means of fans which are
made so that they can maintain an air pressure in the stokehold above
that of the outside atmosphere. This is the system almost universally
adopted in war vessels, and it is used also in some fast passenger
ships. The air pressure usually adopted in large vessels is that
corresponding to a height of from 1 to 1-1/2 in. of water, whilst so much
as 4 in. is sometimes used in torpedo-boats and similar craft. This is,
of course, in addition to the chimney-draught due to the height of the
funnel. In the closed ashpit or Howden system, the stokehold is open,
and fans drive the air round a number of tubes, situated in the uptake,
through which the products of combustion pass on their way to the
chimney. The air thus becomes heated, and part of it is then delivered
into the ashpit below the fire and part into a casing round the furnace
front from which it enters the furnace above the fire. In locomotive
boilers the draught is produced by the blast or the exhaust steam. With
natural draught a combustion of about 15 to 20 lb. of coal per sq. ft.
of grate area per hour can be obtained. With forced draught much greater
rates can be maintained, ranging from 20 lb. to 35 lb. in the larger
vessels with a moderate air pressure, to as much as 70 and even 80 lb
per sq. ft. in the express types of boiler used in torpedo boats and
similar craft.

_Performance of Boilers._--The makers of several types of boilers have
published particulars regarding the efficiency of the boilers they
construct, but naturally these results have been obtained under the most
favourable circumstances which may not always represent the conditions
of ordinary working. The following table of actual results of marine
boiler trials, made at the instance of the British admiralty, is
particularly useful because the trials were made with great care under
working conditions, the whole of the coal being weighed and the
feed-water measured throughout the trials by skilled observers. The
various trials can be compared amongst themselves as South Welsh coal of
excellent quality was used in all cases.

In experimental tests such as those above referred to, many conditions
have to be taken into account, the principal being the duration of the
trial. It is essential that the condition of the boiler at the
conclusion of the test should be precisely the same as at the
commencement, both as regards the quantity of unconsumed coals on the
fire-grate and t