A minor correction is worth publishing. In the article Myers states:
We’ve got a pretty good handle on the outline of limb development in multiple tetrapod lineages now, and they all use the same tools. Contrary to Sarfati’s implication, they all have apical ectodermal ridges (with some rare exceptions in a few highly derived, direct-developing frogs) and zones of polarizing activity, they all use the same set of molecules, including FGF-4 and FGF-8 and the same Hox genes and retinoic acid and BMPs. If there’s one thing we know, it’s that limb development is dazzlingly well conserved.
Despite the fact that I think most of Myers's article is a wonderful vulgarization of limb development, this last part is an overstatement.
It is true that many aspects of limb and digit development in tetrapods is remarkably well conserved throughout tetrapods, but it is not true that all tetrapods have an AER, it is also not true that they all use apoptosis to create the interdigital spaces. All salamanders that have so far been studied grow their digits from buds and lack an AER, much like Sarfati explained for frogs (only Sarfati attributes it to the wrong animal). Apoptosis has only been detected in the developing digits of one salamander Desmognathus and its role in limb development is not clear (Franssen et al. 2005). It is also the case that salamanders begin digit development with digit 2, which is also in stark contrast with the majority of other tetrapods which develop digit 4 first. This is an interesting and perplexing problem for developmental biologists.
So does this mean that Sarfati is right? Does this mean that the limbs of salamanders cannot be homolgous to those of other tetrapods? The problem comes from this statement by Sarfati which exemplifies a rather deep ignorance of developmental biology (and I'm not even an expert!):
If the birds evolved from dinosaurs, then one would expect common genes. These in turn would code for a common development in the embryo.
This is both overly simplistic and fallacious, since we already know that most of the "genetic toolkit" for animal development is highly conserved. The fact is, the same genes are used in nearly all tetrapods. In fact, vast numbers of the same genes are involved in similar processes in all animals. Some genes have been duplicated and modified to produce different functions -- which is precisely how animal form evolves!
As Myers points out beautifully, and what is the most important point here, is that the patterning mechanisms are conserved. We know that developing tetrapod limbs use a special hierarchical pattern of Hox genes that lay out the "construction zones" in a developing embryo. We see remarkable conservation in both the Hox genes and the placement of special region known as the zone of polarizing activity (ZPA) which plays a role in setting up anterior-posterior polarity of the developing limb.
The emerging picture of salamander limb development is that there are a striking number of "mechanical" differences in the way the construction takes place. The order of digit development and many of the processes that make the digit protrusions are different. However, what doesn't appear to have changed, is precisely what we would expect if the digits are homologous: the genes that control the identity of the digits.
Animal development doesn't work in the same way as humans construct buildings. It's not as logical as we would think it is. In fact, it's rather perplexing and enigmatic and often proceeds along seemingly very illogical courses. The process of making a digit (as in, a protrusion of the distal part of a limb) is under different controls from the process that says what those digits are (the thumb vs. forefinger, etc.). Sarfati's oversimplification is academically irresponsible. It is flat-out wrong making the entire thesis of his argument baseless.
Franssen, R.A., Marks, S., Wake, D. and Shubin, N. 2005. Limb chondrogenesis of the seepage salamander (Desmognathus aenus(Amphibia: Plethodontidae). Journal of Morphology 265:87-101