It is a big honour to be invited to give this lecture in memory of that great engineer, Sir Henry Royce, but particularly so for anyone who happens to work in the Company of which he was co-founder and which of course represents the best memorial for which he could have wished.

I am very conscious of all this tonight but also, and with a distinct unease, that the Grand Old Man might be something less than sympathetic to my theme. The traditions of quality, craftsmanship and meticulous attention to detail design for which Royce is rightly famous, find, of course, their extreme justification in the construction of complex space vehicles, with their critical dependence on reliability. (There were about 15 million parts which had to work in the whole Apollo system, nearly half in the vehicle itself.) There is no doubt that he could have made immense contributions to space propulsion, as he did for land and air.

We have grown used to the idea that in the Middle Ages inventors produced several forerunners, either real or imaginary, of modern aerodynamic devices. The helicopter was designed, in principle, before Leonardo, while the rigid aerofoil, in the form of the plane-surface kite, achieved a remarkable degree of sophistication before the end of the sixteenth century. It has commonly been assumed that among these mediaeval inventions we should include some sort of three-dimensional kite, a precursor of Hargrave's box-kite, either in the form of a windsock equipped with wings, or in that of a primitive hot-air balloon, with or without additional lifting force provided by wings. I wish to examine the evidence for the existence of these objects, and to make a fresh assessment of the likelihood of their ever having flown.

The interest of wings with variable sweepback springs directly from pure commonsense and appeals to no profound knowledge of aerodynamics for its justification.

To realise the advantage of variable geometry, it is enough to know that only a wing of small relative thickness is capable of good performance at supersonic speeds and that by increasing the sweepback from 20° to 70° the thickness of a wing is divided by about 2.

In the advanced position, the wing offers its full span to the airstream and with high-lift devices in action (leading-edge slats and trailing-edge flaps combined), the aeroplane can develop the considerable lift necessary for take-off and landing as well as for break-through and for slow approach. Wings still advanced but slats, flaps and undercarriage retracted, the aeroplane is in excellent maximum fineness condition for protracted cruising at subsonic speed or for a long wait. As soon as transonic (Mach No of more than 0-8) or supersonic speeds are in question, the wings are progressively folded back.

Much of the Gloster story has already been told by H. F. King in Flight for 27th May 1955. The same edition contains a generous contribution by Air Commodore Alan Wheeler which most vividly portrays a pilot's point of view on the Grebes, Gamecocks and later on, the Gladiators and Meteors.

When I left school with a Scholarship to Cambridge, subsequently transferred to Manchester School of Technology because I thought it would enable me to acquire a more practical knowledge of electrical engineering, the Wrights had only just made their first flight and the idea of a career in the aircraft world never entered my head. However, after the final exams at the end of the three years at Manchester and four years experience in Works and in a Consulting Engineers office, I had the opportunity in 1909 of following my Chief, Mervyn O'Gorman, to Farnborough when he was appointed by Lord Haldane as Superintendent of the Balloon Factory, later rechristened the Royal Aircraft Establishment. My salary was £3 a week and my first job was Personal Assistant to the Superintendent and later I was given the resounding title of Technical Assistant in Charge of Supplies, mostly the special materials which were being developed for aircraft and engines. At the outset I had to make hydrogen for the Airships which were the sole product of the Factory for more than a year. Making hydrogen was more interesting than you might think because occasionally the gasometers, where the hydrogen was stored before compression into steel bottles, used to blow up. We made no less than four new Airships before Lord Haldane gave permission for aircraft design to start and we had some leeway to recover to catch up with A. V. Roe, Sopwith, Blackburn and several others who had started earlier.

When I was young and in my prime,
The Boats were flying all the time.
But now that I am old and gray,
Those Boats, like me, have had their Day.

The first time I saw flying-boats close-up was in 1928. The first was an old Curtiss HS2L being used in Tima-gami for forest fire patrol in northern Canada. Designed in about 1916, this antique was at the end of its tether, but was doing a job that no other aircraft could do until a replacement became available. The second flying-boat I saw was this replacement, the Canadian Vickers Vedette, another single-engined pusher biplane, but a much better one.

