John Whitlock

Background: Tooth replacement rate can be calculated in extinct animals by counting incremental lines of deposition in tooth dentin. Calculating this rate in several taxa allows for the study of the evolution of tooth replacement rate. Sauropod dinosaurs, the largest terrestrial animals that ever evolved, exhibited a diversity of tooth sizes and shapes, but little is known about their tooth replacement rates.
Methodology/Principal Findings: We present tooth replacement rate, formation time, crown volume, total dentition volume, and enamel thickness for two coexisting but distantly related and morphologically disparate sauropod dinosaurs Camarasaurus and Diplodocus. Individual tooth formation time was determined by counting daily incremental lines in dentin. Tooth replacement rate is calculated as the difference between the number of days recorded in successive replacement teeth. Each tooth family in Camarasaurus has a maximum of three replacement teeth, whereas each Diplodocus tooth family has up to five. Tooth formation times are about 1.7 times longer in Camarasaurus than in Diplodocus (315 vs. 185 days). Average tooth replacement rate in Camarasaurus is about one tooth every 62 days versus about one tooth every 35 days in Diplodocus. Despite slower tooth replacement rates in Camarasaurus, the volumetric rate of Camarasaurus tooth replacement is 10 times faster than in Diplodocus because of its substantially greater tooth volumes. A novel method to estimate replacement rate was developed and applied to several other sauropodomorphs that we were not able to thin section.
Conclusions/Significance: Differences in tooth replacement rate among sauropodomorphs likely reflect disparate feeding strategies and/or food choices, which would have facilitated the coexistence of these gigantic herbivores in one ecosystem. Early neosauropods are characterized by high tooth replacement rates (despite their large tooth size), and derived titanosaurs and diplodocoids independently evolved the highest known tooth replacement rates among archosaurs.

Tooth replacement is a common trait to most vertebrates, including mammals. Mammals, however, have lost the capacity for continuous tooth renewal seen in most other vertebrates, and typically have only 1–2 generations of teeth. Here, we review the mechanisms of tooth replacement in reptiles and mammals, and discuss in detail the current and historical theories on control of timing and pattern of tooth replacement and development.

Background As gigantic herbivores, sauropod dinosaurs were among the most important members of Mesozoic communities. Understanding their ecology is fundamental to developing a complete picture of Jurassic and Cretaceous food webs. One group of sauropods in particular, Diplodocoidea, has long been a source of debate with regard to what and how they ate.

The rise of computational methods and widespread availability of phylogenetic software packages has resulted in a dramatic increase in the number of phylogenetic analyses published each year. We anticipate that as the number of workers and the number of analyses continues to increase, interest in so-called 'meta-analyses’ that compare not just the resultant trees, but the data themselves, will also increase. Here, we demonstrate a simple method for authors to facilitate such comparisons by visually representing the character data included in their analyses, the Character Distribution Map (CDM).

Sauropod dinosaur bones are common in Mesozoic terrestrial sediments, but sauropod skulls are exceedingly rare—cranial materials are known for less than one third of sauropod genera and even fewer are known from complete skulls. Here we describe the first complete sauropod skull from the Cretaceous of the Americas, Abydosaurus mcintoshi, n. gen., n. sp., known from 104.46±0.95 Ma (megannum) sediments from Dinosaur National Monument, USA. Abydosaurus shares close ancestry with Brachiosaurus, which appeared in the fossil record ca. 45 million years earlier and had substantially broader teeth. A survey of tooth shape in sauropodomorphs demonstrates that sauropods evolved broad crowns during the Early Jurassic but did not evolve narrow crowns until the Late Jurassic, when they occupied their greatest range of crown breadths. During the Cretaceous, brachiosaurids and other lineages independently underwent a marked diminution in tooth breadth, and before the latest Cretaceous broad-crowned sauropods were extinct on all continental landmasses. Differential survival and diversification of narrow-crowned sauropods in the Late Cretaceous appears to be a directed trend that was not correlated with changes in plant diversity or abundance, but may signal a shift towards elevated tooth replacement rates and high-wear dentition. Sauropods lacked many of the complex herbivorous adaptations present within contemporaneous ornithischian herbivores, such as beaks, cheeks, kinesis, and heterodonty. The spartan design of sauropod skulls may be related to their remarkably small size—sauropod skulls account for only 1/200th of total body volume compared to 1/30th body volume in ornithopod dinosaurs.Electronic supplementary materialThe online version of this article (doi:10.1007/s00114-010-0650-6) contains supplementary material, which is available to authorized users.

