An investigation is described which shows the type of skin reaction which may be expected from the intracutaneous injection of 0·1 c.c. normal horse serum in the normal and sensitized human subject. 150 normal individuals and 100 who had received previous serum injection were skin-tested. Of the normal subjects 30% showed a reaction of skin sensitivity, while 88% of the sensitized group gave a sensitive response, the reactions in the latter being more intense and of longer duration than those of the normal group. It is also shown that general skin sensitivity, demonstrated by a sensitive skin reaction in the opposite arm 2 weeks later, can be produced by intracutaneous injection of 0·2 c.c. of normal horse serum.

The value of specific skin tests in the prognosis of serum reaction is considered and a series of tests on 200 diphtheria cases is recorded. Since it has been shown that diphtheria has no appreciable effect in diminishing skin sensitivity (Westwater, 1935) the present series probably gives an accurate estimation of serum skin sensitivity even in the presence of acute diphtheritic infection.

The incidence of serum reaction was found to be only 6% higher in those cases which gave a sensitive skin reaction to horse serum prior to therapeutic serum injection than in those showing a normal response. It is therefore concluded that the demonstration of skin sensitivity is of no value in predicting the subsequent development of general serum reaction, and skin testing with horse serum, although it may demonstrate the presence of specific antibodies in the skin cells, is not an index of general sensitivity, which depends upon individual reactivity and power to produce antibodies following injection of foreign serum.