Casey Luskin Takes on Homology

This is an excessively recursive post: my review of Casey Luskin's review (on the Evolution News & Views website) of Henry Gee's, Rory Howlett's and Philip Campbell's review (in Nature) of a scientific paper by Toshiyuki Matsuoka et al.  This is not "blogging on peer-reviewed research;" this is blogging on tendentious carping about a popularization of peer-reviewed research.

The paper describes Matsuoka's research into which parts of the skeleton originate from the neural crest cells formed in the outermost of the three layers (the ectoderm) of the developing embryo, and which parts originate from the mesoderm, the middle layer.  This, in turn, the Nature editors state, gives us insights into the evolution of the various parts of the skeletons of modern vertebrates.

Luskin makes two points.  The first is that Matsuoka's research doesn't depend on the assumption of common descent, and could be done simply on the basis of comparative anatomy.  I'm pretty sure this is true.  The second, which takes up most of his article, is that Matsuoka's research throws a monkey wrench into the whole idea that homology is a result of common ancestry (actually, it's not quite clear whether Luskin means to attack the whole idea of homology -- which was identified by creationist biologists before evolution was invoked to explain it -- or just "Darwinian" explanations of it).

Luskin notes that Matsuoka notes a bizarre feature of embryonic development across species: the development of muscles from particular groups of cells is far more consistent ("more conserved") than are the actual bony sites to which  they attach.  Matsuoka, rather than reject the idea that the detailed similarity in form, location, and embryology of the neck muscles are homologous, proposes a startling idea:

The rather counterintuitive "scaffold model" perceives muscle connectivities as the basic units (because they precisely correspond to cell populations) but considers the bones that everyone can see as mere epiphenomena and subjects of change.

Luskin mocks this as a "bizarre" conclusion, forced by Matsuoka's slavish adherence to the dogma of common ancestry.   But it's not entirely clear what is so "bizarre" about it.  Vertebrates, after all, don't literally inherit their bones, or muscles, or nerves, or other body parts; they inherit genes, which interact with cellular mechanisms to grow all these parts.  Genes are not literally blueprints; there is no gene for a scalpula located right next to the genes for the clavicle or cervical vertebrae.    Matsuoka is suggesting that rather than trigger bone development directly, genes trigger muscle development in such a way as to promote bone development, with some interspecific variation in exactly how they accomplish this.  Interestingly, David Klinghoffer has a post on EN&V today that makes just this point: the genome isn't literally a "code" or blueprint (how exactly this makes the case for intelligent design is not quite clear).  Luskin can't quite get his head around the point that just because bones are more likely to be preserved than muscles or genes, they don't have to be directly inherited rather than effects of things that are directly inherited (of course, as Upton Sinclair noted of some other people, his salary depends on not understanding certain things).

Luskin goes on to note some further oddities of the homologies in vertebrate skeletons, such as the apparent loss, on multiple occasions, of the cleithrum, the major shoulder bone of primitive tetrapods, in different tetrapod groups (frogs still retain it, though salamanders don't, and many ancient amniotes did, though no modern amniotes do):

You've got to love the last sentence, which is worth repeating: "We speculate that a common, as yet unknown, genomic cis-regulatory architecture governing neck ossifications in tetrapod ancestors might have predisposed different descending tetrapod lineages to similar parallel trends." This is almost starting to sound like teleological evolution--but of course that's not allowed in Darwinian thinking, so it's just by the luck of the draw that all of these taxa independently arrived at the same bone structure. 

Now, Stephen J. Gould used to argue against the instinctive resort to "adaptionist" explanations; not every feature of biology or evolution, he argued, was an adaption.  One alternative that he mentioned more than once was developmental constraints: species descended from a common ancestor inherited developmental pathways that made some possible evolutionary changes difficult or impossible, while making others very easy and likely to arise as a side effect of other changes, whether they were directly beneficial or not.   This is not "teleological" thinking, as though evolution meant to lose this bone in advanced tetrapods; this is a suggestion, very common in evolutionary thinking for a century or more, that some changes are easier and more likely than others, and thus are likely to occur independently in multiple lineages.

Luskin includes in the article a quote from the creationist textbook Explore Evolution:

In sharks, for example, the gut develops from cells in the roof of the embryonic cavity. In lampreys, the gut develops from cells on the floor of the cavity. And in frogs, the gut develops from cells from both the roof and the floor of the embryonic cavity. This discovery -- that homologous structures can be produced by different developmental pathways -- contradicts what we would expect to find if all vertebrates share a common ancestor. (Explore Evolution, p. 44)

The book is right: we would, as a first rough estimate, expect that homologous structures develop through homologous pathways.  On deeper reflection, though, it's not quite clear whether we ought to invariably expect that.  On the one hand, if genes and tissues can evolve, presumably developmental pathways can evolve also, and still remain homologous.  In principle, one might find that novel genes had taken over some part of a developmental pathway from the ancestral form, though given that the same genes are found in all cells of the embryo, this is hardly necessary; the phenomena cited by Explore Evolution could result  from changes in where various gene cascades were triggered, without much change at all in the genes that actually build the gut.

There's a point that keeps going through my mind as I read Luskin's articles on the Nature series on "Evolutionary Gems," and that is, how does this support design?  If he's denying that homologies exist, he's basically telling us to believe him and not our own (and several centuries' worth of comparative anatomists') lying eyes; if he's arguing that homologies represent "common design," then how is this helped by arguing that in fact the designs are not common at all?  The variations in detail of how vertebrate bodies are built with conservation of basic elements (across disparate body shapes and lifestyles) fits at least as well with our expectations of evolution as with intelligent design.  But then, we're not really entitled to have expectations of intelligent design; it's sole testable hypothesis is that somewhere, somehow, evolutionists have got some stuff wrong.  Unfortunately, that testable hypothesis doesn't really differentiate intelligent design from evolutionary theory, whose proponents have generally followed Charles Darwin in realizing that they don't understand everything.