In this paper the authors communicate the first results of an attempt to discover the relation which must exist between the chemical constitution and the physiological action of a substance. As the chemical constitution of the majority of physiologically active substances is unknown, they investigate the subject by examining the physiological action of a substance before and after the performance upon it of a definite chemical operation, introducing a known change into its constitution. The question is thus reduced to a problem in what may be called a chemico-physical calculus of finite variations.