The unusual skulls of Diplodocus and its relatives, with their protracted, blunt snouts and anteriorly restricted dentition have spawned a number of theories but no consensus regarding the method by which these megaherbivores gathered food. Recent work has shown a considerable variation in snout shape throughout the group, ranging from the square-snouted Nigersaurus to the more gently rounded Dicraeosaurus. These differences in snout shape are coupled with variations in dental microwear, suggesting that snout shape is related to diet. A broad snout shape and microwear patterns consistent with high dietary grit indicate that diplodocoids like Diplodocus and Nigersaurus were low-level browsers. Others, like Dicraeosaurus, were more likely to have been selective low-to mid-height browsers, based on their comparatively pointed snout and low-grit content diet. Jaw elements referred to Tornieria indicate that it was also a selective browser. In order to study the evolution of herbivory of diplodocoids, a new phylogenetic analysis of the group was conducted. This analysis attempts to resolve the interrelationships of the group by creating a new character-taxon matrix generated by first hand observation of specimens. Most interesting are the relationships of the highly specialized rebbachisaurids, the most derived members of which can be divided into two clades: a South American clade including Limaysaurus and Cathartesaura; and a more widely distributed clade including the South American Zapalasaurus as sister taxon to the African Nigersaurus and an unnamed rebbachisaurid from Europe. Of additional interest is the recovery of Suuwassea as a basal dicraeosaurid, making it the only Laurasian representative of that group. Mapping dietary information onto the best-fit cladogram indicates that feeding behavior was independent of phylogeny. Instead, it appears that herbivorous adaptations were more directly tied to regional ecology. The phenotypic plasticity expressed between taxa is on par with the variation in snout shape recognized during the ontogeny of Diplodocus, and suggests that feeding specializations may have been easier to evolve than previously recognized.

Although relatively common by sauropod dinosaur standards, few diplodocoid skulls have received thorough description in the post-war era. One of the first specimens to benefit from a modern treatment was Carnegie Museum (CM) specimen 11162, a large, nearly complete skull assignable to Apatosaurus louisae. With the recent resurgence in interest in diplodocoid evolution and ecology, this specimen is of great significance: it is the most complete described skull of Apatosaurus, and is a primary source of information regarding the facial skeleton of flagellicaudatans other than Diplodocus and Dicraeosaurus. Here, we revisit this specimen following additional preparation and the recent publication of several descriptions of diplodocoid skulls that have highlighted new characters and regions of interest.
We identify or confirm six size-independent cranial character states that distinguish Apatosaurus from Diplodocus. We confirm that Apatosaurus lacks a basipterygoid recess, and that the basipterygoid processes are autapomorphically flared distally, as previously published. We also confirm the presence of globose basal tubera, as opposed to the sheet-like tubera of Diplodocus; we further observe that the tubera of Apatosaurus face posteriorly, not posterolaterally as in Diplodocus. In lateral view, the squamosal and quadratojugal of Apatosaurus do not appear to closely approximate each other as they do in Diplodocus. Finally, the supraoccipital crest of Apatosaurus is massive, prominent, and flanked by deep nuchal fossae, unlike the relatively low crest and shallow fossae of Diplodocus. We are unable to confirm the absence of a sharp-lipped fossa surrounding the preantorbital fenestra; this fossa has been proposed as an autapomorphy of Diplodocus, but the poor preservation of this region of CM 11162 obscures the condition in Apatosaurus.
In addition, we document the presence in CM 11162 of two character states that were previously identified as autapomorphies of Diplodocus: no contact between the vomer and premaxillae, and pterygoid at least partially medial to ectopterygoid on transverse palatal hook. The presence of these character states in Apatosaurus suggests that they are synapomorphies of a more inclusive diplodocoid clade, probably Diplodocidae."

Diplodocid sauropods are characterized by cranial characters that may be directly related to herbivory: the maxilla and premaxilla are elongated, with the teeth located far anterior of the quadrate/articular joint, and the quadrate is inclined, such that it lies beneath the orbit ventrally and posterior to the orbit dorsally, bringing the jaw joint itself forward. Preliminary investigation of the morphology of juvenile Diplodocus indicates that younger individuals do not completely share this morphology.
The present study represents an empirical approach to quantifying ontogenetic change in the skull of Diplodocus through geometric morphometrics. Individual skulls of varying size were reconstructed in lateral, dorsal, and anterior view. Landmark and semi- landmark based morphometric analyses on these reconstructions indicate allometric growth, particularly in the facial regions. Of particular significance are the changes that occur in the tooth-bearing regions. Such changes include the relative repositioning of the last tooth, which moves anteriorly relative to the rest of the face with increasing skull size. The anterior margin of the skull, formed by the premaxillary and maxillary bones, broadens, forming the truncate, flattened shape of the adult. Little relative change occurs in the region of the braincase. The observed changes represent a shift in shape from a more conventional oral morphology to the divergent shape typified by Diplodocus adults. Moreover, the largest excursions occur between the large sub-adult and adult specimens. This may be concurrent with the plateau observed at the upper range of growth rates, but a lack of definite age control hinders comparisons of growth rate and skull morphology. The nature and timing of these changes, therefore, have important consequences for interpretations of the life history and feeding mechanics of Diplodocus.

