Erratum: Hydrodynamic simulations of the performance landscape for suctioneeding fishes reveal multiple peaks for different prey types (Integrative and Comparative Biology DOI: 10.1093/icb/icaa021)

A coding error was found in the script in which the force exerted on the prey was calculated. This error only applied to attached ('mollusk-like') prey. The error arose because the vector that defined the location of the prey at each time step did not keep the position of the prey fixed. The consequence was that the forces exerted on the prey, i.e. the forces in play at the position the prey occupied, were not correctly calculated. Thus, the simulations carried out for mollusk-like prey did not produce the correct maximum force associated with the specific combination of phenotypic traits. We have corrected this in the updated article, and re-done the simulations and the subsequent analyses. Figures 3, 4, and 5, and the supplementary material have been updated respectively. The fitted gam revealed that all smooths and tensor smooths were significant, with r² ¼ 91.6%. The gradient ascent analysis revealed a less rugged landscape with a small number of peaks (Fig. 3). In the PCA carried out to characterize the trait space associated with each peak, the first dimension explained 68% of the variance with strong positive loadings on time to peak jaw protrusion and strike distance, while the second dimension explained 27.5% of the variance with strong positive loadings on ram speed and peak gape. In the subsequent cluster analysis, the peaks were classified into four clusters, which differed from one another in terms of performance, i.e. the maximum force exerted on the prey, with cluster 1 outperforming the other three. (Fig. 4). The first cluster peak generally described close strikes with short strike distance, while the remaining three were associated with longer strike distance but varied in terms of speed of mouth and jaw movements. All clusters were associated with performance levels higher than the average of the landscape. The basins of attraction were of roughly equal size and partly overlapping, especially clusters 2, 3, and 4 (Fig. 5). For clusters 2 and 4, the basins of attraction encompassed the peaks, while for clusters 1 and 3, the peaks were located outside the basins. We thus conclude that feeding performance when targeting mollusk-like prey is highly dependent on the specific kinematic phenology (combination of kinematic traits) of the predator. We also note that some of the identified performance peaks differ considerably both in terms of maximum force exerted as well as in the trait profile associated with them. This indicates that predators chiefly targeting prey that cling to a holdfast may exhibit different kinematic strategies while feeding.