> to C, Delamlre. 367 



the water, and the impossibility of doing it on account of the 

 ferocity of the caymans, the danger of being pricked by the 

 rajas and the teeth of the small carib-fish — youth and a great 

 deal of resignation are reijuired to endure all these. The 

 evil is passed, and I have reaped more than I durst venture 

 to hope. 



It is believed {see the map of father Caulin, the best ex- 

 tant, though all the names are wrong,) that the Spanish pos- 

 sessions of Guyana extend to the equator. But I have found, 

 by very good observations o£ the stars called the Cross and 

 Canopus, which I made among the rocks of Culimacaii, that 

 San Carlos del Rio Negro, the most southern establishment, 

 is in 1*^ 53' of north latitude; and that the line passes through 

 the government of Great Para, near St. Gabriel-de-las-Ca- 

 chuellas, where there is a cataract, but not so considerable a$ 

 the two famous ones of Atures and Maypura. 



At Cumana, before the earthquake, which we experienced 

 on the 4th of November 1799, the magnetic inclination, 

 measured with Borda's compass, was found to be 44° 20' of 

 the new division : after the earthquake it was 43° 35'; the 

 needle made 229 oscillations in the course of ten minutes. 

 Experiments have proved that the magnetic charge has 

 changed in this part of the world, and not in the needle. 



At Calabozo, in the centre of Uana, lat. 8' 56' 5fj", long, 

 from Paris 44° 40' 18", the inclination was 39^ 30': number 

 of oscillations 222. 



At Atures, one of the cataracts of the Orenoquo, in lat, 

 5"" 39', long. 44° 42' 19'', the inclination was 32° b5': num- 

 ber of oscillations 221. 



At St. Fernando d'Atabapo, a mission at the mouth of the 

 Guaviara, lat. 4° 9' 50', the inclination was 30' 30' : num- 

 ber of .oscillations 219. 



At St. Carlos de Rio Negro, lat. 1° 53', the inclination 

 was 23° 20' ; number of oscillations 216. 



According to the rules given by Messrs. Cavendish and 

 Dalrymple, care was always taken, while observing, to turn 

 the compass to the east and west to find the mean inclina- 

 tions, and to correct the error which takes place when the 

 axis of the needle does not pass exactly through its two points. 



During 