Diplodocoid sauropods are notable not only for their large average size but also for their remarkable skulls, the use of which in gathering food has long been a source of confusion. Particularly noteworthy is the shape of the snout, which is commonly described as broad and blunted. Snout shape has been correlated with feeding behavior in mammals, both living and extinct. Blunt snouts are most often found in grazers and low-croppers, whereas rounded or pointed snouts are associated with more selective browsers: in some instances relatively small absolute differences in shape correlate with highly distinct diets. This relationship may have important consequences for determining the habits of diplodocoid sauropods. Prior work has shown that snout shape differs between adults and juveniles within at least one genus, Diplodocus; it also appears likely that this is true for other diplodocoid taxa as well. Snout shape also varies between taxa, both drastically, as in the disparate shapes of Nigersaurus and Diplodocus, as well as subtly, such as between Diplodocus, Apatosaurus and Dicraeosaurus. Snout shape can be quantified by a metric that is here termed the premaxilla-maxilla index (PMI). Comparisons of PMI and enamel microwear (an indicator of diet) in a sample of diplodocoids allow for the relationship of shape and diet to be tested. Adult skulls with blunt snouts (e.g. Diplodocus) typically have teeth with unidirectional scratch patterns and increased pitting, both indicators of low-height grazing behavior. Preliminary investigations of skulls with rounded snouts indicate an association with less regular wear patterns, which may be related to browsing rather than grazing behavior. Shape and wear data from some diplodocoid and dicraeosaurid sauropods are examined in the context of an ongoing phylogenetic analysis, building towards the development of a comprehensive hypothesis of the evolution of variability in feeding behavior in diplodocoid sauropods.

The diplodocoid sauropods have been defined in large part by their blunt, shovel-snouted skulls and their narrow-crowned, anteriorly sequestered dentition. This arrangement of the snout and teeth has also been the focus of some controversy, and hypotheses abound to explain the function of this anatomy for feeding. Here we describe a nearly complete juvenile skull of Diplodocus (CM 11255) which does not conform to the morphology expected based upon larger specimens. Although somewhat laterally crushed, the skull is still largely intact, and CT imaging has revealed a nearly complete palate, allowing an accurate reconstruction of the lateral dimensions of the skull. In dorsal view, the reconstructed CM 11255 is narrower than would be expected and quite rounded anteriorly, in contrast to the characteristic squared-off snout of adult Diplodocus. The dentition in CM 11255 is also not restricted to the extreme anterior portion of the skull, and in fact extends posteriorly to near the preantorbital fenestra. A rounded snout in dorsal view is more typical of basal sauropods, and ontogenetic change towards a square snout is potentially related to feeding behavior and specialization in adults. Work comparing snout shapes and dental microwear features between different diplodocoid sauropod genera (including Diplodocus) suggests that snout shape is closely linked to feeding strategy. Although no microwear features have been recovered from CM 11255, the shape change suggests ontogenetic niche partitioning between adults and juveniles is suggested—smaller juveniles accessed nutrients through highly selective browsing behaviors, and large adults obtained nutrition through less selective, bulk browsing behaviors.

