265 



posed of a certain number of rays, whose refrangibility is a discon- 

 tinuous one. The refraction of each ray is determined by the angle 

 between the middle lines of the image of the slit, seen directly, and 

 of the corresponding refracted band. This angle is independent of 

 the aperture of the slit, and remains the same if the slit be reduced 

 to a physical straight line. 



There cannot exist a deflected band smaller than 3' ; i. e. no such 

 band appears under an angle less than the angle under which the 

 aperture is seen directly, without the interference of the prism. This 

 law holds through all my numerous observations of electric spectra. 

 [I find difficulty in applying it to the case of the common solar 

 spectrum.] There are, in many instances, bands observed which are 

 seen under an angle greater than 3' ; but generally such bands are 

 resolved into two or more bands of single breadth. In some instances 

 where the angle in question does not reach the double of 3', there 

 appears in the middle part of the band a bright line ; larger bands 

 are generally divided by small, well-defined, dark lines. 



In order to distinguish the rays of different refrangibility in the 

 different gases, I denoted such a ray by adding to the symbol belong- 

 ing to the gas the Greek letters a, (3, y, indicating the succession of 

 the corresponding bands in each spectrum. Accordingly, for instance, 

 the band Ny, appearing under an angle of 6', is divided by a dark 

 line into two ; the bands NS and N0, having each a breadth of 10', 

 are divided by two dark lines into three single bands. The bands 

 Hga and Sn Cl 2 y, 5' broad, show a bright middle line. 



The space between two bright bands is either absolutely black, or 

 of a greyish tint, or of a faint tint of that colour which is indicated 

 by the position of the space within the spectrum. With regard to 

 these faint tints, the eye, by the effect of contrast, is commonly a very 

 bad judge of colour, and there may easily be a deception, admitting 

 a succession of colours in the spectrum which in reality does not 

 exist. 



I cannot enter here into the details by which I obtained exact 

 measures of the minimum refraction of the different rays. All 

 measures were taken without deranging the adjustment of the gonio- 

 meter ; for ascertaining its constancy, each spectrum was compared 

 with the spectrum of hydrogen. From the angles of refraction, I 

 deduced the indices of refraction, and hence the corresponding lengths 



