178 



PKOCEEDINGS OF THE AMERICAN ACADEMY 



TABLE IV. 



From an examination of the residuals of the preceding tables, the following 

 conclusions are drawn : — 



I. That the magnitude of the accidental errors, in all observations prior to 

 Maskelyne (in their present state of reduction), prevents any safe inference with 

 respect to periodic errors of single period. In general, a certain degree of accu- 

 racy must be reached, before it is possible to distinguish with certainty between 

 accidental and periodic errors ; and, as I shall presently show, wherever there 

 is an accumulation of accidental errors, periodic errors will be developed in any 

 system of observations, in which there is an irregular distribution of the clock 

 stars in Eight Ascension and Declination. 



II. That in absolute determinations of Eight Ascension, there is no positive 

 evidence of the existence of these errors. An examination of Maskelyne's sep- 

 arate observations will show that the slight periodicity in his later catalogues is 

 apparent rather than real. When we come to more modern observations, we 

 have several instances in which the evidence on this jjoint is quite decisive 

 Airy's Cambridge observations were nearly absolute determinations. For 

 his clock errors, he used stars in the same parallel; and notwithstanding the 

 decisive evidence of periodicity in his provisional places, it is nearly eliminated 



