OBSERVATIONS WITH TUB ZENITH TELESCOPE. 



387 



From which 



X = — 



.1119 ± 1255 



y = — 



.1090 ± 0587 



z = — 



.0029 ± 0723 



u — — 



.0572 ± 0642 



v> — 



.103G ± 0092 



w = — 



.0497 rfc 0919 



The probable error of a single equation of weight unity is ± 0".182. 



It was obvious from the beginning that the 12 and 14 months terms could not be 

 separated with any certainly in a series of this character, hut it was thought best to 

 include both for the sake of completeness. The indeterminatiOn is shown very conspicu- 

 ously by the large probable errors of x, y, z and u. This however does not extend to v', 

 the probable error of which is quite small. 



Romemberioo- the significance of v' we have for the correction to Struve's value of 



the aberration constant 



20.4451 

 20 



Strnvc's value 

 Resulting value of constant 



: .1059 

 20.4451 

 20 ".551 ± 0094 



A second solution of the normal equations, excluding the terms depending on the 

 annual period, gives 



x — — .0568 ± 0097 



y - — .1381 ±0074 



„/ ;;= _.1048 ± 0090 



w = — .0177 ±0227 

 Correction lo Btruve's constant =1 .1071 

 Value of constant 20 ".552 ± 0002 



Finally, if we exclude from our equations every unknown quantity except v' we 



find 



Correction = .0844 

 (Constant 20".580 



If the latitude variation can he represented by a single periodic term, and the 

 observations were distributed uniformly throughout a full period, both with respect to the 

 rieht ascensions of the stars employed and the date of observations, we should expect the 

 last two values to he practically equal. 



The value 20".5- r >l is that adopted in the reduction of the present series. The 

 necessary correction to the latitude is found by multiplying the computed reduction for 

 aberration by the factor .00518. This correction is applied to the individual values of 

 the latitude before beginning the process of adjustment which follows. 



