OP THE UNITED STATES. 373 



'I'roiiwhton makes use ol'the spring of the metal itself for the 

 aeljiistment of his iiistriinients, as has lieen seen. 



The comparison might be made to a greater extent, but 

 such would not be consistent with the object of the present 

 papers. It will be proper, however, in any given case to pay 

 attention to the nature of the establishment, anil the particular 

 branches in which each artist is most distinguished, that a 

 preference may be given to such as will be likely to give 

 most satisfaction. 



Similar considerations apply to the ditterent kir»ds of time- 

 pieces. Chronometers arc of such importance to the navy, 

 that every attention has been directed to their improvement, 

 sit:ce their invention by Mr. Harrison. On the contrary, the 

 improvements in clock making have been comparatively 

 limited, and the artists employed in this branch are not so 

 able as those employed in the construction of chrono- 

 meters. 



Astronomical clocks are not kept for sale as chronome- 

 ters arc. There are but a few constructed, and those agree- 

 ably to order only. 



As to chronometers, it is hardly to be expected, that among 

 the great numbers kept on hand by dilVerent artists, there 

 should not be some of them preferable to others. If there- 

 fore several chronometers be required, as was my case, it 

 will be proper to take them from difterent artists, in order to 

 have a greater chance of success, which no artist can inva- 

 riably attach to his work. 



It must be ol)served, from the principles upon which chro- 

 nometers are constructed, that they will go differently at 

 sea from what they do on land. This I have proved by se- 

 ven chronometers wliieh I kept going on my passage from 

 London to Philadelphia, and which I compared logelher dai- 

 ly. It is evident that the constant motion of the ship must 

 aflect the heavy balance, which is 'never an'ecled when the 

 chronometer is completely at rest. The (|ii;intity of fh's ef- 

 fect is of course both micertain. and piuidiar to each chro- 

 nometer, both in respect to quantity and direction. 

 VOL. n. — 3 c 



