A Population-based Threshold Model Describing the Relationship Between Germination Rates and Seed Deterioration

Abstract
The time required for germination increases prior to the loss of viability as seeds deteriorate during storage. We have developed a population-based threshold model that can quantify and predict seed germination times and percentages after various ageing periods under constant conditions. The model assumes that each seed has a maximum potential lifetime, the values of which vary in a normal distribution among individual seeds, and that the time to germination of a particular seed is inversely proportional to the remaining difference between the accumulated ageing period and the maximum lifetime of that seed. This model described with reasonable accuracy the germination time-courses of lettuce (Lactuca sativa L.) seeds after increasing periods of controlled deterioration at 10% moisture content (fresh weight basis) and 40 °C. In seeds of high initial quality, two phases of deterioration were detected, a relatively slow phase before significant loss of viability occurred and a more rapid phase once viability began to decline. Prehydration and redrying treatments that shortened the time to germination also reduced or eliminated the initial lag phase before the loss of viability began. Seed germination rates after different ageing periods can be normalized on a common ‘ageing-time’ scale, emphasizing the continuous and progressive nature of deteriorative processes in seeds. The model predicted, and the data confirmed, that there was a critical imbibition time after which no additional normal seedlings would germinate. This critical imbibition time for a given seed lot was independent of the ageing period. The maximum lifetime threshold model, in conjunction with previous population-based models of seed viability during storage, can provide a quantitative description of seed germination rates and normal or abnormal seedling percentages throughout the storage life of a seed lot.

This publication has 0 references indexed in Scilit: