2*74 Prsesepe Group; Measurement and Reduction 



measuring the plate as to errors in the meridian places. To ob- 

 tain more precise information on this point, let us correct the 

 meridian places of each of the comparison stars by the mean of 

 the residuals for that star, and suppose we have effected thes leat- 

 square solutions anew, using now the corrected meridian places. 

 It can easily be shown that the new solutions would lead to ex- 

 actly the same values of the constants as had been first obtained, 

 but now each residual will be altered by a certain quantity, 

 namely, the amount of the corresponding correction to the mer- 

 idian place. We may then subtract at once the mean of the re- 

 siduals for a star, from the corresponding residual in each least- 

 square solution and then compute the probable errors of the 

 constants. The results of such a computation are as follows : 



Plate 





Probable Error of 



Probabl e 



Error of 





P 



or r. 



k or c 



I 



±0 



.000013 



±0 // .026 



II 





24 





.049 



III 





16 





.032 



IV 





16 





.032 



V 





15 





.030 



VII 





08 





.016 



VIII 





II 





.022 



IX 





20 



±0" 



.041 



eans, 



±o. 



000015 



.031 



,ns were 













dzO. 



000032 



±0" 



.065 



and these must be regarded as indicating the uncertainty in the 

 absolute values of the constants ; if the constants which we have 

 obtained are in error, then there will be a decided tendency to 

 error in the same direction on different plates, and the smaller 

 probable errors given above indicate how much we should expect 

 the adopted values of the constants to differ from each other as 

 obtained for different plates. Consequently only a small part of 

 the discrepancy between the probable errors of the measured co- 

 ordinates and of the right ascensions and declinations can be due 

 to uncertainties in the adopted constants. 



The discrepancy is probably caused by inaccuracies, and in 

 some cases neglect, of instrumental corrections. For example, 

 the difference between the two complete measurements of a coor- 



