86 



SCIENCE 



[N. S. Vol. XXXIX. No. 994 



of the water and earth. The former usu- 

 ally occurred in chains or small irregular 

 groups, reacted positively to the Gram 

 strain, formed a meager or only fair sur- 

 face growth on solid media, and produced 

 considerable acid in carbohydrates. The 

 cocci of the water and earth occurred in 

 large cell groups or packets, never in 

 chains, were usually Gram negative, grew 

 abundantly on solid media and generally 

 failed to ferment carbohydrates. There 

 were exceptions of course, as there always 

 must be in an unstable group like the 

 bacteria. Some organisms which a general 

 consideration of all their characters would 

 place in the latter group were found on the 

 skin, while others were Gram positive or 

 fermented the sugars. Yet on the whole the 

 relation seemed a sufficiently definite one 

 to warrant the division of the spherical 

 bacteria into two subfamilies, the Para- 

 coccaceag and the Metacoccaceaa. The next 

 thing which was apparent was that the 

 color of the pigment produced, instead of 

 being a minor varietal character, was fun- 

 damentally correlated with other proper- 

 ties which were apparently of sufficient 

 importance to deserve generic rank. It 

 appeared that the orange and white staphy- 

 lococci, along with the diplococci and 

 streptococci, all shared the properties of 

 the Paracoccacese just enumerated, while 

 the yellow and red pigment formers (in 

 spite of the occasional presence of the 

 former on the skin and even in connection 

 with pathological processes) exhibited the 

 characters of the Metacoccaceae (Fig. 4). 

 The white and orange forms further 

 differed from each other in the fainter 

 surface growth of the former and in the 

 important fact that liquefying members of 

 the orange series liquefy twice as rapidly 

 as do the liquefying white strains. Hence 

 we distinguished these groups as the genera 

 Aurococcus and Albococcus. Among the 



Metacoccacese the yeUow and red forms 

 were sharply separated by the much higher 

 proportion of strains which reduce nitrates 

 to nitrites and by the absence of ammonia 

 formation in nitrate pepton broth and by 

 the rarity and slowness of liquefaction, 

 among the red chromogens, for which we 



HAHTAT CRISTA" 3Mia^^(}<m C^ffiE LyoS 



PARACOCCACEAE 



I 



METACOCCACEAE 



I I 



Fig. 4. Group Dippekences between the Paka- 

 coccACE^ AND THE Metacoccace^. The Upper 

 blocks show for 221 strains of coeoi whose type of 

 ehromogenesis (white or orange) would place them 

 with the parasitic subfamily, the per cent, of 

 strains fulfilling each of five individual criteria of 

 that group. The lower blocks show for 279 strains 

 of cocci whose type of ehromogenesis (yellow or 

 red) would place them with the saprophytic sub- 

 family the per cent, of strains fulfilling each of 

 five individual criteria of that group. 



suggested the generic name Rhodococcus. 

 The important point brought out by these 

 studies was that ehromogenesis and the 

 Gram strain, characters which we all had 

 believed to be comparatively unimportant, 

 proved to be correlated with a number of 

 other properties and therefore highly sig- 

 nificant. On the other hand the Migula 



