise) have related adaptation allowing them to figure out where to poke through the bark to find their prey. This is parallelism. But other cases of parallel evolution involve several different populations (which may or may no be different subspecies or different species) undergoing Natural Selection from a base population with some genetic variability such that the smaller sub populations all have a high incidence of the same genetic allele (variant). The loss of bony plates in various freshwater sticklebacks seems to be a case of (mainly) this sort of evolution.
The case of the sticklebacks is especailly interesting because the genetic variant that causes this trait in most (but not all) of these fish is also associated with a behavioral trait. Intrigued? You should be! As you know, I'm guest blogging for a couple of weeks at Smithsonian.com, and I've just put up a post examining the sticklebacks: What Robot Fish Can Tell Us About Parallel Evolution.
Go have a look and leave nice comments so the Smithsonian likes me!