Specimens of Peronopora Nicholson, 1881, are abundant in Upper Ordovician rocks of the North American Midcontinent. Based on the positions of units in the Composite Conodont Standard Section, we have sampled 211 specimens over a stratigraphic interval of 9.1 million years. The average duration of sample spacing is 61,664 years but is commonly as small as 32,800 yr.Thirty-four morphometric characters were measured in each specimen and were converted into multistate characters; character-state breaks were established based upon each character's ability to discriminate between phenetic groupings. Each character was subsequently weighted based on the number of derived states, degree of independence from other attributes, and estimated heritability.Cladistic analysis of these data indicate that there are eight species in Peronopora each consisting of an optimally defined crown group and a basal stem group (or paraclade). Character states shared by stem and crown groups define species but, within species, stem and crown groups also differ in some character states. The species are, in ascending order from the base of the tree, Peronopora decipiens (Rominger, 1886), P. compressa (Ulrich, 1979), P. pauca Utgaard and Perry, 1964, P. milleri Nickles, 1905, P. horowitzi new species, P. vera Ulrich, 1888, P. sparsa Brown and Daly, 1985, and finally P. dubia (Cumings and Galloway, 1913). Diagnostic keys permit the unique assignment of each specimen to a species and the separation of members of stem groups from those of crown groups. Thirty-one characters are required to discriminate between all 211 specimens. This contrasts with previous studies of Peronopora where eight or fewer characteristics were used. Of the ten characters most useful in discrimination, only three had been used in the conventional species literature. This accounts, largely, for only 29.8 percent (51 of 171) of previously identified specimens being classified as members of the same species in this analysis. Discriminant function analysis of original measurements, using species identity as the grouping criterion, produces statistically significant separation of species.It appears that stratigraphic position had an explicit and undue effect on previous concepts of species many of which could not be recognized independently of stratigraphic position. All species of Peronopora appear, or are inferred to have appeared, within the Lexington Limestone between the base of the Grier Member and the top of the Millersburg Member. The cladogram indicates that species evolved in a sequential order, but their first appearance datums have been stratigraphically punctuated. Three species have ranges terminating in the Early to Middle Maysvillian, one in the Middle Richmondian, and four in the Late Richmondian. The latter four (or five) of these species died out in the extinction associated with the unconformity at the top of the Richmondian.

Environmental conditions affect both the character and variability of developmental patterns (=astogeny) within each of four species of Ordovician bryozoans. Regressions between pairs of stereological measurements for populations from both high-and low-diversity habitats differ significantly in 79 percent of all comparisons. Deviations from a rigid pattern of development, measured as dispersion from regression, were greater in species populations from low diversity settings in 77 percent of comparisons. Therefore, both developmental patterns and their variability differ intraspecifically along a diversity gradient in representatives of four bryozoan families. Additionally, dispersion values were larger in younger rather than older colonies in two species irrespective of diversity level, thus suggesting an age-related reduction in variation.Changes in developmental trajectories indicate that colonies from low-diversity settings are generally paedomorphic relative to conspecific populations from high-diversity habitats. The heritability of these characters and developmental patterns, estimated using variance partitioning techniques, is greater in high-diversity associations. These findings suggest that character state modifications dependent upon astogeny, or a consequence of astogenetic modifications, are more heritable in high-diversity settings. However, if the environment can cause facultative heterochrony, the possibility of fixing such patterns in subsequent generations is increased, although the mechanism for accomplishing this is presently unknown.

A paradoxical relationship exists between the genetic and morphologic adaptive strategies of benthic marine invertebrates; morphologically variable species from unstable environments have been shown to possess less genetic variability than species with more constant phenotypes from stable habitats. The mode of growth of Ordovician bryozoans provides an insight into this paradox. These bryozoans exhibit morphologic gradients within zooidal subcolonies that change throughout colony development. Entire clusters of zooids begin and cease growth as a function of their spatial position with respect to neighboring clusters. A comparison of the within-colony and among-colony components of developmental and morphologic variability was made from populations inhabiting Ordovician environments of differing stability. In four stratigraphically persistent species, the level of developmental variability is homogeneous within species, but varies significantly across taxa. The two species with the highest levels of developmental variability (less canalized development) fit the concept of r-selected opportunistic species and are most abundant in communities of lowest diversity. The other two species have much lower levels of variability (more canalized development), fit the concept of K-selected equilibrium species, and are most abundant in the communities of highest diversity. Within-colony morphologic variability is also higher in the opportunistic rather than the equilibrium species, indicating that the higher morphologic variability observed in unstable environments is the product of within-genotype deregulation, and not the result of higher genetic polymorphism. The equilibrium species in stable environments have lower levels of morphologic deregulation and correspondingly greater variation among genotypes than within genotypes in fossil populations.

Area cladograms produced by parsimony analysis of endemicity illustrate historically developed biogeographical associations among Caradocian, Ashgillian, Llandoverian, and Wenlockian bryozoans. Areas in North America, Siberia, and Baltica were organized into three provinces and 12 biomes over a time interval of 35 million years. Six of these biomes belonged to the North American-Siberian Province and became extinct during the Ashgillian. Three biomes represent a successional series of biogeographical associations in the Late Ordovician of Baltica, and the middle biome of this succession is most closely related to that of the Wenlockian platform in North America. All four Silurian biomes are represented in Late Ordovician local areas, indicating that the associations important in the recovery radiation were already in existence prior to the extinction events. Three of these four biomes expanded their geographic extent in the wake of the Late Ordovician extinctions. Several biome extinction and replacement events took place during lowstands of sea level, suggesting that biogeographic reorganizations took place as a consequence of habitat loss in epeiric seas. Biome development largely depended on the extent of major litho-topes and their intersections with deep ocean and climatic barriers. The loss of regional habitats, associated with marine regression, was a key factor in biome extinction and reorganization, and indicates that biogeography played a significant role in the Late Ordovician mass extinctions and Silurian recovery radiations. Vicariance hypotheses are needed to account for the development of barriers subdividing ancestral areas, whereas hypotheses of congruent dispersal are required to explain the appearance of biomes in geographically disjunct areas.