Disentangling fluctuating asymmetry from otolith shape

Fluctuating asymmetry (FA), defined as random deviations from ideal, perfect symmetry, has been hypothesized to be negatively associated with fitness and positively associated with stress. Fish otoliths deserve special attention in regard to this topic because their FA is relatively easy to measure, unlike the fitness of wild fish themselves. In addition, otoliths are convenient for estimating FA, because they are systematically extracted for aging purposes, and otolith collections might allow the use of FA as a bioindicator of change at different spatio-temporal scales. Here we demonstrate 2 key methodological tools: the use of shape instead of previously used univariate (metric or meristic) traits, and the use of a statistical protocol capable of disentangling true FA from other types of asymmetry (namely, directional asymmetry and antisymmetry) and from measurement error. In the case of the razor fish Xyrichthys novacula; fluctuation asymmetry of the otolith shape of 180 individuals from 5 populations accounted for 23.2% total shape variability (5 fold larger than the measurement error). Moreover, FA clearly differed between populations. In contrast, FA measured using a univariate trait (radius at the sulcus) did not detect between-population differences. Moreover, in a simulation experiment, we demonstrate that FA measured using shape is able to reveal a true pattern, while FA measured using a univariate trait extracted from the same data is unable to do so. Therefore, the analysis of FA using shape paired with an adequate statistical model may increase our knowledge of the potential links between otolith FA and fitness or stress.