THE COLON GROUP OF BACILLI 125 



gas in 24 hours, 73 per cent after 48 hours, and the 

 remaining 27 per cent only after 72 hours; but the 

 Committee on Standard Methods (1912) believes that 

 the forms which fail to develop in 48 hours are attenuated 

 forms of little practical significance. Sawin (1907) 

 reports comparative results with dextrose broth and 

 bile on different classes of waters, the most striking 

 of which are tabulated on p. 124. 



Like other enrichment methods which eliminate com- 

 peting forms it is no doubt true that the lactose bile 

 test cuts out some weak colon bacilli. As a presumptive 

 method, however, it is far superior to dextrose broth, 

 giving a higher proportion of positive tests with polluted 

 waters and a lower proportion of erroneous positive 

 tests with waters of good quality. In an examination 

 of 176 surface waters in eastern Massachusetts, carried 

 out under our direction, B. coli was isolated 70 times. 

 The dextrose-broth test was positive 120 times, an error 

 of 70 per cent; while the bile test, alone, was positive 

 78 times, an error of only n per cent. The tabulated 

 results of these experiments indicates fairly the merits 

 of the bile medium for preliminary enrichment and as a 

 presumptive test. 



PRELIMINARY AND COMPLETE RESULTS OF DEXTROSE 

 BROTH AND BILE TESTS. 176 SURFACE-WATERS 



