5IO NOTES OF A BOTANIST 
deep ravines, furrowed in soft alluvial sandstone 
rock, wherein blocks and pebbles of quartz are inter- 
spersed, or interposed in distinct layers. Towards 
their source they are obstructed by large masses of 
quartz and other rocks ; but as we descend the 
stones grow fewer, smaller, and more rounded, until 
towards the mouth of the Bombonasa, and thence 
throughout the Pastasa, not a single stone of the 
smallest size is to be found. The beaches of the 
Pastasa consist almost entirely of powdered pumice 
brought down from the volcano Sangay by the river 
Palora. When I ascended the Bombonasa in the 
company of two Spaniards who had had some 
experience in mining, we washed for gold in the 
mouth of most of the rivulets that had a gravelly 
bottom, as also on some beaches of the river itself, 
and never failed to extract a few fragments of that 
metal. All these streams are liable to sudden and 
violent floods. I once saw the Bombonasa at Puca- 
yacu, where it is not more than 40 yards wide, rise 
18 feet in six hours. Every such flood brings down 
larofe masses of loose cliff, and when it subsides 
(which it generally does in a few hours) the Indians 
find a considerable quantity of gold deposited in the 
bed of the stream. 
The gold of Canelos consists almost solely of 
small particles (called "chispas," sparks), but as 
the Indians never dig down to the base of the wet 
gravel, through which the larger fragments of gold 
necessarily percolate by their weight, it is not to be 
wondered at that they rarely encounter any such. 
Two attempts have been made, by parties of 
Frenchmen, to work the gold-washings of Canelos 
systematically. One of them failed in consequence 
