quire
any repairs, and when they do, the process is not difficult. If any of
the rivets be damaged, as many must be taken out as will make room for
the free admission of the hand. A small flat mandrel being introduced
into the hose, the new rivets are put into the leather, and riveted up
the same as new pipe; the mandrel is then shaken out at the end.

If the leather be damaged, it may be repaired either by cutting out
the piece, and making a new joint, or by riveting a piece of leather
upon the hole.

The manner of attaching the hose to the coupling-joint is also a
matter of very considerable importance. If a joint come off when the
engine is in operation, a whole length of hose is rendered useless for
the time, and a considerable delay incurred in getting it detached,
and another substituted.

To prevent this, the hose ought to fit as tightly as possible to the
coupling-joint, without any packing. In riveted hose, a piece of
leather, thinned down to the proper size, should be put on to make up
the void which the thick edge of the leather next the rivet
necessarily leaves; the hose should then be tied to the coupling-joint
as firmly as possible with the best annealed copper wire, No. 16
gauge.

When the hose are completely finished in this manner they are proved
by a proving-pump, and if they stand a pressure of two hundred feet of
water they are considered fit for service. I may also add, that when
any piece of hose has been under repair it is proved in the same
manner before it is deemed trustworthy.

The proving of the hose is of very considerable importance, and the
method of doing so which I have mentioned is greatly superior to the
old plan of proving them on an engine or fire-cock. By the latter
method, no certain measure can be obtained by which the pressure can
be calculated. In the first place it must depend on the relative
height of the reservoir from whence the water is obtained and that of
the fire-cock where the experiment is made; and as the supply of water
drawn from the pipes by the inhabitants may be different on different
days of the week and even in different hours of the day, it is quite
evident that by this method no certain rule can be formed for the
purpose required, the pressure being affected by the quantity of water
drawn at the time.

The method of proving by an engine is considerably better than this;
but when a proving-pump can be obtained it is infinitely better than
either. One