Even Huygens when trying to detect the presence of striæ in polished glasses viewed them under oblique illumination, usually at a considerable distance, so as to give full scope to the aberrations, and had recourse for greater exactitude to a telescope. But the method was carried to its highest pitch of perfection in 1867 by Toepler who employed the following procedure: A small luminous source _a_ (Fig. 49) illuminates a lens _L_ which throws an image _b_ of the luminous source. If the eye be so placed that the image falls on the pupil, the entire lens, if perfect, will appear equally illuminated, for the reason that all points of it send out rays to the eye. Coarse imperfections of form or of homogeneity are rendered visible only in case the aberrations are so large that the light from many spots passes by the pupil of the eye. But if the image _b_ be partly intercepted by the edge of a small slide, then those spots in the lens as thus partly darkened will appear brighter whose light by its greater aberrations still reaches the eye in spite of the intercepting slide, while those spots will appear darker which in consequence of aberration in the other direction throw their light entirely upon the slide. This artifice of the intercepting slide which had previously been employed by Foucault for the investigation of the optical imperfections of mirrors enhances enormously the delicacy of the method, which is still further augmented by Toepler's employment of a telescope behind the slide. Toepler's method, accordingly, enjoys all the advantages of the Huygens and the Foucault procedure combined. It is so delicate that the minutest irregularities in the air surrounding the lens can be rendered distinctly visible, as I shall show by an example. I place a candle before the lens _L_ (Fig. 50) and so arrange a second lens _M_ that the flame of the candle is imaged upon the screen _S_. As soon as the intercepting slide is pushed into the focus, _b_, of the light issuing from _a_, you see the images of the changes of density and the images of the movements induced in the air by the flame quite distinctly upon the screen. The distinctness of the phenomenon as a whole depends upon the position of the intercepting slide _b_. The removal of _b_ increases the illumination but decreases the distinctness. If the luminous source _a_ be removed, we see the image of the candle flame only upon the screen _S_. If we remove the flame and allow _a_ to continue shining, the screen _S_ will appear uniformly illuminated. [Illustration: Fig. 50.]