New mid-Cretaceous stable isotope (δ18O and δ13C) records of multiple planktonic foraminiferal species and coexisting coccoliths from Blake Nose (western North Atlantic) document a major depth-ecology reorganization of planktonic foraminifera. Across the Albian/Cenomanian boundary, deep-dwelling Praeglobotruncana stephani and Rotalipora globotruncanoides adapted to living at a shallower depth, while, at the same time, the population of surface-dwelling Paracostellagerina libyca declined. Subsequently, the opportunistic species Hedbergella delrioensis shifted to a deep environment, and the deep-dwelling forms Rotalipora montsalvensis and Rotalipora reicheli first appeared. The primary paleoenvironmental cause of the observed changes in planktonic adaptive strategies is uncertain, yet their coincidence with an earliest Cenomanian cooling trend reported elsewhere implicates the importance of reduced upper-ocean stratification. Although there has been an implicit assumption that the species-specific depth habitats of fossil planktonic foraminifera were invariant through time, planktonic paleoecology is a potential variable. Accordingly, the possibility of evolutionary changes in planktonic foraminiferal depth ecology should be a primary consideration (along with other environmental parameters) in paleoceanographic interpretations of foraminiferal stable isotope data.

In considering the history of biodiversity paleontologists have focused on exploratory investigations of empirical data derived from the fossil record. Starting with the pioneering work of Philips (1860), and continuing at an increasing pace through today, this inductive approach has dominated diversity research. In contrast, deductive theoretical considerations that focus on the expected history of biodiversity, and develop independently of empirical knowledge, have remained under-explored. Appreciating the need for a nomothetic paleobiology (Gould 1980), we here reconsider the history of biodiversity, using deductive models constrained by a few, self-evident parameters. This analysis centers on the marine fossil record, the primary target of most previous empirical studies on the geological history of global biodiversity (e.g., Valentine 1969; Raup 1972, 1976; Sepkoski et al. 1981; Alroy et al. 2008).

The most abundant and diverse graptolite assemblages are found in offshore, deep-water black shales—the classical “graptolite fades” (deep-water or isograptid biofacies). The mean duration of Ordovician graptolite species confined to the deep-water facies (here referred to as “group 1” species) is 2.19 Myr, significantly shorter than the mean duration of species in the deep-water facies that are also known in sediments of the shallow-water shelf or platform (“group 2” species) −4.42 Myr, indicating a significantly higher extinction probability (p = <0.001). These figures are based on the precise age ranges of species derived from the time-calibrated composite sequence of 1446 Ordovician to early Devonian graptolites, built by the constrained optimization procedure (CONOP) from 256 measured sections worldwide, and exclude the effects of the Hirnantian mass extinction. The difference between groups cuts across families, morphological types, and pandemic/endemic distributions. An environmental influence is strongly suggested, and although both groups were planktonic, they were unlikely to have shared the same habitat in the water column. The new duration measurements therefore are interpreted as favoring a depth-stratification of graptolite habitats in the water column.

When assessing the timing of branching events in a phylogeny, the most important tools currently recognized are a reliable molecular phylogeny and a continuous, relatively complete fossil record. Coralline algae (Rhodophyta, Corallinales, and Sporolithales) constitute an ideal group for this endeavor because of their excellent fossil record and their consistent phylogenetic reconstructions. We present the evolutionary history of the corallines following a novel, combined approach using their fossil record, molecular phylogeny (based on the 18S rDNA gene sequences of 39 coralline species), and molecular clocks. The order of appearance of the major monophyletic taxa of corallines in the fossil record perfectly matches the sequence of branching events in the phylogeny. We were able to demonstrate the robustness of the node ages in the phylogeny based on molecular clocks by performing an analysis of confidence intervals and maximum temporal ranges of three monophyletic groups of corallines (the families Sporolithaceae and Hapalidiaceae, as well as the subfamily Lithophylloideae). The results demonstrate that their first occurrences are close to their observed appearances, a clear indicator of a very complete stratigraphic record. These chronological data are used to confidently constrain the ages of the remaining branching events in the phylogeny using molecular clocks.

