PROCEEDINGS OF THE SOCIETY. . 481 



wrong because Naegeli and Schwendener write as if the images of the 

 source of light (dioptric beam and spectra) were in a true plane, and 

 yet all in the same phase. He points out that, in order to interfere 

 according to theory at a point in the image plane, they must lie on a 

 concave surface directed towards that point ; but that the concave sur- 

 face on which they actually lie can be varied at will to give any desired 

 small alteration of flatness to the surface without affecting the result. 

 Now it is perfectly true that the dioptric image and spectra lie on a 

 concave sm-face (attention to the same matter has been drawn by 

 Dr. E. Strehl in his Theorie der allgemeinen Milcroskojrischen Abbildung, 

 1900); but it is a mistake to suppose that an alteration of flatness of 

 the surface can be attained in the region of the upper focal plane of the 

 objective without the image plane being affected. If the image plane 

 is readjusted by lengthening the tube, then, indeed, the degree of flat- 

 ness of the curve is altered, and for the reason that in that case the 

 objective would have been slightly racked down, thereby altering its 

 distance from the actual light source, the condenser, and the diaphragm, 

 the position of which relatively to the objective has an important bearing 

 on the position of the spectra. It will be observed that I use the ex- 

 pression " in the region of" the upper focal plane of the objective. This 

 is done advisedly, since the expression " upper focal plane " has been 

 used in such a loose fashion by Naegeli and Schwendener and other 

 writers, and they have assumed that visual spectra when present are 

 always formed just there. But though under the special conditions 

 of the experiments which they discuss they are formed in that region, 

 we shall shortly see that the visual spectra are by no means always 

 found there. 



This brings us to Mr. Gordon's remarks that any number of diffrac- 

 tion spectra can be crowded into an objective by the expedient of 

 racking the condenser up and down, and altering the curve of the 

 wave-front as it passes the grating, and that, therefore, a narrow-angled 

 objective can be made to yield the diffraction image of a wide-angled 

 immersion one. This I think I can also show to be based on a miscon- 

 ception of the position of the spectra. 



If the wave-front passes the grating as a curved surface, then the 

 whole basis of the experiment is altered. For one thing, rays emanat- 

 ing from any point of the light source must be passing a great number 

 of lines in the diffraction plate in different phases ; but apart from this 

 we no longer get the diffracted rays of light focussed and in the same 

 phase in the upper focal plane * of the objective, for the conjugate focal 

 surface of the light source will, by the very nature of the conditions named, 

 be somewhere else, higher or lower, in the tube. That this is actually 

 the case, and that the srjectra apparently emanating from the upper 

 focal plane are really formed elsewhere, can be proved by catching them 

 on a screen of ground glass or celluloid fixed at the end of a tube, which 

 can be inserted and pushed down into the ordinary Microscope tube. 



The crowding together of the spectra, which Mr. Gordon speaks of, 

 is merely the overlapping of the spectra, which are seen altogether out 

 of focus as diffused surfaces. It is by no means a closing together of 



* As it is customary to speak of " planes,'' this word is used here and elsewhere 

 to denote the more or less curved surface on which the spectra are formed. 



Aug. 21st, 1901 2 K 



