---

期 刊 ：植物遗传资源学报 2014年 15卷 1期

关键词：种子寿命；生活力；贮藏环境；交互作用；

Keywords:Longevity and viability of seed, Storage conditions, Interaction,

摘 要 ：为研究利用存活比、影响种子贮藏寿命和生活力的主要贮藏环境因子来预测高粱、苋属、甜荞麦和莴苣种子批的寿命和生活力，3个含水量水平的高粱和苋属的8个种子批被贮藏在20℃、30℃和40℃的贮藏温度下长达810d。以概率回归和存活比线性回归方法，贮藏实验获得的高粱、苋属及文献报道的甜荞麦、莴苣种子批的发芽率和贮藏时间的成套数据，被用于估算不同种子批的平均贮藏寿命（p50）和种子死亡时间分布的标准差（σ），而两种回归方程的拟合性也同时被比较；两种方法估算出的p50和σ则分别被用于对贮藏环境因子：贮藏温度（t℃）及其平方、种子含水量（m%）和它的对数（logm）以及t 和m的交互作用（t×m和t×logm）作逐步回归分析，结果表明，用存活比来估算p50、σ是可行的，而且估算出的p50和σ较用概率估算的有统计学意义；贮藏温度和种子含水量的交互作用（t×m）是使4类作物种子快速丧失生活力，减少寿命的主要贮藏环境因子，包含这个因子和存活比的数学模型：Gi/G0=A-P/10K-CWT(t×m)，预测甜荞麦种子批的发芽率较Vi=Ki- p/10K-CWT(t×m)有更高的准确性；所以，建议用前者来预测初始发芽率是已知的种子的贮藏寿命和生活力，而用后者预测初始发芽率未被测定的种子的寿命和生活力。

Abstract:To predict seed longevity and viability of Sorghun vulgare, Agopyrum escutentum, Lactuca sativa, and Amaranthus by survivor proportion and main storage condition factors, seeds of 8 seed lot of Sorghun vulgare and Amaranthus with 3 levels of water content were stored at 20℃, 30℃ and 40℃ for lasting 810 d. Data sets of germination percent and storage period of seed from storage experiments of Sorghun vulgare and Amaranthus seed lot and from literature reported for Fagopyrum escutentum and Lactuca sativa seed lot were used to estimate mean viability period (p50) and standard deviation of seeds death distribution in time (σ) for different seed lot using regression equations of probability and survivor proportion in linearity, and fitting of the equations were also synchronously compared. Moreover, p50 and σ by two equations estimated were used to analyze storage condition factors with stepwise regression method, including storage temperature(t℃)and its quadratic term（t2), water content of seed (m%)and its logarithm(logm), and interaction between storage temperature and water content of seeds (t×m and t×logm), respectively. The results showed that estimated p50 and σ with survivor proportion to estimate was feasible and had more statistical significance than probability estimated. Seed viability was quickly lost and its longevity was decreased by interaction between storage temperature and water content of seed (t×m). The mathematic model Gi/G0=A-P/10K-CWT (t×m) containing the interaction factor and survivor proportion had higher accuracy to predict initial germination percent of Lactuca sativa than the model Vi=Ki- p/10 K-CWT (t×m). Thus, the former and latter models were recommended to predict initial germination rate with longevity and viability of seed known and unknown, respectively.

---