BACTERIA DEVELOPING AT DIFFERENT TEMPERATURES 249 



216 is well below 20, probably about 10; while the ratio for other 

 filtered waters appears to be somewhat higher, lying between 15 

 and 20. The pond waters are characterized by very low ratios, 

 the normal ratio being about i, 41 per cent of the samples having 

 ratios below i, and 31 per cent between i and 20. 



The ratios between the bacteria and acid-producers at jo are higher 

 than the corresponding ratios at 20. The normal ratios for the 

 river water and Applied 216 lie between 40 and 80, probably being 

 about 60 in each case. A majority of the samples from Filter No. 

 216 had ratios above 80, with 32 per cent of the samples having 

 ratios between 20 and 40, the remainder of the samples being divi- 

 ded in two equal groups, having ratios between i and 20, and between 

 40 and 60. With the effluents from the slow sand niters we again 

 find a majority of the samples having ratios above 80, with 36 per 

 cent of the samples evenly divided into two groups, with ratios 

 between 20 and 40, and between 40 and 60, respectively. Seventy- 

 five per cent of the pond- water samples have ratios above 80, while 

 the remaining 25 per cent have ratios between 40 and 60. A dis- 

 tinction between the different classes of water is seen in the number 

 of samples having ratios above 80, about 25 per cent of the pol- 

 luted water samples, 50 per cent of the filtered waters, and 75 per 

 cent of the pond waters being so characterized. 



Normal ratios between 40 bacteria and acid- producers. In study- 

 ing the ratios between the bacteria and acid-producers at 40 we 

 have available a large variety of sources and more samples from 

 each source, so that our results are more conclusive. The normal 

 ratio for the river water appears to lie just above 60, while that for 

 Applied 216 appears to be between 70 and 80. Over 55 per cent 

 of the Filter No. 216 samples have ratios above 80, thus placing 

 the normal ratio for this water just above 80. With the effluents 

 from the slow water filters we find the ratios more uniformly distrib- 

 uted, and although the normal ratio appears to be about 60, its 

 location is not distinctively marked. The pond water normal ratio 

 is located between 70 and 80, and the ratios are distributed less widely 

 all being above 20, as was the case with the samples from Applied 

 216 and Filter No. 216. The sea waters are again divided into two 

 groups, 19 per cent of the samples having ratios less than i, while 



