158 ANNUAL OF SCIENTIFIC DISCOVERY. 



at 90 degrees from the former, and at 180 degrees from each other, at 

 which the deviations are equal and minimum. The positions of the 

 analyser, which for a given point of the heavens procure the maxi- 

 mum and minimum transmissions, correspond to those of greatest 

 intensity of the direct bands and inverse bands of the polariscope of 

 Savart. They are thus determined without much difficulty. 



The atmospheric heat can be depolarized by means of a plate of 

 mica placed near the extremity of the exterior tube, and perpendicu- 

 lar to the incident rays. The analyser being in the position of the 

 minimum of transmission, the deviation of the index experiences no 

 serious diminution when the principal section of the interposed rnica 

 coincides with the plane of polarization, while the deviation is aug- 

 mented when the rotation of the mica in its own plane brings its prin- 

 cipal section to an angle of 45 with the primitive plane of polariza- 

 tion. 



The phenomena of the polarization of atmospheric heat are much 

 less apparent in winter than in summer. The difference is doubtless 

 due to the want of sufficient sensibility in the apparatus, to the greater 

 difficulty of experimenting at low temperatures, and to the small pro- 

 portion of polarized rays which on the most favorable days accompany 

 the natural heat. 



The serenity of the air exercises a marked influence on this pro- 

 portion, which becomes probably null when the heavens are obscured. 

 Finally, it is easy to satisfy one's self, particularly if the atmosphere 

 be calm and without clouds, that the polarization augments from the 

 environs of the sun up to a certain limit, from which forward it de- 

 creases. I have found it inappreciable in the regions occupied by 

 neutral points. 



NON-POLARIZATION OF THE AURORA BOREALIS. 



Although the results which I have obtained are purely negative, it 

 may be useful to record the fact that, having on several occasions dur- 

 ing the last eight months, examined the light of the aurora borealis 

 with a Nichol's prism, I liave never detected any trace of polarization. 

 To show that this did not arise from the faintness of the light, I may 

 mention, that on the last occasion when I observed it, the polarization 

 of the same light produced by reflexion from the surface of a river, 

 was distinctly visible, although the direct light was evidently free from 

 all sensible polarization. 



This fact is adverse to the supposition that the light of the aurora 

 borealis is reflected from crystals of ice. Prof. Rankine, Lond., Ed. 

 and Dub. Phil. Mag. 



LONGITUDINAL LINES OF THE SOLAR SPECTRUM. 



The following is an abstract of some recent investigations made by 

 Prof. Ragona-Scina on the lines exhibited in the solar spectrum. 

 Heretofore the longitudinal lines of the solar spectrum have attract- 



