Assessments of seed storage physiology among Arecaceae (palm) species are often inconclusive because seeds exhibit diverse responses to low temperature and moisture conditions. Interrelationships between dry matter accumulation, cell structure and water relations during seed development of the endangered Hawaiian endemic palm, Pritchardia remota, suggest that damage from drying results from mechanical strain. Endosperm and fruits accumulate dry mass through most of the 400 d gestation period, but embryos reached maximum dry mass about 250 d post-anthesis (DPA). Mostly sucrose and some triacylglycerols accumulated in the cytoplasm and vacuoles of embryo cells, and organelles in mature embryo cells de-differentiated. Water content and water potential decreased as embryos matured and embryos contained about 0.45 g H2O (g dry mass)− 1 ( − 26 MPa) at shedding. Mature embryos survived drying to 0.16 g g− 1 ( − 49 MPa), but further drying was lethal. A model of allowable cell shrinkage is consistent with the substantial, but incomplete, desiccation tolerance acquired in P. remota embryos, and provides a new framework to explain variation in critical water contents as embryos develop. We suggest that desiccation tolerance, which distinguishes recalcitrant and orthodox physiologies among seeds, can be quantified by mechanical strain when embryo cells shrink during drying.

The drying rate of whole seeds of Ekebergia capensis (Meliaceae) was shown to influence the response to desiccation, with rapidly dried seeds surviving to lower water contents. Short-term rapid drying (to water contents higher than those leading to viability loss) actually increased the rate of germination. The form of the time course of decline of axis water content varied with drying rate; slow drying could be described by an exponential function, whereas with rapid drying initial water loss was faster than predicted by an exponential function. These observations suggest that slow drying brought about homogeneous dehydration and that the rapid drying was uneven across the tissue. This raised the possibility that the different responses to dehydration were a function of different distributions of water in the axis tissue under the two drying regimes. However, ultrastructural observations indicated that different deleterious processes may be occurring under the different drying treatments. It was tentatively concluded that a major cause of viability loss in slowly dried material was likely to be a consequence of aqueous-based processes leading to considerable membrane degradation. Uneven distribution of tissue water could not be rejected as a contributory cause of the survival of rapidly dried seeds to low bulk water contents. The differential response to dehydration at different drying rates implies that it is not possible to determine a ‘critical water content’ for viability loss by recalcitrant seeds.

The contents of the cytokinins zeatin (Z), zeatin riboside (ZR), and isopentenyladenine (IPA), the combined contents of gibberellins1 + 3 (GA), and the contents of indoleacetic acid (IAA) and abscisic acid (ABA) were measured during the development of the desiccation-sensitive seeds of Avicennia marina (Forssk.) Vierh. During the stage of histodifferentiation the amounts of these plant growth regulators (PGRs) were measured on whole fruits. During the phase of seed growth and reserve accumulation measurements were made on the embryonic axis, cotyledons and pericarp separately. Patterns in the amounts of PGRs present during histodifferentiation were similar to those reported for desiccation-tolerant seeds and it suggested that this process is under similar ‘hormonal’ control in A. marina as in orthodox seeds. Very high contents of cytokinins, particularly ZR, were present in both axes and cotyledons during reserve accumulation. This is thought to be related to the nature of the reserves accumulated (soluble sugars), rather than to the phenomenon of desiccation sensitivity. With the exception of ABA, embryonic contents of PGRs were relatively high at seed shedding, consistent with the rapid germination of this highly recalcitrant seed. ABA contents in the embryo were low during reserve accumulation, but concentrations in the pericarp increased throughout this development stage. ABA in the perticap could act to prevent precocious germination. The low concentrations of ABA in the embryo could be related to the desiccation-sensitivity of the seeds of A. marina.

The interactions between water content and temperature on freezing and desiccation damage were examined for Zizania (wild rice) embryos at several stages of maturity. The water content of excised embryos was manipulated by flash drying at 35°C or room temperature to between 2.5 g H2O/g dw and 0.05 g/g. Embryos were then exposed to temperatures ranging from 5 to −50°C. Viability following the drying and cooling treatments was assayed by leakage of electrolytes and germination in culture. Viability of embryos decreased when embryos were dried below a critical water content. The critical water content was greatest for the least mature embryos. Critical water contents were also temperature dependent and increased with decreasing temperature. Even though the critical water content varied with developmental status and temperature, the water activity corresponding to the critical water content appeared to be constant at 0.90. The most mature embryos survived temperatures as low as −50°C while the least mature embryos survived only to −18°C. These trends were predicted by ‘phase diagrams’ based on the physical properties of water in embryos at different stages of maturity (Vertucci et al., 1994a). Our results confirm the earlier prediction that long term preservation of Zizania grains is possible at −20°C and the degree of success will be related to the maturity status of the embryos.

