192 
BLUEFIELDS. 
but subordinate roots proceed from one to another, 
anastomosing in all directions (if I may use such a 
term), so as to make a most elaborate network of a 
multitude of meshes of various angular forms and 
sizes. These cross-roots are at each extremity united 
with the larger roots, and look as if the whole net- 
work had been skilfully carved out of one solid plank 
of wood, by cutting out the areas or meshes, and 
rounding the component bars ; the very bark that 
covers the w^hole is continuous, where the roots unite, 
as if they had been always integrally one. 
The only mode in which I can account for this 
singular phenomenon is the following hypothesis. The 
seed of the tree was originally deposited on the sum- 
mit of the wall, beneath the eaves. As it germinated, 
the roots ran down towards the earth, some per- 
pendicularly, some diagonally ; but all creeping along 
the surface of the wall, no roots having shot out from 
its perpendicular. As these roots increased, they 
sent out side rootlets, which, still running on the face 
of the wall, by and by came in contact with another 
of the primary roots. Then, instead of creeping over 
it, as the roots of other trees would have done, the 
soft tip of the rootlet actually united with the sub- 
stance of the root at the point of contact, the fibres 
of the two becoming interlaced, and their united sur- 
faces gradually becoming covered with a common 
bark. The repetition of this process has produced 
the very curious wooden net which I have attempted 
to describe. 
The tendency of the roots of the Fig to envelope 
closely any surfaces that they meet with, produces 
