My experimental work in flight refuelling goes back over a period of ten years, when I became interested in its possibilities as a means of removing the hazard' of a heavily loaded take-off. My idea was that an aircraft should have a normal and safe take-off, and that when it was comfortably in the air, it should receive its load of fuel, thus being given both an extended range and an increased1 pay-load.

In 1932, initial experiments were carried out on two D.H. 9's, and early in 1933, we continued experiments with a D.H.9 and a Handley Page W.iotransport machine.

The main object of this paper is to put forward a routine method for taking acccount of the variation in section in the stressing of rigid jointed frames. The formulæ in general use which mainly apply to members of uniform section are suitably modified so as to be applicable to cases where members may be of variable section. The underlying feature of the method is to base the treatment of a rigid jointed frame on the properties of cantilevers loaded at their free ends. Certain deflection coefficients are involved and as shown may be found separately as required for each member. These coefficients may be computed either analytically or graphically. The importance of taking account of variation in section is brought out by practical examples. In certain cases it is shown that very considerable errors may arise if instead of having regard to the actual variable section a mean value is assumed as being sufficiently close from the point of view of the resulting loads. From experience already gained in the application of the method to an actual structure the labour involved is not much in excess of that for the case of the same structure with the sections averaged.

The method does not take end load deflection effects into account and an extension of the method to do this seems desirable. There are, however, many cases in which end load deflection effects are not serious, and in any case it seems useful to have a solution in which these are neglected in the absence of means for obtaining a more comprehensive analysis.