1 56 PRACTICAL BACTERIOLOGY. 



air. Thus germs of species, which before were not 

 present in the water, may make their way into it."^ 



With care, of course, this can only rarely occur, 

 still, when accurate analyses are required, a method 

 must be employed, which excludes the possibility of 

 such accidents. 



Suitable vessels, which satisfy all requirements, may 

 be very easily prepared. Thin test-tubes are selected 

 and are heated strongly at about their middles in a 

 gas flame or better in a blow-pipe flame. The heated 

 part is then drawn out into a thin tube of about 5 to 

 10 cm. long. The test-tubes are then stopped up 

 with plugs of wadding, and are sterilised in the hot- 

 air steriliser. By means of this simple arrangement 

 we may dispense with all the various kinds of vessels 

 which are recommended for the reception of the 

 samples of water. 



If the sample is to be taken from a conduit pipe, 

 the water must be allowed to run for a few minutes; 

 if a spring water is to be examined, it must be pumped 

 up for at least five minutes, to get rid of all the 

 germs, which are sure to be present in the pipes. If 

 the sample is to be taken from a draw well or from 

 standing water, it must not be derived from the sur- 



' That the flasks if properly rinsed are uninjured by the 

 sublimate, has been proved by the author by means of 2,000 

 experiments. The best methods may fail in inexperienced 

 hands. 