In a seminal paper in 1975, Gould proposed that postcanine occlusal area (PCOA) should scale metabolically (0.75) with body mass across mammals. By regressing PCOA against skull length in a small sample of large-bodied herbivorous mammals, Gould provided some marginal support for this hypothesis, which he then extrapolated as a universal scaling law for Mammalia. Since then, many studies have sought to confirm this scaling relationship within a single order and have found equivocal support for Gould's assertion. In part, this may be related to the use of proxies for both PCOA and body mass, small sample sizes, or the influence of a “taxon-level effect,” rendering Gould's scaling “universal” problematic.

Our goal was to test the universality of Gould's prediction and the impact of the taxon-level effect on regressions of tooth size on body mass in a large extant mammalian sample (683 species spanning 14 orders). We tested for the presence of two types of taxon-level effect that may influence the acceptance or rejection of hypothesized scaling coefficients. The hypotheses of both metabolic and isometric scaling can be rejected in Mammalia, but not in all sub-groups therein. The level of data aggregation also influences the interpretation of the scaling relationship. Because the scaling relationship of tooth size to body mass is highly dependent on both the taxonomic level of analysis and the mathematical methods used to organize the data, paleontologists attempting to retrodict body mass from fossilized dental remains must be aware of the effect that sample composition may have on their results.

In the modern world, biotic diversity is typically higher in low-latitude tropical regions where there is abundant insolation (light and heat) and low thermal seasonality. Because these factors broadly covary with latitude, separating their possible effects on species diversity is difficult. The Eocene was a much more equable world, however, with low temperature seasonality extending into lower-insolation higher, cooler latitudes, allowing us to test these factors by comparing insect species diversity in (1) modern, temperate, low-insolation, highly seasonal Harvard Forest, Massachusetts, U.S.A., 42°29'N; (2) modern, tropical, high-insolation, low-seasonality La Selva, Costa Rica, 10°26'N, and; (3) Eocene, temperate, low-insolation, yet low-seasonality McAbee, British Columbia, Canada, above 50°N paleolatitude. We found insect diversity at McAbee to be more similar to La Selva than to Harvard Forest, with high species richness of most groups and decreased diversity of ichneumon wasps, indicating that seasonality is key to the latitudinal diversity gradient. Further, midlatitude Eocene woody dicot diversities at McAbee, Republic (Washington, U.S.A.), and Laguna del Hunco (Argentina) are also high, similar to modern tropical samples, higher than at the modern midlatitude Harvard Forest. Modern correlations between latitude, species diversity, and seasonal climates were established some time after the Eocene.

