Mutation of the COMATOSE locus in Arabidopsis results in a marked reduction in germination potential. Whilst the morphology of comatose (cts) embryos is not altered, physiological analysis reveals that mature cts seeds do not respond to gibberellin. Prolonged chilling of imbibed seeds only partially restores germination potential, and seeds do not after ripen. Genetic analysis shows that the cts phenotype is expressed in the embryo and phenotypic differences between wild-type and mutant plants were not observed during other stages of plant growth and development. Therefore cts represents a new class of mutant, with a specific lesion that results in severely impaired germination potential. Genetic interactions were analysed between cts and loci that regulate embryo maturation, and abscisic acid biosynthesis and perception. Results from these studies showed that the cts mutant phenotype required the wild-type action of these loci, and suggested that CTS exerts a repressive function on these loci. A model is presented postulating that CTS promotes increased germination potential, and represses embryo dormancy. These functions of CTS may result in the removal of embryo dormancy as a prerequisite to germination.