Numerous studies have demonstrated the ecological and social relevance of different aspects of animal locomotor performance, as locomotion is frequently required to capture prey, escape from predators, find mates and defend territories. Typically, maximal performance is quantified using two alternative methods: across distance intervals or across time intervals, the later by extracting the speed observed between a specified number of video frames. Here, we analyze how this choice may influence statistical inference and the derived biological interpretations. Our results indicate that data obtained using 10-cm intervals are categorized, not normally distributed, exhibit lower individual repeatabilities and have four times more variance compared to the data obtained by quantifying speeds across short time intervals. These results have important methodological implications, as they suggest that the choice of method of speed quantification substantially influences the quality and bias of maximal performance descriptors. This may in turn have a profound influence on ecomorphological inference, as it affects our capacity of detecting variation in performance within and across groups, and associations between locomotor performance and other traits.
展开▼
