对两个亚热带优质蛋白玉米群体的适应性混合选择研究 Ⅱ.相关响应

Effects of Mass Selection on the Adaptive Improvement of Exotic Quality Protein Maize Populations Ⅱ. Correlated Responses

在两个源于亚热带的优质蛋白玉米(QPM)群体,中群13和中群14中,对早开花性状进行4轮混合选择,籽粒产量和植株形态特征发生显著的相关响应。1989至1991年在北京作的3年评价试验中,中群13和中群14平均每轮每亩籽粒产量分别增加41.8(28.7％)和41.5kg(29.0％),与线性回归响应(b=37.4和40.4kg)相吻合。1991年在北京、济南、武功、成都和海南三亚作的5点联合试验中,平均每轮增加产量13.3(5.4％)和12.5kg(5.1％),与线性回归响应(b=12.7和1O.1kg)吻合。籽粒产量增长的主要原因是空株率减少,单株有效穗数增多。本试验中产量与单株穗数有高度正相关(r=0.96)。经过前3轮混合选择,单株结穗数分别从0.55增加到0.98和0.99,平均每轮增益26.1％和26.7％。籽粒产量增加的第二个原因是提高了经济系数,经过前3轮选择,分别从0.29和0.28提高到0.44和0.42。同时,单株穗粒重也有所提高,平均每轮增加6.1和7.3g。在多点联合试验中,改良群体的株高和穗位高度呈下降趋势,但统计上差异不显著。单株叶片数则显著减少,平均每轮减少0.49和0.35片。多点联合方差分析表明,籽粒产量和叶片数表现出显著的基因型×环境互作效应,这可能与各轮群体对光周期敏感性的差异有关。

Four cycles of mass selection for early silking were conducted in two subtropical quality protein maize (Zea mays L.) (QPM) populations, C, Pop 13 and C. Pop 14. The grain yields of populations presented significant response to the selection. The evaluating experiments in Beijing (long day condition) showed that the average gains (over 3 years) for yields per cycle of C. Pop 13 and C.Pop 14 were 626 kg · ha-1 (28.7%) and 623 kg· ha-1 (29.0%), and the linear responses were 561 kg· ha-1 and 606 kg · ha-1 respectively. A multi-located evaluating experiment (Beijing, Shaanxi, Shandong, Sichuan and Hainan) in 1991 showed that the average yield gains per cycle were 199 and 188 kg · h-1, respectively.The yield gains are associated with the increase of available ears per plant and grain weight per ear. The ears per plant were 0.98 and 0.99 for the advanced cycles vs. 0.55 for the original populations, giving an average response of 26.1% and 26.7% per cycle, respectively. The harvest index were concommitantly increased from 0.29 and 0.28 for CO to 0.44 and 0.42 for C3 as a result of remarkable increase of grain yield per plant.The plant height and ear height of advanced cycles were reduced but not significant statistically. Leaves per plant were decreased significantly with an average gain of-0.43 and -0.35 per cycle. The pooled variance analysis showed somewhat that the genotype×environment interactions existed in most characters under evaluation, which maybe were caused by photosensetivity of different cycles.