'4 
THOMPSON YATES LABORATORIES REPORT 
to be absent in all 15 samples, in some absent in approximately .1 gram, in others 
in .02 gram. 
(/) Streams and Rivulets. — The following experiments have reference to the 
water from sources which form the main Liverpool water supply. These are two in 
number ; Lake Vyrnwy acts as one reservoir, while the rest of the water of the city is 
taken from the Rivington Reservoirs. The water is filtered through sand filter beds at 
Oswestry and Rivington respectively, and the water supplied to the city consists of 
a mixture of the two supplies. A very complete analysis was made of the streams and 
rivulets of the watershed supplying Lake Vyrnwy. In the foregoing sets of analyses 
the material examined was obviously unpolluted, and B. coli was found absent in every 
case in the quantities analyzed ; of the previous history of these streams and rivulets 
one cannot be so sure. The water of thirty-eight streams was examined, 1 c.c. being 
taken for analysis in each case and three plates poured. The B. coli was found to be 
absent in thirty-one cases and its presence was demonstrated in seven. In these, 
however, it was present in very small quantities ; one organism per c.c. in one case, and 
less than this amount in the other six cases. The details of the sources, etc., of 
the water examined are given in Table V, and it is very interesting to notice that in 
nearly all the cases in which B. coli was found, there is very good evidence of the 
possibility of sewage pollution, and in many instances this had previously been 
suspected and the analyses were made with a view to clearing up this point. 
It was found that, although on the whole the rivulets of the Vyrnwy watershed 
were free from the B. coli, yet that the sand of certain of the sand filters at Oswestry 
(see Table VI), used for the subsequent filtration of the Vyrnwy water, contained the 
bacillus to a considerable extent, so it was thought desirable to examine larger 
quantities of the water. About n\ litres were concentrated by filtration through a 
Chamberland and Berkefeld Filter, the filter brushed with a measured quantity of 
water and 1 c.c. plated. These experiments have no great numerical accuracy, the 
object being rather to demonstrate the mere presence or absence of the colon bacillus. 
The surface water of Lake Vyrnwy, at Masonry Dam, was analyzed in this way 
two days after having been drawn. The water contained approximately 158 B. coli 
per litre. The water of the Cedig River, analyzed four days after having been 
drawn, showed 133 per litre. Probably these figures are rather lower than they 
should be, as the B. coli tends to decrease in numbers when kept in fairly pure 
water in a closed bottle. This has been pointed out previously in this paper and 
is further confirmed by the following experiment. Water from the surface of the 
lake near the Tower was examined 25 days after having been taken ; B. coli was 
found to be absent in at least 514 c.c. Unfortunately the same water was not 
analyzed at the time it was drawn, or shortly afterwards ; but it is probable it would 
then have contained the B. coli, though probably not in 1 c.c. 
The rivulets of the watershed of the Lower Rivington Reservoir were 
