The whirling arm provided most of the systematic aerodynamic data gathered up to the end of the nineteenth century. Its flaws, however, did not go unnoticed. Test results were adversely influenced as the arm's eggbeater action set all the air in the vicinity in rotary motion. Aircraft models on the end of an arm in effect flew into their own wakes. With so much turbulence, experimenters could not determine the true relative velocity between the model and air. Furthermore, it was extremely difficult to mount instruments and measure the small forces exerted on the model when it was spinning at high speeds. Something better was needed.
That something better was a "wind tunnel.'' This utterly simple device consists of an enclosed passage through which air is driven by a fan or any appropriate drive system. The heart of the wind tunnel is the test section, in which a scale model is supported in a carefully controlled airstream, which produces a flow of air about the model, duplicating that of the full-scale aircraft. The aerodynamic characteristics of the model and its flow field are directly measured by appropriate balances and test instrumentation. The wind tunnel, although it appears in myriad forms, always retains the five identifying elements italicized above. The wind tunnel's great capacity for controlled, systematic testing quickly rendered the whirling arm obsolete.
The unique role and capabilities of a wind tunnel can best be appreciated by recognizing the aerodynamic  forces and moments acting on an aircraft in flight. The three basic forces are lift, drag, and side force as measured in an axis system referenced to the direction of flight of the aircraft. The drag force is along (but reversed to) the flight path; the lift and side forces are at right angles to it. In a wind tunnel the axial centerline of the test section defines the direction of the oncoming wind-the aerodynamic equivalent of the flight path. The ease of measuring aerodynamic forces relative to the tunnel axis on a model held stationary in the airstream opened a new era in aerodynamic experimentation.
Frank H. Wenham (1824-1908), a Council Member of the Aeronautical Society of Great Britain, is generally credited with designing and operating the first wind tunnel in 1871. Wenham had tried a whirling arm, but his unhappy experiences impelled him to urge the Council to raise funds to build a wind tunnel. In Wenham's words, it "had a trunk 12 feet long and 18 inches square, to direct the current horizontally, and in parallel course.'' A fan-blower upstream of the model, driven by a steam engine, propelled air down the tube to the model.
Wenham mounted various shapes in the tunnel measuring the lift and drag forces created by the air rushing by. For such a simple experiment, the results were of great significance to aeronautics. Wenham and his colleagues were astounded to find that, at low angles of incidence, the lift-to-drag ratios of test surfaces could be surprisingly high-roughly 5 at a 15 degree angle of attack. Newton's aerodynamic theories were much less optimistic. With such high lift-to-drag ratios, wings could support substantial loads, making powered flight seem much more attainable than previously thought possible. These researches also revealed the effect of what is now called aspect ratio: long, narrow wings, like those on modern gliders, provided much more lift than stubby wings with the same areas. The wind tunnel idea was already paying off handsomely.
With the advent of the wind tunnel, aerodynamicists finally began to understand the factors that controlled lift and drag, but they were still nagged by the question of model scale. Can the experimental results obtained with a one-tenth scale model be applied to the real, full-sized aircraft? Almost all wind tunnel tests were and still are performed with scale models because wind tunnels capable of handling full-sized aircraft are simply too expensive.
In a classic set of experiments, Osborne Reynolds (1842-1912) of the University of Manchester demonstraded that the airflow pattern over a scale model would he the same for the full-scale vehicle if a certain flow parameter were the same in both cases. This factor, now known as the Reynolds number, is a basic parameter in the description of all fluid-flow situations, including the shapes of flow patterns, the ease of heat transfer, and the onset of turbulence.