204 Sir John Convoy on the Distribution 



p. 337), Franz (Pogg. Ann. cxv. p. 266), Knoblauch (Pogg. 

 Ann. cxx. p. 177, and cxxxvi. p. G(}), Fizeau and FoucauJt 

 (Comptes Rendus, xxv. p. 447, and reprinted in the Annates de 

 Chimie, 5th series, xv. p. 363), Desains (Comjytes Rendus, lxvii. 

 p. 297, and lxx. p. 985), Lamansky (Pogg. Ann. cxlvi. p. 200). 

 Similar measurements were made with the limelight by Des- 

 ains (loc. cit.) and Lamansky (loc. cit), and of the. electric 

 light by Professor Tyndall (Phil. Trans. 1866, p. 1). They 

 all found but slight indications of heat in the violet and blue 

 regions of the spectrum, the amount increasing in the green, 

 yellow, and red, and attaining a maximum at a point beyond 

 the end of the visible spectrum. 



The experiments were all made by placing a thermometer 

 (one of the ordinary construction being used by Sir W. Her- 

 schel and MM. Fizeau and Foucault, and a thermopile and 

 galvanometer by the other observers) in various parts of the 

 dispersion-spectrum formed by prisms of either glass, rock-salt, 

 or sylvine. As Dr. Draper points out in the paper already 

 referred to, this method appears to be an essentially defective 

 one, as, owing to the unequal dispersion by the prism of rays 

 of different refrangibility, a greater number of undulations of 

 different wave-lengths must have been incident upon the sur- 

 face of the thermometer when it was placed in the red and yel- 

 low portions of the spectrum than when placed in the green, 

 blue, or violet portions ; and the amount of heat indicated 

 by the instrument being in proportion to the amount of radiant 

 energy incident upon its surface, the unequal dispersion of the 

 prism would be sufficient to account for some difference in 

 the heating effects produced by different portions of the spec- 

 trum 



A graphical method appearing to afford the readiest means 

 of determining the probable effect produced by the unequal 

 dispersion of the prism, a tracing was made, on paper divided 

 into squares of ^ inch, of the curve representing the intensity 

 of the heat in different portions of the visible spectrum, as deter- 

 mined by MM. Fizeau and Foucault (Ann. de Cliim. 5 ser. xv. 

 p. 377) — the position of the fixed lines in the spectrum, as 

 given by them, being marked on one edge of the paper, which 

 was taken as the x axis, and a scale of wave-lengths in " tenth- 

 metres " laid down at right angles to this, and the curve for 

 the dispersion of the prism constructed in the ordinary manner. 

 At nineteen equidistant points in the spectrum the ordinates 

 of the dispersion-curve were measured in wave-lengths ; the 

 difference between any two of them gave, approximately, the 

 dispersion of the prism for that portion of the spectrum. A 



