116 TYPICAL QUEENSLAND LAGOON 
found in the effluents of these filters at the same period was 
very small, the oxygen in the applied water being practically 
exhausted before the water reached the outlet of the filter. 
As the amount of dissolved oxygen in the lagoon water 
at this time of the year is comparatively low, aerators were 
fixed at the inlets of both filters. These had the effect of 
practically saturating the applied water with oxygen, increas- 
ing the amount from 75 per cent. of the possible quantity 
to about 98 per cent. Notwithstanding the increased amount 
in the applied water, however, the oxygen still continued 
to be almost entirely exhausted in its passage through the 
filters, though a slight increase took place in the amount con- 
tained in the effluents. 
In order to investigate this phenomenon, arrangements 
were made whereby samples of water might be drawn from 
different positions in the sand bed. Several determinations 
of the number of bacteria and the amount of dissolved 
oxygen in the water ai different depths in the sand were made. 
It was expected that the rates of decrease of bacteria and 
dissolved oxygen would be somewhat similar, or that the 
decrease in oxygen would cease in those portions of the filter 
which contained practically no bacteria; this was expected 
on the assumption that the decrease in the amount of oxygen 
dissolved was entirely due to its consumption by the bacteria 
in the process of oxidation of the organic matter in the water. 
It was found that the bacteria were practically all 
removed in the upper 6 inches of the sand, but that whilst 
about one-half of the total absorption of oxygen by the filter 
took place in the first § inches, the amount continued to de- 
crease at a fairly regular rate throughout the remaining depths 
of sand where the water at the same time contained prac- 
tically no bacteria | am not able to satisfactorily explain 
this phenomenon, but intend if possible to investigate 
further. 
INTERMITTENT SAND FILTRATION. 
Following a line of thought suggested by the last 
mentioned experiments, an intermittent filter was constructed 
with a view to testing the effect of more efficiently aerating 
the sand bed. The filter is 76 square feet in area, and is 
constructed on the lines of the City filter at Lawrence, 
Massachusetts, in the United States of America. The bed 