Plants undergo a series of physiological, biochemical and molecular changes in response to adverse environmental conditions or stresses such as drought, low temperature or high salt. Several genes and their corresponding proteins have been described that may play a role in withstanding water-deficit-related stresses or full desiccation. In particular, sugars and late-embryogenesis-abundant (LEA) proteins have received the most attention. Plant responses to water-deficit and desiccation have been well-characterized at the molecular level; however, pinpointing the precise roles of the gene products in protecting cells under conditions of water deficit remains a challenging task. While few plants are capable of withstanding full desiccation, most seeds undergo this event as a pre-programmed and final stage in their development. These are the so-called ‘orthodox’ seeds. In contrast to seeds of orthodox species, those of recalcitrant species do not acquire desiccation tolerance during their development and are shed from the parent plant at relatively high water contents. The essential components of desiccation tolerance of seeds are likely to involve the ability to effect repair upon subsequent rehydration as well as the ability to accumulate protective substances that limit the amount of damage which otherwise would be caused by water loss. Studies have begun to examine whether the desiccation sensitivity of recalcitrant seeds is at least partially the result of an insufficient accumulation of LEA-type proteins, or whether other factors (including a lack of protective sugars) are more important. This review assesses some of these studies as well as recent research to understand gene and protein function using transgenic host plant systems.

Water contents, desiccation tolerance, respiratory rates and subcellular characteristics of three contrasting seed types were studied during development. Avicennia marina (a tropical wetland species) and Aesculus hippocastanum (a temperate species) produce recalcitrant seeds and Phaseolus vulgaris produces orthodox seeds. During development, A. hippocastanum and P. vulgaris seeds showed a decline in water content and respiration rate with a concomitant increase in desiccation tolerance. These parameters did not change during the development of A. marina seeds once they had become germinable. There was a decrease in the degree of vacuolation and an increase in the deposition of insoluble reserves in A. hippocastanum and P. vulgaris seeds, while A. marina seeds remained highly vacuolated and did not accumulate insoluble reserves. Mitochondria and endomembranes degenerated during the development of A. hippocastanum and P. vulgaris seeds, but remained unchanged in A. marina seeds. The data are consistent with the hypothesis that extensive vacuolation and high metabolic rates contribute to desiccation sensitivity. However, the development of recalcitrant A. hippocastanum seeds is similar to that of orthodox P. vulgaris seeds. These data are in accord with the concept of seed recalcitrance being a consequence of truncated development. The results suggest that there may be three categories of seeds: orthodox seeds which develop desiccation tolerance, seeds which show similar development to orthodox seeds, but are shed before desiccation tolerance is well developed, and seeds which show no developmental trends giving rise to increased tolerance.

The relationships among water content, relative humidity and temperature were documented in both seeds and excised embryos of Coffea spp. using water sorption isotherms. Isotherms were constructed at 5, 15 and 25°C and calculated for lower temperatures. There were no apparent differences in sorption characteristics among whole seeds of several cultivars of C. arabica and among different species of Coffea. Excised embryos of genetically diverse Coffea germplasm also exhibited similar sorption characteristics, though there were substantial differences observed between embryos and whole seeds. The shape of isotherms of coffee seed tissues was intermediate to the reverse sigmoidal shape observed for orthodox seeds and the monotonic shape observed for desiccation intolerant plant tissues. The heats of sorption calculated for RH ≤ 25% for whole seeds of Coffea spp. were similar to orthodox seeds. In contrast, the heats of sorption calculated in the same RH range for excised embryos were intermediate between those of orthodox and recalcitrant embryos. Our observations are consistent with earlier observations that desiccation sensitivity or poor longevity is linked with low levels of water sorption at relative humidities less than 25%. An explanation for this remains elusive.

There is a substantial literature on the basic physiology and response to desiccation of recalcitrant seeds, but little is known about their ecology and even less of their evolutionary status. It is difficult to assess the response of early land plants to dehydration, but it is likely that desiccation tolerance in vegetative tissue arose concomitantly with the invasion of the land. Similarly, from the fossil record it is not possible to assess the desiccation response of early seeds, and furthermore, it is difficult to see phylogenetic relationships among species producing recalcitrant seeds. A consideration of the available evidence, however, suggests that the first seeds were desiccation-sensitive, but tolerance evolved early and probably a number of times, independently. The desiccation sensitivity and short life span (generally shorter than the interval between flowering) of recalcitrant seeds have implications in terms of regeneration ecology. A long-term soil seed bank as such does not exist; rather the seeds germinate and form a seedling bank. However, there is a wide range in post-shedding physiology among recalcitrant seed species, and although species producing recalcitrant seeds are common in the humid tropics, they do occur in habitats with more marked seasonal variation. Here regeneration strategies may be more specialized.

