82 The Formation of Coal . [February, 
the annual layers curiously loosened and separable from each 
other, like successive rings of bark. This continued till I 
had stripped the stick to about half of its original thickness, 
when it became too compact to yield to further stripping. 
This structure apparently results from the easy decompo- 
sition of the remains of the original cambium of each year, 
and may explain the curious fadt that so many specimens of 
fossilised wood exhibit the original structure of the stem, 
although all the vegetable matter has been displaced by 
mineral substances. If this stem had been immersed in water 
capable of precipitating or depositing mineral matter in very 
small interstices, the deposit would have filled up the vacant 
spaces between these rings of wood as the slow decomposi- 
tion of the vegetable matter proceeded. At a later period, 
as the more compact wood became decomposed, it would be 
replaced by a further deposit, and thus concentric strata 
would be formed, presenting a mimic counterpart of the 
vegetable structure. 
The stick examined appeared to be a branch of oak, and 
was so fully saturated with water that it sunk rapidly upon 
being released. 
On looking around, the origin of this sub-aqueous forest 
was obvious enough. Here and there the steep wooded 
slopes above the lake were broken by long alleys or down- 
ward strips of denuded ground, where storm-torrents or 
some such agency had cleared away the trees and swept 
most of them into the lake. A few uprooted trees lying at 
the sides of these bare alleys told the story plainly enough. 
Most of these had a considerable quantity of earth and 
stones adhering to their roots ; this explains the upright 
position of the trees in the lake. 
Such trees falling into water of sufficient depth to enable 
them to turn over would sink root downwards, or float in an 
upright position, according to the quantity of adhering soil. 
The difference of depth would tend to a more rapid penetra- 
tion of water in the lower parts, where the pressure would 
be greatest, and thus the upright or oblique position of many 
of the floating trunks would be maintained till they absorbed 
sufficient water to sink altogether. 
It is generally assumed that fossil trees which are found 
in an upright position have grown on the spot where they 
are found. The fafts I have stated show that this inference 
is by no means necessary, not even when the roots are 
attached and some soil is found among them. In order to 
account for the other surroundings of these fossil trees a 
very violent hypothesis is commonly made, viz., that the 
