188 



SCIENCE. 



[N. S. Vol. XII. No. 292. 



ciable but still large enough to leave open 

 the question whereabouts during that period 

 of 10 or 20 seconds occupied in making the 

 bisection, the planet actually was at the 

 time assumed for the bisection . There arises 

 an uncertainty which may be systematic. It 

 may be diiierent for different observers and 

 for different places. This uncertainty seems 

 to preclude dependence upon the process of 

 micrometi'ic measurement. In the case of 

 the satellites of Saturn we know that results 

 are obtained differing enormously. 



S. J. Brown. And for the satellite of 

 Neptune the difference is still greater. 



Professor Newcomb. There are some sys- 

 tematic differences attending micrometric 

 observations, but these can be diminished if 

 we take our observations to avoid them. 

 "We have the method of difference of right 

 ascension and declination, when we take 

 the right ascension entirely free from the 

 declination, taking the right ascension at 

 the time of passing a certain thread. If the 

 star is chosen of about the same magnitude 

 as the planet, it will be free from one source 

 of systematic error, but there may be a great 

 many others. If each set requires a minute 

 or two, and if two threads are used, you may 

 make 60 comparisons in an hour, which I 

 think will be more than can be done with 

 the micrometer. There is then much room 

 for systematic error. 



G. W. Hough. I think the most accurate 

 work can be done ■n'ith the micrometer. 

 With an 18^-inch telescope there is no 

 trouble in seeing an 11th or 12th magnitude 

 star ; and it is only necessary to get the time 

 of the bisection and the ]30sition angle and 

 distance. The time can be known within 

 a few seconds, and ten observations can be 

 made as readily as one. So we can get a 

 final result free from systematic error. 



Simon Newcomb. That may be quite 

 possible for a 12th magnitude star. 



G. W. Hough. The method of differ- 

 ences of right ascension and declination I 



use very often ; but then we must depend 

 upon the stability of the telescope. It may 

 require IJ to 2 minutes to pass from the 

 star to the planet, and there is a liability 

 to error from the action of the wind or 

 from other causes. In photographic work 

 the star will be near the horizon at the time 

 required for its observation, and the time 

 required will be very long, especially if we 

 use isometric or other such plates ; and that 

 will be a great objection. 



W. W. Campbell. It seems to me that 

 both the methods suggested should be sub- 

 mitted to actual trial before the entire cam- 

 paign is based upon either method. It 

 ought to be possible by means of known 

 asteroids moving rapidly to obtain photo- 

 graphs, by allowing the stars to trail or the 

 asteroid to trail, and we could learn with 

 what accuracy the position of the asteroid 

 can be obtained upon a certain plate. 

 Furthermore it will be easy to test the 

 method of observing faint stars near the 

 asteroid. In large telescopes it will be pos- 

 sible either to get a star of the 12th or 13th 

 magnitude within 1' of arc, and by select- 

 ing stars in different directions from the 

 asteroid it will be possible to eliminate 

 practically all the personal element entering 

 into the question. In my experience in 

 micrometric measures there is never an un- 

 certainty greater than 2 or 3 seconds as to 

 the exact instant of bisection, and the mo- 

 tion of the planet in that time, for that 

 particular observation, will be small as 

 compared with the accidental error probably 

 made. If we take ten observations to- 

 gether the motion of the planet will be 

 practically unimportant. My main point 

 is that the method fixed upon should be 

 actually submitted to test before the cam- 

 paign is started. It would be a serious 

 matter if the campaign should be carried 

 on, and afterwards it should be ascertained 

 that in consequence of some oversight the 

 results are of very little importance. 



