Annual Address to the Members. liii 



The Cape Catalogue of 1,905 Stars for the Equinox 1865 from 

 observations 1861 to 1870. 



All these catalogues to which I have referred are technically 

 known as Catalogues of precision, and the stars which they con- 

 tain are, as a rule, well-known stars. They are thus no mere list of 

 stars of which the places are given merely with precision sufficient 

 for their identification, but they are the mean results of observations 

 of high accuracy repeated on several different nights, and, in the 

 case of important stars, on many different nights, in order to pro- 

 vide those reference points in the heavens which are required for 

 the purposes of astronomy of precision. 



But it has become now necessary to go farther and to devise and 

 erect an instrument capable of excluding some of those sources of 

 error which may still affect our observations. Thus, if we inter- 

 compare the results of nearly simultaneous observations at Cordoba, 

 Melbourne, and the Cape, we find small systematic differences which 

 are too persistent in character to be the result of accidental errors of 

 observation. When observations of a particular star made on a 

 number of different nights, with the same instrument at the same 

 Observatory, agree pretty closely inter se we can derive from the 

 agreement of the results the probable accidental error of observation 

 and the probable accidental error of the mean result. Similarly we 

 can derive the probable error of the position of the same star from 

 observations made at another Observatory. Now if these two results 

 disagree beyond the limits which their probable errors would lead us 

 to expect, then there is reason to suspect some systematic cause for 

 their difference. This luay have its origin in the instrument itself, 

 in the condition of its surroundings, or in some peculiarity of the 

 observer. We can trace out these sources of error in various ways ; 

 we find with the same instrument personal peculiarities between 

 different observers which are affected by the brightness of the star 

 or the direction of its motion. There exist means for tracing out, 

 determining, and eliminating these. Or there may be errors due to 

 the installation of the instrument ; for example, the Cape transit 

 circle is placed in a room which is part of the Observatory main 

 building, with thick walls and so on, so that it is impossible to 

 equalise the temperature inside the room with that of the external 

 air. This produces small errors due to abnormal refraction. Then 

 there may be instrumental errors — due to mechanical defects, such 

 as errors of graduation of the divided circle and errors in the form of 

 the pivots. No workmanship of human hands is perfect ; the only 

 course is to provide the most perfect means for the determination of 

 the errors. There may be also other instrumental errors due to 



