f paper, upon which is
printed a dead black disc, subdivided by concentric circles and radii,
printed in white. In Jeffery's counter (Fig. 207), each subdivision has
an area of 1 square centimetre; in Pakes' counter (Fig. 208), radii
divide the circle into sixteen equal sectors, and counting is
facilitated by concentric circles equidistant from the centre.

[Illustration: FIG. 207.--Jeffery's disc, reduced.]

[Illustration: FIG. 208.--Pakes' disc, reduced.]

(a) In the final counting of each plate, place the plate over the
counting disc, and centre it, if possible, making its periphery coincide
with one or other of the concentric circles.

(b) Remove the cover of the plate, and by means of a hand lens count the
colonies appearing in each of the sectors in turn. Make a note of the
number present in each.

(c) If the colonies present are fewer than 500, the entire plate should
be counted. If, however, they exceed this number, enumerate one-half, or
one-quarter of the plate, or count a sector here and there, and from
these figures estimate the number of colonies present on the entire
plate. In practice it will be found that Pakes' disc is more suitable
for the former class of plate; Jeffery's disc for the latter. It should
be recollected however that unless the plates have been carefully
leveled and the medium is of equal thickness all over it is useless to
try and average from small areas--since where the medium is thick all
the bacteria will develop, where the layer is a thin one, only a few
bacteria will find sufficient pabulum for the production of visible
colonies.

It will be noted that the quantities of water selected for addition to
each set of tubes of nutrient media have been carefully chosen in order
to yield workable results even when dealing with widely differing
samples. Plates prepared in agar with 0.1 c.c. and in gelatin with 0.02
c.c. can be counted even when large numbers of bacteria are present in
the sample; whereas if micro-organisms are relatively few, agar plate 4
and gelatine plate 1 will give the most reliable counts. Again the
counts of the plates in a measure control each other; for example, the
second and third plates of each gelatine series should together contain
as many colonies as the first, and the second should contain about half
as many more than the third and so on.

2. Qualitative Examination.--

_Collection of Sample._--The water sample required for the routine
examination, w