Claude Grahame-White, who died on 19th August 1959, in Nice, after a short illness, was perhaps the outstanding British pilot in those golden years of flying 1909-1914. He did more than anyone to stimulate interest in flying in those early days and, in the 1916 edition of the then well-known “Representative Subjects of the King” published by Dod's Peerage Publishers, was described as “a pioneer of aviation, not only as an aeroplane pilot but as a constructor, and one of the chief originators of this new Industry in England—an industry which now promises to equalise in importance, and eventually eclipse, even such a great industry as that of shipbuilding”.

World interest in the development of a satisfactory fully automatic all weather landing system has increased considerably in the past year or so. Statistics on accidents reveal the disquieting fact that a very large percentage occur in the landing phase, either in poor visibility or in conditions where limited or misleading visual clues conspire to affect a pilot's judgment of the required vertical descent path. Factors such as the changed performance characteristics of large turbo-jet aircraft, higher approach and landing speeds and differences in the approach attitude of new aircraft all add to the pilot's difficulties during the landing operation.

When considering the problems of take-off and landing from the pilot's point of view, the most important fact to remember is that he is a human being. This drawback immediately limits his capabilities in the operation of an aeroplane because the control of a pilot's actions is through the computer we call the brain. This computer, although extremely efficient in almost all functions, has nevertheless serious limitations when it comes, first, to speed of computation, and secondly, the number of functions it can compute at one time.

Until now the speed of operation of aircraft has been compatible with the speed of computation of the human brain that is governing the pilot's reactions. Also the size and complexity of the aircraft and the operation have been compatible with the number of functions the brain can compute at one time

I claim to have no expert knowledge on landing aids, but the Aeronautical Research Laboratories in Australia have done some work in this field. We were invited, or encouraged, to take up this branch of work by a Meeting of the Commonwealth Advisory Aeronautical Research Council in Australia in 1955, when the difficulties of the landing phase and the need for some form of aid for the pilot were emphasised.

I need say nothing about the difficulties of the pilot: they have already been amply emphasised. An early example of operational research in this field was an investigation made by Professor Collar in 1941 on the problem of the dark night take-off, when the pilot tends to mistake forward acceleration for a climb in the absence of sufficient external visual cues.

The paper I have presented to you here is a brief account of the work we have been doing in the past six years at Farnborough and B.L.E.U. In studying visual judgments during the past few years, we have been driven to one conclusion which is pretty well the same as that which Capt. Prowse put before you, namely, that we are reaching the limit of what the human being can do. Every visual task has a certain failure rate, and I think this rate depends on the value of V|ak, where V is the approach speed, a is the acceleration which the pilot is able and willing to apply during a corrective manoeuvre, and k is an index representing the goodness of the visual stimuli. The tendency is for V|a to increase as aircraft get larger and heavier, but we have hitherto managed to counteract this by improving the visual aids, i.e. by increasing k.

In both civil and military flying take-off and landing accidents represent a major part of all accidents. In both take-off and landing, the problems are becoming more acute with the current change to larger and faster aircraft, and even more significantly with the change to jet engines.

The majority of accidents occurring in these two phases of flight are attributed to “ human error,” or “ pilot's error of judgment,” but recent human engineering studies have shown that in many cases the real cause has been due to the engineering situation overloading the pilot in some way, or requiring him to make judgments based on insufficient information.

Before any aircraft type is granted a certificate of airworthiness its performance characteristics are accurately determined and, for large aircraft at least, scheduled in an approved flight manual. Superimposed on these data are certain basic performance criteria which usually limit the operational weight with altitude and temperature, and these limitations apply under all operating conditions.

In order to ensure a satisfactory level of safety for daily operation under varying conditions it is necessary to apply an operations code which takes into account other parameters such as :

(a) runway surface, gradient and length,

(b) obstructions in the take-off and landing flight paths,

(c) obstructions en route.

—Since flow separation occurs readily from a highly swept leading edge, but gives rise in general to a steady flow, it is proposed that the rational approach to the aerodynamic design of slender wings is to attempt to control, rather than to suppress, these separations. This leads to the suggestion that the leading edges should be sharp, and that the wing should be shaped so as to make them attachment lines at one attitude (the design attitude) at which classical wing theory can be applied. It is argued, further, that if the leading edge separations are to develop regularly with change of attitude of the wing, separation must occur only from the trailing edge at the design attitude; and the velocity field favourable to boundary layer development without separation forward of the trailing edge is discussed. Subject to the restrictions thus imposed on the design, low drag is sought at the design attitude. This leads to the consideration of a particular class of doubly curved mean surfaces satisfying the leading edge condition, onto which thickness distributions are superposed so as to provide favourable velocity fields together with low drag. A number of examples are considered, using slender thin wing theory for flexibility, to illustrate the manner in which plan form and thickness distributions affect the pressure distribution, and to indicate the relatively high lift/drag ratios which seem feasible. Some consideration is given to the limitations of the theory used and to the further developments which seem desirable.

It has been found that there is a critical height for “sandpaper” type roughness below which no measurable disturbances are introduced into a laminar boundary layer and above which transition is initiated at the roughness. Braslow and Knox have proposed a method of predicting this height, for flow over a flat plate or a cone, using exact solutions of the laminar boundary layer equations combined with a correlation of experimental results in terms of a Reynolds number based on roughness height, k, and local conditions at the top of the elements. A simpler, yet more general, method can be constructed by taking additional advantage of the linearity of the velocity profile near the wall in a laminar boundary layer.

The Coefficient of contraction of a jet of fluid flowing through an orifice has been widely studied in the field of hydraulics, using a jet of water discharging through an orifice into the atmosphere and measuring the area of the jet at the vena contracta. the coefficient of contraction cr is then defined as:

Recent Researches have led to some possible explanations for thin aerofoil stalling behaviour. Apart from the Owen Klanfer criterion these are the reverse flow breakdown hypothesis of McGregor and Wallis's turbulent separation theory.

This note describes simple theoretical boundary layer calculations which indicate the feasibility of Wallis's hypothesis. In addition the results of some experiments on a thin two-dimensional aerofoil with various leading edge configurations with Reynolds number, based on model chord, of 1.8 million and 1 million support either of these hypotheses, depending on the leading edge configuration. It is concluded that thin aerofoil stall can occur broadly, through either of the suggested mechanisms, depending on conditions in the nose region.

The note in the July Journal by H. Pearson and A. B. McKenzie (p. 415) shows that the flow into an axial compressor is nearly at a constant velocity at all points in the entry plane; and that if a wake of low energy air is fed into it, the compressor reduces the static pressure locally in this wake, in such a way as to tend to preserve constant velocity over the whole face.

I would like to point out a possible ambiguity in the paper by J. Dunham, in the October issue of the Journal. When considering the relative stresses in similar steel and aluminium alloy blades the author states that for blades “ doing the same duty ” the maximum stress in the steel blade is three times that in the aluminium type. This would seem to be true only if both blades are made to vibrate with the same tip amplitude, whereas the term “ doing the same duty” implies withstanding the same air loading, and consequently the steel and aluminium alloy blades would suffer different deflections.