The first public account of the methods and principles of performance testing of aircraft is, it is believed, that presented to this Society in a paper read by Mr. H. T. Tizard in 1917. Since that paper appeared there have been few changes in testing methods—a fact which indicates how very well the requirements had been met by the proposals therein described. During the period of nearly ten years which has intervened, however, attention has been directed to the analysis of the data provided by the tests. At that early date, the designer was no doubt content to check his predicted performance by the information as to rates of climb and speeds made available by the work of the R.F.C. Testing Squadron; and if he found discrepancies, they were probably attributed to scale effect or to some other of those conveniently indefinite agencies which sometimes enable the theoretician who has been rash enough to predict a performance to evade the consequences.

The original title put down for this paper was “The Design and Operation of Commercial Aircraft.” I hope it will be observed that I have actually taken a very much more modest title than that. A paper with so broad a title as the original one might well have been read five or six years ago, but at the present stage of development of commercial aviation, he would be a bold man indeed who would attempt to deal with two such subjects in a single one–hour paper.

It happens that the Society has had two very valuable papers dealing with commercial air transport comparatively recently. I refer in the first place to Sir Sefton Brancker's paper of October, 1925, which gave a very comprehensive review of the progress of civil aviation during its first six years, including a general survey of the development of commercial aircraft design. A little later on we had a most interesting paper by Mr. Cowley dealing with the more theoretical aspects of aircraft transport economy.

In no branch of structural engineering is there a rigorous definition of the term “secondary stress.”

If the primary stresses in a structure are determined, that is, the stresses in the members due to an external load system, assuming that all the members are joined together by perfect pin- or ball-joints, then the secondary stresses are in general taken to be the additional stresses due to the rigidity of the actual joints used in practice.

In such a highly redundant structure as an airship hull the labour involved, in determining even the primary stresses, precludes the use of the normal methods of stressing. It is usual to make use of generalised methods which give approximate results. These generalised methods imply that the external loads are applied to the structure in a certain distribution. Though this is rarely achieved, yet the results obtained are in most cases sufficiently accurate if suitable bracing is supplied to redistribute the external loads over the cross-section; the effect of the initial wrong distribution being then merely local.