These recollections start a long time after those in the memorable Centenary issue of the JOURNAL, but will add a page in the history of the RAE which I joined in 1939—and found myself in the Engine Research Flight.

Life there was very peaceful before the War and aeroplanes were new and exciting. Looking back now, some of the work done there was very primitive with none of the elaborate test and recording equipment available today. I recall flying for hours in a Harrow, looking for icing conditions to test an automatic carburettor de-icing system; endurance flights on new plug types, high octane fuel and tests of an early Merlin in a very odd aeroplane where the observer sat in a box under the pilot's seat. The pilots then in C Flight are now almost legendary figures—“Spinner” White, killed in the prototype Spitfire K 5054, Rex Hayter, “Tiger” Hawkins, Roly Falk, “Bruin” Purvis, Hubert Broad, Clouston and H. J. Wilson, to name but a few.

Human air transport has beenwith us for centuries— in fact since time immemorial. As in the classical case it is really quite modern in its conception.

There is a partially retractable landing gear. Although there are no seats, there is good leg room and the passenger has his air ticket. Take-off was easy. The flight, being slow, takes anything up to nine months. The landing will be laborious and its successful completion will be accompanied by a fanfare of joyous publicity on the part of the parent company together with the usual noisy complaint from the passenger. Emergency oxygen is occasionally administered.

In applying the numerical correction method to the case where two natural frequencies of the system under test are equal, it is necessary to ensure that the matrix

is diagonal in order to eliminate singularities. (In this equation δA is the displaced mass matrix; is the partitioned matrix containing the two modal columns appropriate to the equal frequencies and β is a transformation matrix (2 x 2), chosen to ensure that δαb is diagonal.)

The linear buckling of circular cylindrical shells is considered with particular attention to the cantilever shell subjected to either a pure bending moment (M) or transverse load (P)—see Fig. 1. It is believed that the conclusions reached have wider application to more general loading cases.

The planar librational motion of a dumb-bell satellite has been studied by several authors. In most cases the analysis has been restricted to circular or nearly circular orbits. In an elliptic orbit, the variation of the angular velocity of the satellite around the central body together with the change in the local gravity gradient provides the satellite with a mechanism for exchanging energy between the librational and orbital modes. In general, this imposes a limit on the orbit eccentricity for which stable (non-tumbling) librational motion is possible. This note investigates the bounds that must be placed on a disturbance applied to a slender gravity gradient stabilised satellite so that it will librate and not tumble. The determination of the stability boundary as presented here was arrived at by the numerical solution of the exact equation of motion, obtained using an IBM 7040 computer.

The investigation of axisymmetric states of stress by the matrix displacement method may, in principle, be carried out as an ordinary exercise of three-dimensional analysis, using TETRA 4, TETRA 10 or other suitable elements. However, this approach is clearly not the most efficient by not taking advantage ab initio of the inherent simplifications arising in an axisymmetric state; this leads to an unnecessary inflation of the number of unknowns and complexity of the mesh. A number of techniques to deal with this limiting condition of a three-dimensional state have been developed at the ISD. The theory for one specific element, TRIAX 6 which evolves most naturally from the tapered TRIM 6 element analysed in the preceding note 8 is presented.

The preparatory work of the preceding technical note is followed up by deriving the elastic stiffness and the initial load vector of an arbitrary TRIAX 6 element subject to an axisymmetric state of stress or strain.