82 Mr. H. H. HofTert on a new Apparatus 



Mr. Mitchell, to whom I am also indebted for several valuable 

 suggestions. 



We first tried a modified form of Maxwell's colour-box, 

 using eventually six equilateral glass prisms, arranged in two 

 sets of three symmetrically with respect to the eyepiece ; so 

 that if the latter be made a source of light, the rays are bent 

 through about 150°, and being then received on to a lens, are 

 focused on to planes situated on either side of the eyepiece, 

 this arrangement being found very convenient for manipu- 

 lation. 



Movable slits of fixed width, sliding in a frame with bellows- 

 screens of opaque silk between, were used as sources of light, 

 lamps being placed outside them. The intensity of the light 

 was regulated by means of Nicol prisms, one being attached 

 to each slit, so as to be capable of rotation, while an analyzing 

 Nicol was fixed in the eyepiece. -The intensity could thus be 

 readily altered for each slit separately. 



This arrangement was found to answer fairly well, but its 

 necessarily large dimensions were a great inconvenience. A 

 very long spectrum was needed, as the slits could not be 

 brought very close on account of the size of the Nicol prisms; 

 hence also there was great loss of light. 



Movable slits, whether of fixed or adjustable width, are 

 very troublesome to manipulate, and do not admit of very 

 rapid and easy adjustment ; they were therefore abandoned 

 altogether, and in their place the light from small lengths of 

 platinum wire, heated to incandescence by an electric current, 

 was employed. Preliminary trials having proved satisfactory, 

 the apparatus was constructed which I now bring before your 

 notice. 



The prisms, lenses, and sources of light are enclosed in a flat 

 box of irregular hexagonal shape, at the smaller end of which 

 the six prisms (P, PI. V.) are placed. These are arranged, as 

 already described, in two sets of three, each prism being set 

 to minimum deviation. The first prisms have their refracting- 

 edges in contact; and by means of a screen (S), in which is a 

 small rectangular aperture, small equal strips of the adjacent 

 faces of these prisms are visible from the eyepiece (E), which 

 is a small brass tube with a narrow slit, about ^ inch wide. 



Supposing light to come from the eyepiece, the prisms 

 would deflect it through about 150° ; it would then fall on to 

 two lenses (L), of about 10 inches focal length, which focus 

 the spectra on to the two sides of the box immediately to the 

 right and left of the eyepiece. Here are placed the incan- 

 descent wires (W); so that, conversely, if these be the sources 

 of light, the rays follow an inverse course, and the correspond- 



