How Did It Become Obvious That An Airplane Should Be Inherently Stable?
The rise of an engineering truth from a quagmire of human needs, experience, and technological compromise
When I was a graduate student in aeronautical engineering, in the late 1930s, my professors told me categorically that a properly designed airplane should be inherently stable aerodynamically. That is, the airplane, if disturbed from equilibrium by a transitory occurrence such as a gust, should return to that condition without any corrective action by the pilot—even if he has let go of the control stick. The need for such built-in stability was simply taken for granted.
Now when you stop to think about it—as we students were not invited to do—the need for inherent stability in a practical vehicle is not at all obvious. A ship, of course, needs to be stable in roll if it is not to capsize, with possible dire consequences to pilot and passengers. On the other hand, a bicycle, though inherently unstable, can be ridden with ease by an average child, who continually applies the necessary corrective control actions without even being conscious of doing so. In fact, taking steps to make the bicycle stable—for instance, by adding so-called training wheels at the rear—so compromises the vehicle’s maneuverability and control that the extra wheels are discarded as soon as they have served their instructional purpose. The more stable a vehicle is— the more it tends to maintain a consistent pattern of behavior—the less controllable it is, for the harder it will be to change its behavior. As with the boat and the bicycle, so with the airplane, human beings had to learn somewhere along the way whether or not it was necessary, or even advisable, to make the vehicle inherently stable.
The story of how that happened is the tale of the growth of a fundamental piece of technological knowledge. What began as an ill-defined engineering problem involving a subjective human element—the pilot’s feeling of ease or difficulty in flying—evolved into a crucial, objective, well-defined concept for the airplane designer. The tale is complex and difficult to trace, for the decision was made over more than a century, by gradual, mutual consent of the aeronautical community as a whole and with a minimum of explicit, recorded deliberation. But it’s worth painting a picture of what happened, if only with a broad brush, because it vividly illustrates how technology can proceed as an almost inchoate group activity, and it gives a shining example of how far different engineering knowledge, and technological learning, can be from pure or even applied science. It’s about invention not as the brainstorm of the genius or the teamwork of the laboratory but as the uncentralized work of the great many people in a field groping toward a consensus about how their kind of machine can best serve almost undefinable human needs and preferences.
Until more than a decade after the Wright brothers’ first powered flight, two philosophies—prostability and antistability—divided the aeronautical community. The historian Charles Gibbs-Smith has nicely characterized the adherents of the two opposing schools as the chauffeurs and the airmen. The chauffeurs conceived early on that the airplane should be extremely stable, a kind of winged automobile that simply required steering. This view grew up gradually in Europe over a century, beginning with the attempted flights of Sir George Cayley in England in the early nineteenth century and continuing with the studies and experiments of Alphonse Pénaud in France in the 1870s and Frederick Lanchester in England in the 1890s. These men carried out their experiments almost entirely with unpiloted models, which clearly cannot fly successfully without inherent stability.
Well aware of the necessity for inherent stability in unpiloted experimental aircraft or gliders, Europeans as a whole apparently found it easy to assume that a high degree of stability would be essential even when a pilot was aboard. Their preoccupation with stability, and consequent blindness to crucial problems of control, proved counterproductive in early European attempts at piloted flight.
The airmen, by contrast, focused primarily on control. Realizing that an ability to maneuver accurately under a myriad of possible conditions would be essential for safe human flight, they sought initially to ride the air in piloted gliders that they could closely control. And they understood, or at least sensed, that stability and control work at cross-purposes: that the more inherent stability an airplane has, the more difficult it is to provide the essential control, to change the plane’s movement. The preeminent early airmen were Otto Lilienthal in Germany in the 1890s and the Wright brothers in America.
Recent engineering analyses have confirmed that the airplane in which the Wrights achieved powered flight in 1903 was, in fact, unstable both longitudinally and laterally. Its lateral, or sideways, instability was apparently largely deliberate, conceptually linked to the sideways instability of the bicycle, with which the Wrights were well familiar [see “How the Bicycle Took Wing,” by Tom D. Crouch, Invention & Technology , Summer 1986]. The longitudinal instability, which has no counterpart in the bicycle, was apparently accidental; evidently the Wrights did not understand longitudinal stability as we do today and simply could not formulate precise ideas about it.
