Intelligence and Miscellaneous Articles. 333 



monia. I intend to prepare and analyze it further. — Coniptes Hin- 

 dus de VAcademie des Sciences, vol. lxxxii. pp. 223-226. 



The lines described by M. Schuster * have not been found again 

 by MM. Steam and Wiillner, who, after verifying the disappear- 

 ance of the band spectrum, saw the well-known spectrum app?ar 

 of the oxide of carbon or acetylene. Schuster's tubes more care- 

 fully prepared gave me, after a certain time, the bright lines of 

 hydrogen only. What, then, was the origin of the lines described 

 in the memoir of 1872 ? It is permissible to attribute them to the 

 vapour of sodium. 



In fact, we remark that the published numbers were not got by 

 direct measurement, but refer to the nitrogen lines of Pliicker's 

 spectrum which appear to coincide with the lines observed ; and 

 these, moreover, are not the most characteristic. Now, if sodium 

 be heated in a Greissler tube, or in an apparatus like that described 

 in §1, and the spark be made to pass, a bright greenish yellow light 

 is produced, the spectrum of which is composed of the following 

 lines : Spectrum of sodium. 



615-5 double . . ' 515-3 



589-2 „ (D) 498-3 



568-7 „ 467 



These numbers are very near those of M. Schuster; only the 

 line 498-3 does not figure in his Plate, it is replaced by 489-4. 

 Conversely, we find in that Plate three extreme lines — 628*8, 

 421-4, and 418*4, which I have not been able to produce, nor are 

 they indicated by MM. Thalen and Lecoq de Boisbaudran. A 

 curious thing, these lines nearly coincide with the characteristic 

 lines of rubidium (629*6, 421*6, and 420*2). I would not by any 

 means say that this metal was present in M. Schuster's sodium, 

 especially with the uncertainty of the wave-lengths ; but in any 

 case it must very readily give a spectrum in a G-eissler tube ; for 

 potassium offers much better facilities for this kind of experiments 

 than sodium. It furuishes, without difficulty, a very pure spectrum, 

 of which the principal lines are the following : — 

 Spectrum of potassium. 

 583 535*3 511 404*4 



580 533-5 509 



578-3 532 

 It must be possible, employing the same method, but under 

 slightly different conditions of experiment, to produce not only the 

 secondary spectra of the alkali-metals, as we have done in the 

 present researches, but also their primary spectra, for the inter- 

 esting discovery of which we are indebted to Messrs. Eoscoe and 

 Schuster. — Comptes Rendus de VAcademie des Sciences, vol. lxxxii. 

 pp. 274, 275. 



ON CERTAIN REMARKABLE POINTS IN MAGNETS. BY R. BLONDLOT. 



If a very short magnetic needle, the centre of gravity of which 

 is supported, be brought near the surface of a magnet, the direction 

 * Pogg. Ann. vol. cxlvii. p. 106. 



