Because of the immaturity of the seeds, the selected model for AS seeds was not a monotonic function of the storage duration (Fig. 1, Table 1). During storage, the viability of seeds appeared to increase in the first two to 19 months, depending on temperatures and MCs, and then started to decrease. This relationship was modeled by function y = β0 + β1x + β2x1/2 where y was the logit transformed germination percentages and x was the number of storage months. Following the same interpretation of a slope in a linear function, the first order derivative β1 + β2 / (2x1/2) (Table 1, Fig. 1) described the rates of the viability change during seed storage.
At both 5 °C and −20 °C, the viability of seeds stored with 7.7% MC increased and decreased faster than that of seeds stored with 5% MC (Table 1, Fig. 1). At the same MC level of 5% or 7.7%, the viability increased at similar rates between the two temperatures but declined much faster at −20 °C than at 5 °C (Fig. 1). Seeds with 63.8% MC stored at 5 °C quickly reached a germination percentage of 56% and then lost all the viability quickly (Fig. 1, Table 1). Seeds with 63.8% MC stored at −20 °C lost all viability before the second germination test.
The time required to reduce germination to half of the maximum germination (P50) ranged from 0.5 (MC = 65.7% and stored at −20 °C) to 221.9 (MC = 4.5% and stored at −20 °C) months for YA and ranged from 2.3 (MC = 63.8% and stored at 5 °C) to 56.7 (MC = 5% and stored at −20 °C) months for AS (Table 1). Under the same storage conditions, P50 for seeds from AS was lower than that for YA seeds (Table 1).