OBSERVATIONS OF LUMINOUS METEORS. 273 



this extreme supposition, is no greater than 21' 5. Thence the mean dis- 

 tance 1*3785 corresponding to a periodic time of 591 days, the other 

 elements being — 



q = -9859 

 « = -2848 

 <p = - 2° 29' 

 ir = 68 19 

 n = 245 50 

 i = 21° 9' 



It is not worth while to consider the case of the duration having 

 possibly been much greater than 15 seconds ; for had it been so great 

 as 30 seconds, the elements of the orbit would be sensibly the same as 

 for a duration of 15 seconds. 



It is remarkable that the elements of the orbit of this meteor, with the 

 exception of the inclination (i), are determined with a degree of accuracy 

 equal to those of a well-observed comet. The node is, of course, given, 

 by the mere record of time, within 3" : the perihelion distance is accurate 

 to the fourth place of decimals : the anomaly ((j>) and the longitude of 

 the perihelion are within a few minutes, while the mean distance and 

 eccentricity must be very approximate. Such favourable conditions, 

 however, as the present will rarely happen. 



Many observers have recognised a radiant point of shooting stars very 

 near to that of this fireball, and about the same date [see Mr. Greg's 

 comparative list, in ' B. A. Report,' 1874] ; but there are as yet no records 

 of the apparent velocity of the meteors. If the radiant be persistent and 

 the apparent velocity of the meteors be slow, there must exist a meteoric 

 ring of nearly circular form occupying the position in space defined by 

 the elements given above. 



So far as I am aware, the assumption of a parabolic orbit has satis- 

 fied, within probable limits of error, all previous observations of this 

 character. Hyperbolic orbits for fireballs have been deduced, but only on 

 the assumption that the observed durations were accurate. Experience 

 proves that the most reliable observers cannot avoid errors of 10 or even 

 20 per cent., and that in favourable cases of long duration. 



The orbit deduced above for the fireball of 1877, Nov. 27, is inde- 

 pendent of any reasonable error in estimating the duration. It is suffi- 

 cient for the establishment of a short periodic time (such as 500 days) 

 that the meteor " moved slowly " from a fairly well-determined radiant 

 point, distant about 90° from the point in the heavens towards which the 

 earth's motion was directed. 



On the Perseids I. 1876, August 9, 10, 11. By Captain G. L. Tupman. 



On the three evenings above mentioned, while residing at St. Moritz, 

 in the Engadine, I set myself the task of determining the position of the 

 Perseus radiant as accurately as possible. I watched the immediate 

 neighbourhood of the radiant, and only took account of those paths 

 which from falling favourably among stars could be mapped with great 

 accuracy. The number of such paths obtained was of course very limited, 

 but they are more valuable for the purpose than a large number of 

 roughly recorded paths. I made use of Mr. Proctor's ' New Star Atlas,' 

 first edition. It is not to be supposed that the determination of the exact 

 point of beginning and ending of each track was attempted — that would 



1878. t 



