428 SPECIAL RESULTS IN THE URANOLOGICAL 



speaking, only a few came from the constellation of Leo ; 

 by far the greater number from that of Perseus. Hence it 

 appears to me to follow, that the great November pheno- 

 menon of 1799 and 1833 did not reappear at that time 

 (1841). Olbers also believed the maximum effect in the 

 November phenomenon to have a period of 34 years (Kos- 

 mos, Bd. i. S. 132 ; English edit. p. 117). If we consider 

 the directions of the paths of the meteors in all their com- 

 plication, and have regard to their periodical return, we 

 find that there are certain points or centres of radiation 

 which always recur, and others which appear only sporadi- 

 cally and in a variable manner." 



Whether the different points of departure vary from year 

 to year, ^which, if we assume the existence of "closed 

 rings," would indicate an alteration in the situation of the 

 rings in which the meteors move, cannot as yet be certainly 

 determined from the observations. A fine series of such 

 observations, by Houzeau (in the years 1839 1842), appears 

 to testify against a progressive variation ( 693 ). Eduard Heis 

 has shewn very justly ( 694 ) that in Greek and Roman anti- 

 quity, attention had already been drawn to a certain tempo- 

 rary uniformity in the direction in which the falling stars 

 shot across the celestial vault : this direction was then re- 

 garded as the effect of a wind already beginning to blow in 

 the higher parts of the atmosphere, and was thus believed 

 to announce to navigators an approaching gale from the 

 same quarter, which might be expected to descend from the 

 upper to the lower regions. 



If the periodical streams of shooting stars are distinguished 

 from sporadical ones by the general parallelism of the paths, 

 or by their radiating from one or more determinate points of 

 departure, a second criterion is also afforded by the num- 



