OP AKTS AND SCIENCES. 153 



with the enveloped pile, was fixed to a slide, b d', so that it could be 

 moved through the spectrum ; the position of the pile being read off 

 on a scale at d'. The tin case and pile also had a vertical motion, so 

 that it could be raised and replaced by a simple slit, which was directly 

 under the face of the pile. This slit was used in reducing the spec- 

 trum to that of the sun, by a method to be presently described. 



Of course the heat radiated by the wire is very feeble, particularly 

 after being dispersed by the prism, and we know that even the spec- 

 trum of the sun is not easy to measure. The ordinary short-coil 

 Thomson galvanometer, when used with the pile, gave scarcely any 

 deflection, even at the point of maximum heating effect of the spec- 

 trum, and in the luminous spectrum it gave no deflection at all. 



Accordingly the Thomson galvanometer was modified as follows. 

 The brass rod and magnet were removed, a hole made in the top, 

 and a glass tube, about 25 cm - long, placed over this. Inside this was 

 suspended an exceedingly fine silk thread, to the bottom of which was 

 suspended the needle. It was necessary to form a suspension of this 

 length, in order to eliminate the viscosity, which, even in a needle as 

 delicate as the one used, was quite noticeable. The needle was made 

 of eight small bars, about 8 mra - long, l mm - wide, and l mm - thick, placed 

 four in the centre of the coil, and the other four, with poles reversed, 

 below the coils. They were as nearly as possible in the same plane, 

 and the system was found, when adjusted, to be so nearly astatic as to 

 swing in fifteen seconds. The slight adjustment that was necessary 

 was made by means of a small magnet, arranged so as to move in any 

 plane, and placed at the side of the galvanometer at m. 



This galvanometer was placed on a stone shelf at G", and its 

 deflections read off by means of a scale and a spot of light at d. C" 

 was a commutator for reversing the current through the galvanometer. 

 This galvanometer was, of course, exceedingly delicate ; the passing 

 of a finger by the pile was enough to throw the spot of light off the 

 scale. It was exceedingly sensitive to changes in the earth's magnet- 

 ism, so much so that on many days no observations could be taken at 

 all. This, however', was avoided in general by reading on both sides 

 of the point of the scale, and by making the observations late in 

 the evening, when the magnetic disturbances were at a minimum. Its 

 sensitiveness, however, enabled the observer to detect the radiations 

 of the violet part of the spectrum when the wire was heated to a 

 white heat. 



One great advantage of using so slight heating effects of the pile, 

 and determining these by a low-resistance galvanometer, is that we 



