BACTERIAL PURIFICATION 227 



polluted with chemicals, that of Lichfield contains a large 

 amount of brewery refuse. I cannot see how the action of coal 

 is different from that of other media, though Dr. Bostock Hill 

 contended that effluents from coal filters show a greater loss of 

 organic carbon as compared to organic nitrogen than in filters 

 made of other materials, and that this is a characteristic property 

 of coal, but Mr. Garfield, now of Bradford, to whom the sugges- 

 tion is due, has not adopted this material for his present works. 



Dr. Fowler, in his report of the Davyhulme experiments in 

 1897, confirms the results of previous observers that coal and 

 burnt clay filters, when worked continuously, rapidly become 

 clogged, and that improved results are obtained with intervals 

 for rest and aeration. He also considers coal to be superior to 

 burnt clay. 



The table in my last edition of the comparative nitrifica- 

 tion effected by different filters, ranks the Garfield filter as 

 lower than other forms, but, as already mentioned, this result 

 is more likely due to the difference in the mode of working and 

 aeration of the filter than to the material. 



Partly for the sake of cheapness, and also because it was 

 expected that coke would in time disintegrate, the use of more 

 compact materials has been suggested. Broken slate^ or shale 

 has been much used in the North, and Thudichum even made 

 laboratory experiments with pounded glass, and found a certain 

 amount of efficiency. Burnt ballast, clinker, cinder, slag, 

 polarite and iron sand have their advocates. Non-porous 

 materials might be expected to have a lower capacity, but it is 

 mainly on the surfaces, and not in the interior of the masses, 

 that the bacterial action occurs. At Exeter, Mr. Cameron 

 expresses a general preference for clinker : at Southampton and 

 other places assorted clinkers from the dust destructors are 

 used. But coke, if available, seems the best material for nitri- 

 fication, and has shown no noticeable disintegration in nine 

 years. Burnt ballast must be carefully made, as many kinds 

 crumble and block up the filter. Road granite, crockery, and 

 old iron and tins are stated to have given satisfactory results, 

 and broken saggers from pottery kilns have been successfully 

 used at Hanley in 1903, and in other places in the Midlands. 



Durability of ^^ Ballast.'' If clay be thoroughly burnt, its 



1 In Nov., i8g8, rough state filters were proposed at Festiniog, and in Marcli, 

 1899, Chambers Smith reported to the Sutton Council his trials of various 

 materials. See San. Record^ Aug. 9, 1906. 



15—2 



