XO. l6 METEOR-ORBITS IN THE SOLAR SYSTEM — VON NIESSL 1 3 



interval of time in seconds required for the passage along this portion, 

 then we have 



which is the average relative velocity of the meteor as deduced from 

 this particular observation, but affected with the great uncertainty 

 that the estimate of duration almost unavoidably brings with it. An 

 average velocity, because we must assume that the velocity cannot 

 remain quite constant during the motion through the atmospheric 

 strata. 



If many associated data of L and t are at hand, we have to care- 

 fully consider whether, say, a single apparently very reliable estimate 

 of t may be retained alone in connection with the value L correspond- 

 ing to this observation, or whether an average value should be 

 deduced, but of course always only in reference to associated data. 

 We do not recommend combining the greatest length of orbit deduced 

 from the earliest apparent visibility with the average of all the esti- 

 mates of duration as is sometimes done, because generally many esti- 

 mates occur among these that relate only to very short bits of the 

 path. 



In general it is best to determine the velocity for each pair {L, t) 

 by themselves and then to take the average with or without giving 

 attention to weight, but even the results of careful efforts to determine 

 the weight are very often doubtful because of the frequent large 

 overestimates of the duration, so that the sources of error are almost 

 invariably on one side. 



If there is no estimate of duration belonging to any of the segments 

 of the path, then perhaps the average of the times may be combined 

 with the average of all the Us. 



Frequently long paths, passing chiefly through the higher layers of 

 the atmosphere and also having their ends at great altitudes, give 

 larger values for the velocity than shorter paths recorded in the lower 

 strata. The answer to the question whether such results also repre- 

 sent quantitatively the retarding influence of the resistance of the 

 atmosphere, or whether they result rather in an overestimate of the 

 duration, which overestimate alters the result for short paths more 

 than for long paths, needs further special study. The fact that indi- 

 vidual results of observations certainly refer to particular well- 

 ascertained definite limited higher or lower located parts of the orbit 

 is therefore always very important and consequently must be most 

 thoroughly investigated. 



