THE STEAM TURBINE—PARSONS. 105 
in economy, which in the compound turbine amounts to 1 per cent 
for every 10° F. of superheat. The second loss, which is from leakage, 
is present in the compound and the sinuous types but not in the De 
Laval type. The amount of this loss decreases as the size of the engine 
increases. It is also chiefly consequent on the coefficient of expansion 
of metals, which is a bugbear to the turbine designer. If a metal 
with a much smaller coefficient of expansion than steel and iron could 
be obtained at a reasonable price and of suitable qualities for the con- 
struction of turbine cases, drums, and shafts, a considerable increase 
of economy could be obtained, as it would allow of smaller working 
clearances and less leakage. The third loss, from insufficient blade- 
velocity, is not present to a material extent in the larger compound or 
sinuous course turbines, but is present, as already explained, to a con- 
siderable extent in the single-wheel type. 
Reviewing more closely the motion of the steam through the blades 
of a compound turbine, we see that the portion of its course during 
which it is traveling at relatively high velocity, and in close prox- 
imity to the blades, is short in comparison with the total length of its 
travel within the turbine. The passageways between the blades con- 
stitute virtually jets of rectangular cross section, but having easy 
curves, and the frictional losses are consequently small. After leav- 
ing the blades, it traverses the intervening space in the form of an 
annular cylinder with a spiral motion, the angle of pitch being about 
30° to a plane normal the axis; and, as the succeeding blades are 
moving in a similar direction to this flow, we see that the velocity 
with which the steam is cut by their frontal edges is much less—in 
fact, less than one-half the velocity at which the steam has issued 
from the previous blades. From this we see how small is the loss 
due to the cutting of the steam by the frontal edges in the compound 
turbine, and also how small is the velocity with which drops of water 
strike the metal of the blades. : 
This is an important feature. 
It has been shown by experiment that if drops of pure water, 
arising from the condensation of expanding steam, impinge on brass 
at a greater velocity than about 500 feet per second there results a 
slow wearing away of the metal. It is very slow, and would require 
about ten years to erode the surface to a depth of 3; mch. In the 
compound turbine the striking-velocity is much below this figure, 
and the preservation of their form and smoothness of surface has 
been found to be practically indefinite. 
It appears that the erosive power of drops of pure water moving 
at high velocity increases rapidly with the velocity, it may probably 
be as the square. Experiment has shown that if saturated steam at 
100 pounds pressure be allowed to flow through a divergent jet into a 
