790 The American Naturalist. [September, 
be taken from the normal scale, the wire may be cut, and the 
plane will rest on the air pressure beneath its surface, in which 
event pressure is substituted for tension on the normal line, 
otherwise the stress being unchanged, either in direction or 
magnitude. 
If the plane be suspended at the soaring angle of 5°, and 
wind is strong enough to lift it to that angle, all tension will be 
taken from the normal scale, the plane resting on air pressure as 
at the 45° inclination, and so on through all degrees of the arc. 
During all this time the parallel scales indicate the weight 
recorded by them at the various angles in calm air, so that they 
are not influenced by wind inthe least degree, however great its 
force might be. 
If at any angle of obliquity the plane be held by a wire from 
each corner, perpendicular to it, and fastened to the platform, no 
force of wind will vary the parallel scales, the tension of the four 
wires holding the entire wind resistance, however great, In 
short, there is no way to get wind to affect the parallel scales. 
If variation in them occurs, it shows at once some error in the 
experiment, either the wires being out of adjustment, or some 
` warpage of plane or other disturbance having taken place. 
The rotating arms for trying this experiment in calm air are 
60 feet radius, their ends describing a circle of 120 feet diameter. 
The plane is hung under the arm, and all the conditions of the 
other experiment complied with. No change could be noted in 
the results, excepting that motions of plane were made with 
greater smoothness, on account of greater steadiness in artificial 
wind. By no means could air resistance be brought to bear on 
the parallel component of gravity. It acted the same in wind as 
in calm air, undisturbed by atmospheric resistance. 
Before anything is inferred from these experiments it will be 
best to state the facts of the case. So completely subversive of 
the usual conceptions of the action of air on surfaces is this trial 
_ that it seems imperative to get the facts right before attempting 
to range them in the order of mechanical sequence. 
For convenience, I will number a few of the conspicuous ones. 
1. The scales, being marked in half pounds, do not register 




