rcest gales that blow.

The structure of the cocoanut seems well fitted to endure winds of
almost any force, and that a remarkably abundant and strong circulation
of air is essential to its best development is well shown by comparing
a tree subjected to it with the wretched, spindling specimen growing
in a sheltered glen or ravine.

Strong confirmation of this may be found within the artificial
environment of a plant conservatory, where it is feasible to reproduce,
in the minute detail of soil, water, temperature, and humidity,
every essential to its welfare except a good, strong breeze. As a
consequence, the palm languishes and it has long been deemed, on
this account, one of the most rebellious subjects introduced into
palm-house cultivation.

THE SOIL.

The soils for cocoanut growing are best selected by the process of
exclusion. The study of the root development of the palm will prove
to be an unerring guide to proper soil selection.

The roots of monocotyledons, to which great division this palm belongs,
are devoid of the well-defined descending axis, which is possessed
by most tree plants, and is often so strongly developed as to permit
of rock cleavage and the withdrawal of food supplies from great depths.

The cocoanut has no such provision for its support. Its subterranean
parts are simply a mat-like expanse of thick, fleshy, worm-like
growths, devoid of any feeders other than those provided at the
extreme tips of the relatively few roots. These roots are fleshy (not
fibrous) and can not thrive in any soil through which they may not grow
freely in search of sustenance. It then becomes obvious that stiff,
tenacious, or waxy soils, however rich, are wholly unsuitable. All
very heavy lands, or those that break up into solid, impervious lumps,
and, lastly, any land underlaid near the surface with bed rocks or
impervious clays or conglomerates, are naturally excluded. All other
soils, susceptible of proper drainage, may be considered appropriate
to the growth of the palm. Spons (Encyclop.) advocates light, sandy
soils. Simmonds (Trop. Agric.) names nine different varieties suitable
for this purpose, describing each at tedious length, and laying more
or less emphasis upon a sandy mixture. These might all have been
covered by the single word "permeable."

As a matter of fact every grain of sand in excess of that required to
secure a condition of perfect permeability is a positive disadvantage
and must be