WATER FILTRATION. XXVII. 



report of Professor Boyce, f.r.s. of Thompson- Yates 

 Laboratories of University College. The water is obtained 

 from Lake Vyrnwy in Wales, and Rivington District near 

 Liverpool, and several wells, all being filtered. The aver- 

 age number of bacteria per cc. for the 365 daily samples is 

 30 per cc. The samples were taken from the fountain in 

 Monument Place, London Road and from a tap in Ashton 

 Hall :— 



Vyrnwy water ... 26 average number per cc. 

 Rivington water ... 20 ,, ,, 



Mixed well water ... 33 ,, ,, 



The first two are surface waters, the last being a mixture 

 of three large wells. 



It is unnecessary to go into details of construction as 

 these will vary according to circumstances. It may be 

 necessary to provide subsiding reservoirs, and it would 

 appear that in dealing with river waters this provision is 

 absolutely necessary. Further it may be necessary to 

 interpose preliminary filtration through roughing filters 

 before final treatment, such provisions have had to be made 

 in countries where the slow system is in vogue. The Puecli 

 system of filtration combines the above. The idea is to 

 introduce a preliminary roughing filter by means of which 

 the coarser particles in suspension are intercepted, with the 

 result that the slow sand filters do not require scraping so 

 frequently, thereby saving a great amount of labour and 

 the time otherwise lost in the deposit of the film of mud 

 before filtration can commence. This system is in use in 

 Paris, and it is stated that the slow filters will run three 

 or four times longer without scraping. 



The depth of sand necessary for filtration appears to be 

 adopted as 3 feet, although there are cases where a lesser 

 depth is used, but the universal practice is not to remove 

 more than 12 inches by scraping, thus leaving 2 feet as the 



