Intelligence and Miscellaneous Articles. 399 



of the solar rays, and it exhibited a well-defined maximum ; in the 

 position relative to this maximum, the pile received so much heat 

 that the deviation obtained often amounted to 20°. On the daik 

 side the spectrum extended to a distance at least equal to that which 

 separated the maximum from the luminous region. The whole of 

 this second part of the spectrum was completely absorbable by a 

 layer of water 2 millims. in thickness ; and the rays forming the maxi- 

 mum were almost completely absorbed. Scarcely a third part of the 

 rays escaped absorption, and it was necessary to approach the lumi- 

 nous region to rediscover well-marked transmissions. Thus in a 

 spectrum formed by rays which have not had to traverse an absorb- 

 ing layer of appreciable thickness, rays are found which are analo- 

 gous to those ordinarily proceeding from the sides of heated en- 

 closures. 



The apparatus I used in these spectrum-studies enabled me to 

 measure to a quarter of a degree the deviation of the pencil by the 

 prism of rock-salt, which w T as always placed in the position in which 

 the deviation of the red was a minimum. In making these mea- 

 surements I soon ascertained that pencils of the same breadth, and 

 the same mean refrangibility, were very differently absorbable by 2 

 millims. of water, according as I took them in the solar spectrum 

 or in a spectrum formed by rays emanating from platinum. In this 

 latter case the absorption is stronger than in the other. 



The differences are very precise, and cannot be attributed to the 

 presence of rays quite absorbable coming from the extremities of the 

 spectrum and which, owing to a defective arrangement of the appa- 

 ratus, were mixed with the rays to be operated upon. If, in fact, 

 there were such a flux of absorbable rays, there ought to be another of 

 rays much more transmissible coming from the most refracted por- 

 tion of the spectrum, and the effects would almost balance each 

 other. Moreover the following new experiment could leave no 

 doubt. A moderator lamp was taken as source of heat, and a very 

 pure spectrum was formed with the rays proceeding from it. In one 

 of my experiments the deviation of the red rays was 40°*5, that of 

 the violet 43°; and I obtained a few tenths of deviation by placing 

 the index which supported my pile at a distance of 38° from the di- 

 rect rays. When the alidade was 39°*9 from the incident ray, the 

 pile received no luminous ray, but was covered by a luminous band 

 near the extreme red. A trough of fluor-spar containing a layer of 

 water of 2 millims. thickness allowed about a third of the rays form- 

 ing this band to pass. Thus, for instance, 33° of deviation were ob- 

 tained with the direct, and 11° with the transmitted. Placing the 

 index at 39°' 2 from the direction of the incident rays, strong deflec- 

 tions in the direct were still observed ; but the transmissibility through 

 the same trough was singularly diminished ; it did not amount to 

 more than yVtj ' tnat i s > the mean deviation for the direct amounted 

 to 19°, and for the transmitted to 3°'l. 



Having made these observations, I modified the nature of the rays 

 by causing them, before falling on the prism, to pass through a layer 

 of water of 0*003 metre in thickness contained between two glass 

 plates. For the two positions which I have mentioned, the trans- 



