Structural studies in seeds with physical dormancy (PY) are important to better understand its causes and release when subjected to treatments for dormancy breaking. The aims of this study were to (1) characterize the PY break; (2) examine the role of different seed structures in water uptake; and (3) identify the water gap in Senna multijuga seeds. Imbibition patterns of dormant and non-dormant (subjected to dormancy breaking treatments) seeds and the morphological changes during dormancy breaking and germination were evaluated. To identify the water gap, the micropyle and lens were blocked separately, and the water absorption by seed parts was determined. Structural characteristics of the seed coat were also examined. Immersion in water at 80°C was efficient in breaking seed dormancy and imbibition occurred first at the hilar region, through the lens. Water was not absorbed through the micropyle or the extra-hilar region. S. multijuga seeds have a testa with a linearly aligned micropyle, hilum and lens. The seed coat consisted of a cuticle, macrosclereids, one (hilar region) or two (extra-hilar region) layer(s) of osteosclereids and parenchyma cell layers. The lens has typical parenchyma cells underneath it and two fragile regions comprised of shorter macrosclereids. Heat treatment stimulated the lens region, resulting in the opening of fragile regions at the lens, allowing water to enter the seeds. It is concluded that short-term exposure to a hot water treatment is sufficient for the formation of a water gap in S. multijuga seeds, and only the lens acts in the imbibition process.

Mature seeds of many legume species are normally characterized by water-impermeable seed coats, a form of physical dormancy. However, observations have suggested that the incidence of mature but permeable (non-dormant) seeds is sometimes substantial. Yet, the ecological processes associated with this non-dormancy have received little attention by plant ecologists. In white clover (Trifolium repens), we therefore studied: (1) the occurrence of initially permeable seeds in wild populations; (2) the relative performance of non-dormant and dormant seeds in plant establishment and reproduction in a field-sown experiment; and (3) the extent to which the trait is affected by humidity and plant genotype in a greenhouse experiment. No less than 35% of all viable seeds from the wild populations proved to be water permeable at maturity. The proportion of permeable seeds within inflorescences ranged from 0 to 100%. In the field-sown experiment, autumn-germinated non-dormant seeds had almost equally good chances of establishing as spring-germinated dormant seeds. Due to a marked head start in growth, the former yielded more flowers (and thus seeds) in the first flowering season. However, the greenhouse experiment proved that variation in the proportion of permeable seed between inflorescences represented a plastic response to humidity conditions during seed ripening, rather than variation among clones (broad-sense heritability ≤ 0.025). Thus the trait is not easily subject to selection.