402 
GUATEMALA. 
steep that they can be descended only at certain points by 
means of ladders and steps cut in the lava rock. Finally 
there are many pits, sometimes no more than a hundred 
feet in diameter, but of very great depth, and filled some¬ 
times with fresh water, but more commonly with saline 
waters so strongly impregnated as to be undrinkable. 
The great lakes of Amatitlan and Atitlan are not cer¬ 
tainly volcanic, although their shores are dotted with 
hot-springs and guarded by volcanoes, — they are not, 
that is, actual craters; but the former seems to he the 
result of a subsidence caused perhaps by the removal 
of material from lower layers by eruptions of Pacaya, 
and it is of no considerable depth, while good authority 
has considered the Lago de Atitlan the result of damming 
up a valley and streams by the masses of the volcanic 
group of the same name. A glance at the map of this lake 
(p. 154) as given by the French geologists whose opinion 
is quoted, will show that the volcanoes occupy a position 
not far from the geometrical centre of the Lago, or where 
they should be if the lake was an ancient crater. Com¬ 
pare with this, if you will, the plan of an undoubted 
volcanic lake, that of Ilopango in San Salvador. This 
body of water is not only the seat of volcanic eruptions, 
as is also the Great Lake of Nicaragua, but probably fills 
a depression that has been the result of the coalescence 
of several points of eruption. I have before me the 
interesting report to the Guatemaltecan Government by 
my friend Edwin Roekstroh of his observations made on 
the eruption of one of these craters in 1880. The lake is 
9,200 metres wide from east to west, and 7,300 metres 
from north to south, with an area of 54.3 kilometres. 
Completely surrounded by precipitous mountains, inter- 
