摘要
就多因素肥料试验及多元肥效模型研究中存在的问题,提出了以因素水平在因子空间分布的专业合理性、主效应的可比性和吸收回归设计的优点作为多元肥效试验的设计原则:认为可用拟合度检验-计算机模拟试验-模型极大值存在判定作为多元肥效模型的选模方法.而且,在多因素肥效试验设计中,少设或不设重复而代之以扩大处理数,可使这种方法更为有效.采用上述方法,在金华地区低丘红壤上建立了超甜玉米-甜玉2号氮、磷、钾三要素肥效反应模型,根据模型,采用产量频率分析法给出了试验地区条性下,甜玉米2号产量达625～730kg/亩,经济效益达300元/亩左右的推荐施肥方案.
For the problems existing in the study of multilactor fertilizer experiment and multi-variable fertilizer response model, it was suggested that the rcasonal distribution of factor's level in the factor' s space, the comparison of the main factors and the advantadges of regressive design should be taken as the designing principle of fertilization field experiment; a test of parameter fitness ( a test of experimentprecision)-judgement on the existance of model maximum ( a test of fertilizer response correspondingto the law of diminishing return ) was put forward as a method of selecting multivariable fertizer response model, and by which a fertilizer response model of supersweet corn i.e. sweet corn Ⅱ to the three essentialfertilizers-N, P, K. on the red soil of low hill in Jinghua region, Zhejiang Province was established.According to ths model, yield frequency analysis worked out the prepositional applying value of the fertilizers in sweet corn Ⅱ with the plant yield up to 625-730 kg/mu and the economical benifit up to 300 yuan/mu or so.
