PEOFESSOR TYNDALL ON CALOEESCENCE. 3 



elements 1*2 inch in length being fixed to the screen behind the aperture. Connected 

 with the latter are two moveable side pieces, which can be caused to approach or recede 

 so as to vary the width of the exposed face of the pile from zero to y^th of an inch. 

 The instrument is mounted on a slider, which, by turning a handle, is moved along a slot 

 on a massive metal stand. A spectrum of a width equal to the length of the thermo- 

 electric pile being cast at the proper elevation on the screen, by turning the handle 

 of the slider the vertical face of the pile can be caused to traverse the colours, and also 

 the spaces right and left of them. 



To produce a steady spectrum of the electric light, I employed the regulator devised by 

 M. FoucAULT and constructed by Duboscq, the constancy of which is admirable. A com- 

 plete rock-salt train was constructed for me by Mr. Becker, and arranged in the fol- 

 lowing manner. In the front orifice of the camera which surrounds the electric lamp 

 was placed a lens of transparent rock-salt, intended to reduce to parallelism the divergent 

 rays proceeding from the carbon-points. The parallel beam was permitted to pass 

 through a narrow slit, in front of which was placed another rock-salt lens, the position 

 of this lens being so arranged that a sharply defined image * of the slit was obtained at a 

 distance beyond it equal to that at which the spectrum was to be formed. Immediately 

 behind this lens was placed a pure rock-salt prism (sometimes two of them). The beam 

 was thus decomposed, a brilliant horizontal spectrum being cast upon the screen which 

 bore the thermo-electric pile. By turning the handle already referred to, the face of 

 the pile could be caused to traverse the spectrum, an extremely narrow band of light 

 or radiant heat falling upon it at each point of its march f . A sensitive galvanometer 

 was connected with the pile, and from its deflection the heating-power of every part of 

 the spectrum, visible and invisible, was determined. 



Two modes of moving the instrument were practised. In the first the face of the 

 pile was brought up to the violet end of the spectrum, where the heat was insensible, 

 and then moved through the colours to the red, then past the red up to the position 

 of maximum heat, and afterwards beyond this position until the heat of the invi- 

 sible spectrum gradually faded away. The following Table contains a series of measure- 

 ments executed in this manner. The motion of the pile is measured by turns of its 

 handle, every turn corresponding to the shifting of the face of the instrument through a 

 space of one millimetre, or -^5 th of an inch. At the beginning, where the increment of 

 heat was slow and gradual, the readings were taken at every two turns of the handle ; 

 on quitting the red, where the heat suddenly increases, the intervals were only half a 

 turn, while near the maximum, where the changes were most sudden, the intervals were 

 reduced to a quarter of a turn, which corresponded to a translation of the pile through 

 Y^th of an inch. Intervals of one and of two turns were afterwards resumed until the 

 heating-power ceased to be distinct. At every halting-place the deflection of the needle 

 was noted, the value of the deflection, referred to the first degree as unit, being placed 



* The width of the image was about O'l of an inch. 

 t The width of the linear pile was 0-03 of an inch. 

 b2 



