(43pp.) The failure of Hume's "constant conjunction" to describe apparently causal relations in science and everyday life has led to various "probabilistic" theories of causation of which Suppes (1970) is an important example. A formal model that was developed for the analysis of comparative agricultural experiments in the first quarter of this century can be used to give an alternative account of "probabilistic causality" that does not take a stand as to the stochastic or deterministic nature of the causal connection. The approach has many applications in social, behavioral, and medical science. This paper discusses the formal model in detail, applies it to "probabilistic causation," and compares the resulting theory to Suppes' theory.