The Corbulidae are one of a handful of a primarily marine bivalve clades that exhibit a remarkable radiation, marked by increased species richness and divergent morphologies, within a long-lived lake. For corbulids, this diversification occurred within the lower to middle Miocene Pebas Formation of western Amazonia. Only one taxon associated with this radiation (Anticorbula) remains extant. We conducted a series of phylogenetic analyses to characterize diversification of Corbulidae within the Pebas Formation and relate that diversification to geologically older freshwater corbulids from the Paleocene Fort Union Formation of the northern Great Plains (United States). We used these results, as well as a quantitative examination of morphospace occupation, to infer whether Pebasian corbulids represent a true species flock, and whether the lacustrine system represented by the Pebas Formation represents a cradle of, or reservoir for, freshwater corbulid diversity. We conducted two sets of phylogenetic analyses using shell morphology characters. A genus-level data set incorporated type species of freshwater corbulid genera, any Paleocene representatives of these genera, and selected brackish and marine corbulid genera. A species-level analysis added all described freshwater corbulid taxa to the genus-level matrix. Our results were highly resolved (few most-parsimonious trees), but not particularly robust (low branch support). For the genus-level matrix, we used a taxon jackknife procedure to explore the effects of taxon sampling on tree stability and topology. Jackknife results recover a subclade of freshwater taxa (including both Anticorbula and Pachydon species and the Paleocene Ostomya sp.) in 92.4% of trees, although placement of this subclade across the ingroup varies, as do the topologic positions of other freshwater species. Freshwater and marine corbulids also are morphologically distinct from each other, a factor that likely reduced the robustness of our phylogenetic results. By combining these results with paleoecologic, stratigraphic, and morphologic data, we infer that freshwater corbulids arose once within the family, prior to the Cenozoic, with three distinct freshwater lineages present at their first appearance in the late Paleocene of North America. Within the Miocene Pebas system of South America, we reconstruct supralimital morphologic evolution within three lineages as freshwater taxa became variously adapted to the fluid, dysoxic muds characterizing lake-bottom facies representative of the Pebas lacustrine system. In addition, corbulids apparently successfully coped with high predation pressures from co-occurring shell-crushing predators. Finally, we consider that freshwater Corbulidae were primarily fluvial taxa throughout their geologic history, with a relatively ephemeral radiation within the Pebasian lake system, thus making the Pebasian system a cradle of diversity for several corbulid lineages.

This paper presents a method for constraining the age of a clade with the ages of the earliest fossil specimens in that clade's outgroups. Given a sufficiently deep, robust, well-resolved, and stratigraphically consistent cladogram, this method can yield useful age constraints even in the absence of specific information about the fossil preservation and recovery rates of individual taxa. The algorithm is applied to simulated data sets to demonstrate that this method can yield robust constraints of clade ages if there are sufficient fossil outgroups available and if there is a finite chance that additional outgroups may be discovered in the future. Finally, the technique is applied to actual fossil data to explore the origin of modern placental mammals. Using data from recently published cladograms, this method indicates that if all Mesozoic eutherians are regarded as outgroups of Placentalia, then the last common ancestor of modern placental mammals and their Cenozoic allies lived between 65 and 88–98 million years ago, depending on the assumed cladogram and the number of outgroups included in the analysis.

Fundamental to the interpretation of bone-bearing faunal deposits is an understanding of the taphonomic processes that have modified the once living fossil community. An often neglected source of bias is that of climate-averaging, which occurs when the duration of bone accumulation exceeds the duration of an individual climatic episode. Tropical and subtropical climate change is dominated by precessional cyclicity (~21,000 year cycle), which controls monsoon rainfall intensity and thus plant communities over time. Under a climate-averaging scenario, the paleoecological characteristics of a faunal deposit represent an amalgamation of more than one phase of the precessional cycle. We investigate the degree of climate-averaging in Plio-Pleistocene bone breccias from South Africa by comparing stable isotope measurements of fossil enamel with the evidence from high-resolution speleothem paleoclimate proxies. We conclude that each of the four faunal assemblages studied are climate-averaged, having formed over a time period in excess of one-third of a precessional cycle (~7000 years). This has implications for the reconstruction of hominin paleoenvironments and estimates of Plio-Pleistocene biodiversity. We hypothesize that climate-averaging may be a common feature of tropical terrestrial vertebrate assemblages throughout the Cenozoic and Mesozoic.

