Bacteria Developing at Different Temperatures 239 



collected at least three times a week throughout the year from four 

 different polluted sources; these being Merrimack River at the Intake 

 of the City Filter, and, from the North Canal at the Experiment 

 Station, river water which has been treated by coagulation and 

 sedimentation, and river water in which the pollution has been in- 

 creased by the addition of more sewage. The samples from the 

 Intake and Canal showed very similar results, taken month by 

 month, as is shown in the foregoing tables, and for this reason the 

 Intake samples have been omitted. The samples of sea water exhibit 

 certain peculiarities, and the samples from the three other sources, 

 while similar in character, represent different degrees of pollution 

 and through them we may gain some insight into the relative fluc- 

 tuations in the bacteria and in the bacterial ratios. 



As the sea waters included samples of varying degrees of pol- 

 lution, some division of the samples into groups becomes advisable. 

 A number of methods of grouping these samples have been tried, 

 the most satisfactory from the standpoint of the subject-matter of 

 this paper being to place all samples having similar numbers of 

 bacteria in one group, and to average all the results in each group. 



The results shown in Table 6 have been obtained in this manner, 

 from which it is seen that the average numbers of bacteria at 20°, 

 the numbers at 40°, and the numbers of B. coli — i. e., acid-producers 

 show a similar increase until the numbers of bacteria reach 5,000 

 per c.c, when the 40° bacteria and the B. coli drop to XQvy low num- 

 bers and again increase gradually with increasing numbers of bac- 

 teria. The ratios between the 20° and 40° bacteria show a corre- 

 sponding increase until the numbers of bacteria reach 1,000 per c.c, 

 a decrease occurring when the numbers of bacteria are between 

 1,000 and 5,000, and the ratios becoming extremely small as the 

 numbers of bacteria increase above 5,000. The same peculiarity 

 is noted for the B. coli ratios, with the exception that the ratio for 

 an average bacterial content below 100 is greater than the ratio 

 for a bacterial content between 100 and 500. The ratios between 

 the 40° bacteria and the B. coli are fairly uniform, fluctuating 

 between 57 and 83. The reasons for the peculiarities above noted 

 cannot be assigned without a careful study of the sources of the 

 various samples and a consideration of all the various factors 



