TABLE 4.

Phenotypic selection for honey bee with relative fitness calculated using either the proportion of flowers visited or the proportion of flowers tripped by honey bees.

N = 153		Proportion of flowers visited					Proportion of flowers tripped
		Log transformed		Bootstrap			Log transformed		Bootstrap
Trait	Selection coefficient	Estimate	P	Estimate	Lower 95%	Upper 95%	Estimate	P	Estimate	Lower 95%	Upper 95%
Racemes per Stem	S	0.314	0.004	0.314	0.108	0.548	0.292	0.011	0.292	0.084	0.579
Racemes per Stem	C22	−0.268	<0.0001	−0.268	−0.443	−0.157	−0.256	0.002	−0.256	−0.455	−0.132
Racemes per Stem	γrcpsrcps	−0.234	0.01	−0.234	−0.450	−0.057	−0.223	0.026	−0.223	−0.465	−0.009
Stems per plant	S	0.907	<0.0001	0.907	0.605	1.262	0.951	<0.0001	0.951	0.649	1.379
Stems per plant	β	0.890	<0.0001	0.890	0.580	1.273	0.947	<0.0001	0.745	0.489	1.174
Stems per plant	C22	0.186	NS	0.186	0.012	0.400	0.196	NS	0.196	0.017	0.349
Stems per plant	γstppstpp	0.230	NS	0.230	0.025	0.536	0.246	NS	0.246	0.028	0.598

The statistical significance of the selection coefficients was determined using either regression models with log transformed relative fitness or using bootstrapping. The methodology used to obtain the different selection coefficients is summarized in Table 1. The letter P stands for probability and NS for not statistically significant. Rcps stands for racemes per stem and stpp stands for stems per plant.