Atmospheric Friction on Even Surfaces. D9 
long by 6 feet square, and has a five-foot suction-fan at 
one end, a cheese-cloth screen, or two, at the other end, to 
straighten the inflowing air. A boy, with a rheostat and 
tachometer, holds the fan at any desired speed, accurately to 
a fraction of one per cent., thus giving an even flow of air 
of like constancy. As the wind-friction moves the plane 
endwise, the displacement is determined by the motion of a 
sharp pointer attached to one suspension-wire, and travelling 
over a fine scale lying on top of the tunnel. The swing of 
the plane can be measured accurately to five thousandths of 
an inch, and the force on the plane is exactly proportional to 
the seale-readings. The wind-speed is usually measured by 
a carefully standardized * pressure-tube anemometer, though 
other kinds have been employed for comparison. ‘The tunnel 
is sometimes narrowed, as shown in fig. 1, to increase the 
speed. 
The planes employed were similar to those commonly nsed. 
to determine the skin-friction of water. The first was 
pine board 4 feet long, 254 inches wide, and 1 inch thick, 
carefully trued and varnished, and suspended in the wind 
tunnel by steel wires tw enty -five thousandths of an inch 
thick. It was provided, fore and aft, with seven-inch pine 
ends whose “ lines” were circular ares tangent to the board, 
the ends being held on by dowel pins. Hach end terminated 
in a sharp edge from the centre of which a steel pin pro- 
truded along stream between guides, to steady the plane 
against wabbling. As the doweling was carefully executed, 
straight planes of any length could be made by adding 
extra boards ; the lengths most employed being 2, 4, 8, 12, 
16 feet. 
The method of using the planes to determine the surface- 
friction was as follows :—The total force was measured at 
various velocities, with the ends first on the sixteen-foot 
board, then successively on the other lengths; and finally 
with nothing between them. Subtracting this last force 
from each of the others gave the friction on each of those 
five lengths. It may not be absolutely true that the end 
resistance was the same for each of those lengths ; but the 
error of this assumption is regarded as very slight for 
several reasons: (1) the end resistance is but a small part 
of the total; (2) the stream-lines are so slightly disturbed 
that the flow about the ends must be practically the same in 
all cases ; (3) the results harmonize very well with those 
obtained by other methods. 
* “ Measurement of. Air Velocity and Pressure,’ Physical Review, 
December 1905. 
