As I have noted before, the core findings of evo-devo were never predicted or anticipated by conventional evolutionary theory:
For there are two subtle, but crucial points to grasp: 1) The New view involved a paradigm shift that was not predicted by the theories of the Old view and 2) the New view, known as Evo-Devo, is much more friendly to front-loading.
Well, I just ran across a paper than provides another illustration of point 1.
Rapid morphological evolution is compatible with classic neo-Darwinian evolutionary mechanisms of random genetic change followed by natural selection, whereby existing structures are modified in their size or shape (Weiner, 1994; Losos et al., 2006). Nonetheless, the discovery of the phylum-level diversification (with no evident intermediates) that occurred in the early Cambrian is not anticipated by this theory (Müller and Newman, 2005). Even more surprising from the viewpoint of the standard evolutionary model has been the realization that the Metazoa have used a highly conserved set of gene products, the so-called developmental-genetic toolkit, to generate diverse body and organ forms since their inception more than half a billion years ago (Carroll et al., 2005). If morphological change is supposed to be driven by, and to track, genetic change, why were there not dramatic changes in gene content corresponding to innovation of new organismal architectures (e.g., the dorsal location of nerve cord and heart in chordates vs. the ventral location of these structure in annelids and arthropods; Gerhart, 2000), or of new developmental mechanisms for generating morphologically equivalent structures (e.g., segmentation in beetles and fruit flies; Salazar-Ciudad, 2001b)?
The unexpected gives way to the paradoxical in the recognition that disparate organisms use homologous genes when making structurally dissimilar but functionally similar structures traditionally categorized as analogous (e.g., eyes in insects, mollusks and vertebrates; fly and bird wings) (Carroll et al., 2001; Wilkins, 2002). Given that there is nothing eye-like about the twin of eyeless/Pax6 transcription factor (Gehring, 2002), or heart-like about tinman/Nkx-2.5 (Schwartz and Olson, 1999), and considering the immense evolutionary and anatomical divergence between Drosophila and the mouse, why are the same gene regulatory proteins used, correspondingly, to initiate the developmental pathways that produce structures that have the same function, but which are evolutionarily unrelated (or have been thought so in the past)? The conflict between the expectations of the neo-Darwinian model and these recent findings from the fields of paleontology, comparative anatomy and genomics, and developmental biology has been termed the “molecular homology-analogy” paradox (Newman, 2006). (emphases added).
From: Newman, SA and Bhat, R. 2009. Dynamical patterning modules: a “pattern language” for development and evolution of multicellular form. Int. J. Dev. Biol. 53: 693-705.
Disparate organisms use homologous genes when making structurally dissimilar but functionally similar structures traditionally categorized as analogous. Hmmm. Perhaps a look at this “pattern language” is in order, eh?