Morphogenesis and homology in arthropod limbs

American Zoologist, Jun 1999 by Williams, Terri A

Morphogenesis and Homology in Arthropod Limbs'

SYNOPSIS. Arthropods exhibit highly diverse limb morphologies ranging from unbranched walking legs to multibranched swimming paddles. Understanding morphogenesis in structurally diverse limbs can be useful for ascertaining homologies between limbs. Structurally similar limbs have been produced by different evolutionary modifications of morphogenesis in certain cases. Whereas it is easy to support the claim that whole arthropod limbs are homologous structures, I demonstrate that it is not always possible to draw well-founded homologies between parts of different limbs. This result is important with regard to general models of appendage development and evolution in arthropods because it clarifies contradictory explanations based exclusively on gene expression data.

DEFINING HOMOLOGIES BETWEEN ARTHROPOD LIMBS

One exciting finding of the last few decades has been the widespread conservation of patterning genes in diverse groups of organisms. Hox genes, which play an important role in patterning the anterior/posterior axis, have been found in groups as diverse as cnidarians and chordates. The functional role of these genes has been difficult to determine in non-model systems, but the initial view based on gene expression data has been that function is also widely conserved. These studies of comparative gene expression have led some researchers to view molecular similarity as a more reliable basis for drawing homologies between organisms than the traditional criteria based mainly on morphology. Thus, for some, the presence of conserved regulatory molecules during development has been assumed to be sufficient to indicate homology in adult structure. This has hardly proved a convincing or satisfying criterion for morphologists. The beguiling nature of in situ hybridizations notwithstanding, molecules do not equal morphology; the two exist at very different spatiotemporal scales within organisms. Most importantly, whereas molecules exist in labile states in a network of pathways, morphology is the outcome of the synthesis of a vast number of such interactions in space and time. Morphology and patterns of morphology are more complex than molecules and patterns of molecules. The use of molecules to establish homology has recently been criticized (Bolker and Raff, 1996; Dickinson, 1995), and we are thrown back to the view that defining homology is difficult (see discussions in Hall, 1994). The strongest statements of homology are those that rely on multiple criteria for support, like embryology, structure, and position within the adult. The general embryological criterion has been refined in recent years-based in part on the recognition of the difficulties of drawing homologies between serially iterated structures-to reflect not simply shared undifferentiated embryological precursors but, more explicitly, shared developmental pathways that generate morphology (Roth, 1984), or, most generally, shared continuity of information (Van Valen, 1982).

Arthropod limb branches, like the digits of vertebrate limbs, have proved problematic with respect to drawing homologies. This general difficulty in drawing homologies between structures with multiple parts was recognized over a century ago. In 1892 Bateson, wrote:

. . the received view of the nature of homologies in teeth assumes that in Variation the individuality of each member of the series is respected. The difficulty in applying this principle is notorious, not only in the case of teeth but in all cases of Multiple Parts, such as digits, phalanges, etc.; and when the actual evidence of variation is before us, the cause of this difficulty will become apparent enough, for it will be found that although Variation may sometimes respect the individual homologies, yet this is by no means a universal rule; and, as a matter of fact, in all cases of Multiple Parts, as to the Variation of which any considerable body of evidence has been collected, there are numerous instances of new forms arising in which what may be called the stereotyped or traditional individuality of the members has been superseded." (p. 104)

This question of a stereotyped individuality has been expanded in recent years and explicitly linked to developmental processes. Attempts to draw homologies between limbs have largely provoked the recent expansion of these ideas. However, such attempts have strictly focused on vertebrate limbs. Differing perspectives address whether homology can be claimed only at the level of the whole limb or down to parts of the limb, i.e., the digits themselves. In the structuralist perspective, structures that result from global patterning mechanisms (e.g., digits) lack individuality since variation in their final form results from variation in the initial conditions of the global patterning mechanisms (Goodwin and Trainor, 1983). Parts generated in this manner simply represent different solutions to a common set of generative rules and thus do not have the attribute of individual homology. However, a similar emphasis on developmental mechanisms led Shubin and Alberch (1986) to claim that homologies could be found between digits since limb development displays repeatable chondrogenetic interactions that produce identifiable local patterns within the limb. This notion of developmental individuality was formalized by Wagner (1989) who identifies homologous structures as those that share a set of developmental constraints.