350 PROCEEDINGS OF THE AMERICAN ACADEMY 



The employinent of Mr. Gill's reversing prism would perhaps be 

 desirable in the continuation of this series. 



The probable error of a single observation of a star within 10° of 

 the north pole has been found to be rather less since the counterpoises 

 were adjusted, in May ; and a still further decrease has been indicated 

 since the object-glass and eyepiece were exchanged in September. 

 The illumination of the field, as before stated, has been more satis- 

 factory. 



I find from a great number of stars the followinji values reduced to 

 the equator : — 



« cos S. 



1882 — 1883, April 17, ± 0'.0267 (64 stars). 



1883, May 9 — 1883, Dec. 31, ± 0».0l96 (GO stars). 



These values, however, are obtained by comparing observations in the 

 same position of the instrument. 



For the whole time I find by 49 stars the value 



€cos8 = ±0'.0238 



from observations in two or three positions indifferently ; a value 

 which is very little larger than the average of the values which 

 precede. 



There are not enough stars beyond 10° from the pole to give an 

 accurate value of the probable error, which in this region is some- 

 what increased ; nor have 1 yet divided those within 10° into zones 

 of declination. 



From 46 stars observed since Sept. 5, 1883, only, I find 



ceosS = ±0'.0161, 



and it is quite possible that this smaller probable error may hold good 

 for present and future observations, so long as the instrument remains 

 in its present normal condition ; the relative weights of the dilTcrcnt 

 periods will, however, be better discussed when more materials are 

 available ; and especially when a sufficient number of stars have been 

 observed in four or more positions. 



Another question deserves investigation : whether the personal equa- 

 tion for polar stars changes from time to time. I conceive, however, 

 that the study of this point must at present be deferred. The simplest 

 and best method to determine it will be to compare observations of 

 successive whole years. I do not imagine that the reversal of object- 

 glass and eyepiece would, )>y bringing into use new parts of the 



