by a _complete
system_ of covering hot-houses and forcing-houses; and this plan would
secure the further advantage of avoiding the undue stimulation of the
plants by a then unnecessary amount of heat, applied solely to prevent the
very evil which covering also prevents, namely, the risk of excessive cold
during the night.

The principle upon which a covering acts most efficiently, is that of
enclosing a complete body or stratum of air exterior to the glass, this
body of air being entirely shut away from the surrounding outer
atmosphere. Air being a bad conductor of heat, the warmth of the interior
is by this means prevented from passing to the exterior atmosphere; or, in
other words, the exterior atmosphere, being prevented from coming in
contact with the glass, cannot absorb from the interior any material
proportion of its heat. To secure this advantage, however, the coverings
_must_ be kept from contact with the glass, and they should extend on
every side where the structure is formed of materials which readily
conduct heat--such as glass or iron. The coverings should in fact form
neither more nor less than _a close outer case_.

One point connected with the application of these coverings, which I
consider would constitute an improvement, and which, as far as I am aware,
has never been acted on, is that of having them to fit so accurately as to
exclude the external air (a matter of no difficulty in the degree
required), and then to have a series of ventilators provided, to stand
open during the night, whereby an interchange of the atmospheric volume
would take place throughout the night, without exposing the plants to
contact with cold air. The stagnation of the internal atmosphere would
thus be prevented, in consequence of the interior air and the air between
the glass and the covering being of different degrees of density, owing to
their being differently charged with heat. By this plan, therefore, I
conceive that direct benefit would accrue to the plants; and it would also
materially assist in preserving that cooler--but not cold--night
temperature, which the fear of injury from frost prevents from being more
fully realised in ordinary cases.

[Illustration]

The annexed diagram represents one of the many ways in which this idea
might be carried into practice. It will be understood that, as here shown,
the side shutters and end shutters (the latter not indicated), fit into
grooves, the upper groove being a