Seed moisture content, storage, viability and vigour

Vertucci and Roos (1990) have proposed that moisture contents in equilibrium with 19-27% relative humidity (RH) are optimal for the longevity of orthodox seeds in storage whereas we have suggested that those in equilibrium with about 10-11% RH maximize the benefit to longevity from seed desiccation (Ellis et al., 1989, 1990a,b; Roberts and Ellis, 1989). These different recommendations are discussed here, not only for scientific interest, but because immediate decisions on conditions for long-term seed storage for genetic conservation cannot be avoided since they are already implicit in current practice. Drying increases the longevity of orthodox seeds in a predictable way (Roberts, 1973), and it has long been known that they can survive considerable desiccation; for example, neither the viability nor the vigour of seeds of five gramineous species was adversely affected by drying to 1 % moisture content (Harrington and Crocker, 1918). Moreover, desiccation to very low moisture contents, typically to 1-2%, can be advantageous to longevity (Evans, 1957; Nutile, 1964). Because of the long periods required for loss of viability in dry seeds, however, knowledge of the moisture content(s) at which orthodox seed longevity is maximal has been lacking. The effect of seed moisture content on longevity can be quantified by a negative logarithmic relation (Ellis and Roberts, 1980a). This relation continued down to 2.1% moisture content in sesame (Sesamum indicum L.) stored at 50°C, but 4 years of experimental storage was required to obtain a seed survival curve in this environment (Ellis et al., 1986). And so in more recent work we have stored seeds at 65°C to obtain complete seed survival curves at low moisture contents within 6-12 months. By so doing we detected a low-moisture-content limit (mc) to the negative logarithmic relation between seed longevity and moisture in air-dry storage and more often than not we found that desiccation below mc (by 1-3%) had no further effect on longevity (Ellis et al., 1988, 1989, 1990a,b). These investigations in diverse species exposed considerable variation in the value of mc which appears to result from differences in seed