Table 1.

Analysis of pollen viability and pollinator responses in the 2011 live plant arrays. Included are the tests of the null hypothesis that there is no variation among our M. guttatus subpopulations (IL, IH, C, OL and OH) in (a) pollen viability, (b) bumble bee arrivals to plants and (c) the number of flowers visited per arrival from the 2011 live plant array. The GLMM used a binomial distribution and logit link to analyze the proportion of viable pollen grains per plant. The repeated-measures mixed model ANCOVAs used an unstructured variance–covariance structure. Fixed effects were tested with F-ratios. Random effects were tested with 1 df log-likelihood ratio tests (G).

Effect	df	ddf	Fixed or random	F	Variance component	G	P
(a) 2011 Pollen viability generalized linear mixed model
Subpopulation	4	400	Fixed	5.20	–	–	0.0004
Block	3	–	Random	–	0.027	0.79	0.1871
Family (subpop)	400	–	Random	–	0.000	0.00	1.0000
(b) 2011 Bumble bee visitation general linear mixed model RMANCOVA
Subpopulation	4	297	Fixed	12.78	–	–	>0.0001
Day	2	1356	Fixed	28.46	–	–	>0.0001
Flowers	1	1356	Fixed	2.16	–	–	0.1417
Subpop × day	8	1356	Fixed	0.85	–	–	0.5596
Block	4	–	Random	–	0.001	1.90	0.0840
Family (subpop)	297	–	Random	–	0.000	0.00	1.0000
(c) 2011 Floral visits general linear mixed model RMANCOVA
Subpopulation	4	246	Fixed	0.56	–	–	0.6949
Day	2	357	Fixed	0.47	–	–	0.6249
Flowers	1	357	Fixed	8.35	–	–	0.0041
Block	4	–	Random	–	0.004	0.90	0.1714
Family (subpop)	246	–	Random	–	0.015	0.40	0.2635