In several recent reports reference is made to spanwise variations in the boundary layers growing in the working sections of various wind tunnels. Systematic investigations by Patel, Bradshaw and de Bray are in general agreement that these variations are caused by the screens which are fitted into the settling chambers of these tunnels to dampen out non-uniformities in the upstream flow. They conclude from their experiments that these variations will be much reduced if the downstream screen (or preferably the two downstream screens) has an open area ratio greater than about 0·57 and that these should be followed by a honeycomb to obtain the most steady and uniform flow, although a honeycomb in this position may increase the turbulence level in the working section.

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.