In previous papers1 I show that hemolyzing streptococci from scarlet fever, which form a specific immunologic group, produce a slightly granular conical colony2 on heated blood agar (chocolate agar), which is not changed in color, in contrast to those peculiar to erysipelas and septic sore throat, which turn this medium green. According to McLeod and Gordon3 the greening of chocolate agar is probably due to the production of hydrogen peroxide. Chocolate agar is prepared by adding 8 cc. of defibrinated sheep blood to 150 cc. of melted and slightly cooled Liebig' s beef extract plain agar, pn 7.0, heated to 90 C, and cooled to 50 C. and pouring the mixture into petri dishes. Freshly poured plates are streaked and either incubated at 37 C. for twenty-four hours and then left at room temperature four or five days, or incubated at 34 C, since the green color is produced slowly by hemolyzing streptococci and as a rule not above 34 C. Dextrose and lack of oxygen inhibit greening. Saelhof4 found that differences in color can be produced by using pure hemoglobin and that a more marked greening of heated hemoglobin agar by hemolyzing streptococci occurred with sheep and rabbit hemoglobin than with horse and human. Some of the colonies of dissociating scarlatinal streptococci, including mucoid, granular convex and conical, turn chocolate agar green but lose this power when reverted to cocci forming conical colonies, immunologically specific for scarlet fever. Stabilized rough colonies do not color chocolate agar. During the course of scarlet fever, not only conical colonies not changing chocolate agar, but non-specific granular, smooth convex and conical colonies are observed which turn chocolate agar green.6 The cocci from these colonies when reverted to S. scarlatinae, colonially and immunologically, lose their ability to change the color of chocolate agar.