of a mass of native iron in Brasil . 273 
the rivulet, but entirely in its bed , which was then dry, and is 
very seldom otherwise. 
I send you a very correct outline of this mass. (PI. XI.) It is 
about 7 feet long, 4 feet wide, and 2 feet in thickness, besides 
a sort of foot on which it now stands, of about six inches in 
height. I he solid contents, however, cannot be inferred 
correctly from these dimensions, since the broad part is hol- 
lowed out underneath very considerably. After making due 
allowance for the cavities, I estimated, on the spot, the solid 
contents of the whole mass to be at least 28 cubic feet, which 
at 5001b. will make its weight to be 140001b. 
Its colour is exactly that of a chesnut, and is glossy at the 
top and sides, but the hollow part underneath is covered with 
a ciust in thick flakes, outwardly of the colour of rust of iron, 
and staining the fingers. The flakes are very brittle, and 
the fresh fiacture is black and brilliant, like some magnetic 
iron ores. 
The glossy surfaces of the block are not smooth, but 
slightly indented all over, as if they had been hammered with 
a rather large round headed hammer. 
There are several cavities in it, from the diameter of a 
i2lh. cannon ball, to that of a musket ball ; the larger ones 
being shallow, but the others much deeper. They all contain 
the same substance as is attached to the great cavity under- 
neath, and some of them also fragments of quartzous stones, 
which I was obliged to break in the holes in order to get 
them out. 
The brown colour of the surface of the block is merely 
a very thin coat of rust, for the slightest scratch with a 
knife produces a bright white metallic streak; and yet, 
