FLIGHT OF INSECTS. 195 



end to (), that is to say, in front of the lo\ver loop. We 

 endeavour to pass this point iiito the interior of the course 

 described by the wing ; but as soou as it enters this region, 

 the rod receives a series of shocks from the wing, vhich rubs 

 on its surface, and wipes off the soot which covered it. When 

 we examine the surface of the glass, we see that the soot has 

 been removed only on the upper part, which shows that at 

 the point (</) of its course, the wing is descending. The 

 same expeiiment being repeated in ('), that is to say, in the 

 hinder part of the orbit of the wing, it is found that the rod 

 lias been rubbed beneath ; so that the wing at a was ascending. 

 In the same manner it may be shown that the wing rises at 

 b and descends al I'. 



Fio. S'2 Detcrtu'nation of the direction of the movements of ;:n iutect's 



winy. 



We now know all the movements executed by an insect's wing 

 during its revolution, as well as the double change of plane which 

 accompanies them. The knowledge of this change of plane 

 was given to us by the unequal brightness of the two branches 

 of the luminous 8. Thus we may feel assured that in the 

 course of ihe descending wing, that is from b' to a in fig. 82, 

 the upper surface of the wing turns slighlly forward, while 

 from a to b, that is, in ascending, this surface turns a little 

 backwards. 



