When the Council of the Society did me the honour to invite me to deliver the Wilbur Wright Lecture I found myself in a dilemma. Many of those whose lives are concerned with engineering or with scientific research must feel, as I do, that after a certain stage they cease to be quite the real thing. The realities of the workshop give place to a daily round of files, reports and estimates, and the laboratory knows one only as a visitor.

When a man has entered upon this ink-stained path and is invited to give a public lecture, there are only two things he can do, except refuse. He must either plagiarise or generalise. Of the two, on grounds of morality, I prefer the latter; but it has the disadvantage common to most moral alternatives, it is much the more difficult. For myself, too, there is a further complication, because the field in which I might feel qualified to generalise would be on some subject relating to engines and their development.

The object of the present report is to make a general survey of the simple formulas and methods of calculation for the determination of the strength of thin-walled structures (plate and shell structures) which are increasing in importance in aeroplane construction, and to facilitate the study of original papers by adding a detailed list of relevant literature. an essential characteristic of the sheet metal covered structures in question is, that the metal skin in addition to the stiffening elements participates in the transmission of force. structures in which the metal sheet serves merely as a covering but which is not loaded in accordance with its strength, do not come within the scope of the present considerations. non-stiffened and stiffened plates and shells, mostly of very small wall-thickness (about 0.5 to 1.2 mm.) may be regarded as structural elements in plate and shell structures (especially shell fuselages, “ monocoques ” and wings).

During the last few years continuous indicating electric exhaust gas analysers have been coming into use more and more for controlling the strength of the fuel/air mixture which the carburettor feeds into the engine.

These instruments have already been described in a publication (vide List of References (i)), but since they are of growing interest, especially for aviation, some further research as to their merits and shortcomings seems to be indicated.

Therefore tests have been run with an aircraft type of instrument (the “ Cambridge (N.Y.) Aero Mixture Indicator ”) and an automobile type of instrument (the “ Lantz Phelps Motor Fuel Combustion Tester ” ).

The rapid progress made in mechanical engineering during the present century has brought many troublesome problems in its wake. One of the most serious of these troubles arises from resonant and such like vibration. It has, therefore, become increasingly necessary for a designer to take account of vibration. A powerful instrument for this purpose was provided by the late Lord Rayleigh, whose method for finding an approximation to the fundamental frequency of vibration of an elastic system is fairly well known. In this paper an attempt is made inter alia to explain Rayleigh's Principle and to clarify, so far as the author is able to do so, the basis of the principle. Other devices for ascertaining approximate frequencies are also given and an appendix has been added on forced vibration experiments.

In a previous issue of this Journal (Vol. 40, pp.663-680, 1936) Mr. N. J. Hoff has investigated the buckling of an elastically encastred strut. The ingenious method used by Mr. Hoff consists in considering portions of various lengths of a rigidly encastred strut which is in a state of neutral equilibrium under the action of its critical compressive load. This load is also the critical load of an elastically encastred strut whose length is equal to the length of the considered portion, if only the rigidities of encastrement are chosen in accordance with the ratios of the slopes of the neutral axis at the ends of this portion to the bending moments there. The total length of the elastically encastred strut does not enter immediately into the formulae derived by this method, there enter only two lengths x1 and xr whose sum is equal to the total length. These quantities are the distances between the ends of the considered portion of the rigidly encastred strut and the middle of its span.