and the high levels of the
city. The reason why the Romans did not build underground aqueducts, as
is done at the present day, has been variously explained. Perhaps they
did not fully understand that water will find its own level over a great
distance. They also would have found great difficulty in overcoming the
high pressure of the water.

In their conduits they built shafts at frequent intervals designed to
relieve the pressure of compressed air in the pipes. The water from the
neighbourhood of Rome rapidly encrusted channels and pipes with
calcareous deposits. Probably the great advantage of accessibility to
leaks and defects gained by building unenclosed aqueducts appealed
strongly to the ancient Romans. They did not fully understand the
technical difficulties involved in the "hydraulic mean gradient." No
machinery was used to pump the water or raise it to an artificial level.
A strip of land 15 ft. wide was left on either side of the aqueducts,
and this land was defined at intervals by boundary stones. No trees were
grown near the aqueduct, to avoid the risk of injuring the foundations,
and any breach of the rules for the preservation of the aqueducts was
severely punished by fines.

Vitruvius gives rules for testing the water, and points out that water
led through earthen pipes is more wholesome than water coming from
leaden ones. He states that the "fall" of an aqueduct should be not less
than 1 in 200. A circuit was often made to prevent the too rapid flow of
the water, and intermediate reservoirs were constructed to avoid a
shortage of water in the case of a broken main. Reservoirs were also
used for irrigation.

The water from the aqueduct was received at the walls of the city in a
great reservoir called _castellum aquarum_, externally a beautiful
building and internally a vast chamber lined with hard cement and
covered with a vaulted roof supported on pillars. The water flowed
thence into three smaller reservoirs, the middle one filled by the
overflow of the two outer ones. The outer reservoirs supplied the public
baths and private houses, while the middle one supplied the public ponds
and fountains, so that, in the event of a shortage of water, the first
supply to fail was the least important. The amount of water provided for
private use could be checked, for purposes of revenue, by means of this
arrangement.

At first the aqueducts were not connected with private houses, but,
later, private persons w