The present paper seeks to apply the ideas of discretisation to time dependent phenomena. As a suitable variational statement we may use Hamilton's principle. In practise this means that the time is discretised into a set of finite elements which are taken to be the same for all structural elements. A finite element in time consists simply of a fixed time interval. In our present discussion we detail in particular the case when at the beginning and end of the time interval the generalised displacements and velocities are given. For dynamic problems this is the minimum of information required, but the technique may easily be extended to account for additional “timewise degrees of freedoms”. Introducing an appropriate interpolation procedure we may obtain the displacement and velocity at any instant of time. It is then possible to carry out in the variational statement the time integration explicitly and to obtain hence a system of linear equations. The method is extremely simple, since the time interpolation of all structural freedoms of an element in space is the same. We also demonstrate that the general case of a multi-degree of freedoms system can be made to depend on the matrices which describe the unidimensional motion of a mass point.

It is evident to all those engaged in the analysis of realistic lubrication problems that none of the existing techniques of computation proves satisfactory for practical purposes. In particular the most extensively used method of finite difference approximations to the relevant Reynolds differential equation suffers from serious disadvantages. Thus, the contour of the lubrication film has to be described by constant co-ordinate lines in special reference systems, unless one is prepared to accept more or less serious inaccuracies. Also auxiliary conditions must be introduced when dealing with discontinuous changes of lubrication data, like film thickness. These and other reasons demand the development of a general method which can deal with all current practical complexities. Clearly, the finite element technique should prove an ideal tool allowing completely arbitrary variation of geometry, film properties, and boundary conditions. The present note intends to explore the theoretical basis of the application of finite elements, using a variational formulation of Reynold's equation. The various elemental matrices are formulated and detailed for triangular TRIC or TRIM-like elements.

In recent years many methods have been proposed for the measurement of the frequency response of linear systems but the method most often used in the study of structures is the steady state vector response method, the natural frequencies and damping ratios being measured from the resulting vector diagrams. This method is tedious and time consuming and often cannot be readily applied outside the laboratory. Some effort has been directed in the past to employ quasi-steady state methods such as the slow frequency sweep technique but errors are introduced because the response at resonance is less than the steady state maximum and the frequency at which the maximum occurs is shifted in the direction in which the excitation frequency is changing. Additional errors are introduced in the analysis of the system response because of the necessary averaging time of the function analyser used to derive the vector diagrams. Some refinements in technique have been proposed by Reed who developed the “λ Law” frequency sweep in which the percentage change in frequency per cycle is constant. This offers some reduction in test time when compared with logarithmic and linear frequency sweeps but a knowledge of the minimum damping likely to be encountered at any resonant frequency within the sweep range is essential so that the method may be employed advantageously.

The purpose of this paper is to summarise some recent results obtained at Standford University regarding the buckling of thin shells. A detailed review of work done by other investigators will not be given here because this was done recently in two papers written by the author. However, a few important steps in the evolution of our knowledge of the buckling of thin .shells will be mentioned to put the new results in proper perspective.

Parts I and II of this series of papers have dealt with the development of ion thrusters for use in an orbit transfer manoeuvre, and of various types of micro thruster for position and attitude control. The present paper deals with the actual satellite electric propulsion system and configuration needed for expansion manoeuvres. The control problems both during the expansion phase and subsequently in the required orbit are discussed. Short analyses for two specific cases (ELDO launcher and Black Arrow) are presented in which payloads and missions are defined.

Burt has shown that all the orbital elements of a spacecraft can be changed by means of a small continuous thrust. He gave particular attention to the problem of expanding from a low orbit to a synchronous one, and the change of orbit inclination and orbit eccentricity was also considered. Such a type of manoeuvre is eminently suitable to electric propulsion because the high exhaust velocities possible lead to a small propellant mass; this mass saving is achieved at the expense of the time taken for the manoeuvre.