Observable evolution today.

Archaeopteryx had an impressive array of features that immediately identify it as a bird, whatever else may be said about it. It had perching feet. Several of its fossils bear the impression of feathers. These feathers were identical to those of modern birds in every respect. The primary feathers of non-flying birds are distinctly different from those of flying birds. Archaeopteryx had the feathers of flying birds,2 had the basic pattern and proportions of the avian wing, and an especially robust furcula (wishbone).

This is strange. Wishbones, or furculae, are known in a host of non-avian dinosaurs. Feathers are known in a host of non-avian dinosaurs. Given some interesting new evidence (e.g. Middleton 2001, 2003), its now pretty clear that Archaeopteryx lacked a reverse hallux (which is what I presume Gish means by "perching feet"). None of these things are suitable to diagnose Archaeopteryx as anything other than a maniraptoran theropod.

Questions about the homologies of certain characters (Gish brings up identity of its manal digits and the degree of opisthopuby) are addressed by myself <edit to remove link>. To quote myself:

Following Alberch and Gale (1983, 1985), Hinchliffe and Hecht (1984), Hinchliffe (1985, 1997) Martin (1991), Tarsitano (1991), Hecht and Hecht (1994), Feduccia (1996, 1998, 1999, 2003), and Burke and Feduccia (1997), it is claimed that the digital identity of the avian manus is II, III and IV, as opposed to the I, II and III traditionally assumed for theropods. Originally proposed as an extension of "Morse's Law," which claimed the patterns of amniote digital reduction always occur in the lateral most digits (I and V; Morse 1872), recent embryological studies have argued essentially the same.

In response, critics have traditionally charged that it was not entirely clear that this amniote "ground plan", as Chatterjee (1998) called it, so essential to the embryological argument's identification of visable anlagen as II, III, IV and V was applicable to birds. As he noted, salamanders first develop digit II, and although Burke and Feduccia (1998) are certainly correct that salamanders are not amniotes, it was not entirely out of the question to assume avian/theropod development begins with digit III (as per Shubin, 1994; Garner and Thomas, 1998), and is similarly extraordinary.

Although these doubts as to the identity of visable anlagen are still expressed (for example, Padian 2001), the position is growing incresingly untenable. Studies by Larsson & Wagner (2002) and Kundrat et al (2002) on Gallus, and the preparation of embryos of Struthio camelus by Feduccia & Nowicki (2002) have all identified the elusive fifth (digit I) anlagen. Other more classical lines of evidence, from the position of the pisiform (Hinchcliffe 1985), to the structure of the stereotypical chondrogenic "Y" (Burke & Feduccia 1997, Feduccia & Nowicki 2002), consistently support the identification of condensations as II, III and IV.

As clear as the above evidence may be, some authors (e.g. Galis et al. 2003) have expressed doubt, not about the digital identity of the avian manus, but of the non-avian theropod manus. Herrerasaurids, recently classified as basal theropods, are so plesiomorphic (specifically with regards to the medial wall of the acetabulum, elongated cervicals three and four, lack of epipophyses, development of MT I and IV, the distally rod-like ischium, and the subnarial foramen) that their phylogenetic position would more properly be described as basal saurischian, if not basal Dinosauria or even some sort of dinosauromorph, making their manal digit reduction possibly independent of Theropoda proper, and their sacral count a case of parallelism (Sues in Dodson et al, 1990; Novas, 1994; Sereno, 1994; Fraser et al, 2002; Galis et al, 2003; contra Feduccia 2003; Larrsson and Wagner 2003). Similar difficulties arise with regards to Eoraptor (Galis et al, 2003). Because there are, as of yet, no clearly unambiguous basal theropods retaining two vestigial manal digits, its entirely possible that the vestige retained in ceratosaurs represents V, not IV as traditionally assumed, having already lost I. In this view, both non-avian and avian theropod digits are numbered II, III and IV.

A third possibility championed by Gauthier and Wagner (1999) is that avian digits indeed develop from condensations II, III and IV, theropod digits are indeed I, II and III, but at some point before the derivation of Coelurosauria a homeotic transformation occurred such that condensation CII developed digit DI, condensation CIII developed digit DII, and so forth. The possibility is intriguing given the disconnect between morphogenesis and identity seen in other body segments (Burke et al, 1995), and recent work by Dahn & Fallon (2000) which has shown that the conserved patterns of development do not necessarly dictate digital identity in particular. Very simply, it shows that the number condensation a digit develops from does not necessarly correspond to a digit homologous with that number. The only major shortcoming is that there is no obvious evolutionary impetus for such a transformation. In any case, there is just not enough evidence to level convincing charges of digital pseudo-homology.

About opistopuby I wrote:

The extreme degree of pubic retroversion restored in Archaeopteryx by Martin (1983, 1991) and cited by others (for example, Ruben et al, 1997), in which the pubis is nearly parallel with the ischium is both unsupported by earlier specimens and demonstrably incorrect in light of newer ones. Because the ***** are disarticulated in the only two specimens (London and Berlin) showing such an extreme degree of opisthopuby, this caudal orientation could very well have been achieved postmortem, as forcefully argued by (among others) Wellenhofer (1985). This was later conclusivly demonstrated by the specimens in which the pelvis was well preserved. According to a recent survey:

"The well-preserved pelvis in the Munich specimen (Wellnhofer, 1993) confirmed that the pubis was directed nearly vertically downward as in Unenlagia (Novas and Puerta, 1997) and Rahonavis (Forster et al, 1998). The pubis was only slightly retroverted, forming a cranial angle of 110d with the long axis of the ilium (Wellnhofer, 1985)." (Elzanowski, 2002, p. 144)

With opisthopuby often exceeding 110 degrees in various dromaeosaurids, its actually quite accurate to say that some non-avian dinosaurs were more opisthopubic than the basal bird itself.

"Protoavis" is addressed by <edit to remove link>

Some other claims included:

1) Archaeopteryx could fly: Yes, it probably could -- but not very well. This hardly establishes your point.

2) Feathers are not highly modified scales: This is correct as well. It does not follow, however, that feathers did not evolve; the current concensus theory, that of Prum 1999, Prum & Brush 2002, etc, does not suggest this is the case. Instead, feathers are seen as morphological novelties.

3) Archaeopteryx is a hoax: You some to disagree with this one, which is a start.

4) The IAC claimed it is not a transitional: This is news to me. I own the proceedings, so be sure to tell me who argued this (preferably with a page number).