32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 66 



average for the length of the visible path 57 km., on the other hand 

 for 120 large and of these about 30 per cent detonating meteors from 

 my own and other computations I find on the average the length of 

 the path 319 km. In some cases, especially when the radiant point 

 lies in the neighborhood of the horizon, there occur still much longer 

 paths. Such an example is here given. 



The meteor of July 7, 1892,^ was visible when it was 74 km. above 

 the neighborhood of Slobozia in Roumania and could be followed 

 without break until it was 158 km. high about 70 km. WNW. of the 

 mouth of the Tiber above the Tyrrhenian sea. There the observa- 

 tions could no longer be made, not so much because of an explosion, 

 as by reason of the apparently gradual recession. It is not unlikely 

 that at this great altitude the remaining part of this mass actually 

 left the atmosphere. The demonstrated length of the path was 

 1350 km. This is the first and hitherto also the only case known to 

 me of an undoubted ascending path. Since the point of perigee was 

 undoubtedly at about 74 km., therefore, to the east of Slobozia there 

 must have been a length of path at least as long. 



It is worth remarking that in 1892, October 18, in lower Austria 

 and Bohemia, a meteor was observed,* especially because of its 

 remarkable long trail which the residuum left behind. So far as the 

 path has been determined with any certainty, this meteor in the course 

 of 1(^30 km. passed from an altitude of 257 km. above a region 70 km. 

 east of Konigsburg in Prussia, until it was 43 km. high somewhat to 

 the iouth of the island of Elba. The streak of light or trail, straight 

 and horizontal as if drawn with a ruler, was at least 634 km. long and 

 remained visible for about three minutes, so that its location between 

 the stars could be easily determined. 



5. The Heliocentric Velocity 



The heliocentric velocity which results from the observed geo- 

 centric velocity in most cases far exceeds the limits for a parabolic 

 orbit and leads to the necessary assumption of a decided hyperbolic 

 orbit. It is certainly remarkable that this almost always occurs in 

 the case of those orbits that have the most favorable location for the 

 most accurate determination of the velocity, viz. : for tliose that are 

 directed from the neighborhood of the anti-apex, as for example in 



^G. V. Niessl. Wien. Ber., 1893, 102, p. 265. 



" G. V. Niessl. Verhdl. d. naturf. Vereins in Briim, IQOI, 39, p. 220. 



