I feel deeply honoured to have received the Society's invitation to deliver the 46th Wilbur Wright Memorial Lecture. I undertook the task with humility in the light of the illustrious company who have preceded me but, more particularly, because of the greatness of the contribution of Wilbur Wright whom we honour tonight which, together with that of his brother Orville, continues to provide a tremendous inspiration to all concerned with aeronautics.

It is never satisfactory to design a structure piecemeal. The whole construction ought to be considered as one unit and designed from the outset to meet each of the several loading actions to which it may be subjected. Part structures designed for one principal loading system and later modified to accord with other part structures or to meet other loading requirements, will seldom prove efficient. Moreover a structure designed from the outset to meet alternative loading cases is necessarily better integrated than a structure designed primarily for one loading case and later modified to meet another.

Semi-empirical methods are described for obtaining the normal forces and centre of pressure positions of isolated cone-cylinders and ogive cylinders with and without boat tails at supersonic speeds. The data for normal force slope and centre of pressure at zero incidence is, in general, accurate to better than 10 per cent and half calibre respectively. The data for high incidence must be considered speculative although reasonable agreement has been found with the very limited experimental data available.

Transition to turbulence in the boundary layer on an aerofoil may occur in several different ways, not the least important of which is by bubble separation of the laminar layer. Thus in establishing an empirical criterion for the limit of existence of a laminar boundary layer the necessary parameters must be able to describe separation and instability, as well as being physically reasonable and simple to calculate and use.

Slender Struts are usually tested with pin joints at both ends, because this is the theoretical condition which is most readily achieved (or closely approximated to) in practice. In conventional testing machines this results in a compression column with two pin joints. The instability of this column makes the accurate alignment of the loading members important, and reliable testing becomes correspondingly difficult.