g they are sometimes used solely for
atmospheric heat, and are found to answer well. But if tanks are
constructed of substantial and enduring materials, they possess little
if any advantage, on the score of expense, over hot water pipes, while
they occupy much more room and are unsightly objects in a well ordered
green-house.

Wooden tanks are frequently used where the heat is required to rise
perpendicularly from them. If constructed of good pine plank, well put
together with white lead, and thoroughly painted inside and out, they
will last for several years. Scarcely any heat will be radiated from the
sides and bottom of a wooden tank. Tanks of brick and cement would
answer better than those made of wood, if it were possible to make them
water-tight when supported by piers above the ground, as they are
usually built. But however carefully constructed, these materials are so
unyielding to the expansion and contraction they are subjected to, that
it is nearly impossible to prevent leakage for any length of time. A
large number of brick and cement tanks have come under our notice, and
we cannot call to mind a single one of them all that has not been a
continual source of vexation and expense to its owner, since its first
construction.

The principle objections to tank heating, as usually employed, are an
excess of bottom heat and a deficiency of atmospheric heat, with a
superabundance of moisture when the vapor from the tank is not properly
excluded from the house. Tanks should be covered with some good
radiating material, as slate or metal. If slate is employed, the joints
should be carefully and effectually cemented. Boards are sometimes used
as a covering, but their radiating power is slight, and their decay
rapid.

Soil or sand, to the depth of six to ten inches, is usually placed upon
the tanks, and used as a plunging bed for pots containing cuttings; or
the cuttings are sometimes inserted in the bed itself.

Any arrangement by which vapor from the tanks is admitted to the roots
of plants is to be avoided, for however desirable a moist bottom heat
may be, it is found from experience that the soil is frequently rendered
a mass of puddle, in which no living roots can exist.

A portion of the covering of the tank may be made moveable to allow
moisture to escape into the house when required.

By means of the tank, bottom heat for propagating or other purposes, can
be very steadily and uniformly maintained, more s