Variations in the orientation and cross-sectional shape of filamentous microfossils provide quantitative measures for characterizing them and probing their native mechanical structure. Here, we determine the tangent correlation length, which is the characteristic length scale for the variation in direction of a sinuous curve, for both a suite of Precambrian filamentous microfossils and six strains of modern filamentous cyanobacteria, all with diameters of a few microns. Among 1.9-2-Ga microfossils, Gunflintia grandis, Gunflintia minuta and Eomycetopsis filiformis possess, respectively, correlation lengths of 360 ± 40 μm, 670 ± 40 μm and 700 ±100 μm in two dimensions. Hundreds of times larger than the filament diameters, these values lie in the same range as the cyanobacteria Geitlerinema and Pseudanabaena, but are smaller than several strains of Oscillatoria. In contrast, the 2-Ga microfossil trichome Halythrix, is found to have a short correlation length of 29 ± 4 μm in two dimensions. Micron-wide pyritic replacement filaments observed in 3.23-Ga volcanogenic deposits also display a modest correlation length of 100 ± 15 μm in two dimensions. Sequences of species in two genera of our modern cyanobacteria possess tangent correlation lengths that rise as a power of the filament diameter D—D3.3 ± 1 for Oscillatoria and D5.1 ± 1 for Geitlerinema. These results can be compared with power-law scaling of D3 for hollow tubes and D4 for solid cylinders that is expected from continuum mechanics. Extrapolating the observed scaling behavior to smaller filament diameters, the measured correlation length of the pyrite filaments is consistent with modern Geitlerinema whereas that of Halythrix lies not far from modern Oscillatoria, suggesting that there may be structural similarities among these genera.

We analyze relationships among a range of ecological and biological traits—geographic range size, body size, life mode, larval type, and feeding type—in order to identify those traits that are associated significantly with species duration in New Zealand Cenozoic marine molluscs, during a time of background extinction. Using log-linear modeling, we find that bivalves have only a small number of simple, two-way associations between the studied traits and duration. In contrast, gastropods display more complex interactions involving three-way associations between traits, a pattern that suggests greater macroecological complexity of gastropods. This is not an artifact caused by the larger number of gastropods than bivalves in our data set. We used stratified randomized resampling of families to test for associations between traits that might result from shared inheritance rather than ecological trait interactions; we found no evidence of phylogenetic effects in any associations examined. The relationships revealed by our study should serve to constrain the range of possible biological mechanisms that underlie these relationships. As previously observed, two-way associations are present between large geographic range and increased duration, and between large geographic range and large body size, in both bivalves and gastropods. In gastropods, planktotrophic larval type is associated with large range size through a three-way interaction that also involves duration; there is no direct association of larval type and geographic range. Gastropods also display two-way associations between duration and life mode, and duration and feeding type. We note that in gastropods, an infaunal life mode is associated with large range size, whereas in bivalves infaunality is associated with reduced range size.

Temporal patterns in species occupancy and geographic range size are a major topic in evolutionary ecology research. Here we investigate these patterns in Pliocene to Recent large mammal species and genera in Western Eurasia. By using an extensively sampled fossil record including some 700 fossil localities, we found occupancy and range size trajectories over time to be predominantly peaked among both species and genera, meaning that occupancy and range size reached their maxima midway along taxon existence. These metrics are strongly correlated with each other and to body size, after phylogeny is accounted for by using two different phylogenetic topologies for both species and genera. Phylogenetic signal is strong in body size, and weaker but significant in both occupancy and range size mean values among genera, indicating that these variables are heritable. The intensity of phylogenetic signal is much weaker and often not significant at the species level. This suggests that within genera, occupancy and range size are somewhat variable. However, sister taxa inherit geographic position (the center of their geographic distribution). Taken together, the latter two results indicate that sister species occupy similar positions on the earth's surface, and that the expansion of the geographic range during the existence of a given genus is driven by range expansion of one or more of the species it includes, rather than simply being the summation of these species ranges.

