H Á VOYAGE TO Book I. 



clouds are to be fecn, though part of thefe are gra- 

 dually diiperfed by the heat of the tun, fo that fome 

 parts are quite clear, others obfcure; but a general 

 darknei's during the whole day is never known. 



It is well known and allowed, that, through the 

 whole extent of tlie ocean, not the leaft current is per- 

 ceivable, till we arrive within the iflands, where in 

 fome parts they are lb ftrong and irregular, that, 

 without the greateü vigilance and precaution, a fliip 

 will be in great danger among this archipelago. This 

 fubjeel, together with the winds peculiar to this coail, 

 ihall hereatler be confídered more at large. 



In the track to Martinico and Dominica there is a 

 fpace where the v/ater, by its white colour, viñbly 

 dittincfuiihes itfelf from the reft of the ocean. Don 

 George, by his eftimate, found this fpace to terminate 

 loo leagues from Martinico; whereas, according to 

 my reckoning, it reached only to within io8 leagues; 

 it may therefore, at a medium, be placed at Í04. This 

 fmall difference, doubtlefs, proceeds from the diiii- 

 culty of difcovering where this whitifli colour of the 

 water terminates, towards Martinico. It begins at 

 about 140 leagues from that ifland, which muil be 

 underfiood of the place where the different colours of 

 the water arc evident ; for, if we reckon from where it 

 begins to be juft difccrnible, the diftance is not lefs 

 than 180 leagues. This trail of water is a certain 

 mark for dire(iling one's courfe; becaufe, after leaving 

 it, we have the fatisfaélion of knowing the remaining 

 dlflance: it is not delineated on any map, except the 

 new one lately publiilied in France ; though it would 

 doubtlefs be of irreat uie in them all. 



Nothing farther remains, than to give an account 

 of the variation of the needle in different parts in 

 which we found the fhip by her latitude and longi- 

 tude; a point of the utmoft confequence in navigation, 

 not only with regard to the general advantage to ma- 

 rijBers in knowing the number of degrees intercepted 



betweea 



