Multivariate Analyses of Small Theropod Dinosaur Teeth and Implications for Paleoecological Turnover through Time

Isolated small theropod teeth are abundant in vertebrate microfossil assemblages, and are frequently used in studies of species diversity in ancient ecosystems. However, determining the taxonomic affinities of these teeth is problematic due to an absence of associated diagnostic skeletal material. Species such as Dromaeosaurus albertensis, Richardoestesia gilmorei, and Saurornitholestes langstoni are known from skeletal remains that have been recovered exclusively from the Dinosaur Park Formation (Campanian). It is therefore likely that teeth from different formations widely disparate in age or geographic position are not referable to these species. Tooth taxa without any associated skeletal material, such as Paronychodon lacustris and Richardoestesia isosceles, have also been identified from multiple localities of disparate ages throughout the Late Cretaceous. To address this problem, a dataset of measurements of 1183 small theropod teeth (the most specimen-rich theropod tooth dataset ever constructed) from North America ranging in age from Santonian through Maastrichtian were analyzed using multivariate statistical methods: canonical variate analysis, pairwise discriminant function analysis, and multivariate analysis of variance. The results indicate that teeth referred to the same taxon from different formations are often quantitatively distinct. In contrast, isolated teeth found in time equivalent formations are not quantitatively distinguishable from each other. These results support the hypothesis that small theropod taxa, like other dinosaurs in the Late Cretaceous, tend to be exclusive to discrete host formations. The methods outlined have great potential for future studies of isolated teeth worldwide, and may be the most useful non-destructive technique known of extracting the most data possible from isolated and fragmentary specimens. The ability to accurately assess species diversity and turnover through time based on isolated teeth will help illuminate patterns of evolution and extinction in these groups and potentially others in greater detail than has previously been thought possible without more complete skeletal material.

孤立保存的小型兽脚类牙齿在脊椎动物微化石组合中极为常见，且常被用于古生态系统的物种多样性研究。然而，由于缺乏关联的诊断性骨骼材料，确定这些牙齿的分类亲缘关系颇具挑战。诸如阿尔伯塔驰龙（Dromaeosaurus albertensis）、吉尔莫里氏猎龙（Ricardoestesia gilmorei）以及朗斯顿伤齿龙（Saurornitholestes langstoni）等物种，其骨骼残骸仅发现于恐龙公园组（Dinosaur Park Formation）的坎潘期（Campanian）地层中。因此，来自时代或地理位置差异显著的不同地层的牙齿，大概率无法归入上述物种。无关联骨骼材料的牙齿类群，例如湖畔副栉齿龙（Paronychodon lacustris）与等齿里氏猎龙（Ricardoestesia isosceles），也已在晚白垩世多个不同时代的化石产地被鉴定发现。为解决这一难题，研究人员针对北美地区桑托期（Santonian）至马斯特里赫特期（Maastrichtian）的1183枚小型兽脚类牙齿构建了迄今标本量最丰富的兽脚类牙齿数据集，并采用多变量统计方法开展分析：典型变量分析（canonical variate analysis）、两两判别函数分析（pairwise discriminant function analysis）以及多变量方差分析（multivariate analysis of variance）。分析结果显示，来自不同地层的同一分类群牙齿往往在定量特征上存在显著差异；与之相反，产于同时代地层的孤立牙齿则无法通过定量特征加以区分。上述结果支持如下假说：与晚白垩世其他恐龙类群一样，小型兽脚类类群往往专属分布于特定的宿主地层中。本文所述的分析方法在未来全球范围内的孤立牙齿研究中具备巨大应用潜力，或为目前已知的最实用的非破坏性技术，可从孤立且破碎的标本中提取尽可能多的研究数据。基于孤立牙齿准确评估物种多样性及随时间推移的类群更替的能力，将有助于更细致地阐明这些类群乃至其他类群的演化与灭绝模式，这是此前在缺乏更完整骨骼材料的情况下难以实现的。