I7?i' ROUND THE WOULD. 183 



shall not pretend to deny ; we were however of opi- 

 nion that they were isles, and that New Caledonia 

 terminated more to S. E., though this, at most, is but 

 a well-founded conjecture. 



But whether these lands be separate isles, or con- 

 nected with New Caledonia, it is by no means cer- 

 tain that we saw their termination to the west. I 

 think we did not, as the shoals did not end with the 

 land we saw, but kept their N. W. direction farther 

 than Bougainville's track in the latitude of 15° or 

 15-£-° Nay, it seems not improbable, that a chain 

 of isles, sand-banks, and reefs, may extend to the 

 west, as far as the coast of New South Wales. The 

 eastern extent of the isles and shoals off that coast, 

 between the latitude of 15° and 23°, were not known. 

 The semblance of the two countries ; * Bougainville's 

 meeting with the shoal of Diana above sixty leagues 

 from the coast, and the signs he had of land to the 

 S. E., all tend to increase the probability. I must 

 confess that it is carrying probability and conjecture 

 a little too far, to say what may lie in a space of two 

 hundred leagues ; but it is in some measure neces- 

 sary, were it only to put some future navigator on 

 his guard. 



Mr. Wales determined the longitude of that part 

 of New Caledonia we explored, by ninety-six sets 

 of observations, which were reduced to one another 

 by our trusty guide the watch. I found the variation 

 of the compass to be 10° 24/ E. This is the mean 

 variation given by the three azimuth compasses we had 

 on board, which would differ from each other a degree 

 and a half) and sometimes more. I did not observe 

 any difference in the variation between the N. W. and 

 S. E. parts of this land, except when we were at 

 anchor before Balade, where it was less than 10° ; 

 but this I did not regard, as I found such an unifor- 

 mity out at sea ; and it is there where navigators 



* See his Voyage, English translation, p. 303. 

 K 3 



