226 AIDS TO BACTERIOLOGY 



known as the ' safety change.' Perhaps the tendency for 

 these two organisms to rise to the top of a bulk of water, 

 where the influence of light is possible, allows sunlight 

 to materially contribute to their ultimate disappearance. 



Houston considers that the ' safety ' of an adequately 

 stored water may possibly come to be ' accepted so fully 

 as even to afford justification for filtration through mechan- 

 ical filters at a specially rapid rate, with the result of 

 thus compensating to a large extent for the initial cost 

 of the storage reservoir.' Houston's final conclusion is 

 that raw river water should be stored antecedent to 

 filtration, preferably for thirty days. Prescott and 

 Winslow (1913) conclude ' that any water which has been 

 stored for four weeks is practically safe.' 



A pure water when freshly drawn, even if kept in a 

 sterile flask free from aerial contamination, will (after 

 a slight diminution in bacterial content during the first 

 three to six hours) exhibit a remarkable increase in the 

 content of bacteria. Less pure waters, such as those of 

 ordinary rivers, which contain initially a large number 

 of bacteria, exhibit, when similarly treated, a much less 

 conspicuous increase in their bacterial population. Frank- 

 land points out, however, that whilst in sewage the number 

 of organisms only gradually diminishes, in pure waters 

 ' after the rapid increase in numbers follows a correspond- 

 ingly rapid decline, so that the numbers again fall below 

 those found in impure surface waters.' Multiplication 

 after sampling appears to be less in a large sample than 

 in a small one, and less when no air is admitted than with 

 free aeration. The composition of the bottle and the 

 temperature at which it is kept also influence the rate of 

 multiplication. 



Collection of Water Samples. Samples should be 

 collected in sterile glass-stoppered bottles. In sampling 

 a lake or stream, the bottle should be immersed with the 

 neck a foot below the surface and away from the edge 

 before the stopper is removed. A tap should be allowed 

 to run for five or ten minutes before a sample is taken, 

 as bacteria may multiply while water stands in the pipes. 

 When sampling a public supply, a tap directly off the 

 main or on a hydrant should be used, as passage through 

 a cistern introduces a factor of contamination unfair to 

 the water. Samples should only be taken from pumps 



