BY HARDOLPH WASTENEYS. 115 
tion in colour was 66 per cent. of the yellow, and 93;per cent. 
of the red units determined by means of the Lovibond 
Tintometer. The number of bacteria per cubic centimeter 
in thé effluent and unfiltered water were also determined 
from time to time, the average result of nine determinations 
showing 462 per c.c. in the unfiltered water, whilst only 7 
per c¢.c. appeared in the effluent, which is equivalent to a 
per centage reduction of 983}. Throughout the whole time 
the appearance of the effluent was good, its taste was excellent, 
‘no odour was noticeable, and it was to the ordinary observer 
practically colourless. The sand used in this filter was of poor 
quality, being the result of recent decomposition of granite, 
it was very friable and exceedingly difficult to clean, and 
even when washed with apparent thoroughness still con- 
tained a good deal of organic matter. Had better sand, 
such as that obtained from the Brisbane River bed been used, 
the results from this filter would doubtless have been still 
better. 
I have described this filter in preference to the other 
sand filters mentioned, because it is the most typical of the 
group. Even better results were obtained from one of these 
filters in whose section ashes played an important part, but 
this improvement was proved to be due not to the ashes, 
but to the finer grade of sand in the surface layer. 
The last mentioned filter contained besides cinders and 
crushed quartz, 12 inches of sand on the surface of the bed, 
which had an effective size of -28 m.m.s. The sand used 
was of superior quality to that in the first-mentioned filter 
composed entirely of sand, which probably accounts for the 
better chemical results obtained. The Bacteriological results 
were however not quite so good, probably on account of 
the fact that the depth of the sand layer in this filter was less 
than in the former filter. The length of run or period between 
scrapings was also not so good as for the sand filter, averag- 
ing only 31 days. this was to be expected on account of the 
finer grade of sand employed. 
During the warmer weather and on account of the in- 
creased bacterial activity occasioned thereby, it was found 
that the removal of the organic matter and colour by these 
filters was much greater than during the colder months of 
the year. Fora similar reason the amount of dissolved oxygen 
