Approximate periodicity for peak rates of global extinction during the past 250 m.y. may have resulted from delayed recovery following major extinction events. Two components can be envisioned for such delays: persistence of inimical environmental conditions for some time after the onset of the crisis, and slow restoration of vulnerable taxa. This general hypothesis is consistent with statistical evidence of linkage between measured rates of extinction of marine invertebrate genera for contiguous stages and substages of the geologic column. The nine broad valleys between the “periodic” peak rates for the past 250 m.y. exhibit only three trivial secondary peaks, indicating that, if the pattern is not artifactual, trends in global rates of extinction have not readily been abruptly reversed. Moreover, the smooth observed trends reflect the fact that regional crises tend to remove many species but few genera. To some degree, high rates of extinction that precede peak rates must represent bias of the incomplete fossil record (the Signor-Lipps effect). High rates that immediately follow peak rates also may, to a degree, reflect biological legacy: (1) final extinction of weakened genera or (2) extinction of new genera that contain few species or represent failed evolutionary “experiments.” Nonetheless, there is evidence that protracted intervals of stressful environmental conditions contributed to high rates of extinction preceding or following certain peak intervals, including the Scythian, Cenomanian, Early Paleocene, and Early Oligocene. The reef-building rudists, for example, suffered heavy extinction during both Cenomanian and Turonian time and then failed to recover quickly.

The late Neogene record of bivalve molluscs in the Western Atlantic offers a more detailed picture of delayed recovery. Here early intervals of glacial expansion caused heavy extinction, leaving an impoverished, eurythermal fauna that was virtually unaffected by late Pleistocene glacial episodes. The episode of heavy extinction in Late Eocene time exhibits a similar phenomenon on a worldwide scale. Among the planktonic foraminifera, warm-adapted stenothermal species died out, and eurythermal forms predominated throughout Oligocene time; restoration of vulnerable, stenothermal species proceeded gradually during the Miocene Epoch. This example of delayed recovery and others like it following earlier global crises may have prevented such crises from following one another in rapid succession, yielding an appearance of periodicity.

The frequent late Cenozoic glacial ages were accompanied by sea-level falls of 100–150 m amplitude. These falls stranded the complex inner-reef platforms and lagoons of tropical Pacific islands, while outer-reef-slope habitats persisted, although displaced downslope. The effects of glacial regressions on bivalves were studied by examining the zonation of species across reef systems and species composition on tectonically uplifted islands, islands with, effectively, local low sea stands. I show that qualitative habitat loss (the stranding of inner-reef habitats) was responsible for the local extinction of about one-third of the bivalve species that inhabit central Pacific islands during high sea stands, whereas quantitative loss of habitat area and climatic effects were inconsequential. Soft-sediment habitats, and consequently soft-bottom bivalves, were more drastically affected by sea-level fluctuations than were hard-bottom habitats and bivalves.

Although many bivalve species were extirpated in the central Pacific during low sea stands, they survived in the western Pacific, where the different geomorphology of many marine systems provided refugia for lagoonal species. Thus, a large proportion of Pacific bivalve species has very dynamic distributions, undergoing great range reductions and expansions with falls and rises of sea level, and much of the present central Pacific lagoonal fauna is of Holocene age. Several implications of these findings are discussed.

The periosteal cortex in the shaft of limb bones is described histologically in three ichthyosaurian genera, Omphalosaurus, Stenopterygius, and Ichthyosaurus. The primary periosteal deposits are composed of typical woven-fibered tissue that was accreted as spongy bone in young individuals, and more or less compact bone in older individuals. During growth, the bone tissue was extensively remodeled with a quantitative imbalance between resorption and redeposition. As a result, the cortex was made cancellous, if previously compact, or still more spongy, if already cancellous. This pattern of remodeling explains why compact cortices are generally lacking in the long bones of ichthyosaurs. The presence of woven-fibered tissue strongly suggests that the limb bones, and probably also the body as a whole, had a rapid postnatal growth in ichthyosaurs, that might have been related to a high, “endotherm-like” metabolic rate. This hypothesis bears on the ecological interpretation of the ichthyosaurs: they could have been capable of sustained, fast swimming and long-range movements, rather than being slow-moving creatures as commonly supposed.

The spinal quotient (S.Q.) is an osteologically defined estimate of the enlargement of the spinal cord at limb levels over that at interlimb levels. It is an efficient predictor of limb use in living reptiles and birds and may be used to predict limb function in fossil vertebrates. Among living reptiles, this ratio of limb to interlimb innervation is greatest in arboreal genera, followed by terrestrial sprawlers, aquatic forms, and undulatory forms. Birds show a wide range of brachial S.Q. values that are roughly commensurate with flight ability. S.Q. values for the manipulative forelimbs of some dinosaurs fall well above those of locomotory limbs. Dinosaur hind-limb values are either well within ranges predicted by living reptiles and birds (most taxa), or highly inflated (stegosaurs, sauropods). This inflation may be the result of presence of a glycogen body similar to that of birds. In no case does the lumbosacral S.Q. support the presence of a “sacral brain.”

