CELESTIAL MEASURINGS AND WEIGHINGS. IQ9 



has been approached is to select for inquiry bright stars, 

 which have in their immediate vicinity, so near as to be 

 seen with them at the same time in the same telescope, 

 two or three other very much smaller ones ; and without 

 troubling ourselves to determine their absolute places in 

 the heavens (so throwing overboard the enormous diffi- 

 culties which, as we have seen, that determination to a 

 sufficient precision presents), confine ourselves to what 

 may be called a microscopic examination and mapping 

 down of the relative distances and situations of these 

 stars inter se. Repeating this at all seasons of the year, 

 we are enabled to ascertain whether the large star main- 

 tains steadily the same invariable position among the 

 smaller ones ; or is affected by any movements of which 

 they do not partake. There is a general prima facie 

 probability that the brighter stars are nearer than very 

 faint ones : and, near objects being more displaced than 

 distant ones by the spectator's change of place; the 

 large star in the case supposed would appear, by the 

 effect of parallax, to move to and fro among the smaller 

 ones; or rather to describe annually a minute ellipsis 

 among them, the exact counterpart, equal in size and 

 similar in the situation of its longer and shorter dia- 

 meters, to that into which the earth's orbit itself would 

 be seen projected by the effect of perspective from the 

 star. Now no casual movement, or one arising from 

 any other physical cause, could be mistaken for such a 

 motion as this. For, not to mention the completion of 

 the revolution in an exact year, the two diameters of 

 the ellipse ought to stand to each other in a certain den- 



