“…In the first, longevity statistics are used as the criterion for comparing treated and untreated groups, the latter having usually been observed some years previously (11,15,16). In the other, the effects of various dietary measures are compared either in the same patient or in small groups, using serial clinical, functional and, in some instances, histological observations as the basis for comparison (13,14,17,18). There are valid objections to both methods, but serial observations during alternating courses of therapy in the same subject appear to offer the most practical means for conducting experiments under controlled conditions.…”