July 27, 1900.] 



SCIENCE. 



125 



comes out to be on the average about 

 ±0".170. Taking the average weight of 

 the parallax to be 30.0 the average probable 

 error of the values of the parallax found 

 would be ± 0".031. In this no account has 

 been taken of the systematic error of the 

 observer which has not yet been discussed 

 for this problem. It should also be borne 

 in mind that the parallax here found is only 

 the relative parallax to which should be ad- 

 ded that of the comparison stars em- 

 ployed. 



It is our purpose further to classify the 

 results. 1st, according to the magnitudes 

 of the stars, and 2d according to the amount 

 of the proper motion which may perhaps 

 lead to interesting conclusions. The results 

 here given may perhaps be slightly modi- 

 fied in the fuller discussion, but in their 

 present form they may serve to give 

 some idea of a piece of work, which we 

 hope will contribute something to our pres- 

 ent knowledge of the stellar universe. 



The Velocity of Meteors as Deduced from Pho- 

 tographs at the Yale Observatory : By W. 



L. El.KIN. 



The instruments in use at the Yale Ob- 

 servatory for the photographic observation 

 of meteors have been equipped with an ar- 

 rangement for the determination of the 

 velocity of meteors. The idea of using 

 photography for this purpose seems to have 

 first been suggested as long ago as 1860 by 

 J. Homer Lane, the well known physicist 

 and discoverer of ' Lane's law.' In 1885 a 

 well planned attempt in this direction was 

 made by Zenker, in Berlin on the occasion 

 of the expected shower of Andromedids, 

 but apparently without success, and lately 

 the suggestion has again been made by 

 Professor Fitzgerald. 



The Yale apparatus consists of a wheel 

 (a bicycle wheel) rotating in front of the 

 cameras and carrying a number of opaque 

 screens. There are at present 12 of these 



interceptors and the rotation is effected at 

 the rate of 30 to 50 turns per minute by 

 means of a small motor worked by 3 or 4 

 bichromate cells. It will be advisable to 

 increase the number of occultations in the 

 future, however. At each revolution a 

 record is made at the chronograph so that 

 the wheel's velocity at any instant is always 

 known . 



The length of the interruption of a meteor 

 trail and the resulting velocity are easily 

 derived from the plates, if the meteor is 

 also recorded on a plate at our second 

 station at Hamden, distant about 3 km. 

 The first attempt was made at the August 

 period last year, and subsequent ones at 

 the Leonid, Andromedid and Geminid 

 epochs in November and December last. 

 In all so far five such trails have been ob- 

 tained with corresponding records at Ham- 

 den and the time and identification also 

 secured. These have been carefully meas- 

 ured and reduced and the resulting data 

 are brought together in the following table 

 of which the headings explain themselves 

 sufi&ciently : 



July 31 

 Aug. 7 

 Aug. 8 

 Nov. 24 

 Dec. 12, 



Green- 

 wich 

 Mean 

 Time 



14 25 25 



16 32 47 



16 31 25 



21 43 



Approxi- 

 mate 

 Altitude 

 (in km.) 



88 to 75 

 60 to 46 

 101 to 94 

 93 to 90 

 90 to 86 



If we now correct the values for the ap- 

 parent radiant and velocity for the effect oi 

 the attraction of the earth and its diurnal 

 rotation by Schiaparelli's formulae, we de- 

 rive the ' corrected ' radiant and velocity, 

 in the following table and hence the ' true ' 

 velocity of the meteors relative to the Sun. 

 The last columns of this table contain the 

 ' true ' and ' apparent ' velocities which a 

 parabolic orbit, or, in the case of the 

 November 24 meteor, an elliptic orbit of 

 6.62 years period should have produced. 