Species arise and establish themselves over the geologic time scale. This process is manifested as a change in the relative frequency of occurrences of a given species in the global pool of species. Our main goal here is to model this rise and the eventual decline of microfossil species using a mixed-effects model where groups each have a characteristic occurrence trajectory (main effects) and each species belonging to those groups is allowed to deviate from the given group trajectory (random effects). Our model can be described as a “hat” with logistic forms in the periods of increase and decline. Using the estimated timings of rises and falls, we find that the lengths of the periods of rise are about as long as the lengths of the periods when species are above 50% of their estimated maximal occurrences. These latter periods are here termed periods of dominance, which are in turn about the same length as the species' periods of fall. The peak rates of the rises of microfossils are in general faster than their peak rates of falls. These quantified observations may have broad macroevolutionary and macroecological implications. Further, we hypothesize that species that have experienced and survived high levels of environmental volatility (specifically, periods of greater than average variation in temperature and productivity) during their formative periods should have longer periods of dominance. This is because subsequent environmental variations should not drive them to decline with ease. We find that higher estimated environmental volatility early in the life of a species positively correlates with lengths of periods of dominance, given that a species survives the initial stress of the environmental fluctuations. However, we find no evidence that the steepness of the rise of a species is affected by environmental volatility in the early phases of its life.

Ecological ordination can reveal gradients in the species composition of fossil assemblages that can be correlated with paleoenvironmental gradients. Ordinations of simulated data sets suggest that nonmetric multidimensional scaling (NMDS) generally produces less distorted results than detrended correspondence analysis (DCA). We ordinated 113 brachiopod-dominated samples from the Frasnian (Late Devonian) Brallier, Scherr, and lower Foreknobs Formations of southwest Virginia, which represent a range of siliciclastic marine paleoenvironments. A clear environmental signal in the ordination results was obscured by (apparently) opportunistic species that occurred at high abundance in multiple environments; samples dominated by these species aggregated in ordination space regardless of paleoenvironmental provenance. After the opportunist-dominated samples were removed, NMDS revealed a gradient in species composition that was highly correlated with substrate (grain size); a second, orthogonal gradient likely reflects variation in disturbance intensity or frequency within grain-size regimes. Additional environmental or ecological factors, such as oxygenation, may also be related to the gradients. These two gradients, plus the environmental factors that controlled the occurrence of opportunistic species, explain much of the variation in assemblage composition in the fauna. In general, the composition of fossil assemblages is probably influenced by multiple paleoecological and paleoenvironmental factors, but many of these can be decomposed and analyzed.

Key morphological traits reveal changes in functional morphospace occupation of reef fish assemblages over time. We used measurements of key functional attributes (i.e., lower jaw length and orbit diameter) of 208 fossil fish species from five geological periods to create bivariate plots of functional morphological traits through time. These plots were used to examine possible function and ecological characteristics of fossil reef fish assemblages throughout the Mesozoic and Cenozoic. A previously unknown trend of increasing orbit diameter over time became apparent. The Teleostei are the principal drivers of this change. The Eocene appears to mark a dramatic increase in two previously rare feeding modes in fishes: nocturnal feeding and high-precision benthic feeding. Interestingly, members of the Pycnodontiformes had relatively large eyes since the Triassic and appear to be the ecological precursors of their later teleost counterparts and may have been among the earliest nocturnal feeding fishes. Our results highlight potential changes in the roles of fishes on coral reefs through time.

Did organisms diversify in different ways on land and in the marine realm over the Phanerozoic, or do the different diversification curves of continental and marine organisms reflect primarily methodological artifacts? To answer this question, a quantitative assessment of the completeness of the global continental fossil record is indispensable. We used comparisons between continental and marine fossil diversity and between past and present-day patterns of continental diversity to assess the absolute and relative completeness of the continental fossil record. Collector's curves of the number of described families over the past 200 years suggest that the global continental fossil record, and even that of European and North American tetrapods, is still highly incomplete. Nevertheless, relative proportions of major continental and marine taxa, patterns of tetrapod endemism, and familial durations suggest that the family-level continental fossil record is reasonably representative. We found that, although continental fossil richness is correlated with the amount of terrestrial clastic sediment available for sampling, the exponential diversification curve of continental metazoans is unlikely to be an artifact of this rock bias. Diversification of the continental fauna appears to have been essentially exponential since the Devonian, with little evidence of major extinction events.

