OF ARTS AND SCIENCES. 



105 



definite aim for their meridian observations and compilations of star- 

 catalogues, and will thus contribute largely to the knowledge of stellar 

 proper motions, as I shall show elsewhere. I may here say, that the 

 probable error of a star's declination, as compiled from the best authori- 

 ties, can be so estimated, classified according to the quantity and good- 

 ness of the materials, and allowing enough to cover all defects. 



Within 30° of the zenith, there are now stars enough of the 

 classes AA, A, and B, for any American latitude, using Talcott's 

 method ; and the probable error of declination of a single pair of stars 

 will vary from 0". 13 to 0".31. Allowing, then, a p. e. of 0". 43 to each 

 observation, we have the final probable error of latitude from a pair of 

 stars, observed thrice, — 



= \J 0^. 



0.432 

 132 + -— or 



y o.ai'J 



+ 



0.432; 



that is, from ± 0".28 to ± 0".40 ; which only requires from eight to 

 sixteen (say twelve) pairs to give a probable error of 0".l. 



Moreover, observations now in progress, both general and special, 

 will in a year or two raise all the stars of the British Association's cata- 

 logue now classed as C (within our latitude-limits) to the class B, 

 and will doubtless transfer many of this class to a higher. On the 

 other hand, the portable meridian circle, or the prime vertical transit, 

 needs only stars of the two highest classes, and, if other practical 

 difficulties do not intervene, can probably secure this same degree 

 of accuracy with fewer observations ; not necessarily, however, with a 

 less amount of time and trouble. I am inclined to think that the 

 extravagant praises of Talcott's method to which our officials give 

 utterance are about balanced by the steady adherence of the French 

 and Germans to the other way, and that the practical difference 

 between them is one rather of habit than essential. It is quite certain 

 that both give excellent results. 



