Seeds of different species are believed to have characteristic shelf lives, although data confirming this are scarce, and a mechanistic understanding of why this should be remains elusive. We have quantified storage performance of c. 42,000 seed accessions, representing 276 species, within the USDA National Plant Germplasm System (NPGS) collection, as well as a smaller experiment of 207 cultivars from 42 species. Accessions from the NPGS collection were harvested between 1934 and 1975, and had relatively high initial germination percentages that decreased at a variable rate during storage at both 5 and –18°C. Germination time courses, which represent the average performance of the species, were fitted to Avrami kinetics, to calculate the time at which germination characteristically declined to 50% (P50). These P50 values correlated with other longevity surveys reported in the literature for seeds stored under controlled conditions, but there was no correlation among these studies and seed persistence observed in the classic buried seed experiment by Duvel. Some plant families had characteristically short-lived (e.g. Apiaceae and Brassicaceae) or long-lived (e.g. Malvaceae and Chenopodiaceae) seeds. Also, seeds from species that originated from particular localities had characteristically short (e.g. Europe) or long (e.g. South Asia and Australia) shelf lives. However, there appeared to be no correlation between longevity and dry matter reserves, soluble carbohydrates and parameters relating to soil persistence or resource allocation. Although data from this survey support the hypothesis that some species tend to survive longer than others in a genebank environment, there is little information on the attributes of the seed that affect its storage performance.

Although viable seeds of Mediterranean dry-fruited shrubs are found in herbivore dung, the ecological importance of this observation is still not well understood. We analysed seed retrieval percentages, defecation time and germinability after sheep gut passage for the five most common shrub species of an area in central Spain (Retama sphaerocarpa, Cytisus scoparius, Halimium umbellatum subsp. viscosum, Cistus ladanifer and Lavandula stoechas subsp. pedunculata). Five ewes were fed seeds, and their dung was collected regularly during the following week. Seeds were hand-sorted from dung subsamples and tested for germinability. The defecated seeds were clustered in time, with a majority retrieved in the 24–40 h period, although over 1% of the seeds were retained in the gut for more than 72 h. Data suggested a possible link between seed size and retrieval, with medium-sized seeds less damaged (16–23%) than larger and smaller seeds (10–12%), although only a small number of species were studied. Germination results showed an increased percentage of germination after gut passage for H. umbellatum (x2 test, P<0.05) and a marginally significant difference for C. scoparius (P<0.1). Soft-seeded L. stoechas did not germinate after gut passage. The results indicate a potential role of herbivore endozoochory for the long-distance dispersal of dry-fruited shrubs and their potential colonization of distant sites.

This paper describes the dispersal–germination characteristics of seeds of 77 native tree species in a seasonal monsoon forest in Uttaranchal state, northern India. Results indicate that 50% of the species dispersed in the hot, dry summer months, 18% during the rainy season, 23% during the cold season and the remainder in late spring. Germination tests on fresh and laboratory-stored seeds revealed a relationship between morphological features of the fruit and both germination percentage and mean germination time (MGT). Highest mean germination (50%) was for dry-dehiscent fruits with winged wind-dispersed seeds, followed by dry-dehiscent fruits with non-winged seeds (38%) and seeds of dry-indehiscent fruits (37%). Lowest germination (29%) was for seeds from fleshy or pulpy fruits. MGT followed the reverse course. Germination data for seeds stored dry in the laboratory during one seeding cycle indicated six patterns of seed germination: (1) average germination percentage of fresh seeds lower than that of stored seeds, indicating an after-ripening requirement; (2) initial high germination percentage followed by low values, indicating a steep to moderate decline in viability following harvest; (3) no germination after 1 or 2 months of seed storage, due to complete loss of viability, indicating short seed longevity; (4) fresh seed germination in some species equalled the average germination value of stored seeds, indicating constant germination for one whole seeding cycle; (5) germination of both fresh and stored seeds remained consistently low throughout the season, indicating a requirement for some kind of pre-treatment or having poor quality of seeds; (6) initial high germination followed by decline and again increase, showing a seasonal rhythm of germination.

Direct sowing is the simplest method of plant establishment for restoration and remediation purposes, but relatively few plants can establish under high salinity conditions. In this study, the ability of different seashore plants and grass cultivars to germinate in different dilutions of seawater (0–400 mM NaCl) was tested. Highest germination was found in distilled water or seawater dilutions up to 100 mM NaCl. When seawater concentrations were increased from 100 to 200 mM NaCl, a strong decline in germination percentage and rate was observed in less salt-tolerant species, such as Matricaria maritima and Achillea millefolium. The more salt-tolerant species, Plantago maritima, Juncus gerardii, Artemisia vulgaris, Agrostis spp. and Rumex spp., had a threshold salinity, where germination was significantly decreased in seawater dilutions between 200 and 400 mM NaCl. Even among the salt-tolerant species, only two, Agrostis stolonifera and Artemisia vulgaris, germinated at 400 mM. Variation in salinity response was observed among populations of Artemisia vulgaris and among cultivars of Festuca spp. Increasing salinity to 200 mM NaCl delayed germination in most species. Ungerminated seeds of most salinity-tolerant species were still viable after 21 d at the highest salinity (400 mM), and showed a rapid and high germination when transferred to distilled water. These species would be able to survive high salinity and germinate when the salinity of the sediments decreases through dilution or leaching of salts. The experiment revealed species and cultivars that will be of interest in further testing for restoration and remediation in saline habitats.

Physical dormancy (PY) was identified in six genera representative of Australian Rhamnaceae and subsequently was broken, based on identification of key seed dormancy characteristics: (1) isolation and classification of embryo features; (2) imbibition experiments to determine the rate and amount of water uptake in seeds; and (3) determination of optimum temperature regimes for germination. All six species had relatively large spatulate embryos. Imbibition studies showed all species possessed PY (i.e. a water-impervious seed coat) that was broken by a hot-water treatment. Alleviation of PY resulted in high germination (<70%) at 7/18°C, temperatures similar to winter in south-west Western Australia. Germination was suppressed at higher temperatures and in the presence of light. The study adds information to our knowledge of seed dormancy in Australian Rhamnaceae, and highlights the benefits of understanding dormancy states in seeds prior to evaluating dormancy-release mechanisms on wild species used in restoration ecology and horticulture.

In contrast to reports in the literature that seeds of Dodonaea viscosa from China and Pakistan are non-dormant, or nearly so, we found that a high percentage of seeds of this species collected in north-western India have a water-impermeable seed coat at maturity, i.e. physical dormancy. Thus, seeds that were mechanically scarified and boiled (to open a ‘water gap’ in the seed coat) germinated to much higher percentages (84% and 77%, respectively) than did those that were non-scarified (24%). Our results agree with studies of seed dormancy in this species in various other parts of its large geographical range.