SEA-FISHERIES LABORATORY. 149 
it. It will be emptied by a series of sea-cocks just before 
high water, and filled again by the tide flowing over the 
tops of the walls. The water emptying from the tank, 
and containing refuse matter washed out from the 
mussels, will thus be carried away by the flowing tide, 
and clean flood-tide water will take its place. 
The sewer outfall is marked on the sketch chart, and 
a series of arrows showing the direction of the flood- 
stream. The current sets directly from the sewer outfall 
up the channel in the direction of the tank, but when the 
stream is running very strongly in the channel it is 
practically still water at the place of the tank, except for 
a slow eddy circulating in the opposite direction to that 
of the tidal stream. It may be thought that this 
arrangement involves risk of contamination of the tank 
from the sewer, but I do not think that this will be so. 
First of all, the main drift of the flood-stream is past 
the tank, and the eddying water will be greatly diluted. 
This dilution is very great, for Mr. Durlacher’s figures 
(published in the Annual Report for 1913) show at high 
water the density of the water in the estuary is every- 
where practically the same, and about that of the sea 
outside the bar (about 1:026). At low water the density 
‘is also nearly uniform, but is now about 1:007 to 1°010. 
There is therefore a very great rise in density during the 
later part of the flood-stream, indicating an enormous 
dilution of the estuarine water by water from the open 
sea. Only traces of sewage can thus reach the tank in 
any circumstances. 
An extensive series of bacteriological analyses of the 
water over the site of the tank was made in June, both 
from samples which I collected myself, and some sent 
me by Captain Lewis. Thus:—27th June, means of two 
series of analyses. 
