PHENOTYPIC CONSEQUENCES OF FORCING GERMINATION - A GENERAL PROBLEM OF INTERVENTION IN EXPERIMENTAL-DESIGN

Many evolutionary and ecological studies of plants with seed dormancy are plagued by design problems related to seed germination. On the one hand, traits of interest, especially life history traits like early growth rate or time to flowering, may not be independent of dormancy phenotype. On the other hand, germination-inducing treatments are likely to affect these traits of interest. Consequently, researchers often have had to use confounded designs without information about the consequences. To examine this problem, we studied how early growth in California poppy (Eschscholzia californica) is affected by treating seeds with giberellic acid (GA) to stimulate germination. The dose used was sufficient to slightly but significantly advance emergence time in treated seeds, but it appeared to cause variable growth responses among treated seeds. Experiments using GA to stimulate germination may thus be misleading, at least for the populations and dose studied. Perhaps more importantly, the experimental and statistical approach we used can be employed to study the effects of this confounded design for other doses and species.