In 1931, several years before the RAE work on box wing theory which was mentioned (in the January Journal), I joined Armstrong Whitworth Aircraft. They were designing the Atalanta monoplane for Imperial Airways, and were in difficulties with the wing stressing, my chief H. A. Francis having, as he said, applied to RAE for advice and “got the bird”. The late J. K. Reid, their leading “methods man”, had already produced the simple exponential theory of uniform box torsion which RAE later re-discovered, but unfortunately, for our tapered thin-skin wing, we soon found out that this did not work.

If You take a quick look at the subjects chosen for this Colloquium you will see that all except one have been, or will be, given by specialists and on their own subjects. In being invited to comment on the responsibilities of the user I have been presented with the unique task in this Colloquium of telling someone else what his job is. It is hoped, therefore, that the next ten minutes may provoke some lively discussion and will be taken in good part by anyone implicated.

Since the two speakers who follow are dealing with the way in which the user handles defects and defect records, it seems fitting to open with one or two comments on the user's responsibilities for this function, as seen by the manufacturer

The Outbreak of war in August 1914 naturally gave a fillip to the development of aeronautics and to building up research teams at the NPL and at the Royal Aircraft Factory at Farnborough. I was one of the earlier war-time recruits, in September 1914, engaged—after a mere assurance that I did believe in aeroplanes—by the enthusiast E. T. Busk, who was to lose his life so tragically when the BE2c he was flying caught fire on the 5th November 1914. At Farnborough I found a Cambridge contemporary of the Mechanical Sciences Tripos 1914, H. L. Stevens, who had joined before the outbreak of war, R. H. Mayo, Geoffrey Taylor, Melvill Jones, who had abandoned work on airships for Vickers, and F. M. Green from the motor trade on engines.

I was in No. 2 course at Denham in 1916 and flew Longhorns, Shorthorns, Henry Farmans and Vickers gun buses. I was in the last pusher training squadron but was stopped flying before I obtained my wings. I became Navigation Officer at Heliopolis and taught aerial navigation and later went to Palestine on the same job.

I remember plenty of amusing incidents of the First World War days. For example, I was standing with General Brancker when the first SE5 arrived at Palestine airfield. The pilot forgot to close his extra air, overshot, and couldn't pick up and went straight through the line of tents (luckily we dug in about 4 ft for coolness) and a yell went up—“keep down”.

The “displaced frequency” method is widely used in the aircraft industry for estimating the generalised masses and stiffnesses of aircraft, which are used in response and other aeroelastic calculations. The usual procedure is to measure the natural frequencies, ωi, and the associated normal modes, zi , by standard resonance testing techniques. (The number of frequencies and modes actually measured depends, of course, on the accuracy required in subsequent calculations.) A small mass, or group of masses, is then added to the structure at a convenient point or points and the resulting frequency changes, δωi noted. Now if αi and εi are generalised masses and stiffnesses respectively, then in normal co-ordinates

I exhibited the “Silver Streak” before a patent was granted to me because my Patent Agent, the late Mr. Griffith Brewer, had asked me to allow him to make a search, and this took a considerable time. What he found as being the “state of the art” is described in my first patent, 185,992, in clauses 70 to 85 on the first page. (Application date, 10 November 1921; Complete Left 15 February 1922; Complete Accepted 21 September 1922). None of those projects described as being the “state of the art” were ever carried into practice successfully, and a study of statements in 70 to 85 will show that the Inventors had failed to realise that such structures as they visualised could not safely be made, as indeed they were never made, and carried into practice in an actual aeroplane or flying-boat.

A further element for the Matrix Displacement method discussed in the author’s Main Lecture was a parallelogram or arbitrary anisotropy under plane stress. As for the other elements analysed in the preceding technical notes of this series, we assign to the parallelogram a linear variation of the strain along the edges. This is a more general specification of the strain pattern than for the parallelogram in ref. 2. Consequently, it is necessary to introduce additional nodal points in order to account for the increased number of kinematic freedoms. In the present case eight nodal points are provided as against four in the original model. The procedure is best understood in

About Half of the speakers so far have mentioned feed-back of user experience. Unquestionably, if we are to make real advances in reliability, Industry must be fed with comprehensive and timely data on the behaviour of its products in Service, and most of us would agree that the systems for data retrieval which exist today fall short of this requirement. It is inevitable, I suppose, that by the time we disperse from this meeting a number of problems within the whole field of reliability—whether to do with rig testing, responsibilities of sub-contractors, statistical prediction, contractual procedures, etc.—will have been thoroughly aired but not always with general agreement on them and in most cases without any feasible solution or break-through having been proposed. I suggest that the particular field of data feed-back is one where the requirement is almost unanimously agreed and where profitable and constructive steps forward are both identifiable and feasible.

Experiments have been made on a thin two-dimensional aerofoil with sharp leading edge and with a jet blowing through a mid-chord slot on the upper surface in order to control the boundary layer and circulation around it. The measurements were made at low Mach number. The effect of changing the angle of blowing was studied in anticipation that this would help to control the leading edge separation bubble. It transpired that variation of the blowing angle caused little change in lift coefficient but produced an appreciable change in moment coefficient. A low pressure separation bubble was formed downstream of the slot with the jet discharging at an angle and subsequently re-attaching to the surface. This reduced the nose-up pitching moment about the quarter-chord point. The experimental results are in fair agreement with the “jet-flap” and jet entrainment theory.