94 Proceedings oj the Asiatic Society. [March, 



other, in order to eliminate instrumental errors. The moon's transits in 

 altitude were taken over 5 horizontal wires, and corrections for inequal- 

 ity of motion were applied whenever necessary. Each observation 

 was reduced independently after the application of the instrumental 

 corrections, but as it appears from the results that the instrumental 

 errors have not been determined with exactitude, the few single observ- 

 ations which were taken have been rejected, and only the pairs retained. 

 The number of pairs is 101 ; the probable error of a single pair, when the 

 astronomical clock was used, is ±3 04 sec ; with the chronometer it is 

 3*31, showing that the results were very slightly impaired by the loss 

 of the services of the clock. The moon was observed both when north 

 and when south of the prime vertical, and almost as many times when 

 east of the meridian, as when west. The zenith distances range from 

 23° to 64J°; the moon's distance from the prime vertical never ex- 

 ceeded 17J°, and was usually much less ; the azimuths ranged from 

 67° to 125°. 



For the culminations it is only necessary to remark that 29 were 

 observed, that the tabular elements were taken from the section 

 " Moon Culminating Stars" of the Nautical Almanac, that the illuminat- 

 ed pivot of the instrument if pointing towards the east one evening, was 

 usually pointed to the west on the next evening, and that the transit 

 axis was reversed on its pillars six times during the course of the 

 observations. 



The probable errors of the zenith distances have been computed 

 from the differences between the mean of each pair of observations 

 and the general mean of the group to which the pair appertains. 

 Those of the culminations have been computed from the differences 

 between the single observations, and the genaral mean of all. Being 

 calculated on the assumption that the tabular places in the Nautical 

 Almanac are free from error, they are of course smaller than they 

 would be if the probable errors of the tables were taken into con- 

 sideration. But they sufficiently serve the purpose for which they 

 are required, namely, to combine the separate groups of results with 

 weights inversely proportional to their squares. 



The results of the individual Observations are given below and the 

 final results are as follows : — 



