论文部分内容阅读
本文论述了定量化施肥技术的发展过程和施肥模型的地位，重点讨论了我国在应用经验模型推荐施肥方面存在的问题和拟解决的途径。主要的问题是，广泛采用的二次多项式模型使推荐施肥量偏高;试验因素水平数少，不能正确揭示肥料效应变化趋势;边际分析时忽视了函数极值判别和产投比。据此，建议今后应减少试验因素，适当增加因素水平;提倡相交直线施肥模型;边际分析前应对函数极值进行判别。
This paper discusses the development of quantitative fertilization technology and the status of fertilization model, and discusses the problems in proposed fertilization application in China and the ways to be solved. The main problem is that the widely used quadratic polynomial model makes the recommended fertilization rate high; the small number of experimental factors can not correctly reveal the trend of fertilizer effect; and the marginal analysis neglects the function extremum discrimination and the production-input ratio. Accordingly, it is suggested that in the future, the experimental factors should be reduced and the factor levels should be appropriately increased; the intersecting linear fertilization model should be promoted; and the extreme value of the function should be discriminated before the marginal analysis.