No abstract

The history of ostariophysan classification is summarized and it is noted that traditional concepts of relationships have never been supported by characters found to be unique to the taxa. We present a new hypothesis of relationships among four of the five major ostariophysan lineages: Cypriniformes, Characiformes, Siluroidei, and Gymnotoidei (Otophysi). Cypriniforms are the sister‐group of the remaining three (Characiphysi), and characiforms are the sister‐group of siluroids plus gymnotoids (Siluriformes). Placement of the Gonorynchiformes as the sister‐group of the Otophysi is supported by additional evidence. Each of the five lineages is monophyletic. Analysis was concentrated upon species thought to be the least specialized within each lineage; choices of these species are discussed. Chanos is determined to be a relatively primitive gonorynchiform morphologically and the sister‐group of all other Recent members of the order. Opsariichthys and Zacco are found to be morphologically primitive cypriniforms. We propose that a monophyletic group comprising the Citharinidae and Distichodontidae forms the sister‐group of all other characiforms. Within the two families, Xenocharax is the least specialized. We suggest that Hepsetus, the erythrinids, and the ctenoluciids are more derived than the distichodontids and citharinids, and may form a monophyletic group within die characiforms. The traditional hypothesis that Diplomystes is the primitive sister‐group of all Recent siluroids is substantiated. Our evidence suggests that Sternopygus is the most primitive gymnotoid morphologically; but rather than being the sister‐group of all other gymnotoids, it is the primitive sister‐group within a lineage called the Sternopygidae by Mago‐Leccia. Previous explanations of otophysan distribution have been based on notions of relationships which are unsupported by the evidence presented herein. Our own analysis of relationships serves primarily to make clear the extent of sympatry, and therefore the probability of dispersal, among the major ostariophysan lineages. The extent of sympatry, together with the widespread distribution of ostariophysans, suggests that the group is older than previously supposed, and our hypotheses of relationships among the characiforms implies that many of the extent characiform lineages evolved before the separation of Africa and South America. Further understanding of ostariophysan distribution must await phylogenetic analysis within each of the five major lineages so that distributions linked with vicariance patterns and dispersal events can be sorted out.

No abstract

Studies of ontogenetic processes are fundamentally dependent on hypotheses of phylogeny. The model of Alberch et al. (1979) is reformulated in terms of phylogenetics and used to describe how heterochronic ontogenetic processes can be detected in nature. Heterochronic processes producing paedomorphosis can result in morphologies which resemble primitive (retained ancestral) traits; the conditions under which paedomorphic and primitive features can and cannot be distinguished are described. The utility of ontogeny for determination of evolutionary character transformations and character polarity and for detection of convergence and parallelism are considered. The ontogenetic criterion for assessing polarity is independent of hypotheses of phylogeny and may be as effective as outgroup comparison. Ontogenetic analysis may aid in the detection of convergence but not in the detection of parallelism.

Disparity appears to decrease or remain stable over geological time in numerous groups. This pattern is sometimes explained in terms of developmental constraints, but labile developmental systems might produce the same pattern should novelties interact, counterbalancing their individual effects. We test the hypothesis that counterbalancing can reduce disparity by comparing ontogenies of shape among nine species of piranhas to identify the developmental novelties. All three parameters examined change multiple times, sometimes dramatically. By comparing levels of disparity between species at two developmental phases, at the transition from larval to juvenile phases, and at maximum adult body size, we find that disparity decreases significantly and substantially over ontogeny. That reduction occurs because of, rather than despite, novelties of postlarval morphogenesis. Some interacting novelties are historically independent and affect different developmental phases, others are historically independent and affect the same developmental phase, and still others are historically correlated and affect either the same or different developmental phases. By modeling hypothetical ontogenies, constraining developmental parameters mathematically to one of the observed values, we find that variation in each parameter, taken by itself, and combinations of them taken two at a time, tend to increase disparity. It is the interactions among all three that reduce disparity. In this group divergent ontogenies transform disparate larvae into similar adults.