Repeatability and correlation of swimming performances and size over varying time‐scales in the guppy (Poecilia reticulata)

Abstract

Summary 1 Although the repeatability of a trait is of interest for several reasons, few studies have critically examined the repeatability or correlation of metrics of swimming performance. We quantified repeatability of three swimming performances (burst speed during a c‐start escape response, critical swimming speed [Ucrit], maximum speed [Umax]) and size/shape measures over short (within‐day), medium (days and weeks), and long (more than a year) time‐scales in a small poeciliid fish, the Trinidadian guppy (Poecilia reticulata). 2 We found that individual differences in most swimming performance metrics were repeatable on various time‐scales. Burst speed during c‐start was repeatable within a day, as well as over days and weeks, but it was not repeatable over the span of a year. However, the medium term repeatability was only statistically significant when accounting for the variation of individuals within a day and correcting for attenuation. Ucrit,Umax and size were repeatable on medium time‐scales, but only size and Umax were repeatable over the span of a year. Our index of shape was not statistically repeatable over any time‐scales. 3 When examining the correlations of performances and size, only a few traits were related to each other. Ucrit and Umax were significantly positively related to each other for trial 1 only, but not any of the others. Umax and the mean c‐start burst speed were correlated on trial 2 only. No other traits were significantly related to each other, and no performance traits were related to size or shape. Ucrit and Umax also did not differ in their overall mean values, suggesting that Umax, a more recently described assessment of fish swimming performance, may be a reliable alternative. 4 All traits, except our index of shape, shifted in their mean value over the span of a year. Ucrit and c‐start burst speed decreased significantly, consistent with previous findings on the effects of senescence on neuromuscular systems. However, Umax and size increased significantly over a year. 5 The measures of swimming performance used in this study are becoming widely used in ecological and evolutionary studies. Our results suggest that individual differences in these measures are relatively stable on various time‐scales. We found no significant relation of these performance metrics with body size or a simple index of shape, and few relations among the performance measures, suggesting that they measure different aspects of locomotor physiology and/or motivation under the test conditions used. Finally, while Ucrit and c‐start burst speed decreased over the span of the year, consistent with theories of senescence, Umax significantly increased over a year, suggesting that further studies of this recently described performance metric are needed.