Table 3.

Effect of genotype, exogenous GA treatment and flower position on the occurrence of floral abnormalities across the early Arabidopsis inflorescence

Phenotypic class	Level of analysis	Effects
		Genotype	GA treatment	Flower position	Genotype: GA treatment	Genotype: flower position	GA treatment: flower position	Genotype: GA treatment: flower position
All floral abnormalities	Whole flower	0.004**	< 0.001***	< 0.001***	0.024*	0.033*	< 0.001***	0.807
Deviations in organ number	Whole flower	< 0.001***	< 0.001***	< 0.001***	0.096	0.136	< 0.001***	0.855
	Sepals	0.360	0.009**	0.533	0.882	1.000	0.996	1.000
	Petals	0.503	0.008**	0.822	0.836	1.000	0.995	1.000
	Long stamens	0.128	0.036*	0.007**	0.279	0.903	0.118	0.999
	Short stamens	0.006**	< 0.001***	< 0.001***	0.256	0.504	0.200	0.953
Organ fusion	Whole flower	0.573	0.004**	0.242	0.751	0.940	0.843	1.000
	Long stamens	0.377	0.002**	0.210	0.819	0.994	0.876	1.000
Organ homeosis	Short stamens	0.007**	0.378	0.608	0.817	1.000	0.963	1.000

Summary of p values for main effects (genotype, GA treatment and flower position) and interactions between them (denoted by colons) obtained via GLM (see “Materials and methods”). The floral abnormalities dataset was analysed at a series of hierarchical levels, parsing the data by phenotypic class of abnormality (deviations in organ number, organ fusion, organ homeosis) and the position (whorl) within the flower where the abnormality occurred. Where no abnormalities of a specific type occurred within a particular whorl, that analysis is not shown. Organ homeosis was only observed in short stamens. Statistically significant p values (p < 0.05) are highlighted in bold. The numbers of asterisks denotes increasing stringency of statistical significance

*p < 0.05, **p < 0.01, ***p < 0.001