248 
OF THE BANYAN TREE OF INDIA. 
whence roots are wished to he formed ; the wound is covered by a ball 
of moist earth, kept on by a bandage of some kind or other, and con¬ 
stantly moistened by a leaking vessel of water suspended above the ball 
of earth. When the latter is filled with fibrous roots, the branch is 
separated from the mother and potted, or placed where it is to remain. 
A vine planted at one end of a wall may by annually layering the 
lowest side shoots be made to cover the whole in time, and the vine¬ 
yards in France are occasionally re-stocked with young plants by layer¬ 
ing young branches in the vacant spaces between the old roots, which 
last having become debilitated are afterwards grubbed up. 
Roots produced in the air are constitutionally very different from 
radicles produced in the earth or in water. Both the latter are extremely 
delicate, and would be almost instantly destroyed by dry air if exposed 
to it. The former are more robust, hardy, and emit no fibres or sponge- 
oles like those in the earth. They descend by a continual protrusion 
of their central parts, the point ever throwing off and leaving behind 
the cuticle or bark. These aerial roots are evidently produced by and 
from the vital membrane of the plant, and certainly not from either the 
bark or the wood, for neither buds or roots are ever seen to originate in 
those members after they are once formed. 
Many tropical plants produce roots from their stems and branches ; 
not only the different species of tropical orchises already alluded to, but 
many of the Bromelicicecz, and particularly the Vandanece. The greater 
number of these two orders are ever producing sets of roots from the 
joints higher and higher up the stem, the older roots as regularly dying 
off below, so that at last the head is supported by large roots like but¬ 
tresses ranged around. From this constitutional peculiarity of these 
plants it is evident that every new extension of the head is furnished 
with a corresponding set of roots, and this incident allows the practice 
of an apparently very unnatural manipulation, and which would be 
positively fatal to the generality of plants ; this is neither more nor less 
than depriving them of all their old roots, in order to induce a stronger 
growth of the head. That this is not only a practicable, but in some 
cases of their culture a necessary proceeding, is well known to practical 
pine-apple growers, and sufficiently proves that the new growth of the 
leaves and fruit depends entirely on the new roots which are simul¬ 
taneously produced. 
Propagation by layering is often preferable to any other method. 
Stone-fruit trees, whether grafted or budded, are sometimes liable to 
gum at the insertion of the graft or bud, and sometimes defects of the 
stock are, or may be, conveyed to the graft; but a healthy well-rooted 
layer is pure individually as well as true to its kind. We have seen 
