A simple, numerical system for the classification of multiple sclerosis into probable and definite categories is proposed. This system is based upon the analysis of the clinical symptomatology of 111 autopsy-proved cases of the disease from England, Norway and the United States. It is not designed to be used as an aid in the diagnosis and should be applied only to those cases where multiple sclerosis is suspected. It permits the incorporation of data obtained by means of evoked responses studies, psychological evaluations and other methods for confirming anamnestic information or delineating some asymptomatic lesions. The numerical system also provides a means for plotting the temporal evolution of the disease strictly in terms of the appearance of new signs and symptoms since recurrences of previously experienced symptoms cannot be scored; thus it has the advantage of not allowing so-called exacerbation due to physiological or psychological factors to falsely suggest disease progression or activity. The major use of this scoring system is to provide a reasonably reliable means of comparing epidemiologic data from different investigators by more clearly defining the probable and definite categories.