SYSTEMATIC, PHYLOGENETIC AND MORPHOMETRIC ANALYSES OF THE JURASSIC RADIOLARIAN GENUS PERISPYRIDIUM DUMITRICA

The Jurassic genus Perispyridium has previously been recognized as a group of biostratigraphically important but morphologically "bizarre" radiolarian species. Despite their practical utility, a comprehensive analysis of patterns of morphologic variation within the genus has not appeared in the systematic literature. During 1982-1984, an intensive stratigraphic sampling program was undertaken in the Jurassic Snowshoe Formation of east-central Oregon to collect material for this purpose. In addition to providing large numbers of stratigraphically located specimens of known species, these samples also yielded thirteen new species of Perispyridium that are described and illustrated. Phylogenetic analysis of the distribution character states among Perispyridium species indicates that this genus can be regarded as formally monophyletic. Also, a number of apparent instances of iterative evolution were documented during this phase of the analysis. Study of patterns of morphometric variation based on thirteen skeletal dimensions revealed that approximately 50 percent of among-species variation can be attributed to species-specific size differences, and that size-independent shape variation is discontinuously distributed among three clusters of species. Patterns of among-species variation are predominantly isometric while patterns of within-species variation exhibit marked allometries. In addition, patterns of morphological integration between skeletal dimensions differ among the morphometric clusters of species. This possibly reflects the operation of an evolutionary process-level discontinuity that may result from the existence of causal architectural constraints operating within the Perispyridium Bauplan. Finally, analysis of temporal patterns of phenotypic evolution for the six most abundant species indicates that, in terms of the range of phenotypic variation present within a series of stratigraphically partitioned samples, no directional trend in overall morphologic variation of multivariate sample means over time is discernible. Patterns of generalized size and shape variation do exhibit variable "log" responses when plotted against stratigraphic position, but this is most likely an example of species-specific morphologic response to local patterns of ecophenotypic variation. Estimates of the culling ratios and effective population sizes required to account for the observed patterns of shape variation are very low, thus further emphasizing the static nature of morphologic variation for these species.