Distribution keys are used in crop models for the description of assimilates partitioning between several plant organs. Standard approach is to use table data for discrete points as functions of a physiological age and to interpolate these relations between the given points. A universal function can be proposed for the construction of continuous relations for a majority of the known distribution keys. This possibility is discussed in the paper. Another approach concerning the description of a shoot/ root relation in the model is also presented. The main idea is to take into account N:C interaction in plant for the determination of a fraction of assimilates which must be allocated in roots or remains in leaves. Indeed, this fraction is changed during vegetation as influenced by environmental conditions. The last approach can be called "adaptive distribution key".