When I first commenced my experiments with a large machine, I did not know exactly what sort of boiler, gas generator, or burner I should finally adopt; I did not know the exact size that it would be necessary to make my engines; I did not know the size, the pitch, or the diameter of the screws which would be the most advantageous; neither did I know the form of aeroplane which I should finally adopt. It was, therefore, necessary for me to make the foundation or platform of my machine of such a character that it would allow me to make the modifications necessary to arrive at the best results. The platform of the machine is, therefore, rather larger than is necessary, and I find if I were to design a completely new machine, that it would be possible to greatly reduce the weight of the framework, and, what is still more, to greatly reduce the force necessary to drive it through the air. [Illustration: Fig. 89.--One of the large screws being hoisted into position. Its size may be judged by comparison with the man.] At the present time, the body of my machine is a large platform, about 8 feet wide and 40 feet long. Each side is formed of very long trusses of steel tubes, braced in every direction by strong steel wires. The trusses which give stiffness are all below the platform. In designing a new machine, I should make the trusses much deeper and at the same time very much lighter, and, instead of having them below the platform on which the boiler is situated, I should have them constructed in such a manner as to completely enclose the boiler and the greater part of the machinery.[14] I should make the cross-section of the framework rectangular and pointed at each end. I should cover the outside very carefully with balloon material, giving it a perfectly smooth and even surface throughout, so that it might be easily driven through the air. [14] This arrangement of the framework is now common to all successful machines. In regard to the screws, I am at the present time able to mount screws 17 feet 10 inches in diameter (Fig. 89). I find, however, that my machine would be much more efficient if the screws were 24 feet in diameter and I believe with such very large screws, four blades would be much more efficient than two. My machine may be steered to the right or to the left by running one of the propellers faster than the other. Very convenient throttle valves have been provided to facilitate this system of steering. An ordinary vertical rudder placed just after the screws may, however, prove more convenient if not more efficient. The machine is provided with fore and aft horizontal rudders, both of which are connected with the same windlass.