植物育种田间试验设计的研究进展

ADVANCES IN FIELD EXPERIMENTAL DESIGN FOR PLANT BREEDING

概述了近５０年来应用于植物育种田间产量选择试验设计的研究进展。在育种早期选择试验阶段，由于材料众多，常常不设重复，利用对照或若干个邻区估计试验小区肥力差异。而在育种中、后期选择试验阶段，要求试验的精确度高，现有两种途径降低试验误差：（１）根据局部控制原理适当减少区组内小区数，即利用不完全区组设计；（２）根据相邻小区土壤条件的相似性，采用近邻分析。

Field experimental designs for plant breeding developed in last fifty years were reviewed. At the earlyyield testing stage of a breeding plsn,enperimental designs without replications were used for large number of lines,where the yields of test-line plots were adjusted with some neighbouring Polts or control plots. At the late stage, incomplete block designs including tattice and generalized lattice and the nearest neighbour analyses were recommended to reduce experimental error and improve precision in testing a farge number'of lines, while the conventional design, suchas randomized complete blocks and Latin squares, could be used to test relatively not large number of lines.