BACTERIOLOGIC EXAMINATION OF WATER. 171

water to be examined in the proportion of i : 10 or i: 100
with sterile water, mixing well, and making the plate-
cultures from the dilutions.

It is best to count all the colonies if possible, but when
there are hundreds or thousands scattered over the plate,
an average estimation of a number of squares ruled upon
a glass background (Fig. 46), as suggested by Wolf hiigel,
is most convenient. In his apparatus a large plate of glass
is divided into small square di-
visions, the diagonals being spe-
cially indicated by color. The
plate or Petri dish is stood upon
the glass, and the number of
colonies in a number of small
squares is easily counted, and
the total number of colonies es-
timated. In counting the colo-
nies a lens is indispensable.
Special apparatuses have been
devised for counting the colo-
nies in Petri dishes (Fig. 47) , FlG' ^-EsmarcVs instrument
) ° '' for counting colonies of bacteria

and in Esmarcli tubes (Fig. 48). in tubes<

The majority of the water-
bacteria are rapid liquefiers of gelatin, for which reason
it seems better to employ agar-agar than gelatin for
making the cultures.

In ordinary hydrant-water the bacteria number from
2-50 per cubic centimeter; in good pump-water, 100-500;
in filtered water from rivers, according to Giinther, 50-200
are present; in unfiltered river-water, 6000-20,000. Ac-
cording to the pollution of the water the number may
reach as many as 50,000,000.

The waters of wells and springs are dependent for their
purity upon the character of the earth or rock through
which they filter, and the waters of deep wells are much
more pure than those of shallow wells, unless contamina-
tion takes place from the surface of the ground.

Ice always contains bacteria if the water contained