October 26, 1900.] 



SCIENCE. 



619 



second grating, B, is adjusted with the wires 

 parallel to A ; and at a distance, y, from the 

 latter is the focal plane S, visible to the eye 

 behind the lens (or in the distant corre- 

 spondingly focused telescope, looking along 

 L'L in Fig. 2, as will be explained below). 



in which relations of x and y for the case of 

 a = 6 have been inserted as an example of 

 many similar data, will be intelligible at 

 once. 



Naturally these results are crude, but as 

 their import is unmistakable, it is not 



a 



\A 



\'jB 



$ 



L^- 



d 



''P 



f<'r- 



Fig.i. 



>i: 



K 



^ 



-y- 



[S_. 



r 



I 





Jig.Z 



It will be convenient to call the grating 

 space at A, a; the space at B, h ; and the 

 space of the image at 8, s, all being parallel. 

 Then the experimental results of Table 1, 



TABLE 1. — EXAMPLE OF FOCAL PLANES FOR GRAT- 

 INGS WITH EQUAL MESHES. (1^6^.214 CM. 

 AND WIRES .030 CM. IN DIAMETER, LENS 

 FOCUS 15 CM. a =: 6. 



necessary to push the experiment further. 

 The first definite result derived from them 

 is this, that the focal planes are distributed 

 along the axis at distances ^, 1, 2, etc., 

 multiples and submultiples of the distance 

 of the gratings apart, when the two gratings 

 are identical, or a = h. The size of the 

 images is usually directly as the distance 

 y from grating B, and if for a=h, x = y, 

 then a = 6 = s, or image and object are 

 equally large. Eemote focal planes are apt 

 to be diffuse and colored nearly uniformly 

 red and blue in alternate bands. Hence 

 the number of foci accessible in this way is 

 not large. 



If the meshes are unequal, the focal 

 planes are still apt to be distributed at dis- 

 tances varying as 1, 2, 4, etc., along the axis. 

 Corresponding distances, y, are smaller rela- 

 tive to X if the projecting grating is finer. 

 The law of distribution is not easily worked 

 out in this way, however, because it is 

 difiBcult to obtain gratings of different 

 meshes but of the same diameter of wire. 

 Neither is it safe to infer the size of image 

 from these experiments. The problem 

 must be attacked in another way. 



