MODERN METHODS OF WATER PURIFICATION. 337 



water required for washing purposes is not often less than 3 

 per cent, for a turbid water, but varies with the turbidity re- 

 quirements. In New Orleans it ranged from y^ to 6'5 per 

 cent, according- to the state of the water's turbidity- Not only 

 turbidity, organic matter and bacteria are removed, but also 

 colour, although our water problem is not of colour apart frpm 

 turbidity. The alum hydrate unites with the soluble colouring 

 matter in the water and decolorises it to a great extent. 



I have gone fully into the American type, or Jewell system 

 as it might be called, although the method Jewell " continen- 

 tal " has advanced far past the old Jewell which we began to 

 read about years ago. with its cypress tanks and rake agitation, 

 because I have be'en impressed with the fact that far better 

 bacterial efficiency has been attained by well-run plants of this 

 description than "it was generally supposed by those who had 

 not followed developments, or did not have personal know- 

 ledge of the working of an up-to-date, well run filtration plant, 

 as one sees thus in the larger, newer installations in America. 

 There are two Jewell plants in England that I know of — York 

 and Wolverhampton. I did not have time to see the latter, 

 but visited the works at York, the water supply of which is 

 from the river Ouse, which I was told might approximate our 

 type of water: there I found one of the older type of Jewell, 

 which had been got because the water during flood was so 

 turbid with fine clay that the sand filters clogged and blocked, 

 and failed to satisfactorily purify the water. Mr. Humphreys, 

 the Water Engineer of York, was very kind in letting me have 

 a thorough insight into the purification of their water. But 

 as they use no coagulant there, the Jewell filters are not doing' 

 the work they are capable of when run on the proper American 

 method, and they act as a preliminary roughing filter merely, 

 the water going to the old sand beds after this ; therefore there 

 is nothing typical. All the same they removed some 80 to 90 

 per cent, of the impurity in the Jewell. The rate used in the 

 sand filters was i gallon per square foot per hour. There is a 

 Jewell filter system in Delagoa Bay filtering the water of the 

 Umbelosi and giving the Company satisfaction. No full reports 

 were available on my visit to the Bay. 



Of other mechanical filters which I had the opportunity of 

 observing and enquiring into in Britain. I would mention and 

 briefly describe the Bell filters, the Mather & Piatt, both of 

 which are similar to the American in several respects, chiefly 

 that the coagulant action of aluminic sulphate is used, and then 

 the Candy filter, wdiich does not use coagulant. 



The Bell system filter is made at Ravensthorpe. Yorkshire, 

 by Bell Bros. They consist of mild steel shells, 8 feet in 

 diameter, with dished ends. They work under pressure. At 

 the bottom of the shell inside is a floor of strainers and pipes 

 set in concrete. 



The filters use a silicious silver sand as filtering material, 

 and the water is coagulated (excepting where it is free from 

 turbidity) bv aluminic sulphate, lime and alum, and in Glouces- 

 ter they use alum and permanganate of potassium. The same 

 C'elatinous coating described before is formed here on the sand. 



