Assistant Professor Stony Brook University Stony Brook University Stony Brook, New York, United States
Abstract Body : With their long necks and enormous bodies, sauropods are among the most instantly recognizable dinosaurs, and their unique bauplan rests at the heart of public and academic fascination with the group. There is broad consensus that the absolutely and relatively long necks of sauropods were central to their evolutionary success, providing an energetically efficient mechanism for consuming high volumes of relatively low-quality browse. However, it has also been suggested that the exceptionally long necks of the largest sauropods resulted primarily from a developmental constraint on neck scaling, inherited from their bipedal sauropodomorph ancestors. Under this scenario, a positively allometric relationship between neck length and body size evolved early in sauropodomorph history, prior to the evolution of quadrupedality and gigantism, where it provided a novel selective advantage in reaching higher browse; later sauropods were simply developmentally constrained to maintain this positive scaling relationship, regardless of their feeding behavior. Problematically, previous work on neck scaling has not attempted to evaluate where along the sauropodomorph stem this putatively distinct scaling regime arose, nor demonstrated statistically that sauropodomorphs exhibit a scaling pattern that is distinct from that of other non-avian dinosaurs. Here, I revisit axial scaling in non-avian dinosaurs, using reversible jump Bayesian modeling and phylogenetic analysis of covariance to agnostically discover and statistically evaluate evolutionarily distinct scaling relationships between neck length and body size (approximated by trunk length). Analysis of a preliminary dataset of 53 species of non-avian dinosaurs favors a three-regime model of neck scaling (p< 0.001; 𝚫AIC = 28.47). An isometric scaling regime (b=1.03, 95% confidence interval: [0.89, 1.17]; R2=0.88) describes all non-avian dinosaurs with the exception of a subset of sauropodomorphs, which are characterized by a shift to a strongly positive allometric relationship between neck and trunk lengths (b=1.81, 95% confidence internal: [1.36, 2.27]; R2=0.94). Contrary to previous work, the best-fit model places this shift at the origin of Sauropoda, coincident with the evolution of obligate, column-limbed quadrupedality and an increase in the number of cervical vertebrae to 12 from a plesiomorphic count of 10. It is not currently possible, however, to decisively reject evolutionary scenarios that place the increase in slope somewhat earlier in sauropodomorph evolutionary history, either at the origin of Massopoda (𝚫AIC = 1.15) or at the origin of Sauropodomorpha (𝚫AIC = 1.96). Within Sauropoda, additional shifts to a third scaling regime characterized by a larger intercept occurred independently in mamenchisaurids, Rapetosaurus, and Euhelopus, all of which have 17 or more cervical vertebrae. Although the results presented here broadly agree with previous work in recovering a unique, positively allometric regime for neck scaling in sauropods, the repeated co-occurrence of regime shifts with increases in cervical vertebral counts suggests two novel conclusions: 1) that relaxation of developmental controls on vertebral formula may be an important precondition for the evolution of distinct scaling regimes, and 2) that the particularly long necks of mamenchisaurids and some titanosauriforms were not passively dictated by developmental constraint, but were instead selectively advantageous.