Stratigraphic data are compared to morphologic data in terms of their fit to phylogenetic hypotheses for 29 data sets taken from the literature. Stratigraphic fit is measured using MacClade's stratigraphic character, which tracks the number of independent discrepancies between observed order and the order of occurrence that would be expected on the basis of a given phylogenetic hypothesis. Acceptance of a phylogenetic hypothesis despite such discrepancies requires ad hoc hypotheses concerning differential probabilities of preservation and recovery. These stratigraphic ad hoc hypotheses are treated as logically equivalent to morphologic ad hoc hypotheses of homoplasy. The retention index is used to compare the number of stratigraphic and morphologic ad hoc hypotheses required by given phylogenetic hypotheses. Each data set is subjected to five analyses, varying in the constraints imposed on the structure of the phylogenetic tree against which fit is measured. Analyses 1–4 compare the stratigraphic and morphologic retention indices using phylogenetic trees consistent with the morphologically most-parsimonious cladogram reported in the original study. Analysis 5 compares retention indices using the overall (stratigraphically and morphologically) most-parsimonious phylogenetic tree, which may be, but is not necessarily, consistent with the reported cladogram. Proceeding from Analysis 1 to Analysis 5, stratigraphic data are allowed greater influence in determining the structure of phylogenetic trees, with the trees in Analysis 1 derived without reference to the stratigraphic character and the trees in Analysis 5 derived from full interaction of stratigraphic and morphologic characters. Morphologic and stratigraphic retention indices for these 29 studies cannot be statistically distinguished in comparisons 3–5, suggesting very similar degrees of fit. The values of these retention indices are high, indicating a generally high level of congruence under these phylogenetic hypotheses. Significant gains (49%) in stratigraphic fit can be realized without significant loss (4%) in morphologic fit as the stratigraphic and morphologic evidence are both allowed to participate in constraining the structure of phylogenetic hypotheses. These results suggest that arguments based on alleged “noisiness” of stratigraphic data offer inadequate grounds for ignoring stratigraphic order in phylogenetic analysis. In terms of congruence, stratigraphic and morphologic data perform about equally well.

Use of quantitative statistical tests can show that there is generally good congruence between estimated cladistic hypotheses of relationship and observed stratigraphy. A data set of 376 cladograms of fishes, continental tetrapods, and echinoderms was tested using three metrics, Spearman Rank Correlation (SRC), Relative Completeness Index (RCI), and Stratigraphic Consistency Index (SCI), to explore the relationships between the indices and differences in results among the three groups of organisms.

There is a strong relationship between SCI and SRC, since both tests measure the same aspect of the fossil record. There is no relationship between RCI and either SCI or SRC. There is a highly significant relationship, as expected, between SRC coefficients and the number of taxa in a cladogram, but no such relationship for RCI or SCI (except in fishes). There is no significant relationship between any of the indices and either the number of internal nodes or tree balance.

Echinoderms show the best stratigraphic consistency of nodes, while continental tetrapods have the best matching of stratigraphic age and cladistic node order. Fishes have the worst match of age and clade ranks, but they do have the most complete fossil record as measured by the RCI. They are followed by echinoderms, and then continental tetrapods, which have the least complete record. This seems to show that life in an aquatic environment leads, in general, to a more complete fossil record.

Two living forms of Globigerinella siphonifera (d'Orbigny), presently identified as Type I and Type II, can easily be distinguished and collected by SCUBA divers because of differences in appearance, arrangement of the rhizopodial network, and the presence or absence of commensals. Additional biological differences are apparent from laboratory culture experiments; Type I individuals survive significantly longer than Type II under conditions of darkness and starvation and have significantly slower chamber formation rates. Stable isotopic analyses of Types I and II also reveal notable differences, with Type I consistently yielding more negative δ18O and δ13C values. Results of Mg/Ca ratio analyses indicate that Type II specimens precipitated their shells in slightly cooler (deeper) surface waters than Type I specimens. These observations and results from DNA sequencing unequivocally demonstrate that G. siphonifera Types I and II should be regarded as biological sister species.

Contrarily, biometric analysis of the empty shells reveals few significant differences between G. siphonifera Types I and II. Of all the features measured from X-ray and SEM images of serially dissected specimens, only shell porosity yields readily discernible differences, with Type I adult chambers averaging 10–20% porosity and Type II adult chambers averaging 4–7% porosity. Statistically significant differences between Type I and II populations are revealed in maximum test diameter (Type I is typically larger) and coiling (Type I is typically more evolute), but these differences do not justify species level distinction of Types I and II using traditional paleontological species concepts.

On the basis of the above evidence, and since all specimens were collected at the same location at ∼3–8 m water depth, we conclude that G. siphonifera Types I and II are living examples of cryptic speciation, whereby biological speciation has occurred in the absence of discernable change in shell morphology. However, it is not clear when or where this speciation took place. Preliminary study of deep-sea cores from the Caribbean and Pacific sides of the Isthmus of Panama reveals a predominance of specimens with Type II porosity values, with rare occurrence of specimens yielding Type I porosity values. Systematic downcore measurement of shell porosity and tightness of coiling needs to be extended back to the middle Miocene, when G. siphonifera first appeared, to determine the timing of the Type I and II morphological divergence.

