33r, 



IIYDROGRAPHICAL SURVEYING [chap. xiv. 



Change of 

 Tempera- 

 ture. 



Tempera- 

 ture of 

 Maximum 

 Rate. 



Different 

 Con- 

 clusions. 



from injury, etc., goes erratically, while the others maintain 

 their rates pretty steadily, we may fairly conclude it to be 

 inferior. 



The imifomiity of rate of a chronometer \vhile on shore, or 

 when the ship is at rest, cannot be taken as a conclusive test. 



2. Change of Temperature. — A chronometer is supposed to be 

 compensated in such a manner that at two temperatures, a 

 varying number of degrees apart, the rates will be equal. At 

 all other temperatures the rates will vary, reaching a maximum 

 at about the mean temperature between the other two. 



Let us, for brevity, call this temperature of maximum 

 rate T. 



If we then examine the rates of a chronometer, we should 

 find a steady change of rate in one direction (nearly always 

 in the direction of acceleration of gaining) from Ioav tempera- 

 tures to high, until we reach T, when the change of rate should 

 vary in the opposite direction. 



For every chronometer we shall have a different quantity 

 for T, and different coefficients of change. Many clironometers 

 are supposed to be compensated for T = 60°, the mean tempera- 

 ture generally experienced over the globe ; but it would seem 

 that makers cannot command the point T ; anyway, many have 

 T over 90°, so that for such a watch, in practice, the direction 

 of change is invariable, which will result in a great accumula- 

 tion of difference of rate when passing tlu-ough hot and cold 

 climates, and where the coefficient is large, in great absolute 

 change of rate. 



Different observers on the performances of chronometers 

 have come to different conclusions on the subject of the law 

 of change for a degree on all parts of the scale, wliich can only 

 be accounted for by supposing that they have experimented 

 on different classes of time-keepers. 



Some have stated that they vary regularly, so as to have the 

 same rate at an equal number of degrees above or below T, and 

 have established the proportion of variation at the square of 

 the difference of T, and the temperature required. 



Other experiments have shown that the manner in w^hich 

 watches vary is not quite so regidar as this, and that the 

 coefficient of change is generally less at temperatures higher 

 than T than at tho^e below. 



