Lttter from V'm Humboldt to Lalande. 359 



that relates to navigation, the fcllowing obfervations will not 

 be unacceptable. 



I have cucfuilv proved what Dr. Franklin and Captain 

 Jonathan Williams aiiert in the TranfaAions of the American 

 Society * rcfpetting the ufe of the thermometer for difco- 

 vering {lialUnvs at lea; and am able to confirm, in the fulled 

 manner, what they have faid. I was aftonilhed to lee how 

 the water became evidently colder as its depth decreafed, and 

 how the neighbourhood of fliallows, and of the coalls, could 

 thereby be announced. The worii fpirit of wine thermo- 

 meter, if only lenfible, will therefore, in the hands of the 

 mod ignorant mariner, be a very ufeful inftrument in the 

 night-time, and durino- ftorms, or when it is difficult or 

 impoffible to heave the lead. This obfervation I cannot too 

 ftronglv recomnien(^to the attention of the Board of Longi- 

 tude. Our whole crew were altonifhed to fee how fpeedily 

 the thermometer fell, when we approached the large bank 

 which extends from Tobago to Grenada, and eaft from Mar- 

 garita. Thefe obfervations may be made with more eafe, as 

 the water of the fea, in an extent of 12000 fquare miles, has 

 always the fame temperature day and night; fo that the moft 

 fenfible thermometer, during from four to fix days fail, does 

 not rife or fall above 0*3 \ In the neighuourhood of flial- 

 lows it was from a to 3 degrees, and more, colder. This 

 obfervation of Franklin, hitherto forootten, may at fome 

 period be of great ufc to navigation; not that leamen fliould 

 throw afide the lead entirely, and truft to the thermometer, 

 for this would be folly ; but bccaufe the obfervations may be 

 fo ea(ily repeated, and beeaufe the thermometer will an- 

 nounce the danger much fooner than the lead, as the colder 

 water a'j )ve the (hallows leflens the temperature in the water 

 in the neichljourhood. lean aifert that this new mean is 

 not more uncertain than the log, and the helps already em- 

 ployed in navigation. If the thermometer does not fall, na- 

 vigators ought not to depend that they are entirely fecure 

 from ftiallows; but if it tails, they muft be on their guard. 

 This warning is certainly more valuable than our fea charts, 

 where the (hallows are in c'^nc^al laid down in a very in- 

 correct manner; and to in)merle a thermometer in a bucket 

 filled with fea water is certainly much eafier. I have alfo 

 nieafured feveral times the fpecilic gravity and temperature of 

 the fea water, at the furface and at certain depths, by means 

 of Dollond's balance and thermometers, which are placed in 

 cafes fumiflied with a valve. As my inhruments were com- 

 pared with the heft Parifian ones, and as I could be more 



♦ Vol iii. p. 32. 



Z 4 certain 



