_{1}, X

_{2}, . . . . ., X

_{K}represent the levels of k experimental factors and y is the mean response, then the inverse model at quadratic variable polynomial response function is defined by: Y = b

_{o}+b

_{1}X+b

_{11}X

^{-1}in the single nutrient case where (X is replace by X

^{-1}in the inverse model) fitting a response surface for N, P, K by the mathematical form of the inverse response surface gives: Y = b+ b

_{1}X

_{1}+

_{ }b

_{2 }X

_{2}+

_{ }b

_{3 }X

_{3 }+ b

_{11 }X

_{1}

^{-1}+ b

_{22 }X

_{2}

^{-1}+ b

_{33 }X

_{3}

^{-1}+

_{ }b

_{12}X

_{1}X

_{2}+ b

_{13}X

_{1}X

_{3}+ b

_{23}X

_{2}X

_{3}. Arguments are given for preferring these surfaces to ordinary polynomials in the description of certain kinds of biological data. The fitting of inverse polynomials under certain assumptions is described and shown to involve no more labour than that of fitting ordinary polynomials. Complications caused by the necessity of fitting unknown origin to the X

_{i}are described. The goodness of fit and coefficient of variations are used to compare both the ordinary and inverse polynomials to fertilizer recommendation and the inverse kind shown to have some advantages.]]>