of the Atmosphere. 1 73 



The following are all the values of D which I obtained. I 

 have added the values of R or the degrees of polarization, in the 

 zenith and in the horizon, when they happen to have been mea- 

 sured : — 



Observations in 1841. 











Rotation 



Rotation 





Apparent time, 

 h m 



3 





Values of D. 



in zenith. 



in horizon, 



April 28. 





88 16 



o 



o 



30. 



2 5 





79 15 



, . 





May 17- 



1 20 





99 



18* 



15i 



June 6. 



4 45 





+ 90 



16| 



22| 



9. 



4 30 





+ 90 



25| 



24! 



10. 



10 39 





92 35 







10. 



12 





88 26 



30 15 



29 



Sept. 15. 



10 18 





88 4 



27 



26! 



Oct. 26. 



4 30 





93 



28i 



27! 



Dec. 17- 



9 7 a.m. 



Appt. 



+ 90 



27 



24! 



18. 



9 „ 



time. 



+ 90 



28 



27 



Observations in 1842. 

 Jan. 7. 9 a.m. 120 Haze. 24, 



•il 5. 



6 58 



+ 90 



30! 



26\ 



20. 



12 49 p.m. 



84 



25 



17 



20. 



2 25 



86 



25 



19 



20. 



3 35 



88 20 



25 





20. 



4 32 



90 



27 



17 



21. 



11 a.m. 



90 



23 





21. 



12 



90 



23 





21. 



1 10 



90 



, , 



25 



25. 



1 17 



90 



26^ 





26. 



10 53 a.m. 



90 



29 





26. 



11 46 



88 



29 





26. 



3 42 p.m. 



88 



29 





27. 



10 41 a.m. 



87 



29 





28. 



11 34 a.m. 



87 



29 





28. 



1 50 p.m. 



88 



29 





29. 



1 34 



87! 



29 





f 3. 



11 29 a.m. 



89 



20 





Omitting the three extreme values of D, viz. 79° 15', 99° 

 and 120°, the mean of all the other values is 89°; but consider- 

 ing that five of the values of D are marked as more than 90°, 

 we may conclude that 90° is, in the normal state of the atmo- 

 sphere, the distance from the sun of the place of maximum 

 polarization, and 45° the corresponding angle of incidence. 



This determination of the place and angle of maximum polari- 

 zation affords a highly probable explanation of the azure colour 

 of the sky. Sir Isaac Newton considers this colour to be a "Blue 



of the first order, though very faint and little ; for all 



vapours, when they begin to condense and coalesce into small 



