The assumption that seeds imbibe most of the water required for germination as liquid through seed-to-soil contact has been a dominant theme in germination research and seeding technology. Under most conditions, seeds are also exposed to water vapour during imbibition, but the relative contributions of liquid and vapour are difficult to assess. In water uptake models that include vapour, procedures used to estimate potential vapour imbibition have underappreciated the effect of distance on diffusion rate. At the same time, the amount of seed-to-soil contact and the liquid bridge from soil water to the seed tend to be greatly overestimated, considering the soil water contents often found in the field. Most researchers have recorded an approximately equal time to germination at soil water contents ranging from field capacity to nearly permanent wilting point, and little response to bulk density, soil type or seed–soil contact. While hydraulic conductivity decreases by several orders of magnitude as soil water content, bulk density and seed-contact decrease, relative humidity remains near 100%. There are several experiments demonstrating timely germination in water vapour alone. The combined evidence contradicts the assumption that seed–soil contact is important for imbibition of water by seeds. Water vapour should be considered the primary source of water for seeds in unsaturated soils.

Martin's (1946) seed classification system has 10 types based on embryo and endosperm characteristics and two additional types based on seed size: dwarf (0.3–2.0 mm) and micro ( ≤ 0.2 mm). He listed 17 families and 12 genera (in five other families) as having dwarf seeds. Our recent discovery of morphophysiological dormancy in dwarf seeds of several taxa of Campanulaceae and one of Gentianaceae prompted an evaluation of dwarf seeds. Martin's paper contains 37 families with one to several small (0.3–2.0 mm) seeded species that he did not list as being dwarf. Comparison of Martin's dwarf families and the 37 small-seeded non-dwarf families revealed no consistent differences between the two groups in endosperm texture, seed-coat anatomy, embryo morphology, class of seed dormancy or phylogenetic position. Also, Martin's dwarf seeds include a variety of embryo morphologies. Consequently, we have revised Martin's key to seed types. The dwarf category has been removed and the micro category replaced by ‘undifferentiated’ to reflect the state of the embryo in fresh seeds. Further, the key now includes linear fully developed, linear underdeveloped, spatulate fully developed and spatulate underdeveloped seed types, which Martin illustrated but did not include in his key. In the revised key, all seeds are distinguished on the basis of embryo and endosperm characteristics.

Intact seeds (seed+endocarp) from freshly harvested fruits of Prunus campanulata were dormant, and required 4–6 weeks of warm followed by 8 weeks of cold stratification for maximum germination percentage. Removing both endocarp and seed coat, however, promoted germination in a high percentage of non-stratified seeds. Treatment of intact, non-stratified seeds with gibberellic acid (GA3) was only partially effective in breaking dormancy. However, GA3 promoted germination of non-stratified seeds in which the endocarp (but not the seed coat) had been removed. The order of abscisic acid (ABA) concentration in fresh seeds was endocarp > seed coat > embryo, and its concentration in endocarp plus seed coat was about 6.2-fold higher than that in the embryo. Total ABA contents of seeds subjected to warm and/or cold moist stratification were reduced 6- to 12-fold. A higher concentration of GA4 was detected in embryos of non-dormant than in those of dormant seeds. Fluridone, a carotenoid biosynthesis inhibitor, was efficient in breaking dormancy of Prunus seeds. Paclobutrazol, a GA biosynthesis inhibitor, completely inhibited seed germination, and the inhibitory effect could be partially reversed by GA4, but not by GA3. Thus, dormancy in P. campanulata seeds is imposed by the covering layers. Dormancy break is accompanied by a decrease in ABA content of the covering layers and germination by an increase of embryonic GA4 content.

