SHOOTING STARS. 579 



4 or 5 sporadic falls. The meteors appear to be the most 

 seldom in January (calculating from the 4th), February and 

 March. 19 



" Although the August and November periods are justly the 

 most celebrated, still since the shooting stars have been 

 observed with greater accuracy, as to their number and 

 parallel direction, yet five others have been discovered. 



January : during the first days between the 1st and 

 3rd ; probably somewhat doubtful. 



April: 18th or 20th? already conjectured by Arago. 

 (Great streams : 25th of April, 1095, 22nd of April 1800 

 20th of April, 1803 ; Cosmos, vol. i. pp. 113-114. An- 

 nuaire pour 1836, p. 297.) 

 May : 26th ? 



July : 26th to the 30th ; Quetelet. Maximum pro- 

 perly between the 27th and 29th of July. The most 

 ancient Chinese observations gave Edward Biot (unfor- 

 tunately too soon taken away) a general maximum 

 between the 18th and 27th of July. 



August, but before the Laurentius stream, especially 

 between the 2nd and 5th of the month. For the most 

 part no regular increase is remarked from the 20th of 

 July to the 10th of August. 



The Laurentius stream itself, Musschenbrock 



and Brandes (Cosmos, vol. i. p. 112, and note). Decided 

 maximum on the 10th of August ; observed for many 

 years. (According to an old tradition which is diffused 

 among the mountain-regions about Pelion in Thessaly, 

 on the feast of the Transfiguration, the 6th of March, the 



19 1 have, however, myself observed a considerable fall of 

 shooting stars on the 16th of March, 1803, in the South Sea 

 (Lat. 13^ N.). Also 687 years before our era, two meteor- 

 streams were seen in China, in the month of March (Cosmos, 

 vol. i. p. 116). 



VOL. IV. V 



