2147. The first attempt to heat hothouses by hot air appears to have been made by Dr. Anderson as early as 1802, in a greenhouse attached to his house at Isleworth; and a patent was taken out for a nearly similar plan the following year by Mr. Stewart. Neither plan, however, answered; in the one case, from the air admitted into the house being too dry for the plants, and in the other, from a deficiency of heat. Some years afterwards, the hothouses of Messrs. Strutt, at Belper and Derby, were heated by a current of hot air warmed by a cockle stove, and returned to be reheated by the stove after it had been chilled in the house. By this plan, which is a very complete one fresh atmospheric air can be admitted when necessary, and heated before it is allowed to enter the house: the heated air is also made to pass over water, and thus becoming surcharged with moisture, it is free from the defects experienced in the first modes of heating by hot air. In 1825, the large conservatory at the Grange was heated by hot air produced by one of Mr. Sylvester's cockle stoves in combination with steam, a plan which answered very well, but was expensive; in 1840, the late Mr. Penn, an engineer as Lewisham, invented a mode of heating by hot air in combination with hot-water pipes; and in 1841 the Polmaise system was invented. As the last two systems have occasioned considerable discussion, and have been compared with each other, though they are, in fact, essentially different, we shall describe both of them somewhat in detail.