There have been relatively few investigations from which effective help could be obtained in the roentgenologic diagnosis of early scurvy, particularly of the so-called borderline cases, except the excellent contributions of Fraenkel 1 and the recent contribution of Wimberger. 2 The subject merits attention because the number of cases of mild scurvy is certainly far greater than is commonly believed. Undoubtedly, there are instances of transient pains in the extremities, lack of appetite and loss of weight, pallor and even hemorrhagic manifestations which pass unrecognized and which usually clear up when a change is made to a more rational diet, not necessarily including orange juice. Several such cases have come to our attention recently at this clinic, and we therefore decided to determine, if possible, how early a definite diagnosis of scurvy may be made by the roentgenogram and what the essential points in the picture permitting such a diagnosis