Incremental lines of von Ebner are microstructural features that demarcate the daily apposition of dentin, such that the total number of incremental lines in a tooth serves as a measure of that tooth’s age. Differences in incremental line counts between successive replacement teeth within an alveolus provide a quantitative estimate of tooth replacement rate. Tooth formation times and rates have been shown to vary substantially among amniotes. We examined tooth replacement rates in two contemporaneous sauropod dinosaurs from western North America: Diplodocus and Camarasaurus. CT data reveal that each alveolus in Diplodocus has up to 5 replacement teeth, whereas Camarasaurus has a maximum of 3. Incremental lines of von Ebner are on average ~14 microns apart in Diplodocus and ~23 microns apart in Camarasaurus. Tooth formation times are about twice as long in Camarasaurus as in Diplodocus. Average tooth replacement rate in Diplodocus is ~35 days, similar to the replacement time of ~30 days found in Nigersaurus, which possessed a tooth battery. Tooth replacement rates in diplodocoid sauropods appear to be faster than those in any other dinosaurs, including derived ornithischian herbivores with tooth batteries. Preliminary data for Camarasaurus indicate a much slower rate of replacement. In the absence of incremental line counts from thin-sections in other sauropods, successive replacement tooth size was used as a proxy for replacement rate (i.e., a larger size discrepancy between successive replacement teeth indicates a slower replacement rate). Basal sauropods have fewer (1–2) replacement teeth per alveolus, and successive replacement teeth show a larger size difference than in Camarasaurus. In derived titanosaurs, which had narrow-crowned teeth, there are a high number of teeth per alveolus and small size discrepancies between successive replacement teeth. Our data suggest that neosauropods are characterized by faster tooth replacement rates than basal sauropods, and that derived titanosaurs and diplodocoids independently evolved very fast tooth replacement rates.

Sauropods evolved from small, bipedal omnivores (e.g., Panphagia). Over time these basal sauropodomorphs became more reliant on herbivory and achieved larger body sizes, eventually becoming obligate quadrupeds. With this increase in body size and dietary shift came an increase in tooth wear, which reduced dental effectiveness. Sauropods responded to increased tooth wear by increasing crown volume and/or increasing tooth replacement rate. Here, we present data describing both responses and place them in a temporal and phylogenetic context. Replacement rate data are based on histological study when possible. In all other cases, rate is approximated using a cross-checked proxy: the height corrected ratio (HCR), calculated by dividing the height ratio of successive teeth in a family by mature tooth height. Higher HCR scores indicate less time between tooth formation events and correspond to faster replacement rates. Basal sauropodomorphs (e.g., Plateosaurus) had low-crowned, leaf-shaped teeth and one replacement tooth per family. Spatulate tooth crowns with greatly increased volumes (~5x that of Plateosaurus) evolved in basal sauropods (e.g., Tazoudasaurus), which enabled these taxa to cope with more extensive tooth wear and resulted in slower replacement (HCR ~2.0). Replacement rate first increased in more derived eusauropod taxa on the stem leading to Neosauropoda (e.g., Mamenchisaurus). These taxa possessed up to two replacement teeth per family and higher HCRs (~4.0). A second increase in replacement rate occurred in Late Jurassic neosauropods. These taxa (e.g., Camarasaurus, Diplodocus) are observed to have three or more replacement teeth per family, which were replaced at least every 62 days. Within Neosauropoda, at least two secondary reductions in crown volume occurred, in the clades Diplodocoidea and Titanosauriformes. This reduction in volume (~14% that of spatulate crowns) was coupled with an increase in replacement rate in both instances (HCR 12.4+). Narrow-crowned taxa often have five or more replacement teeth per family; in diplodocoids (e.g., Diplodocus), replacement occurred twice as fast (30–35 days) as in spatulate-crowned neosauropods.

Character data in a phylogenetic analysis are often seen as being trapped in a ‘black box’, hidden from outside observers. In addition to the obvious difficulties this poses for independent assessment of the results of an analysis, important patterns within the data may go unnoticed without further examination. Here, we outline a simple graphical method for displaying regional character information. These images provide a convenient visual shorthand that allows a reader to understand at a glance the proportional representation of various anatomical regions in a matrix. This information can be used to identify regions of the body that are phylogenetically important, to identify areas of data paucity, and as a companion to comparative cladistic methods as a way to compare data between different analyses. In addition, these methods can be adapted to display missing data by body region; in tandem, these images provide a simple way to examine the influence of missing data on phylogenetically problematic taxa. Applying these methods to phylogenetic analyses of Sauropoda and its subgroups reveals interesting trends. Large-scale analyses examining the evolution of Sauropoda as a whole are represented subequally by characters from each of the three main body regions (cranial, axial, and appendicular). This is not surprising, as the broad taxonomic scope of these analyses covers multiple alterations to the skull, changes in the complexity of neural spines and internal pneumaticity of vertebral centra, and important locomotor transitions. Finer scale analyses, such as those of the subgroup Diplodocoidea, reduce the importance of certain body regions, such as the appendicular skeleton, and highlight others, such as the dermal skull (20% of all characters). The relative importance of cranial characters is suggested to be a function of innovations focused on the cranial region over the limbs and girdles. Curiously, the cranial region is also the region with the most missing data (41% cells unscorable), which may have been a contributor to a lack of phylogenetic resolution in the past.