136 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 
them—everyone applies them every day in the ordinary aflairs 
of lite, in a practical, though unsystematic way, and there 
is no reason why we who apply them in private matters should 
not systematise, and apply them in public affairs, especially 
when our business is such as to lead us to interfere with the 
industry and liberty of other people. 
One or two particular methods of analysis are taken to 
illustrate the simple statistical precautions considered. These 
statistical devices are, however, general ones, and may be 
modified and extended so as to fit any other methods of analysis 
that may be employed. 
Method of Analysis. 
A shell-fish bed usually includes various, more or less 
marked sub-areas in each of which the lability to sewage 
pollution varies. Hach sub-area must, therefore, be sampled 
and considered individually. 
A sample consists of 5 mussels collected “‘ methodically 
at random” from different parts of the sub-area. Whether 
the sample consists of 5 or 10 or 20 mussels is, of course, a 
matter of practical convenience—a compromise always. 
The mussels are opened and the liquid in the shells is 
drained into a small mortar. The soft parts are cut out, minced 
up with scissors and put into the mortar. The whole is ground 
up into a thin paste, and this is poured into a 250 c.c. flask, 
which is then filled up to the mark with water. The flask is 
well shaken, and the coarser particles are allowed to settle 
to the bottom. A number of samples, of one c.c. each, are 
taken from the supernatant liquid, and each sample is placed 
in a separate Petri dish. Liquid neutral-red, bile-salt, lactose — 
agar (McConkey) is then poured into the dishes (10 c.c. into 
each), and agar and infected liquid are well mixed. “The 
plates are set and incubated. : 
One c.c. of the liquid in the flask is taken as containing 
1/50th part of an average mussel. | | 
