Intelligence and Miscellaneous Articles. 239 



to the corresponding diminution of deviation increases rapidly in the 

 less refrangible part of the spectrum. But the ray A of the extreme 

 red is difficult to work with, on account of the feeble lustre of the 

 solar light in this part of the spectrum ; its wave-length is not known, 

 and it is not mentioned in most of the tables of refraction hitherto 

 published. 



It appeared to me interesting to determine this wave-length by the 

 aid of a grating ; and I used for this purpose, not the solar light, but 

 the least deflected bright ray of potash-salts, which, as the recent 

 experiments of M. Kirchhoff have shown, coincides exactly with the 

 ray A. 



I used a Babinet's goniometer showing 10 seconds, and a grating 

 of about 4 square centimetres superficies divided into fortieths of a 

 millimetre ; I compared the deviation of the line A with that of the 

 bright soda line. To obtain the greatest intensity possible in the 

 luminous source, I used several methods, especially the combustion 

 of hydrogen charged with the vapours of potassium (as MM. Wolf and 

 Diacon have done on the suggestion of M. Foucault), and the volati- 

 lization of chloride of potassium in an oxyhydrogen blowpipe. The 

 latter method, devised by M. Debray, always succeeded best. Spite 

 of these precautions I could only observe the first spectrum, which 

 gave little value to isolated observations ; but the mean of a great 

 number of concordant measures was about 768 million ths of a 

 millimetre. 



For the wave-length of the line A the number 750 millionths of a 

 millimetre is generally received — a number deduced from the theore- 

 tical law of dispersion found by M. Cauchy, or from an interpola- 

 tion formula. When the season is more favourable to this kind of 

 experiments, I propose to determine the refractive indices of the ray 

 A in different substances, and to see if Cauchy's law agrees suffi- 

 ciently with experiment. 



This research has led me to another observation. As I needed a 

 very intense luminous source, I thought of the volatilization of potas- 

 sium between the two poles of a powerful voltaic battery ; but the 

 result did not correspond to my anticipation. I obtained a magnifi- 

 cent spectrum, more complicated than those which have been hitherto 

 indicated for potassium ; the red line corresponding to the solar 

 line B was very intense and quite double : but I sought in vain for 

 the line A ; and examining with care, I saw a feeble red illumination 

 on both sides of an obscure space in the region of the luminous line ; 

 I could even distinguish a brilliant line between two black lines ; 

 that is to say, the extreme double line of potassium was inverted. 

 This partial reversal of the potash lines does not seem to me to be 

 in disaccord with M. Kirchhoff's theory ; for it is to be noted that 

 the line which is reversed is that which is produced at a lower tem- 

 perature. The same phenomenon takes place with sodium as in 

 M. Fizeau's experiment ; the double line D is the only one which is 

 reversed ; but this reversal has a peculiarity which every one has been 

 able to observe : it is that the reversed black line can in certain cir- 

 cumstances increase to a considerable extent, retaining always well- 

 defined limits. The preceding method is advantageous for deter* 



