Quantifying nonlinear temporal effects of ethanol preservation on round goby (Neogobius melanostomus) anatomical...

Geometric morphometrics provides a powerful means of evaluating differences in phenotypic traits among specimens. However, inferences of trait variability can be confounded when measurements are based on preserved samples. We evaluated effects of ethanol preservation on morphology over a 22-week time period for a Laurentian Great Lakes invasive fish, round goby (Neogobius melanostomus, Pallas 1814), using sets of 17 lateral and six dorsal landmarks. We tested whether ethanol preservation affected the magnitude of inter-population variation between individuals collected from lake and river habitats. Generalized least square regression determined that length did not significantly vary through the preservation time series for fish from either population, while mass decreased significantly. Body shape variation was summarized using principal component analysis, which revealed that most preservation-associated changes occurred in the first 14 days. The lateral shape experienced a large magnitude change during the first 24 h in ethanol then only minor changes for the remainder of the study. The dorsal shape began to revert to pre-preservation measurements about 14 days following preservation. Additionally, differences in shape were apparent between the two populations throughout the experiment; however, the magnitude of differences between populations varied depending on whether dorsal or lateral landmarks were considered. Our study demonstrates that tissue responses to ethanol preservation can be more complex than a simple loss of mass, resulting in difficult to predict consequences for geometric morphometric analyses, including variable responses depending on the anatomical region being analyzed.