MULTIVARIATE MORPHOMETRICS AND ALLOMETRY OF THE TRIASSIC REPTILE TANYTRACHELOS AHYNIS
Analysis of corresponding right and left variables (e.g., right femur length and left femur length) showed no consistent bias, with mean differences between sides near 0. This finding justified the combination of left and right side variables for subsequent analyses to augment sample size. However, reducing the number of variables, also in an effort to increase sample size, drastically amplified the noise within analyses and increased operator error relative to identical analyses conducted on larger variable sets. Various combinations of variables all yielded similar results with no significant separation of specimens when grouped by morphotype (present versus absent) or stratigraphic horizon (lake cycle 2 versus lake cycle 16). Allometric trajectory (positive or negative) of individual skeletal elements was assessed relative to overall body size as approximated by femur length. Skeletal elements humerus length, tibia length, and fibula length have positive allometric coefficients while radius length and ulna length have negative allometric coefficients. The lack of discrete groupings in the multivariate morphospace, morphometric congruency of the two morphotypes, and lack of discernible differences between the two compared stratigraphic horizons suggest consistently that all analyzed specimens belong to a single morphogroup (morphospecies?). This makes the resulting data particularly suitable as a reference standard for morphometric comparisons with other taxa within the family Tanystropheidae.