244 Mr Galbraith on Kater's Azimuth Circle. 



five or six observations, come within 5" of the truth, when made 

 successively in the same evening on Polaris, or perhaps with 

 more certainty on stars to the north and south of the zenith. 

 From the unfavourable state of the weather during these few 

 months past, it was difficult to obtain a favourable opportunity 

 to try this effectually here, and this is the reason why Nos. 8 

 and 9 of the preceding series differ so much from the others. 

 In fact; two sets made at that time should have been properly 

 rejected, as they were made too hurriedly, and the bubble of 

 the level was too near the extremity of the level to be confi- 

 dently relied on. On the 28th of November last, five series, in 

 which the circle was only once reversed each series, gave 55° 

 5& 5T.5 N., from observations on the pole-star, which agrees 

 very nearly with the general mean previously obtained. Whe- 

 ther so close a result could be always calculated upon, my expe- 

 rience with this circle does not enable me to decide. It is 

 common even with instrument makers to estimate the accuracy 

 of these too high. They frequently divide the value of one 

 division of the vernier by the number of readings. In the pre- 

 sent circle, which has three verniers each shewing ten seconds, 

 if six readings be recorded, the instrument being once reversed, 



the probable error is taken at — = 1"-, instead of -y^ = — = 



4" nearly, or about three times greater. Still, however, when 

 the observations are repeated a considerable number of times, 

 the errors of reading, division, and pointing (pointe as the 

 French call it) must be greatly diminished, if not completely 

 destroyed, as we know from experience. Indeed, with a similar 

 circle from a series of ten days' observations, John G. Kinnear, 

 Esq. determined the obliquity of the ecliptic by observations 

 taken in June 1834 to be 23° 2T 41".5, which exceeds the result 

 that I derived from the Greenwich observations by 1".2 only, 

 and Bessels' by T.2 for January 1, 1834. 



The deductions from these small but compact instruments are 

 much more accurate, therefore, than from their size we had any 

 reason to expect, and may be advantageously employed in many 

 geodetical and astronomical operations with great success. 

 Hence, the smaller sized ought to recommend themselves to 

 scientific travellers and to medical officers attached to foreign 

 stations, who are anxious to distinguish themselves in geogra- 

 phical and astronomical researches. 



