Bacteria Developing at Different Temperatures 253 



ings from cultivated fields, etc., when averages of a large numijcr 

 of samples extending over a considerable period are included. The 

 ratios between the bacteria and B. coli on samj)k's from the two sources 

 do not agree, the highest ratios being obtained for the canal samples 

 when the river was low, and the lowest ratios when the river was 

 high. With the Intake samples, however, the lowest ratios occurred 

 at a time of medium high water, and the highest at extreme high 

 water. 



TABLE 19. 

 Relation BETWEEN Volume OF Flow AND the Bacterial Content op Merrimack River Water. 



CONCLUSIONS. 



The apparent discrepancies which have occurred between the 

 results of bacteriological and of chemical analysis of water have 

 caused a reasonable doubt in the minds of many persons, having 

 occasion to use such results, as to the practical value of the bacterio- 

 logical determinations. That such discrepancies do exist cannot 

 be denied, and that the bacteriological results instead of the chemical 

 results should be doubted is natural, considering that the complete 

 chemical analysis is composed of a number of individual factors, 

 each of which has received long and careful study, while the bacterio- 

 logical procedure is confined to a determination of the numbers of 

 bacteria, and of the presence or absence of one specific type of bac- 

 teria, i. e., the colon type. If the chemical analysis of water were 

 confined to a determination of the total nitrogen content, instead 

 of dividing that nitrogen content into its constituent parts — free 

 ammonia, albuminoid ammonia, nitrates, and nitrites — and only a 

 qualitative test were made for chlorine, the interpretation of the char- 

 acter of the water from the chemical results would be as frequently 

 in error as when a similar interpretation based on the usual bacterio- 

 logical results is attempted. If, on the other hand, complete and 



