THE EQUATORIAL FORESTS ii 
conveniences as bridges were almost unknown in 
that region, we should not have been able to 
get across if our guide had not swum over and 
brought us a canoe from the other side. A few 
steps beyond stood the four or five cottages we 
were in quest of, embosomed in a grove of orange 
trees and plantains. I surveyed them with interest, 
for they were the first abodes of the dwellers of the 
forest I had seen, although there were some of 
mongrel character (like their inhabitants) in the 
Nazare and other suburbs of Para. They wore an 
air of neatness and comfort, and made me think of 
Will Atkins's house on Robinson Crusoe's island. 
The walls were of palm leaves, closely woven into 
a sort of matting. The roofs were covered with a 
sort of shingles, made by tying several of the 
broad flat fronds of a small palm called Ubi'm 
(Geonoma) on to a stick so as to closely overlap 
each other. A roof of Ubfm looks pretty, keeps 
out the rain well, and lasts a long time. At a 
short distance was the essential mandiocca planta- 
tion, covering several acres. An old Indian pointed 
out to me eight or nine varieties of that most useful 
vegetable (the Manihot utilissima of botanists), each 
grown in a plot kept carefully separate from the 
rest ; he professed to distinguish them by the 
leaves, but I confess I was unable to do so ; how- 
ever, there is no doubt that the roots vary much in 
shape and colour, some being whitish, others deep 
yellow ; that some kinds ripen sooner than others, 
and that some suit best for making farinha de agua, 
others for farinha secca. Farinha de agua is made 
by macerating the mandiocca roots in water till 
they are soft enough to be broken up by hand. 
