In Figs. 1 and 2 are shown two cantilevers, one acted on by a force which maintains its direction as the cantilever deflects, and the other acted on by a follower force which remains tangential to the end of the cantilever. It can be shown, by a simple demonstration, that a follower force is non-conservative (see Bolotin), and therefore has the capability, if it exceeds a certain critical magnitude, of causing oscillatory instability. Thus, the follower force system in Fig. 2 can become unstable in the oscillatory sense while the Euler Strut in Fig. 1, being a conservative system, can become only “statically” unstable, i.e. with the strut diverging monotonically from its equilibrium position. The follower force system can, of course, also become statically unstable. The overall stability of this system has been examined by Bolotin and Beck. Herrmann and Bungay have investigated the stability of a system that combined the main characteristics of the two above in any chosen proportion.

This section deals with the application of models developed in Part I to particular UK domestic and international routes. Most of the information used is in time series, a form of data which brings with it many problems. As well as the usual ones such as multicollinearity, serial correlation and simultaneity, the civil aviation market has its own characteristics which bring with it estimating problems. Special attention is paid to short- and long-run responses to income and price changes, to means of explaining secular trends in growth, and causes of fluctuations. When possible, generalisations concerning the size of parameters and the appropriateness of certain model forms will be made. The approach however is to treat routes and route groups as unique, with different passenger characteristics, development and histories. Such differences require different approaches and different models. Much of the discussion will be concerned with discussing the appropriateness of one or other model to routes of certain characteristics.

At the beginning of manned flight, aircraft needed some length of ground run in order to take off and fly. The Wrights took under 60 ft, and with a wing loading of 1.47 lb/sq ft and a CL max of approximately 0.9 they achieved first manned flight.

Time went on, and man grew more familiar with the hazards of increased take-off length and take-off speed. His growing knowledge of aerodynamics allowed him to increase the CL max values on wings. These facts combined to allow him to use increased wing loadings to improve his flight efficiency in terms of payload, cruise speed and range. The following graphs show the growth of these characteristics.

The appearance of the Kármán vortex street in the wake of bluff bodies has been well-known for sixty years. But the simple question of whether it appears also in the wakes of two cylinders in tandem at the same speed range, seems to have been unanswered as yet. The main reason for this is presumably the lack of practical interest. The only exception was in the early years of aviation in the era of biplanes. The use of streamlined or cylindrical struts was common and stimulated the investigations of Panell et al 1915, Wong & Yu 1917, Frazer & Simmons 1917, Biermann & Herrnstein 1933, and others.

Demand forecasting is a hazardous business, particularly in civil aviation. Any technique which can improve the accuracy of forecasts will aid the development of the industry, for demand fluctuations are a prime source of aviation's instability. If these fluctuations could be predicted accurately the consequences of wide demand fluctuations could be minimised, but this rarely happens. Many of the causes of the fluctuations are outside the control of the forecaster. Otherwise, ignorance of the relationship between cause and effect is the source of inaccurate predictions. An approach to forecasting, which is designed to overcome the latter, is the application of econometric models. Many of the econometric models in civil aviation are still primitive, yet, much has been expected of them.