Report on the Bacteriology of Water. 



183 



through paper nor through porcelain exerts any material effect on 

 the chemical composition of the water, such differences as appear 

 being almost within the limits of experimental error, especially when 

 it is borne in mind that (a), (6), and (c) were taken from different 

 bottles. 



Filtration of the Water through Porcelain. — The sterilisation of the 

 water without application of heat was conveniently effected in the 

 following manner : — 



A glass cylinder (A) open at both ends is tightly fitted to an 

 india-rubber stopper (CC), which is also fitted into the strong 



Fm. L 



Porcelain Filter. 



glass bottle (D), a porous cylinder of biscuit-porcelain of the con- 

 struction shown in section in (BB) also passes through the same 

 india-rubber stopper. The bottle (D) has a lateral tubulure (E) 

 which is plugged with sterile cotton- wool. The whole of this ap- 

 paratus is sterilised in position by placing it bodily in the steam- 

 steriliser for several hours on three successive days, after which it is 

 ready for use. The water to be filtered is poured into the glass 

 cylinder (A), and the tubulure (E) is connected by means of press are 

 tubing with a water-pump. On thus reducing the pressure in- 

 side (D), the water in (A) is forced by atmospheric pressure through 

 the porous cylinder (BB). In order to remove the filtered water 



o 2 