The variation in time-averaging between different types of marine skeletal accumulations within a depositional system is not well understood. Here we provide quantitative data on the magnitude of time-averaging and the age structure of the sub-fossil record of two species with divergent physical and ecological characteristics, the brachiopod Bouchardia rosea and the bivalve Semele casali. Material was collected from two sites on a mixed carbonate-siliciclastic shelf off the coast of Brazil where both species are dominant components of the local fauna.

Individual shells (n = 178) were dated using amino acid racemization (aspartic acid) calibrated with 24 AMS radiocarbon dates. Shell ages range from modern to 8118 years b.p. for brachiopods, and modern to 4437 years for bivalves. Significant differences in the shape and central tendency of age-frequency distributions are apparent between each sample. Such differences in time-averaging magnitude confirm the assumption that taphonomic processes are subject to stochastic variation at all spatial and temporal scales. Despite these differences, each sample is temporally incomplete at centennial resolution and three of the four samples have similar right-skewed age-frequency distributions. Simulations of temporal completeness indicate that samples of both species from the shallow site are consistent with a more strongly right-skewed and less-complete age-frequency distribution than those from the deep site.

We conclude that intrinsic characteristics of each species exert less control on the time-averaging signature of these samples than do extrinsic factors such as variation in rates of sedimentation and taphonomic destruction. This suggests that brachiopod-dominated and bivalve-dominated shell accumulations may be more similar in temporal resolution than previously thought, and that the temporal resolution of multi-taxic shell accumulations may depend more on site-to-site differences than on the intrinsic properties of the constituent organisms.

A growing body of work has quantitatively linked many macroevolutionary patterns, including short- and long-term changes in biodiversity, rates of taxonomic extinction and origination, and patterns of extinction selectivity, to temporal variability in the sedimentary rock record. Here we establish a new framework for more rigorously testing alternative hypotheses for these and many other results by documenting the large-scale spatiotemporal intersection of the North American sedimentary rock and fossil records. To do this, we combined 30,387 fossil collections in the spatially explicit Paleobiology Database with a comprehensive macrostratigraphic database consisting of 18,815 sedimentary lithostratigraphic units compiled from 814 geographic regions distributed across the United States and Canada. The geological completeness of paleontological sampling, here defined as the proportion of the available sedimentary rock record that has been documented to have at least one fossil occurrence, irrespective of taxonomy or environment, is measured at four different levels of stratigraphic resolution: (1) lithostratigraphic rock units, (2) hiatus-bound rock packages, (3) regional stratigraphic columns, and (4) sediment coverage area (km2). Mean completeness estimates for 86 Phanerozoic time intervals (approximately stages; median duration 5.3 Myr) range from 0.18 per interval in the case of lithostratigraphic rock units to 0.23 per interval for stratigraphic columns and sediment coverage area. Completeness estimates at all four levels of stratigraphic resolution exhibit similar temporal variation, including a significant long-term increase during the Phanerozoic that is accentuated by an abrupt Campanian–Maastrichtian peak. This Late Cretaceous peak in completeness is approximately five times greater than the least complete Phanerozoic time intervals (Early Cambrian, Early Devonian, late Permian, and Early Cretaceous). Geological completeness in the Cenozoic is, on average, approximately 40% greater than in the Paleozoic. Temporal patterns of geological completeness do not appear to be controlled exclusively by variation in the frequency of subsurface rock units or an increase over time in the proportion of terrestrial rock, but instead may be general features of both the marine and terrestrial fossil records.

Patterns preserved in the fossil record are of the highest importance in addressing questions about long-term evolutionary processes, yet both the description of pattern and its translation into process can be difficult. With respect to gradual phyletic change, we know that randomly generated sequences may exhibit characteristics of a “trend” apparent patterns, therefore, must be interpreted with caution. Furthermore, even when the claim of a gradual trend can be statistically justified, interpretation of the underlying mechanisms may be challenging. Given that we can observe populations changing rapidly over tens or hundreds of years, it is now more difficult to explain instances of geologically gradual (as opposed to punctuated) change.

