Table 2. Phylogenetically-controlled multiple regression analyses of coefficient of variation (CV) of sperm dimensions in relation to body mass and testes mass.

Dependent variable	Predictor	Adjusted R 2	Slope	F	P	Lambda (λ)	Effect size (r)	Effect size CLs
CV HL	Body mass	0.33	0.20	0.007	0.93	0.47*,*	0.01	(−0.39 to 0.42)
	Testes mass		−0.23	14.69	<0.001		0.62	(0.32 to 1.14)
CV HW	Body mass	0.06	0.13	0.02	0.88	0.79*,†	0.03	(−0.37 to 0.43)
	Testes mass		−0.14	3.69	0.06		0.37	(−0.01 to 0.79)
CV HA	Body mass	0.20	0.24	0.72	0.40	0.38†,*	0.17	(−0.23 to 0.58)
	Testes mass		−0.20	7.77	0.01		0.50	(0.14 to 0.96)
CV MPL	Body mass	−0.07	0.06	0.003	0.95	0.68*,†	0.01	(−0.39 to 0.41)
	Testes mass		−0.06	0.27	0.60		0.10	(−0.30 to 0.51)
CV PPL	Body mass	0.05	0.03	0.97	0.33	0.95*,†	0.20	(−0.20 to 0.61)
	Testes mass		−0.13	2.44	0.13		0.31	(−0.08 to 0.72)
CV TFL	Body mass	0.37	0.17	2.42	0.13	0.17†,*	0.30	(−0.09 to 0.72)
	Testes mass		−0.31	14.82	<0.001		0.62	(0.32 to 1.14)
CV TSL	Body mass	0.36	0.19	1.86	0.18	0.16†,*	0.27	(−0.12 to 0.68)
	Testes mass		−0.31	14.53	<0.001		0.62	(0.31 to 1.13)

Data have been processed as the logarithm of CV. All tests were conducted with 23 df. The superscripts following λ value indicate significance levels (^† n.s.; *P<0.05) in a likelihood ratio tests against models with λ = 0 (first position) and λ = 1 (second position). The effect size r was calculated from the F values; its noncentral 95% confidence limits (CLs) are also given. Confidence intervals excluding 0 indicate statistically significant relationships. P values and CLs that indicate statistical significance are shown in bold. Abbreviations: HL, head length; HW, head width; HA, head area; MPL, midpiece length; PPL, principal piece length; TFL, total flagellum length; TSL, total sperm length.