Segmentation is an aspect of anatomy which does not lend itself to the Darwinian conception of gradual evolution. First, there is the question of how new articulations could have arisen through Darwinian evolution. How could a useful new jointed bone evolve in a fortuitous place with its complementary anatomical structures? Darwinian evolution operates "insensibly" (Darwin 1859); it is gradual and smooth change. But articulations are intrinsically discontinuities, at least where rigid structures are involved. Bones are either fused or articulated; there is no middle ground.
Second, whether segments include articulated hard parts or not, their symmetry is a difficulty for the Darwinian model of gradual change through random variation. Repeated similar segments, such as the vertebrae, are referred to as 'serial homologs', suggesting that each is built with the same genetic information. But if each segment gradually evolves in turn through the natural selection of morphological variants, then it has its own evolutionary history and its own independently evolved genetic information. The symmetry of segments must therefore be due to evolutionary convergence or parallelism, rather than to their sharing the same genetic information. This view must be rejected, since the number of identical segments may be so great, and the symmetry of segments may be so detailed, including many anatomical features and cell types, not merely the shape of a bone. There must be a mechanism through which segments are added which are built with the same genetic information expressed in pre-existing segments; but this mechanism is not explained by the general theory of natural selection.
Those are theoretical problems; there is also a significant problem with the physical evidence for gradual evolution in segmented organisms: the fossil evidence illustrating the course of evolution in segmented organisms shows that the number of segments--in those lineages where data are available--is decreasing. There is an unexpected lack of fossil series illustrating a process of gradual elaboration in segmented organisms, in the sense of an increase in the number of discrete skeletal parts. This is the case in the paleontology of both arthropods and vertebrates. There is, on the other hand, considerable evidence in the form of fossil series illustrating the process of reduction in number of segments.
By "fossil series" I do not refer to cartoon charts which show fish-to-man in a few upward steps. I refer to the hard evidence for evolution, fossils showing evolutionary change among indisputably related animals. The series showing the evolution of the horse is the most familiar example. The textbooks which introduced this evolutionary sequence to us did not point out that the number of parts in the skeleton actually decreased as the modern horse developed. This is because the authors were promulgating the idea of evolution as the natural development of complex organisms. In this context, the generalization that the number of parts is decreasing is not helpful and is not recognized as meaningful. Only advanced texts in evolution and anatomy recognize the reduction and specialization of homologous segments--the loss and distortion of skeletal parts--as an established principle. The fossil evidence shows that the number of segments in a given lineage is being reduced over time, while the segments that remain tend to lose their symmetry and to take on specialized roles.
How is this pattern of reduction in number of parts to be reconciled with a broad conception of evolution as the gradual development of complex organisms? The answer implicit in the literature is that the observed reductions are secondary phenomena, and that increase in skeletal complexity within segmented organisms does in fact take place, but that it happens that we have no evidence illustrating it directly. "Complexity" is not a clear scientific term; but here let us make it so by restricting it to a quantifiable meaning: the number of discrete skeletal parts.
Evolution is supposedly a process that can build complexity anytime, anywhere, in any sense. But in the cases where we can almost observe the process, through series of fossils, we see reduction in the number of skeletal parts. Let us put aside the idea of gradually increasing complexity in every sense, and take seriously the ubiquitous pattern of reduction and specialization among skeletal elements. Extrapolating this pattern of change back into the past suggests a progenitor with many parts, and this implies a model that has considerable explanatory power.
To say that the number of parts in the skeletons of segmented organisms is decreasing is not to claim that evolution is necessarily degenerative in any other sense, any more than a classic sculpture is a degeneration of a block of marble. Although the number of parts is decreasing, the parts that remain may be distorted greatly into specialized forms. Consider the wing of the pterosaur, which is a distortion of the last digit or 'little finger'.
But there is an unfamiliar suggestion of finitude in this evolutionary model. If evolution involves occasional loss of parts, while no new parts are added, the morphological raw material--parts which are free to be adapted to new purposes--must eventually run out. My answer is that there was a mechanism for adding parts, but it is a mechanism not likely to be effective in our era.
The evidence for evolution through loss of parts implies that there was a relatively brief period during which the basic types of our fauna were formed, and that these organisms included segmented animals consisting of many symmetrical parts, whose subsequent evolution has been a matter of reduction and distortion of these parts.
S.J.Gould has complained of the meaninglessness of the generalization that serially homologous parts are being distorted and reduced in number (Gould1980a). With the proposed model, this pattern is no longer merely an inductive result without consequences or explanatory value. Under the proposed model, this pattern is a consequence that can be deduced; random variation acting upon a set of symmetrical elements should produce asymmetry, not new symmetrical elements.
Although the pattern of reduction in the evolution of segmented organisms was recognized by the end of the 19th Century, it has remained an obscure and difficult matter, because evolutionary biology tends to accept elaborative developments contrary to the principle of reduction, such as the transformation of fins into limbs with more complex skeletons.
forward to Ideal Serial Homology