ON SHOOTING STARS. 



obtain a number which may be used for the altitude of the middle point. This average 

 descent according to the table is about twenty geographic miles. But the average length of 

 track will be found to be between twenty one and thirty-four miles. If the points from 

 which the motion of the meteors is directed are uniformly distributed over the visible hemi- 

 sphere, then it may be shown that the average descent would be one-half the average length. 

 When, therefore, only the first or last altitude is given in the table, we ought, according to the 

 first result, to take ten miles from the fiust, or add ten miles to the last altitude, and use the 

 result as the altitude of the middle point. The second result gives a quantity between five 

 and a quarter and eight and a half miles to be added or subtracted. I have used, therefore, 

 eight miles as half the amount of descent. 



When one end only is given, the resulting altitude is counted once. When both ends are 

 given it is counted twice. Again, certain sets of altitudes are computed from base lines that 

 were too short or too long, or they are otherwise not deserving of full confidence. Such are 

 the flights observed and computed by Brandes and Benzenberg in 1798, 1801, and 1802, a 

 part of those by Brandes in 1823, those by the younger Brandes in 1833, those by Bogus- 

 lauski. Erman, and Litteow in 1837 and 1838, by Coulvier-Gravier in 1845, and by Le 

 Verrier in 1856. These are counted once only, while the other sets are counted twice. In 

 the best cases, therefore ,the altitudes are counted four times. Proceeding in this manner, and 

 converting miles into kilometres, we have the following numbers: 



Altitudes between and 30 kilometres 39 



111 



243 



277 



10G 



57 



"20 



20 



Altitudes over 300 kilometres 

 These numbers are exhibited graphically in figure 1. 



10 



2 





200 



JOO 



(292) 



Gd VO MO jS(FZ. ISO 2U> S40 X70 



