The A.G.P. is required to produce 20 kw. A.C. and 3 kw. D.C. continuous normal full load, at 3,750 r.p.m. The load is required for operating the electrical services on a large aircraft and must be maintained both when it is stationary and when it is in flight up to an altitude of 11,000 feet.

The A.G.P. must run cool in the standard English summer atmosphere (23°C. at ground level) and draw the necessary air from ports cut in one side of the aircraft. These ports must be formed from a 1ft. diameter hole divided on its diameter. The cooling air is to be drawn in through the forward semi-circular hole and, after passing through the installation, expelled through the rear semi-circular hole.

An accurate knowledge of the strength of materials is more essential in aircraft construction than in any other branch of engineering, as every aircraft component has to be reduced in weight to a minimum compatible with safety. Equally essential in aircraft construction is the use of every available refinement in calculation, although it is not possible as yet to estimate with sufficient accuracy from calculations alone the strength of many of the components. At every stage hv aircraft construction, mechanical tests are therefore made, first to determine the qualities of the materials, then to verify the strength of each part and finally to. check the strength of the complete machine.

Owing to the wide range of my subject and to the limited time available, I propose to deal briefly with some of the aspects of mechanical testing of special importance in aircraft research and development, describing more particularly some of the test apparatus and methods of testing I have had opportunities of developing.

Thanks to the pioneer work of Lanchester and Prandtl everyone seriously interested in aeronautics has nowadays sufficiently clear ideas about the power and the associated drag which are essential for the support of an aeroplane in flight. This induced drag, as it is now called, can be calculated with an accuracy which is ample for most practical purposes, and even with the very clean aeroplanes of the present time it contributes, at and above cruising speed, a minor part only to the total drag. We are equally familiar with the idea that the remainder of the drag—the drag involved in getting the aeroplane through the air as opposed to holding it up—can be reduced, by careful design, nearly to the mere skin-friction of the air rubbing over exposed surfaces; the reason being, of course, that the pressures on the surface of a well streamlined aeroplane are so distributed that their horizontal components form an almost completely balanced system.

In view of the recent notes in this Journal by Professors Jacobs and Duncan, it may be of interest to give some additional references to investigations of the flexural centre of beams, which otherwise might be overlooked, and to add some remarks which seem to have escaped the attention of previous writers.

Although the earliest investigation seems to have been published by Griffith and Taylor (Ref. 4 of Duncan's paper), Eggenschwyler gave an independent discussion of the problem for narrow beams. Considerable progress was made by Weber in two papers, published in 1924 and 1926. Similar results were obtained by Schwalbe, who was apparently unaware of Weber's previous investigations.

A study has been made of the effect of inlet conditions on the performance of conical diffusers with 4:1 area ratio and 5 and 10 degrees total angle of expansion. The conditions at entry were varied by using different approach lengths of diffuser inlet diameter, and by means of projecting annular screens of woven wire cloth. With this new technique it was possible to vary the velocity distribution substantially within moderate settling lengths, and to produce velocity profiles with inflections. The suitability of the annular screen method of boundary layer generation for diffuser investigations was confirmed from the comparison with the approach length results.

Energy and pressure coefficients, as well as diffuser energy efficiency and the conversion efficiency, were found to depend on the diffuser angle β and the momentum thickness ratio at inlet θ/D 0. The latter emerged as one of the chief parameters controlling diffuser performance. Variation in the inlet shape parameter H of the order of 20 per cent did not significantly affect the pressure recovery or the losses in the diffuser. For moderately thick boundary layers, θ/D 0 X β<0·1, the diffuser angle could be eliminated as a parameter by plotting the pressure coefficient against θ/D 0 X β. The Reynolds number based on diffuser inlet diameter was, for essentially incompressible flow, 2·5 X 105 in all tests.