NA VI G A TION. 39 



in workmanship, and however carefully adjusted 

 by trial, does not give equable time-keeping at all 

 temperatures through wide natural ranges. It had 

 been sought to remedy this defect by the appli- 

 cation of secondary compensation on various 

 ingenious plans, but with no practical success. 

 Thus the best chronometers of the best makers 

 in modern times are practically perfect only 

 within a range of 5 or 10 Fahrenheit on each 

 side of a certain temperature, infinitely near to 

 which the compensation is perfect in the individual 

 chronometer. 



The temperature for which the compensation is 

 perfect, and the amount of deviation from per- 

 fection at temperatures differing from it are 

 different in different chronometers. Mr. Hartnup 

 finds that at the temperature for which the com- 

 pensation is perfect, the chronometer goes faster 

 than at any other temperature, and that the rate 

 at any other temperature is calculated with 

 marvellous accuracy (if the chronometer be a good 

 one) by subtracting from the rate at that critical 

 temperature the number obtained by multiplying 



