2.7 Change in the magnitude of Dehnel’s phenomenon from 1955 to 1985
We compared the amount of change in the skull height, length, and width from the data subset where we had data from all three size extremes (summer juveniles, winter subadults, regrown adults). Additionally, we tested for the effect of time on seasonal changes in relative skull size (skull height and skull width corrected for skull length). To do this, we compared juveniles (July and August separately) with subadults and subadults with adults in ANCOVA models with age as the factor, year and skull length as the covariates, and the interaction between year and age.
To visualize the change in the magnitude of Dehnel’s phenomenon over time, we divided the years where we had data from all three size extremes (summer juveniles, winter subadults, regrown adults) into three time intervals (1953-1963, 1964-1974, 1975-1985). We compared the seasonal curves of the skull height between these time intervals with a generalized additive model (GAM) using the “time interval” as a parametric term and “month” as a nonparametric term. We used a Gaussian distribution and added a smoothing function to “month” with knots restricted to five (because we had an a priori expectation for the presence of the shape of a seasonal skull height pattern).