predetermined time of
duration of test has expired, when the total water as measured in the
tanks, and representing the amount of steam condensed during that time,
is easily found by adding together the quantities given at each
individual measurement.

All that are necessary to insure successful results from a plant similar
to this are care and accuracy in its operation and construction.
Undoubtedly in most cases it is preferable to weigh the condensed steam
instead of measuring the volume passed, and from that to calculate the
weight. If dependence is being placed upon the volumetric method, it is
advisable to lengthen the duration of the test considerably, and if
possible to measure the feed-water evaporated at the same time. Such a
course, however, would necessitate little change, and none of a radical
nature, from the arrangement described. Where, however, the measuring
method is adopted, the all-important feature, requiring on the tester's
part careful personal investigation, is the graduation of the tanks. It
facilitates this operation very considerably when the receptacles are
graduated upon a weight scale. That is to say, whether or not a vertical
scale showing the actual hight of water be placed inside the tank, it is
advisable to have a separate scale indicating at once to the attendant
the actual contents, by weight, of the tank at any time. It is the
tester's duty to himself to check the graduation of this latter scale by
weighing the water with which he performs the operation of checking.

Apart from the foregoing, there is little to be said about the measuring
apparatus. As has been stated, accuracy of result depends in this
connection, as in all others, upon careful supervision and sound and
accurate construction, and this the tester can only positively insure by
exhaustive inspection in the one case and careful deliberation in the
conduct of the other.

It will be readily understood that the procedure--and this implies some
limitations--of a test is to an extent controlled by the conditions, or
particular environment of the moment. This is strictly true, and as a
consequence it is often impossible, in a maker's works, for example, to
obtain every condition, coinciding with those specified, which are to be
had on the site of final operation only. For this reason it would
appear best to reserve the final and crucial test of a machine, which
test usually in the operating sense restricts a prime mover in c