Chap. LXXX. GRANITE BOULDERS. 261 
to the south of Hombori, where it was called A'gele 
was quite absurd if it be correct that the Galmdu 
which we had crossed the preceding day was really 
the lower course of the river near A^ribmda; but 
it is very difficult to say how these courses cor- 
respond, and nothing is more likely than that the 
same watercourse may join the Niger by several 
openings. As it was, we had a long dispute as to 
the manner in which we should cross this water, and 
the following morning we had to take a tedious 
roundabout way to get over it. 
After a march of two miles from our starting 
point we reached a crossway. We followed the advice 
of the Wadawi, who, having taken the lead at the 
moment, chose the path to our left, though that on 
the right crosses the rivulet at this spot ; but in 
the end it was perhaps as well that we did so, as 
otherwise we should scarcely have been able to ford 
it. We therefore continued our march after my 
companions had finished their prayers, which, as 
we always set out at an early hour, they used to say 
on the road. The open pasture grounds were here 
broken by large boulders of granite, while the rivulet, 
girt by fine large trees, approached on our right, or 
at least one branch of it, the river dividing near its 
mouth into a delta of a great many smaller branches. 
We here changed our direction, keej^ing parallel along 
the shore of the great river, where on a rocky island 
was situated the village of Ay6ru or Airu, from 
whence a troop of about twenty people were just 
s 3 
