PERU. 589 
an inch. They are extremely numerous, often but an inch apart, 
and are almost always vertical, or nearly so, in position. Occasionally 
they branch so as to resemble a small twig. They appear to be concre- 
tions that were formed about the fibres of the roots of some grass. A 
grass is now growing above the cliff whose numerous rootlets might 
give rise, under the same circumstances, to a similar appearance. In 
some instances a vegetable structure or half decomposed wood is 
found at the centre of the concretion: yet, in the greater part, the 
structure is arenaceous throughout, and imperfectly concentric. Oxyd 
of iron is the cementing material in some, and may be in all. The 
yellow and brownish rings of colour, derived from the oxyd of iron, 
often resemble in appearance decomposing wood. A solution of iron 
appears to have trickled down the roots in the soil, and cemented the 
sand around them. 
The wood in this layer often contains gypsum disseminated through 
it in slender transparent crystals. ‘These crystals are occasionally an 
inch long; they are perfect in their faces and extremities, and have 
a brilliant lustre. ‘Some specimens, in which the colourless transpa- 
rency of the gypsum contrasts with the black wood, are quite beauti- 
ful. They usually lie in the direction of the woody fibre, though 
sometimes otherwise when the crystals are clustered. The form 
of these delicate crystals is the simplest presented by the mineral. 
They are often cruciform, as in the annexed figures, but none of the 
Fig. 1. 
arrow-head twins were observed. Similar crystals occur occasionally 
in the clay adjoining the wood. ‘This gypsum is evidently derived 
from sea-water, which always contains a small proportion. The wood 
is frequently wet by the spray of the present seas; and when the 
water dries it deposits its sulphate of lime and common salt. The 
latter is liable to be dissolved out by fresh waters, while at the next 
wetting with salt, the gypsum crystals thus begun would enlarge. 
This process continued for a time would produce an abundant deposit 
of crystals of gypsum. P 
148 
