THE PRESIDENTIAL ADDRESS.
153
sooner or later it allows germs to pass, and may even give a filtrate
which has picked up more organisms in its passage through the
filter than were originally present in the liquid; and thus these
filters require frequent sterilization to make them safe, and less
perfect filters of the familiar household type may easily become foci
for the propagation of bacteria, doing a great deal more harm than
good.
It is therefore evidently hopeless to expect that mere mechanical
filtration on the enormous scale required for the water supply of a
great population can be efficient ; and yet chemical, and still more
biological, analysis shows that the sand-filter beds of the Water
Companies work an improvement in the quality of the water that is
simply marvellous, reducing, according to the last published report,
the "colonies" of bacteria from 1,100 to 29.
The interstices of the filter are clearly too large to produce this
result by mere filtration, and if it were owing to mechanical deposit on
the surface of the grains of sand any slight increase in the speed of
the percolation would carry through the filter the accumulated im-
purity. I think it is clearly proved that the purification effected is
the work of organisms which attach themselves to the sand and
entangle and actually feed on the bacteria contained in the water
passing through the filter.
A similar process is evidently going on along the whole course
of streams which are not so hopelessly polluted as to destroy the
beneficial organisms which can feed on hurtful ones, and
continue in the water the process of assimilation of matter in a
state of noxious decomposition that is so familiar in every cultivated
field.
In applying these results to the question of water supply, we see
that we may adopt various means to obtain that measure of purity
that we need. We may take a water supply from valleys in hard,
primary rocks, so far from centres of population that sewage pollu-
tion cannot exist, and thus obtain a water of great purity and soft-
ness, but generally discoloured by peat, and often showing a danger-
ous tendency to corrode lead. In the case of London, this means
bringing a water supply from the higher valleys in Wales. Again we
may trust for freedom from contamination to natural filtration
through deep beds of rock ; in this case the amount of filtration of
the water depends on the freedom of the rock from flaws and faults ;
nothing short of solid rock appears to be at all effectual. Experience
L