Recalcitrant seeds are shed moist from the plant and do not survive desiccation to the low moisture contents required for prolonged storage. It has been widely hypothesised that during desiccation of these seeds a stress induced metabolic imbalance develops that leads to free radical mediated damage and viability loss. We investigated this hypothesis in a comparison of two sympatric species of Acer during late seed development and post-harvest desiccation: A. platanoides (Norway maple) has orthodox seeds and A. pseudoplatanus (sycamore) has recalcitrant seeds. In both species, respiration rates declined to similar levels at shedding, and the extent of defences against free radicals appears no less in sycamore than that in Norway maple. During drying there was no evidence for the accumulation of a stable free radical, increased lipid peroxidation or decline in free radical scavenging enzymes in either species. In addition, there was a very similar, large increase in total tocopherol in both species. This increase in sycamore was largely of alpha-tocopherol, whereas in Norway maple the increase was largely from its precursor, gamma-tocopherol. Arguably this suggests a similar mechanism in both species, but increased oxidative stress in sycamore. In general, the results suggest that, although damage resulting in viability loss was clearly taking place, the limitation to desiccation tolerance did not result from inadequate free radical scavenging. Soluble carbohydrates and dehydrin-like proteins were also measured during late seed development and drying in sycamore and Norway maple. The greater concentrations of sucrose, raffinose and stachyose and amounts of dehydrins in the radicles and cotyledons of Norway maple compared with those in sycamore was consistent with greater desiccation tolerance in the former. Sycamore seeds are dormant and at the tolerant end of the continuum of desiccation sensitivity among recalcitrant species, and this may account for their different response to that of the seeds of other more sensitive recalcitrant species studied.

A suite of mechanisms or processes that together have been implicated in the acquisition and maintenance of desiccation tolerance in orthodox seeds is discussed in the context of the behaviour of desiccation-sensitive seeds, and where appropriate, parallels are drawn with the situation in vegetative plant tissues that tolerate dehydration. Factors included are: physical characteristics of cells and intracellular constituents; insoluble reserve accumulation; intracellular de-differentiation; metabolic ‘switching off’; presence, and efficient operation, of antioxidant systems; accumulation of putatively protective substances including LEAs, sucrose and other oligosaccharides, as well as amphipathic molecules; the presence and role of oleosins; and the presence and operation of repair systems during rehydration. The variable response to dehydration shown by desiccation-sensitive seeds is considered in terms of the absence or incomplete expression of this suite of mechanisms or processes.

Three categories of damage are envisaged: (i) reduction in cell volume which can lead to mechanical damage; (ii) aqueous-based degradative processes, probably consequent upon deranged metabolism at intermediate water contents. This is termed ‘metabolism-induced damage’ and its extent will depend upon the metabolic rate and the rate of dehydration; and (iii) the removal of water intimately associated with macromolecular surfaces leading to denaturation: this is referred to as desiccation damage sensu stricto. The effects of drying rate and the maturity status of seeds are considered in relation to the responses to dehydration, leading to the conclusion that the concept of critical water contents on a species basis is inappropriate. Viewing seed postharvest physiology in terms of a continuum of behaviour is considered to be more realistic than attempting precise categorization.

Rapid dehydration of excised embryonic axes (or other explants) from desiccation-sensitive seeds permits retention of viability (in the short term) to water contents approaching the level of non-freezable water. This opens up the possibility of long-term conservation, by cryopreservation techniques, of the genetic resources of species producing non-orthodox seeds.

Recalcitrant seeds of Trichilia dregeana were stored at 16 or 25°C, either at the water content at which they were shed or partially dried. Although having been exposed to a short period (approx. 6 h) at temperatures up to 30°C prior to storage, seeds at the original water content maintained viability for several weeks at 16°C. However, storage of undried seeds at 25°C was deleterious within 8 d, indicating a chemical basis for degeneration of hydrated recalcitrant seeds. Seeds that had been mildly dehydrated to the relatively high axis level of 1.68 g H2O g—1 dry mass, while maintaining full germinability immediately after drying, exhibited only 4% viability after 8 d in storage at 16°C and had completely lost viability after the same storage period at 25°C. Ultrastructural features characterizing hydrated seeds included indications of enhanced activity associated with initial exposure to the elevated temperature as well as some signs of stress. However, over an effective 15 d storage period at 16°C, ultrastructural features showed the cells to have retained little damage and to have been in an enhanced state of activity commensurate with ongoing development towards germination. After a longer storage period, however, signs of damage, including indirect evidence for disarray of the cytoskeleton in some axis cells, became apparent in line with the declining seed viability. Immediately following dehydration from an average axis water content of 1.97 to 1.68 H2O g g—1(the sub-imbibed condition), some ultrastructural abnormalities were apparent, but the seeds remained 100% germinable. However, within the 8 d storage period in this sub-imbibed condition, a spectrum of severe ultrastructural degeneration, including indirect evidence of the collapse of the nucleoskeleton and extensive cell lysis, accompanied viability decline of the seeds to 4%.