Postulated mechanisms for reproductive isolation and speciation of Types I and II include alloparapatric, depth parapatric, and sympatric speciation. These models could be tested if further analysis of fossil G. siphonifera shells allows determination of the timing of speciation, the preferred depth distribution, and the history of geographic distribution of Types I and II.

Computer-generated searches through hypothetical fitness landscapes for progressively more fit variants were used to assess the effects of developmental barriers (mimicked by barring specific types of morphological transformations) and abruptly shifting environmental conditions (mimicked by sudden shifts in how fitness was defined) on the number and accessibility of optimal phenotypes. Relative fitness was defined in terms of maximizing light interception, mechanical stability, or reproductive success, or minimizing surface area, or optimizing the performance of various combinations of these tasks. Developmentally obstructed and unobstructed walks located, on average, equivalent numbers of phenotypic optima. The number of optima identified by both kinds of walks increased in proportion to the number of simultaneously performed tasks used to measure fitness. Walks passing from more complex to less complex fitness landscapes located more optima than walks passing through unchanging, stable landscapes. The model thus suggests that there are no a priori reasons to assume that (1) the morphological options available to adaptive evolution become more restrictive as biological complexity increases, (2) “developmental barriers” necessarily restrain a lineage from evolving well-adapted morphologies, and (3) generalist organisms are less successful than specialists. Also, because the number and accessibility of fitness peaks were proportional to the complexity of fitness landscapes, the model predicts that the probability of cladogenesis will increase as landscape complexity increases, while anagenesis will be encouraged when fitness is defined in terms of performing one or a few tasks simultaneously.

The exceedingly rich middle Pleistocene mammalian fauna from the classic Ensenadan Tarija basin in southern Bolivia contains a diversity of medium to large-bodied herbivores consisting of both endemic (†Toxodontia, †Litopterna, Xenarthra) and immigrant (Rodentia, Proboscidea, Perissodactyla, and Artiodactyla) taxa. In order to characterize feeding ecology and niche differences, a suite of morphological characters was measured for each of 13 species of herbivorous mammals from the Pleistocene of Tarija; these were combined with carbon isotopic results from tooth enamel. (The Xenarthra were excluded from this study because they lack tooth enamel.)

Several different bivariate and multivariate combinations of characters can be used to characterize the feeding adaptations, niches, and guild composition of the Tarija mammalian herbivores. During the Pleistocene the browsing guild in the Tarija basin is interpreted to include the tapir (Tapirus tarijensis), extinct llama (Palaeolama weddelli), peccary (Tayassu sp.), and deer (Hippocamelus sp.). The mixed-feeding guild included two horse species (Hippidion principale and Onohippidium devillei), litoptern (Macrauchenia patachonica), and capybara (Neochoerus tarijensis). The grazing guild included the numerically dominant horse (Equus insulatus), two lamine species (Lama angustimaxilla and cf. Vicugna, provicugna), notoungulate (Toxodon platensis), and gomphothere proboscidean (Cuvieronius hyodon). The grazing guild has the widest range of body sizes relative to the two other guilds. Closely related sympatric species within the Equidae and Camelidae differentiate their niches from one another using a combination of body size, feeding ecology, and probably local habitat. Most of the paleoecological reconstructions resulting from this combined morphological and isotopic analysis corroborate previous studies based primarily on morphology; there are, however, some notable surprises.

A cladistic analysis of the three recognized patterns of central nervous visual organization among teleosts reveals that there is a pattern of intermediate complexity representing the plesiomorphic condition for teleosts, and that there is a simple visual pattern in two unrelated teleost groups which can be concluded to be a secondarily reduced derived condition, as well as an elaborate pattern which is present only in acanthomorph teleosts, thus likely representing a synapomorphy for this taxon. The elaborate central nervous visual pattern, therefore, is one of many functional-anatomical advanced features characterizing the acanthomorphs. Furthermore, when neontological and paleontological data is compared with the paleoecological record of early acanthomorph history during the Late Cretaceous, it is consistent with a hypothesis that this acanthomorph synapomorphic functional-anatomical complex arose likely in ctenothrissiforms as an adaptation to the life in the reorganizing reefs of that geologic period.

Morphologic diversification among rostroconch molluscs is explored using multiple phylogenies derived from different methods. In addition to strict parsimony, phylogenetic estimates were derived using four different methods that employ stratigraphic data. The resultant phylogenies are generally very similar to one another and to the phylogeny proposed by Pojeta and Runnegar (1976). All estimates (including the Pojeta and Runnegar estimate) imply much lower morphologic separations among post-Ordovician rostroconchs (measured here as the frequency of character state change per branch) than among Cambro-Ordovician rostroconchs. However, the data do not suggest that morphologic evolution became more constrained among rostroconchs as a whole, but instead suggest a reduced characteristic rate of morphologic change in the clade that happened to survive the end-Ordovician extinction. Likelihood ratio tests provide strongest support for the hypothesis that morphologic evolution was more constrained within a derived subclade (corresponding to previous definitions of the Conocardioidea, minus the Eopteriidae) than it was in a broad paraphylum (corresponding to the Ribeirioidea + Eopteriidae). Estimates from each of the phylogenetic methods lead to the same conclusions. Various metrics indicate that the pattern is not due to poor sampling of Cambro-Ordovician species and thus merits a biological explanation. Nonphylogenetic analyses of morphologic disparity suggest a similar history of morphologic evolution, including the apparent difference in characteristic rates of morphologic evolution between paraclade and subclade.