530 Prof. Trowbridge and Mr. (Jrandall on Groove-Form 



mirror M, slit S 1? mirrors M l and Mj', so that the spectra 

 from the grating fell on the slit of the second spectrometer 

 at S 2 . The first spectrometer is thus arranged as an ordinary 

 grating spectrometer with fixed angle between collimators. 

 Successive orders on each side of the central image are 

 brought onto the slit S 2 by rotating the grating through an 

 angle S such that 



nX = 2b cos 6 sin S, 



where W is the angle between collimators and b the grating 

 space. 



For the second spectrometer the vacuum instrument de- 

 signed by one of the present authors * served to resolve the 

 energy falling on the slit into the different components due 

 to the overlapping of orders of the grating spectra. The 

 two 30° rock-salt prisms P and P' together with the plane 

 mirror L form a Wadsworth combination, which is equal in 

 dispersive power to a 60° prism. The two mirrors M 2 and 

 M 2 ', placed so that the slit S 2 and the bolometer strip B are 

 at their respective foci, complete the optical s}^stem of the 

 second spectrometer. The bolometer strip B was exactly 

 the width of the slit S 2 , namely about 0*7 mm., or about 

 3 minutes of arc in terms of the rotation of the prism system. 

 The current through the strip was about 0*05 ampere, and 

 the sensitiveness of the galvanometer was about 5 x 10 -10 

 ampere per scale-division (millimetre), its resistance being 

 3 ohms. 



In practice the vacuum spectrometer was set so that a 

 particular wave-length interval fell on the strip B, and the 

 grating was turned through the series of angles : thus 

 were obtained a series of deflexions which were proportional 

 to the energy radiated from the Nernst lamp for this wave- 

 length, and diffracted by the grating into the several orders of 

 spectra. The observations for the three gratings are plotted 

 in the curves which constitute the three accompanying 

 Piates. The abscissas represent the angular readings of the 

 grating spectrometer ; the ordinates the corresponding de- 

 flexions in centimetres. There was also taken a series of 

 observations to determine the energy radiated from the 

 Nernst lamp, and undergoing reflexion from a gold mirror, 

 as this is the kind of surface on which the gratings are ruled. 

 In considering the curve obtained for a grating, using any 

 particular wave-length, the relative quantity of energy avail- 

 able at this wave-length due to the above conditions must 

 be kept in mincl ; also there may be noted absorption bands 

 * A. Trowbridge, Phil. Mag. [6] vol. xx. p. 768 (1910). 



