This article attempts to discuss the relations between the nitrogen sources in data from a large field experiment in which various levels of fertilizer nitrogen were applied to various grass mixtures in the presence and absence of white clover. It was seen that a number of effects occurred, among them the strong depression of clover nitrogen by fertilizer nitrogen (to the extent that equations predicting their contribution to grass nitrogen could not be produced directly by multiple regression but had to be constructed separately), a further apparent depression of clover nitrogen transfer effectiveness, a more significant relationship between grass nitrogen increase and the previous year's clover nitrogen than the current year's clover nitrogen, and apparent competition between grass and clover despite the latter's contribution to increased grass yield.

As some of the effects are counteractive, in each comparison made a balance appears to exist. This varies between sites and years in such a way that although the variation in grass nitrogen is always largely dependent on the clover and fertilizer nitrogen the relations are not consistent as shown by the mathematical equation representing them, and cannot be extrapolated or predicted for other sites. It is suggested that a mathematical model be constructed to simulate such an interactive system and that predictions when necessary be made using this, rather than using simple equations.