ng an electrical load at
both the maker's works and upon the site of operation.

One consideration of importance is worth inquiring into, and this has
relation to the largest turbo-generators supplied for power-station and
like purposes. Obviously, the testing of, say, a 7000-kilowatt
alternator by any standard electrical-testing method must entail
considerable expense, if such a test is to be carried out in the maker's
works. Nor would this expense be materially decreased by transferring
the operations to the power-station, and there erecting the necessary
electrical plant for obtaining a water load, or any other installation
of sufficient capacity to carry the required load according to the rated
full capacity of the machine.

Assuming, then, that there exist no permanent facilities at either end,
namely the maker's works and the power station, for adequately procuring
a steady electrical-testing load of sufficient capacity, there still
remains, in this instance, an alternative source of power which is
usually sufficiently elastic to serve all purposes, and this is of
course the total variable load procurable from the station bus-bars. It
is conceivable that one out of a number of machines running in parallel
might carry a perfectly steady load, the latter being a fraction of a
total varying quantity, leaving the remaining machines to receive and
deal with all fluctuations which might occur. Even in the event of there
being only two machines, it is possible to maintain the load on one of
them comparatively steady, though the percentage variation in load on
either side of the normal would in the latter case be greater than in
the previous one. This is accomplished by governor regulation after the
machines have been paralleled. For example, assuming three
turbo-alternators of similar make and capacity to be running in
parallel, each machine carrying exactly one-third of the total
distributed load, it is fair to regard the governor condition, allowing
for slight mechanical disparities of construction, of all three machines
as being similar; and even in the case of three machines of different
capacity and construction, the governor conditions when the machines
are paralleled are more or less relatively and permanently fixed in
relation to one another. In other words, while the variation in load on
each machine is the same, the relative variation in the governor
condition must be constant.

By a previously mentioned sys