Consider a thin-walled tube fixed at one end and subject to torque of known distribution. The problem is to find how the thickness of the wall should vary along the length so that the weight is a minimum for a given angle of twist at the free end.

At distance x from the fixed end, let T be the applied torque, A the enclosed area of the tube, P its perimeter and y the wall thickness.

This note suggests a method, suitable for stressing purposes, of obtaining a fuselage air-load distribution, for a swept-back wing aircraft. The method is basically that given in Ref. 1. Mach number effect on the aerodynamic centre position is discussed and an approximate method of allowing for this is suggested. A numerical example is given.

Aerodynamic theory may be used with reasonable confidence to estimate the low-speed value of the stability derivative Ip for wings of most plan forms under quasi-steady conditions so long as viscous effects are unimportant. Wind tunnels tests are made to cover other cases such as the partially stalled swept wing, but the writer has heard of a difficulty encountered by a firm when using the free-rolling method. A variant of that technique in which this difficulty is avoided and which has been used at Imperial College for some years may be of interest.

This note describes brief tests made to check a proposed extension to the Royal Aeronautical Society Stressed Skin Data Sheets on stresses and deflections in flat panels subjected to normal pressure. Simple modifications to the test rig, which greatly improved the edge fixing, are also included.

The problem of deducing resonance modes of vibration of an aircraft in free space is a concomitant of flutter calculations if the number of degrees of freedom used is to be small. When the structure is complex in that it involves wings, fuselage and tailplane, each of which possesses infinitely many normal modes, it becomes apparent that the number of point masses which must be considered, in constructing a dynamical equivalent to give a sufficient coverage of the frequency range in which flutter is likely, is very large. For example, it may be necessary to use four rods per wing and per tailplane and four point masses on the fuselage. This would involve 8 + 8 + 4 degrees of freedom, and if the usual technique of characteristic roots is used it would be necessary to consider a characteristic root matrix of order 20 × 20.

Reading the paper by Tye and Cundall in the June issue of the journal, I was started on a train of thought on the same lines.

I agree with the paper but feel that the argument has not been taken far enough. The main conclusion is:–

So long as a random failure is no more expensive and troublesome than a controlled replacement, it is always cheaper to wait until an item fails before replacing it. This is quite independent of the law of failure with time.