XX] LONDON: VOYAGE TO SINGAPORE 319 



of these two points on the count's map is 2° 22', whereas on 

 mine it is 3° 45' ; my estimate being about 60 per cent, more 

 than his. By measuring carefully with compasses in lengths 

 of five miles, with a little allowance for the minuter bends, 

 his distance is 315 miles, mine 494, mine being thus 55 per 

 cent. more. 



It is unfortunate that Count Stradelli has^ given us no 

 information as to how he estimated his distances. In a river 

 flowing through a densely wooded country, with nowhere 

 more than a few hundred yards of clear ground on its banks, 

 with a very crooked and twisted course, and with a current 

 varying from being scarcely perceptible to such rapidity that 

 a whole crew of paddlers can hardly make way against it, it 

 is exceedingly difficult to ascertain the rate of motion in 

 miles per hour. 



Canoes of different sizes do not travel at very different 

 rates, when each has its complement of men, and I had 

 taken many opportunities to ascertain this rate in still water. 

 Then, by noting the time occupied for a particular distance, 

 say between two of the cataracts, both during the ascent and 

 descent of the river, the mean of the two would be the time 

 if there were no current. Making a little allowance for the 

 load in the canoe, the number or the quality of the rowers, 

 etc., this time multiplied by the rate of travel in still water 

 would give the distance. This was the plan I adopted in 

 making my map of the Uaupes. It is, of course, a mere 

 approximation, and liable to considerable errors, but I did 

 not think they would lead to such a large difference of 

 distance as that between the Count's map and my own. We 

 have no doubt erred in opposite directions, and the truth lies 

 somewhere between us ; but until some traveller takes a good 

 chronometer up the river with a sextant for determining local 

 time, or a telescope of sufficient size to observe eclipses of 

 Jupiter's satellites, the true length of the river will not be 

 settled. 



In one of the latest atlases, "The Twentieth Century 

 Citizens' Atlas," by Bartholomew, the position of the Jurupari 

 fall is 62 per cent, further from the mouth of the river than 



