PROCEEDINGS OF SECTION I. 669 



a thinning down of the scum except in No. 7, where an increase is 

 recorded. This, however, may be due to one of the possible errors 

 of observation mentioned above. A gradual reduction occurred up 

 to June, but the next month the rise referred to below took place. 

 Since then solvent action has again progressed. 



A curious phenomenon is revealed by the records of July, where 

 an all-round increase is observed. At the end of June for five days 

 during heavy rain all the tanks were used, and undoubtedly the incre- 

 ment must be due to this : it could not be an accidental coincidence. 



The conclusions that present themselves from a consideration of 

 the above history are these : — 



1. That with a slow flow through the tanks there is a tendency 

 to heavy scum formation at tjie inlet end and very little at the outlet. 



2. That in resting a tank containing scum the scum shows a 

 gradual but irregular disposition to become thinner. 



3. That with the most rapid flow so far employed (less than six 

 hours through the tanks) the scum not only shows no tendency to 

 excessive accumulation at the inlet, but, on the other hand, is more 

 evenly distributed over the whole surface of the tank contents. 



These conclusions, it is hardly necessary to state, cannot be held 

 to be established from the above experiences as applicable to all sewages 

 in all biological installations. They may be so, but, as far as we can 

 say here, they are only true for the conditions under which they were 

 observed. 



I shall show later in what way purification of the sewage was 

 efiected by the changes made in working the tanks. 



The tanks have been twice cleaned out to make alterations. On 

 neither occasion was there any accumulation of silt of importance. 



In the filter beds the second stage of purification is effected by a 

 process which is in reality one of the classical varieties of land-treat- 

 ment, viz., " intermittent downward filtration." The beds are worked, 

 if not to their utmost, at all events constantly, and cultivation is not 

 attempted. 



Seven out of the eight filter beds are in constant use, the eighth 

 being kept in reserve — as already mentioned — for stormwater and 

 other emergencies. Each bed is used in rotation, and receives a day's 

 flow once a week. The tank effluent is run over the surface from the 

 main carrier until the whole is submerged several inches. The length 

 of time that elapses before the liquid disappears varies from a few 

 hours to several days, according to the looseness or otherwise of the 

 surface soil and to existing meteorological conditions. From time 

 to time the beds are subsoil ploughed, and this operation permits the 

 effluent to sink from sight in a few hours. As the beds become con- 

 solidated and clogged from use, the time extends to as much as two 

 and a half days. In times of rain, of course, percolation and evapora- 

 tion are much retarded. As soon as a bed is dry enough to permit 

 it, harrowing operations are commenced and the upper few inches 

 turned over. 



