NAVIGATION. 41 



January, 1874, and on her voyage to Calcutta the 

 chronometer was subjected to variations of temper- 

 ature ranging from 50 to 90. The chronometer 

 was tested by the Calcutta time-gun on the 26th of 

 May. The time reckoned by it, with correction 

 for temperature on Hartnup's plan, was found 

 wrong by 8J seconds. Another chronometer, 

 similarly corrected by Mr. Hartnup's method, and 

 from his rating, gave an error of only 3^ seconds. 

 The difference between the reckonings of the two 

 chronometers was thus only 5 seconds, and the 

 error in reckoning by taking the mean between 

 them only 6 seconds. This corresponds to an 

 error of only a mile and a half in estimating the 

 ship's place in tropical regions. The reckonings of 

 Greenwich time from the two chronometers, 

 according to the ordinary method, differed actually 

 by 4 minutes 35 seconds, corresponding to 68 J 

 geographical miles of error for the ship's place. 



From Mr. Hartnup's investigations, it is 

 obvious that one important point for a good 

 chronometer is, that the temperature of maximum 

 rate should be as nearly as may be the mean 



