Variable enamel growth rates in hippopotamid canines: Implications for seasonality reconstructions using inverse modeling of intra-tooth isotope data

Seasonal variations in climatic variables, and the resulting changes in vegetation, are strong factors governing ecosystem dynamics in modern and ancient times. Stable isotope ratios recorded in tooth enamel document isotopic variations in the environment at the time of enamel formation and thus reveal the intensity and duration of seasonal dietary and climatic variations. However, the long and multi-phased process of enamel mineralization causes a dampening of the original input signal. An inverse model previously developed for ever-growing canines of Hippopotamus amphibius proposes to recover the original input signal and assumes constant enamel growth rate, appositional angle, and maturation length. The present study aims to test these assumptions. To do so, we integrated data from histological thin sections, microtomodensitometric analyses, and stable isotope analyses on teeth of extant H. amphibius specimens (3 upper canines, 1 lower canine, 1 third molar) to quantify the geometric and temporal patterns of enamel mineralization. To estimate enamel extension rates (EER, in m/increment), we counted the number of increments representing the position of appositional front for each segment of 5 mm along the enamel-dentine junction in thin sections made along the growth axis of each tooth. We used microtomodensitometry to determine the pattern of enamel maturation using grey values profiles of X-ray radiographies as a proxy for enamel mineralization degree. Serial sampling along one upper canine of an individual from Chad, coming from an environment with one rainy season per year, allowed us to document the intra-tooth d13C and d18O variations over 6 years and thus provided an independent temporal control on histological variations. The histological study showed that the enamel apposition phase is strongly irregular over time within the canines, with no clear temporal trend. EERs vary strongly among teeth and within each tooth (50-200 m/increment, 100-350 m/increment, and 80-200 m/increment for the 3 upper canines; 150-550 m/increment for the lower canine; 70-130 m/increment for the third molar). The median EER value from the upper canine of the juvenile individual (ca. 180 m/increment) is significantly higher than median EER values from the upper canines of two adult individuals (ca. 110 m/increment). Similar variations are also observed in apposition angles (3-8, 2.5-4.5, 3-7 for the 3 upper canines; 2-8 for the lower canine; 6-18 for the third molar). The enamel mineralization parameters vary with age and tooth type (canine vs. molar). Based on strongly correlated seasonal variations in d13C and d18O, we also confirm cyclic dietary variations with higher proportions of C4 plants consumed during the dry seasons. Using the range of enamel mineralization parameters observed within one single hippo canine, we conducted sensitivity tests on the inverse modeling method, producing different modeled input signals that suggest a wider range of uncertainty. In conclusion, the documented intra-canine variability of EER, as well as other histological parameters (apposition angle, maturation length), reveals challenges when applying the current inverse model to wild populations. Future work would benefit from a systematic histological investigation into the sources of variation of enamel growth and mineralization patterns.