The conception that erythema nodosum is an expression of rheumatism dates back to Mackenzie1 (1886). This theory, which was mainly supported by the occurrence of arthralgias, has found most of its advocates among Mackenzie's countrymen; in England it was the prevalent opinion for at least two or three decades. But during the last fifteen or twenty years the theory of a rheumatic cause has lost ground proportionately as the evidence has become more and more convincing that erythema nodosum is, as a rule, caused by tuberculosis. It is now fairly generally held by pediatricians in Scandinavia that tuberculosis is the cause of the erythema in from 75 to nearly 100 per cent of children afflicted. This view seems to be shared by a number of pediatricians in the United States and in England.
However, investigators have been disinclined to give up the idea that in certain cases there may