Larval marine perciform fishes have significant swimming abilities for much of their pelagic larval dispersal stage (PLD). This gives them the ability to directly influence their dispersal. However, there are several ways of measuring swimming ability, and in the context of dispersal, it is not clear which is the most appropriate. It is important to use the right swimming measure not only for modelling dispersal, but also for use in predicting both demographic and genetic connectivity. Recently, Nanninga & Manica (2018, MEPS, 589:1-12) concluded that for marine, demersal fishes “larval swimming capacities could serve as a powerful indicator of a species’ long-distance dispersal potential”, and were a better predictor than PLD. As their metric of swimming capacities, Nanninga & Manica used mean critical speed (U-crit), which is measured in a laboratory raceway by incrementally increasing flow until the larva cannot maintain position. I will examine the suitability of U-crit versus other swimming metrics, in particular in situ speed. The among-individual variability of swimming performance will be stressed, as will the ontogeny of swimming ability.