Limitations of p50 as a measure of seed longevity and the way forward

PremiseComparative studies of seed longevity assess how environmental conditions, species identity, populations, or genotypes affect seed viability loss in stored seeds. These studies commonly use the time it takes for seed viability to drop to 50% - known as p50, - as a measure of seed longevity. However, p50 is influenced by the initial seed viability. To allow comparisons regardless of the initial seed viability, standard protocols for comparative studies of seed longevity recommend using seed lots with "similar and high" initial viability, typically between 85 and 100%. However, even such range of initial seed viabilities might result in substantial variation in p50. MethodsHere, we use a modelling approach to illustrate how variation in initial viability affects p50 estimates, and to propose alternative, more robust measures of seed longevity in storage. ResultsWe show that for hypothetical accessions with identical rates of seed viability loss, variation in initial viability between 85 and 100% leads to a threefold variation in p50 estimates. Most of the unintended variation is introduced by accessions with very high viability (>95%). Restricting the initial viability to a narrower range (e.g., 85-95%) reduces but does not eliminate this bias. Alternatively, p50 can be recalculated to a standardized value of initial viability (e.g., 90%), which makes it proportional to the rate of probit viability loss without any noise. However, the most straightforward measure of seed longevity for comparative studies is the rate of viability loss itself, represented by {sigma} (sigma) from the viability equation. ConclusionThe common measure of seed longevity, p50, is affected by unintended, substantial variation caused by varying initial viability of the seed lot, and is thus suboptimal for comparative studies of seed longevity among seed accessions. More robust measures of seed longevity include the rate of viability loss, or p50 standardized to a certain initial seed viability.