Entelodonts are medium to large (perhaps 150–750 kg) Oligocene–Miocene bunodont artiodactyls with unique crania but typical artiodactyl postcrania. Functional and ecological interpretations are difficult because there is no clear modern analog to an entelodont in size, dentition, and cranial morphology. Entelodont crania combine primitive and derived features, including laterally expanded zygomatic arch/large temporal fossa (suggesting a large temporalis muscle), unreduced dental formula, long premolar row, fused mandibular symphysis, isognathy, and subcylindrical dentary condyles. Furthermore, the cranium has unique architectural and ontogenetic aspects unparalleled in extant mammals. Canines show heavy, carnivoran-like, apical wear in old individuals, suggesting regular contact with food. Conical premolars and associated diastemata dominate the tooth row. Premolars are often apically worn, in a fashion somewhat like those of carnivorans (e.g., Crocuta, Borophagus). Molars are low-cusped crushing teeth. Wide gape, indicated by the form of the coronoid process and temporal fossa, facilitated canine and premolar use, probably both in feeding and in social behavior. Jaw mechanics, tooth morphology, and tooth wear are compatible with omnivory and probable scavenging, an intriguing proposition for the huge Dinohyus hollandi.

All erect, branching (ramose) organisms adhere to the same fundamental geometric growth law: the rate of distal growth decreases away from the branch axis. Regardless of the phylogenetic history of an organism, the formation of cylindrical branches requires adherence to this law. In colonial ramose organisms such as trees, corals, and bryozoans, this law poses a problem. How do colonies coordinate the growth rates of the individual modules to produce an integrated branch? This question is addressed in the context of three Ordovician and three Devonian species of ramose trepostome bryozoans. Using remnant growing tips in the endozone as isochronous surfaces, relative rates of skeletal secretion among zooids were measured. Measurements of skeletal and void spaces across a colony branch enabled calculation of the volume of skeletal material secreted by zooids between successive remnant growing tips. Results indicate that rate of skeletal secretion systematically decreases from the branch axis outward into the exozone. This suggests that zooid morphogenesis is controlled to a certain degree by the colony. Colonial control over zooidal growth rates in turn regulates the shape of the colony.

The Paratethyan basins of eastern Europe and western Asia became isolated from marine influence in the Late Miocene, and were the sites of several remarkable endemic radiations of brackish and freshwater organisms. Here I describe the patterns of tempo and mode before and during the radiation of the gastropod Melanopsis in the Pannonian basin of eastern and central Europe, and I explore the underlying mechanisms of evolutionary change.

The most ancient melanopsid species in this area, M. impressa, was present in freshwater areas marginal to the basin well before the radiation. Widely spaced samples of M. impressa indicate that this species underwent a period of stasis lasting at least 7 m.y. The end of stasis corresponded with the extinction of the last of the normal marine fauna in the basin, suggesting that the lack of other fauna and/or reduced salinity in the basin permitted expansion of the melanopsids from the basin margins into the basin proper. Stasis ended with the onset of changes in size, shouldering, and ontogeny, which led eventually to M. fossilis. Change occurred over a 2-m.y. interval; a series of intermediates is present for all three characters. Within-sample correlations provide no evidence that the three characters are constructionally linked; instead they appear to be changing independently. The mode of change in the M. impressa–M. fossilis lineage appears to have been anagenetic. Alterations in the rate and direction of selection (and/or genetic links between characters) are probably required to explain the overall slowness of the change.

Most Pannonian basin melanopsid species arose by rapid cladogenesis in the Middle Pannonian Stage. Physical factors in the basin probably influenced the timing of this diversification; contrasting patterns of variation and diversity between two melanopsid clades suggest that intrinsic factors influenced the extent of diversification.

Both fossil preservation and sampling methods affect perceived patterns of biotic diversity. Artificial range truncations, for example, may lead to incongruences between apparent- and actual-diversity curves. Thus, a catastrophic extinction event may appear gradual. Recent advances in biostratigraphic-gap analysis provide models for the distribution of gap lengths between fossil occurrence horizons and provide methods to place confidence intervals on local taxon ranges and remove the biases caused by artificial range truncations. Confidence intervals for a set of local taxon ranges may then be evaluated collectively to test a hypothesis of co-extinction/co-emigration or co-origination/co-immigration. In the case of terminal Cretaceous ammonites from Seymour Island, range-chart data are compatible with an abrupt extinction event, although the test statistic is not minimized at the stratigraphic horizon that was suggested by Macellari (1986).