Since John Lynch mentioned it first, why yes, I have read this article on limb loss in vertebrates.
Some of the complicating features of developmental genetics are pleiotropy and multigenic effects: that is, that the genes required to build an organism are all tangled together in an intricate web, with multiple genes required to properly assemble each character (that's the multigenic part), and each gene having multiple effects on multiple characters (that's pleiotropy). One might think of the organism as a house of cards, each card supporting all of the cards above it, so that tinkering with any one piece leads to catastrophic collapse. This isn't the case, of course. While developing systems are all elaborately interlocked, they also exhibit modularity and surprisingly robust flexibility. One recently published example can be found in Shapiro et al. (2004) which describes the developmental flexibility of the regulation of the pelvic appendages in sticklebacks, and ties it all neatly to patterns of evolution.