and on some Phenomena connected therewith. 205 



contact to be too close, though I do not remember to have 

 met with this in practice ; and theoretically it is what would 

 be expected when we consider that the original does not act 

 by opacity. According to this view a different result should 

 be obtained in copying an opaque grating ; and such I have 

 found to be the case. For this purpose I employed a copy of 

 the same 6000 Nobert, taken some years ago on a tannin 

 plate, and prepared the photographic film on the same lens as 

 before. When the resulting photograph was examined, the 

 spectra of the first three orders showed bright centres. The 

 diameter of the first dark ring in the first spectrum was '44 

 inch — smaller than before. 



With the 3000 Robert in place of the 6000 the ring-system 

 is formed on a larger scale. The centres for the first four 

 spectra are black, with the exception of the actual place of 

 contact, where evidently the collodion film was impressed 

 mechanically. The diameter of the first dark ring in the first 

 spectrum is '90 inch, not quite the double of *54 inch, although 

 the same lens as before was used. In the second spectrum 

 the diameter of the first dark ring is "56 inch, and in the third 

 spectrum *40 inch. 



Interesting as these bands may be in theory, they are to be 

 avoided as much as possible in the practical reproduction of 

 gratings, not merely because a part of the area is lost, but 

 also on account of the reversal which takes place at every re- 

 vival of brightness. Without having examined the matter 

 very closely, I had generally found the performance of gratings 

 which showed these bands to be inferior ; and now it would 

 seem that the explanation is to be found in the above-men- 

 tioned reversals, which could not fail to interfere with the 

 resolving-power. 



During my early experiments it happened once that in the 

 course of printing an accidental shifting took place, leading to 

 the impression of a double image. A more perfect result was 

 afterwards obtained by intentionally communicating to the 

 plates a slight relative twist in the middle of the exposure* 

 When a spectrum from such a grating is thrown upon the 

 eye, parallel bars are seen perpendicular to the direction of 

 the grooves; but the number and position of these bars de- 

 pend upon the order of the spectrum. In one case twenty- 

 five bars were counted in the first spectrum, and twice that 

 number in the second. But it is unnecessary to dwell further 

 upon these observations, as they correspond exactly to what 

 the ordinary theory of gratings would lead us to expect. 



January 29. 



Phil Mag. S. 5. Vol. 11. No. 67. March 1880. Q 



