si 6 Dr. Herschel's Experiments for investigating 
splintered defect, this is no more than what must happen to the 
appearance of every object which is seen through a distorting 
medium. 
I laid the convex side of a plano-convex lens of 2,8-inch 
focus with a diameter of 1,5 upon a plain mirror, and when I 
saw a set of rings I tilted the lens so as to bring the point of 
contact to the very edge of the lens, both towards the light 
and from the light, which, on account of the large diameter of 
the lens, gave a. great variety in the angle of incidence to the 
rays which formed the rings; but no difference in their 
size or appearance could be perceived. This seems to prove 
that no modification of the first surface in which the angle of 
incidence is concerned, such as refraction and dispersion, has 
any share in the production of the rings, and that it acts merely 
by the intromission of light ; and though even this is not with- 
out being influenced by a change of the angle, it can only 
produce a small difference in the brightness of the rings. 
A more forcible argument, that leads to the same conclusion, 
is as follows. Laying down three 54,-inch double convex 
lenses, I placed upon the first the plain side of a plano-convex 
lens of f inch focus ; upon the second, a plain slip of glass ; 
and upon the third, the plain side of a plano-concave lens also 
inch focus. I had before tried the same experiment with 
glasses of a greater focal length, but selected these to 
strengthen the argument. Then, as nothing could be more 
different than the refraction of the upper surfaces of these 
glasses, I examined the three sets of rings that were formed 
by these three combinations, and found them so perfectly alike 
that it was not possible to perceive any difference in their size 
and colour. This shows that the first surface of the incum- 