Seeds vary widely in the sensitivity of germination to oxygen (O2) partial pressure, depending upon the species, temperature, dormancy state and physiological status of the seeds. Most analyses of the O2 sensitivity of germination have focused on final germination percentages and estimated the O2 percentage in air that is required to reduce germination to a given percentage (usually 50%). In contrast, we have applied a population-based threshold model utilizing time courses of germination to quantify three parameters related to seed germination sensitivity to O2 availability: the median base (or threshold) O2 percentage, the standard deviation of O2 thresholds among seeds in the population, and an oxygen–time constant that relates O2 percentage to germination timing. The model fits germination responses accurately across a wide range of O2 concentrations. The response to O2 was logarithmic in all cases, with the O2 percentage required for 50% germination ranging from 21% to as low as 0.005%, depending upon the species, the temperature and the seed dormancy level. Modelling indicated that some seeds can adapt to low O2 percentages and shift their thresholds to lower values over time. Lower temperatures decreased the minimum O2 threshold, as did after-ripening. Seed priming generally reduced the oxygen–time constant and increased the standard deviation of germination responses, but had relatively little effect on the O2 sensitivity per se. The population-based threshold model can be used to quantify the O2 sensitivity of seed germination and to predict germination rates and percentages when O2 availability is limiting.

Substances that could play a role in maintaining seed viability (phenolic compounds, α-tocopherol, sterols, ascorbic acid, glutathione and soluble proteins) were estimated in beech (Fagus sylvatica L.) seed lots that had been stored for 2, 5, 7 and 10 years at − 10°C. Germination capacity was strongly and positively correlated with amounts of total phenolic compounds, ultraviolet (UV)-absorbing phenols and soluble proteins. Moderately strong relationships appeared between germination capacity and α-tocopherol and ascorbic acid contents. Total sterol and glutathione contents were not correlated with germination capacity. A strong, negative correlation was found between germination capacity and reactive oxygen species (ROS), such as superoxide radical () and hydrogen peroxide (H2O2), as well as with lipid hydroxyperoxides (LHPOs). The putative role of these compounds in the maintenance of beech seed viability during long-term storage is discussed.

The effect of flooding on the dormancy level of Setaria parviflora seeds from a non-flooded upland and a seasonally (winter–spring) flooded lowland in the Pampa grasslands of Argentina was investigated. Seeds from both communities were subjected to reciprocal burial treatments in the two habitats, and exhumed during and after the flooding season. Effect of immersion in water at 5°C was compared to incubation of seeds on the surface of water-saturated paper at the same temperature. After exhumation of the buried seeds or immersion treatments, germination was assayed at 25°C and at 20/30°C in the dark or in combination with light. Burial in the lowland, which was flooded in winter–spring, significantly reduced germination, while burial in the non-flooded upland did not reduce germination. Similarly, immersion in water at 5°C significantly reduced germination compared to non-immersed controls. During summer, seeds buried in the lowland showed increased capacity to germinate, particularly when exposed to fluctuating temperatures or light. Thus, flooding induced secondary dormancy in S. parviflora seeds, and it was broken during the non-flooding season. These responses of the seeds would prevent germination until there was no further risk of flooding. Remarkably, in S. parviflora seeds harvested from both habitats, we observed essentially the same germination requirements after flooding. However, some slight differences were detected between the seed populations exhumed from the lowland site, indicating that flooding had larger effects on the seeds from the upland community. This suggests some differentiation of these populations evident only after flooding in the field.

Observing the embryo in dust-like seeds with a rigid coat presents a particular challenge, because it is not possible to use conventional techniques applied to normal-sized seeds. A new combination of softening and clearing techniques, which is remarkably effective for direct observations of embryos in minute seeds, is reported here. Moreover, no special seed treatment is required prior to application of this combination of techniques, and good results are obtained from living and preserved material. Dust-like seeds of Cytinus hypocistis and Cytinus ruber with a rigid coat were treated successively with Franklin's and with Jeffrey's softening fluids, and then with Herr's clearing fluid. The combined action of the two softening fluids caused sufficient seed coat rupture to allow Herr's fluid to clear seed tissues, permitting effective observation of the embryo. This technique should be applicable to a wide range of species with dust-like seeds, and its possibilities for biological studies are numerous.