on the Eleme7itary Colours of the Spectrum. 491 



affected by the dilatation of the pupil, and the consequent 

 confusion of vision, M. Melloni repeats the experiment which 

 he calls mine, but which I disavow, with a square aperture in 

 place of a circular one. What would Fraunhofer, WoUaston, 

 and Young think of a spectrum produced from a square aper- 

 ture inscribed in a dwcXo. four -tenths of an inch in diameter*! 

 But waiving this objection, let us examine the experiment 

 itself He goes on to say, that he sees " a red rectangle 

 almost square-^, followed by a broad dark zone, and then by 

 a very brilliant yellow rectangle, of which the longer sides were 

 directed vertically and parallel to the length of the spectrum % ; 

 there came then a deep indistinct colour, then the blue." This 

 i/ellow rectangle was found to be more elongated horizontally 

 in the elementary spectrum, or that produced by a narrow 

 prism, than in the spectrum produced by a wider prism ; and 

 hence Melloni concludes that the increased elongation arises 

 from the overlapping of the rays produced by the wide prism, 

 and not from coif used vision §, because there is no increase in 

 a vertical direction. He takes it for granted that I have used 

 a wide prism, and consequently infers that I studied a spec- 

 trum in which the colours were overlapped. I deny the as- 

 sumption as well as the inference. 



Viewing the experiments of Melloni by themselves, and 

 admitting the longitudinal expansion of the luminous rect- 

 angles to be a fact, I ascribe it principally to the width of his 

 aperture, and in a secondary degree to irradiation ; or it may 

 arise from an indistinctness of vision, or from an unusual quan- 

 tity of filaments floating in the vitreous humour of his eye. * 



In describing the colours of his spectrum as altered by 

 absorption, he mentions a aw/ rectangle, a dark space, a yellow 

 rectangle, then a deep indistinct colour, then the blue. 

 Now we ask him what he means by a deep indistinct colour'^ 

 Is \i yellow, or green, or yellowish-green ? It must be one of 

 these. II it h yellowish-green ov greenish-yellow, then certain 



* " If," says Dr. Young, " the breadth of the aperture viewed through a 

 prism be somewJiat increased, each portion encroaches on tlie neighbouring 

 colours and is mixed with them," &c. — Lect. on Nat. Phil., vol. i. p. 439. 

 M. Melloni increased the aperture of Fraunhofer from one-fiftieth to twenty- 

 fiftieths, and hence he useil ;i mixed spectrum. 



f 'I'he word almost is not known in geometry. The prismatic image of 

 ?. square aperture could only be a square when every ray of light passing 

 through that aperture had the same index of refraction. 



X This is to us quite unintelligible ; but we presume the author means 

 that the yellow rectangle was longer in the direction of the length of the 

 spectrum than in a vertical direction. 



§ M. Melloni does not seem to be aware that the apparent magnitudes 

 of luminous spaces depend upon irradiation, or an expansion of the invage 

 on the retina, which varies with the degree of illumination. 



