The von Liebig Hypothesis
用Heady和Pesek的玉米对氮磷反应的经典数据，通过非嵌套假说检验，证明冯·李比希模型（含米切利希区制）最能解释实验数据，并扩展至转换回归模型，为农业经济学者提供生产函数选择依据。
Abstract Using a celebrated sample of corn response to nitrogen and phosphorus collected by Heady and Pesek, it is shown that a von Liebig model with Mitscherlich regimes is the best interpreter of the experimental data. This assertion is based upon the results of nonnested hypothesis tests among five alternatives. An extension of the test to switching regression models reinforces the superiority of the von Liebig hypothesis for this sample. Cost functions, dual to von Liebig production functions, have continuous derivatives with respect to the parameters and may be easier to estimate when input prices are available.