There is some tendency toward reliance on clinical findings for early diagnosis of primary tuberculosis. Practitioners of both pediatrics and general medicine are guilty of this unjustified approach.
Many symptoms and symptom complexes have been mentioned. In the Scandanavian countries it has been noted that 95% of the cases of erythema nodosum are associated with tuberculosis.1 In the United States, however, erythema nodosum occurs much oftener in association with rheumatic fever and occurs infrequently under any conditions. At a recent round-table discussion on tuberculosis conducted by the American Academy of Pediatrics, a group of approximately 35 participants was polled as to the frequency of erythema nodosum. Only one member reported the incidence of as many as two cases each year, and his residence was in northern Michigan, a climate comparable to that of the Scandanavian countries.
In the past, presence of phlyctenular conjunctivitis has been considered almost pathognomic of