Germination of winterfat ( Eurotia lanata (Pursh) Moq.) seeds at reduced water potentials: testing assumptions of hydrothermal time model

Abstract

The hydrothermal time model not only quantifies seed germination progress as affected by temperature and water potential, but also has ecological and biological significance. Assumptions of the hydrothermal time model were tested using two non-dormant seed collections of winterfat ( Eurotia lanata) with two seed size classes. Winterfat is a native shrub with superior forage quality. Germination rates (GR) of subpopulations were estimated from germination time courses over a water potential range from 0 to −1.33 MPa at 2, 5, 10, 15, 20, and 25 °C. Parameters of the hydrothermal time model were estimated from the relationships between GR and temperature at various water potentials and between GR and ψ at various temperatures. Model assumptions were tested using these estimated parameters. Results indicate that base temperature ( T b(50)) and base water potential ( ψ b(50)) of the 50% subpopulation were not independent of temperature and water potential. The ψ b(50) was lowest at intermediate temperatures between 10 and 15 °C, while T b decreased linearly with increasing water availability. The estimated shift rates of T b(50) with ψ were between 2.18 and 3.81 °C MPa −1 for the two collections and large seeds had a greater shift rate than small seeds. Hydro time ( θ H) was constant among subpopulations only at optimal temperatures. A linear increase of θ H with subpopulation was found at lower temperatures, especially at 2 °C. There were no significant differences in ψ b(50) between large and small seeds, but significant differences were observed in hydrothermal time requirement ( θ HT(50)), which was lower at intermediate temperatures than at either lower or higher temperatures. The predictability of the hydrothermal time model was improved especially at low temperatures when θ HT(50) was allowed to change with temperature as measured by a modified R 2 value. Changes in other parameters with temperature or water potential did not further improve the predictability of the hydrothermal time model. Therefore, further efforts in improving the hydrothermal time model should focus on variations in θ HT.