From the above table it is evident that the most frequent 

 direction of the cirrous bands nearly coincides with the magnetic 

 meridian ; whilst the maximum of the directions of their mo- 

 tions is (as appears from the foregoing tables) nearly at right 

 angles thereto. In the finest displays of cirri, I have almost 

 invariably found the direction of the parallel bands to be at right 

 angles to the line of motion. Cirri are often pennated, the fine 

 cirrous fibres forming angles with the lines of the principal 

 bands. They are also frequently dendriform. This latter 

 arrangement appears due to the same cause as that which pro- 

 duces similar forms on a pane of glass by the freezing of the 

 deposited moisture. True cirri are always composed of ice- 

 crystals (probably four-sided prisms), as appears from their 

 invariably producing halos of 47°, often accompanied with par- 

 helia or paraselense, by the refraction of the rays of the sun or 

 moon, whenever they intervene between us and these luminaries. 



The following table exhibits the directions of the cirrous fibres, 

 1841 to 1847 inclusive, no observations having been made upon 

 them in 1840. 



.0*81 



