260 Prof. Liveing, On a new spectroscope. [May 1&, 



greater for orange light. So the green part of the spectrum is in 

 comparison with the red even more drawn out by the mercury 

 solution than by carbon disulphide. The figures above given 

 depend only on one set of observations, and are therefore open to 

 correction by future examination. 



(2) Professor LiVEiNG, On a new spectroscope. 



The aim of this instrument is to obtain considerable dispersion 

 with as little loss of light as possible, and at the same time good 

 definition in all parts of the spectrum. Young has pointed out 

 the economy of using one half-prism attached to the collimator 

 with one face perpendicular to the axis, and another similar 

 half- prism attached in a similar way to the telescope. A beam 

 of light can thus be used larger, in proportion to. the size of the 

 face of the prism, than when a single prism with an angle equal 

 to the sum of the angles of the half-prisms is used. Thollon has 

 given a mathematical investigation to show that when two half- 

 prisms are thus used and the incident and emergent light is normal 

 to the first and last faces of the pair respectively, the deviation is a 

 minimum. It is, however, easy to see without analysis that two 

 prisms of 30" placed as in Fig. 1 (Plate VII.), so as to be equivalent 

 to a single prism of QQf at minimum deviation, have the angles of 

 incidence and refraction the same as when they are placed as in 

 Fig. 2, only these angles occur in a different order. The deviation 

 will be the same in each case, and there will" be equal clearness of 

 definition. In the spectroscope now described, two half-prisms are 

 attached to the collimator and telescope respectively, and there are 

 two whole prisms between them. These however are all compound 

 prisms, so as to get considerable dispersion without the loss of light 

 involved in a very oblique incidence. 



With so great a dispersion as these two whole and two half- 

 prisms give, it is necessary to adjust the prisms to minimum 

 deviation, or nearly so, in order to get good definition, or even to 

 see faint lines at the extremities of the spectrum. In order, 

 therefore, to be able to use it for the kind of work in which the 

 author has, in conjunction with Professor Dewar, been for some 

 time past engaged, which often involves a rapid transition from one 

 end of the spectrum to the other, some automatic adjustment of the 

 prisms was desirable. This is effected by a simple system which 

 cannot easily go wrong. • The collimator L, see Fig. 3, is fixed to 

 the table of the instrument so that the prolongation of the first 

 face of the half-prism attached to it may pass through the centre 

 of the table ; a broad arm E, turning on an axis G at the centre of 

 the table, carries the telescope T and a vernier by which the angle 

 through which the telescope is turned can be read upon the 

 graduated rim of the table. The two whole prisms are each carried 



