E. W. Stevick's study of Ganda (1969) is an important contribution to our understanding of a complicated tone system, both through the new data on conjugation which it makes available and through the rules which allow of a unified treatment of nominal and verbal units as they appear either alone or in the sentence. The present paper is an attempt to make use of some of the possibilities offered by Stevick's work, especially by reducing the unanalysed types of verb stems to sequences of underlying formatives to which appropriate rules can apply, and by interpreting junctures in phonological terms. Several questions, e.g. rules pertaining to enclitics, had to be left out of account; this is due to the fact that the present author had no possibility of working with an informant for this analysis (no more than for any of his previous work on Ganda).