iSyi.] Light 271 



of the spectrum will also rotate the disc and cause the prisms to assume their 

 proper positions for that particular refrangibility of light under examination. 



And, finally, a point, g, in the limb of the sector is coupled to a fixed centre, 

 g', in the bed plate by another connecting rod,g-, g'. This is for the purpose of 

 compelling the sector to travel in a direction parallel to itself; that is to say, 

 if in any position of the instrument a line were drawn upon the sector parallel 

 to the collimating telescope, it would also be parallel to it in every other 

 position. Hence, a little consideration will show that the readings of the 

 sector give actual differences of deviation. 



In fact, if the two telescopes were placed parallel to each other, and that 

 point marked 180 on the sector and the other parts numbered up and down 

 from that, the instrument would then read actual deviations. 



In practice, however, such a system of division will probably not be found 

 as convenient as some others, as it is only differences that are required ; but 

 these differences, it will be seen, will not be correctly given unless the sector 

 has that particular motion imparted to it described above. 



Two-Prism (Compound) Spectroscope on Automatic Principle. — The two-prism 

 spectroscope for Dr. Huggins's star observations is constructed exactly on this 

 same principle, modified, of course, to suit the two instead of three prisms. 

 The prisms are compound, and admit of a pencil of light, i£" x 1" to pass to 

 objectives of telescopes which are i|" diameter. The power of the spectroscope 

 is, therefore, equal to that of four large prisms of 6o°. There is one common 

 collimator used for several spectroscopes; this collimator is 1 £" aperture and 

 4^" focus. The great shortness of focus is obtained by making the object- 

 glass of that quadruple form which Mr. Grubb applied to the 7-inch circle 

 telescope in Armagh. The advantage of having the collimator of large angular 

 aperture is that a more powerful cylindrical condensing lens can then be used 

 and greater brilliancy obtained. 



Compound Prisms. — The compound prisms devised by Mr. Grubb, F.R.S., 

 for spectroscopic investigation are composed of three elements. The centre 

 one being of dense flint glass of an angle varying from go to ioo°, according 

 to the exact nature of the glass, and crown prisms cemented on each side of 

 about one-fourth that angle reversed in the direction of their apices. By 

 cementing these crown prisms on the centre flint, the light which otherwise 

 would not be capable of passing either in or out of the prism is enabled to do so, 

 and a dispersion is by this means obtained equal to about two 6o° prisms, and 

 with a deviation equal to about one 6o° prism. Various prisms have been 

 made from time to time on the compound principle, and by various artists, 

 but in most cases they have been tied with the condition of " direct vision," 

 which is by no means necessary, except for special purposes, and sometimes 

 even undesirable. 



In other cases the centre flint has been made of extravagant angle ; this is 

 also useless, as crown prisms of great angle must then be attached, which again 

 brings down the dispersion ; and the inside surface being so much inclined to 

 the course of rays much light suffers internal reflection. In the proportion 

 arrived at by Mr. Grubb, the required qualities have been found both theoretically 

 and practically to be the best balance. The advantages in using them can be 

 summed up as follows : — Double the dispersion is obtained by them as compared 

 with an equal number of ordinary 6o° prisms with the same number of surfaces 

 (the inside not counting, being cemented), a less absorption from the glass, one- 

 half the complication of mechanism (particularly if automatic spectroscopes 

 be used), and freedom from danger of injury to surfaces of flint glass. 



Mr. Browning has contrived a direct vision spectroscope, for use in the 

 Bessemer process. It has great dispersive power, with very low magnification, 

 the result being a very luminous spectrum, which fills the whole field of view. 

 The jaws of the slit are opened or closed by turning a milled ring just behind 

 them. There are no adjusting screws or movable parts to this instrument. It 

 is made with what are known as fixed adjustments, so that in an ordinary 

 workman's hands there is nothing to get out of order. The instrument contains 

 ten prisms, and will divide the d lines in the solar spectrum quite easily. 



