104 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
of higher level through a series of rapids and intervening pools to a 
lake of lower level. The boiler corresponding to the lake of higher 
level and the condenser to that of lower level. 
In the flow through the turbine the steam is repeatedly gathering 
a little velocity from the small falls of pressure, which is as soon 
checked and its energy transferred to the blades, over and over again; 
fifty to one hundred times is this repeated before it is fully expanded 
and escapes into the condenser. 
The number of blades in a steam turbine is very great; in a 2,000- 
horsepower engine it may be from 20,000 to 50,000 and the surface 
speed of the several barrels of the turbine will be from 150 to 300 feet 
per second. In such an engine it is arranged that the lineal velocity 
of the blades will approximate to one-half that of the tangential com- 
ponent of the steam issuing from the guide blades. The blades, as we 
have seen, are curved, with thickened backs, and are smooth; the 
steam therefore flows around them, and past them, without much loss 
by shock or eddy current or frictional loss. The proportions of tur- 
bines as regards diameter, height of blade, and blade openings are 
calculated so that, under average working conditions, the correct 
expansion of the steam shall be attained, and the fall in pressure and 
velocity of steam at each turbine of the series shall be such as to secure 
for it the highest efficiency. 
When a turbine is tested the pressures at many points along the 
barrel are recorded, and the calculated pressures confirmed and veri- 
fied by experiment, and these are usually in close accord. As the 
result of data accumulated from experiments on many turbines, the 
probable horsepower that will be obtained from a given design of 
turbine can be predicted with as much accuracy as in the case of the 
reciprocating engine. The best results that have been obtained from 
large turbines show that about 70 per cent of the available energy in 
the steam is converted into brake horsepower; and where, we may 
inquire, has the other 30 per cent gone to? 
The chief losses of efficiency in all steam turbines are due to three 
principal causes: Firstly, to skin-friction of the steam coursing at 
high temperature through the small openings between the blades; 
secondly, to unavoidable leakages; and, thirdly, to eddy-current losses 
arising from insufficient blade velocity and errors of workmanship. 
The first of these losses, the friction of the steam, is reduced by 
superheating, and thus partially removing the fluid frictional loss 
arising from the drops of condensed water mingled with the steam. 
In some cases this gain in efficiency is worth the extra cost of the 
superheater, but, unless intermediate superheaters are used, initial 
superheat cannot be raised high enough to maintain dryness through- 
out the major part of expansion without destroying the turbine. 
Moderate initial superheat, however, is generally used with some gain 
