. 2013 Sep 19;8(9):e76510. doi: 10.1371/journal.pone.0076510

Table 1. Effects of sperm competition and metabolic rate on spermatogenic traits.

Dependent variable	Predictor	Slope	F	P value	λ	r	CI	n
% of seminiferous tubules	RTS^a	0.19	16.04	0.0002	0.79^{*, n.s.}	0.46	0.25 to 0.76	62
	MSMR	0.37	18.54	0.0001	0.25^{n.s., *}	0.59	0.34 to 1.01	37
Tubule diameter	RTS^a	0.03	1.45	0.23	0.71^{*, n.s.}	0.15	-0.10 to 0.40	66
	MSMR	0.04	0.96	0.33	0.24^{n.s., *}	0.15	-0.16 to 0.46	43
Height of epithelium	RTS^a	0.05	1.05	0.31	0.50^{n.s., n.s.}	0.19	-0.17 to 0.55	32
	MSMR	-0.08	3.80	0.067	<0.01^{n.s., n.s.}	0.42	-0.03 to 0.92	20
Number of Sertoli cells	RTS^a	-0.18	0.11	0.12	0.67^{n.s., n.s.}	0.06	-0.29 to 0.41	35
	MSMR	-0.14	0.91	0.35	0.23^{n.s., *}	0.19	-0.21 to 0.59	27
Efficiency of Sertoli cells	RTS^a	0.29	15.98	0.0004	<0.01^{n.s., *}	0.60	0.32 to 1.05	32
	MSMR	0.13	1.30	0.27	<0.01^{n.s., *}	0.23	-0.18 to 0.65	25
SECL	RTS^a	-0.07	5.48	0.02	<0.01^{n.s., *}	0.29	0.05 to 0.56	62
	MSMR	-0.10	7.59	0.008	0.76^{*, n.s.}	0.38	0.10 to 0.70	47
Spermiogenesis	RTS^a	-0.03	0.34	0.56	<0.01^{n.s., *}	0.12	-0.29 to 0.54	25
	MSMR	-0.06	1.31	0.27	<0.01^{n.s., *}	0.26	-0.21 to 0.74	20
Daily sperm production	RTS^a	0.26	8.58	0.006	0.999^{*, n.s.}	0.45	0.15 to 0.83	36
	MSMR	0.21	1.85	0.19	<0.01^{n.s., *}	0.27	-0.14 to 0.70	25
Sperm in cauda	RTS^a	1.70	116.31	<0.0001	0.59^{n.s., *}	0.86	0.99 to 1.61	43
	MSMR	-0.99	2.94	0.097	0.999^{*, n.s.}	0.30	-0.06 to 0.68	31
Sperm in ejaculate	RTS^a	1.58	32.21	<0.0001	<0.01^{n.s., *}	0.66	0.49 to 1.11	44
	MSMR	-1.22	3.48	0.071	0.999^{n.s., n.s.}	0.30	-0.03 to 0.65	37

Phylogenetically controlled multiple regression analyses revealing the effects of relative testes mass (RTS) and mass-specific metabolic rate (MSMR) on spermatogenic traits. ^a In the RTS analyses, we report the values for the second predictor (testes mass) after controlling for the effect of the first predictor (body mass; see Table S2 for the values of body mass). All variables were log₁₀ transformed (with the exception of the proportion of seminiferous tubules, which was arcsine transformed) prior to analysis. The superscripts following the λ value indicate significance levels (n.s., p > 0.05; * p < 0.05) in likelihood ratio tests against models with λ = 0 (first superscript) and λ = 1 (second superscript). The effect size r was calculated from the F values; we also present the non-central 95% confidence interval (CI), an interval excluding 0 indicating statistically significant relationships. The P values and CI that indicate statistical significance are shown in bold. Abbreviations: n: number of species in each analysis; SECL: seminiferous epithelium cycle length.