348 Proceedings of the Royal Society 
If tho artificial wing be in one piece, it ought to be made to vibrate 
in a more or less horizontal direction ; if in several pieces, it should 
be worked in a more or less vertical direction, as the wing in this 
case acts alternately as a short and long lever, in virtue of its 
closing and opening during the up and down strokes, the acting 
area of the wing being greatly reduced during the up stroke, and 
greatly increased during the down one. 
If a properly constructed artificial wing be made to vibrate in a 
vertical direction, it invariably darts downwards and forwards in a 
curve during the down stroke, and upwards and forwards in a similar 
but opposite curve during the up stroke, the two curves running 
into each other to form a progressive, continuous, wave track. 
If the win. 2 : be made to vibrate from side to side in a more 
or less horizontal direction, it rises zig-zag fashion by a series of 
looped curve movements, each pass of the wing being on a 
higher level than that which preceded it. Whether the wing be 
moved vertically or horizontally, it invariably twists and untwists 
during its action. In twisting and untwisting, it developes figure- 
of-S curves, not only along its anterior and posterior margins, but 
throughout its entire length and breadth. 
The figure-of-8 vertical movement may be converted into the 
figure-of-8 horizontal movement by a slight rotation of the wing 
on its long axis, or by a tilt of the body or frame bearing the 
wing. It is in this way that the wing may act either as an ele¬ 
vator and propeller, or merely as an elevator. Thus it is not 
uncommon to see an insect elevate itself by a horizontal screwing 
figure-of-8 movement, and then, suddenly changing the direction 
of the stroke of the wing and of the body, dart forward in a nearly 
horizontal direction. 
The artificial wing, like the true one, attacks the air at a great 
variety of angles during the down and up strokes. Thus during 
the down stroke the angles which the wing makes with the horizon 
are increased, whereas during the up stroke they are diminished. 
The angles made by the different portions of the artificial wing 
vary as in the living wing, the angles made by the parts nearest 
the root being greater than those nearer the tip. This is occa¬ 
sioned by the manner in which the artificial wing twists and 
untwists during its action, the torsion in question being due to the 
