BY R. GREIG-SMITH. 



535 



reflux condenser. The suspensions were incubated overnight, 

 usually for 20 hours, at 22°, and were then counted by the plate- 

 method. The experimental error was minimised by making four 

 tests of each, and making five plate-counts of each test. 



Experiment i. 



The following numbers for raw filtered water are taken from 

 the control-tests of other experiments. 



100 cells of Bac. prodigiosits became 



May 25th, 1914 

 June 4th, 1914 

 June 12th, 1914 

 June 18th, 1914 

 June 30th, 1914 

 July 2nd, 1914 



74 

 49 

 71 



136 

 86 



112 



The action of unboiled and boiled filtered tap-water upon Bac. 

 prodigiosus is irregular, possibly on account of the water itself 

 varying, but the general tendency is for the boiled and cooled 

 water to retard the growth of the bacteria suspended in it. 