Here we describe morphologic change in two bivalve lineages from the late Miocene Lake Pannon. We evaluate change according to the model-based methods of Hunt. Both lineages exhibit size increases and shape changes over an interval of nearly 4 million years. Size and two shape variables in the conjungens lineage are best fit by a model of directional evolution; remaining shape variables mostly conform to unbiased random walks. Body-size evolution in the diprosopum lineage is also significantly directional but all shape variables are best fit by the unbiased random walk model; the small number of sampling intervals available for this lineage (n = 6) makes determination of the actual pattern more difficult. Model-fitting results indicate that the parallel trajectories of increasing log shell height over time in the two lineages can be accounted for by an underlying trend shared by both lineages, suggesting that the size increases may be a shared response to the same cause. The pace of phenotypic change, measured as Lynch's Δ, is slower than the neutral expectation for all size and shape traits.

Our examples illustrate well the paradox of gradualism; the sequences exhibit significant directional morphological evolution, but rates of change as measured over the long-term are apparently too slow for directional selection or even drift to be the cause. Viewing long-term phenotypic evolution in terms of populations tracking peaks on adaptive landscapes is useful in this context. Such a view allows for intervals of directional selection (during times of peak movement–resulting in the overall trends we can detect) interspersed with intervals of stasis (during times of peak stability–resulting in overall changes that appear to proceed more slowly than the neutral expectation). The paradox of gradualism thus reduces to (1) peak movements and their drivers, which are not restricted in rate as are population-genetic drivers, and (2) the maintenance of stasis, on which no consensus exists.

We can identify no environmental parameter in the central European Neogene that exhibits consistent change across the interval of gradual morphologic change. It may be that in Lake Pannon the long-term persistence of generally ameliorating conditions (plentiful resources and habitat space, few predators or competitors) resulted in geologically slow but consistent peak shifts, which in turn facilitated size increase and shape change in these lineages.

Both the body fossils and trackways of sauropod dinosaurs indicate that they inhabited a range of inland and coastal environments during their 160-Myr evolutionary history. Quantitative paleoecological analyses of a large data set of sauropod occurrences reveal a statistically significant positive association between non-titanosaurs and coastal environments, and between titanosaurs and inland environments. Similarly, “narrow-gauge” trackways are positively associated with coastal environments and “wide-gauge” trackways are associated with inland environments. The statistical support for these associations suggests that this is a genuine ecological signal: non-titanosaur sauropods preferred coastal environments such as carbonate platforms, whereas titanosaurs preferred inland environments such as fluvio-lacustrine systems. These results remain robust when the data set is time sliced and jackknifed in various ways. When the analyses are repeated using the more inclusive groupings of titanosauriforms and Macronaria, the signal is weakened or lost. These results reinforce the hypothesis that “wide-gauge” trackways were produced by titanosaurs. It is commonly assumed that the trackway and body fossil records will give different results, with the former providing a more reliable guide to the habitats occupied by extinct organisms because footprints are produced during life, whereas carcasses can be transported to different environments prior to burial. However, this view is challenged by our observation that separate body fossil and trackway data sets independently support the same conclusions regarding environmental preferences in sauropod dinosaurs. Similarly, analyzing localities and individuals independently results in the same environmental associations. We demonstrate that conclusions about environmental patterns among fossil taxa can be highly sensitive to an investigator's choices regarding analytical protocols. In particular, decisions regarding the taxonomic groupings used for comparison, the time range represented by the data set, and the criteria used to identify the number of localities can all have a marked effect on conclusions regarding the existence and nature of putative environmental associations. We recommend that large data sets be explored for such associations at a variety of different taxonomic and temporal scales.