of the Prismatic Spectrum. 349 



(Inly dilated. In order to obtain a map independent of any 

 particular setting of the prism, it is necessary to readjust 

 to minimum deviation for each ray that is observed. But 

 even this map depends upon the refracting-angle. 



The best representation of the prismatic spectrum for theo- 

 retical purposes is one in which the position of a fixed lino 

 measures its refractive index (which may be calculated in the 

 ordinary way from the observed value of the miniminn devia- 

 tion). If we then erect ordinates inversely proportional to the 

 squares of the wave-lengths of the principal hxcd lines, the 

 points thus determined will be found to lie very nearly on a 

 straight line, whose prolongation cuts the spectrum at the 

 limit of refraction. In the case of an ordinary spectrum the 

 line determined in this way is curved, and its prolongation 

 cannot be made with the same confidence ; nevertheless there 

 is no difficulty in marking approximately the required limit. 



In the ^ Philosophical Transactions ' for 1840 there is a very 

 interesting memoir by Sir John Herschel, in which he de- 

 scribed an ingenious method of obtaining a thermal spectrum. 

 The solar rays were received directly on the prisms without the 

 intervention of a mirror or slit, and after refraction were col- 

 lected to a focus by an achromatic lens of 25 inches focal length. 

 Two flint-glass prisms of 45° in the position of minimum de- 

 viation Avere employed; and the apparatus was driven by clock- 

 work so as to follow the apparent motion of the sun. In the 

 focus of the lens was placed a strip of thin paper, of which the 

 hinder side had been carefully blackened with Indian ink, or 

 (preferably) with the smoke of oil of turpentine. When all 

 w^as ready for an experiment, the clean side of the paper was 

 washed over rapidly with strong spirits of wine, the effect of 

 which was to render the paper transparent so as to allow the 

 smoke to show through. As the alcohol evaporated under the 

 influence of the thermal rays, the paper gradually recovered its 

 whiteness. " After a few minutes a whitish spot begins to ap- 

 pear considerably below the extreme red end of the luminous 

 spectrum, which rapidly increases in breadth until it equals the 

 breadth of the luminous spectrum, and even sometimes sur- 

 passes it — and in length until it forms a long appendage ex- 

 terior to the spectrum, and extends moreover within it, up to 

 and beyond the fiducial yellov/. In this state, and just as the 

 general drying of the paper begins by whitening the whole 

 surface to confuse the appearances, a second, sudden and 

 copious, wash of alcohol from above dovnivv'ards must be ap- 

 plied, without disturbing the spectrum or in any way shaking 

 the apparatus. The superfluous alcohol will have hardly run 

 cff when the phenomena of the thermic spectrum will begin 



