DETERMINATION OF THE PLACES OF THE PLANETS. 329 



This little table shows that if we can fix the time of transit, say 

 of Uranus, to a tenth of a second, which, be it noted, represents 

 no longer actual time, but angular measure of Right Ascension, 

 we have only found the place it really occupies with an uncertainty 

 of some 50 minutes of time. This shows the weakness of using 

 the rotation of the Earth as a means of fixing the positions of the 

 heavenly bodies. The difficulty cannot be entirely overcome as 

 long as our time is based on the rotation of the Earth and the 

 motion of the Sun amongst the stars, as is the case at present. 

 But the fixed stars move so very slowly, as indeed the word 

 " fixed " indicates, that it is possible to obtain their positions with 

 a precision which is at least 10 times greater than the exactitu3e 

 in determining the position of a planet on the meridian. 'Astro- 

 nomers attempt to use this exactitude by adopting other means 

 of observation, such as the heliometric measures of the angular 

 distances of planets from fixed stars which was extensively tried 

 under Sir David Gill's direction at the Cape Observatory, or by 

 means of photographs, by which the planets can be referred to the 

 neiighbouring stars, a means extensively used at the Greenwich 

 Observatory. The convenience of the photographic method is 

 so great that it will ultimately prevail and the exactitude will be 

 very great, but at the present time the places of the comparison 

 stars are in general not very well-known. The progress of the 

 Carte-du-Ciel catalogue will in time remedy this defect. If the 

 observation of a planet occulting a bright star is carefully recorded, 

 it will yield the place of the planet at a given second of time with 

 all the precision of the star-place itself. 



The determination of the Moon's place by meridian-passages 

 is subject to difficulty, insomuch that it is the rough illuminated 

 limb which has to be observed. The error of a meridian lunar 

 observation is at least three seconds of arc corresponding to about 

 6 seconds of time, so that if the star's place is well-known and 

 the occultation has been well-observed, the Moon's place should 

 be found within o".i. As several occultatlons can be observed on 

 one evening by several observatories working in co-operation, it 

 should be possible to attain this accuracy, but again with the pro- 

 viso that the stars' places were known to that degree of precision 

 and this is only the case with a few of the brighter stars. 

 It therefore appears to the writer that the principle of using the 

 Earth's rotation as a means of angular measurement should be 

 used as sparingly as possible. 



Star places should be found by means of photographic plates, 

 which should be referred to each other by angular measurements, 

 such as advocated by Professor Turner. The angular measure- 

 ment due to the rotation of the Earth would then only be required 

 for finding the position of the Sun amongst the stars. 



As a star's Right Ascension and Declination are always chang- 

 ing through the precession, wh'lst the stars themselves are almost 

 " fixed," it is obvious that this choice of co-ordinates is an unfor- 



