152 C.-E. A. WiNSLOw AND Anne F. Rogers 



in the Massachusetts Institute of Technology. In collecting them 

 each subject rubbed the front and back of one hand with a wet wad 

 of sterile cotton, running the wash water into a sterile cup. Finally 

 a small number of cultures were obtained from excreta of man and 

 animals. Under Habitat III cultures were obtained from a wide 

 variety of natural waters — public supplies, streams, ponds, pools, 

 shallow wells, driven wells, and the sea. Samples were taken as 

 far as possible only from sources held to be free from pollution. 

 Under Habitat IV organisms were isolated from various samples 

 of earth, loam, clay, sand, etc., obtained mainly in different regions 

 of eastern Massachusetts. The cultures grouped under Habitat V 

 were taken from plates exposed to the air, indoor and out, and here 

 are also included certain organisms of unknown origin which appeared 

 as contaminations, or for whose previous history we have no record. 

 In each case the sample to be studied was first plated on agar 

 and incubated at 20°. Colonies which looked like cocci (not pos- 

 sessing, that is, the characters of such well-marked forms as B. 

 mesentericus, B. Zopfi, or B. fluorescens) were fished to agar streaks; 

 from each sample only one culture was taken, unless several distinct 

 types of colonies appeared. The agar streak cultures were examined 

 under the microscope and, if apparently cocci, were replated in order 

 to insure their purity, again transferred to agar streaks, and again 

 examined under the microscope. All this preliminary work was 

 carried out at 20°, and the stock cultures finally obtained were kept 

 on agar at the same temperature. There can be no doubt that by 

 this method of procedure we failed to obtain many of the more strictly 

 parasitic streptococci which grew only feebly on solid media and 

 are most active at a temperature of 37°. This fact must be taken 

 into account in interpreting our results. For Micrococcus and 

 Sarcina, however, the series should be fairly representative. 



2. Selection of Characters for Study. 



The characters ordinarily used in descriptive bacteriology are few, 

 particularly in a group of such simple morphology and limited bio- 

 chemical powers as the Coccaceae. This number must be still further 

 reduced, however, when we come to inquire which of them really 

 indicate constant and independent variations. In the first place, it 



