Table 2. Nesting males and sneakers differ in sperm motility, velocity and longevity.

Variable		Sperm motility (% motile)		Curvilinear sperm velocity (VCL)
	df	t value	P (>|t|)	t value	P (>|t|)
Intercept	87	13.93	<0.001	16.74	<0.001
Effect of male type: sneaker relative to nesting male	87	−3.96	<0.001	−1.80	0.076
Effect of time: 50 s relative to 40 s	58	−2.25	0.028	−4.93	<0.001
Effect of time: 60 s relative to 40 s	58	−4.91	<0.001	−7.84	<0.001
Interaction between: the effect of male type (sneaker relative to nesting male), and the effect of time (50 s relative to 40 s)	87	1.19	0.238	1.82	0.071
Interaction between: the effect of male type (sneaker relative to nesting male), and the effect of time (60 s relative to 40 s)	87	2.23	0.028	2.749	0.007

Results are shown for sperm motility and curvilinear velocity. Other measures of velocity yield qualitatively similar results. No differences in straightness or linearity were detected. The P values of significant effects are highlighted as bold font. A linear mixed effects model was fit by maximum likelihood with time (40, 50 and 60 s) as a repeated measure and cross as a random effect (using lme from library nlme in R version 3.2). Fixed effects were: motility∼Type × Repeat; VCL∼Type × Repeat (Repeat=40, 50, 60 s). n=30 in vitro crosses (10 per treatment). Each cross involved a new female, sneaker and nesting male. Fish were not used more than once either within or between experiments.