Whatever the details of their knowledge, the Wrights’ ability to control their airplane laterally by means of coordinated wing warping and rudder movement—an ability enhanced, if not exactly made easier, by that lateral instability—was crucial for the attainment of practical human flight. Moreover, without the extensive mannedflight experience to which the Wrights’ unstable airplane opened the way, the question about inherent stability could hardly have been settled at all. As the historian Gibbs-Smith put it, “Until the pilot has experienced and can anticipate and control the behaviour of his machine in the air, he cannot decide what is necessary (or possible) to leave to automatic mechanisms and built-in qualities.”
Following Wilbur Wright’s epochmaking demonstration of controlled flight in France in 1908, the long and difficult melding of stability and control—and with it the assessment of whether inherent stability was desirable—began in earnest. The trend in the direction of positive stability, however, did not proceed very fast. Since the ideas behind stability and control gradually became understood by just a few theoretically minded people, they could hardly be common knowledge among practical designers. Moreover, flight-test criteria and wind-tunnel methods to measure the stability of a given airplane were still sorely lacking. We now know, from flights of restored aircraft of that period and replicas constructed for motion pictures, that at least several well-known airplanes from the beginning of the 1910s were highly unstable. Their pilots, however, knew how to control them. The situation must have been highly confusing, with both designers and pilots feeling their way dangerously into the unknown.
The question began to get sorted out as the 1910s proceeded, though the learning process was a slow one. In 1911 a British applied mathematician, G. H. Bryan, gave the first rigorous, comprehensive analysis of the dynamic response of an airplane to a departure from equilibrium. His work added valuable knowledge, but it was too sophisticated mathematically to be of much use to the practical designers of the time. Its most important effect was to inspire and guide the first windtunnel experiments on airplane stability. These began at the National Physical Laboratory in Britain in 1912 and 1913 and at the Massachusetts Institute of Technology in 1915 and 1916. The MlT investigation revealed that two popular configurations, including the Curtiss JN-2, forerunner of the famous JN-4 “Jenny,” were quite unstable over a portion of their speed range. Both studies attracted wide attention and gave the aerodynamic community valuable knowledge about stability in general.
Studies like these could reveal the level of stability or instability in a given design, but they couldn’t indicate what level was desirable. Without coordinated flight tests and meaningful pilot opinion, the investigators could only recommend what seemed generally sensible on theoretical grounds. Leonard Bairstow and B. Melvill Jones, at the National Physical Laboratory, wrote that having “a machine flying in still air [that] should tend continually to depart from the condition of normal flight, and require the constant attention of the pilot to bring it back to the correct attitude, will hardly be urged as a characteristic in general desirable.” They added that “the view now generally held is that it is necessary that the machine should be stable, but not too stable.” This was the view from both sides of the Atlantic, repeated in technical articles and books in the late 1910s and reiterated at greater length in the Annual Report of the federal National Advisory Committee for Aeronautics (NACA) in 1918.
Even in 1918, however, and despite the intensive development of aircraft in World War I, the practical situation did not always correspond to the “view now generally held.” As one modernday engineer has put it, “The perplexing fact, for researchers, was that when airplanes encountered the instabilities directly forecast by the mathematicians, the airplanes got along quite well and flew anyway.” Like the machines of the early 1910s already mentioned, a number of successful aircraft from the war—the Sopwith Camel in Britain and the Jenny and the ThomasMorse S-4C in the United States—have recently been found to have been longitudinally and laterally unstable.
The words of Brian Lecomber, a pilot who tested a restored Camel in 1978, are especially illuminating: “The Camel is mildly unstable in pitch and considerably unstable in yaw, and both elevator and rudder are extremely light and sensitive. … The ailerons, on the other hand, are in direct and quite awe-inspiring contrast. … All this results in an aircraft which, initially, feels horribly wrong to a present-day aeroplane driver. … Once the initial shock has worn off [however,] the machine becomes not so much difficult to fly as merely different.”
The similarly unstable Jenny was used as a trainer to teach thousands of pilots to fly. Clearly, whether an airplane should be inherently stable is, within limits, a matter of preference, not necessity.
Of course, the instability of at least some aircraft may have been the result less of designers’ intentions than of their ignorance about how to achieve stable designs; conclusive evidence is hard to find. But instability may have reflected remaining uncertainty about whether stability was really desirable. Jerome Hunsaker, of MIT, reporting in 1916 on the wind-tunnel research there, moderated his recommendation for stability with the observation that “it is well known that the French monoplane pilots demanded at one time a neutral aeroplane with no stability whatever against pitching, on the ground that ‘stable’ aeroplanes were too violent in their motion in gusty air.” He concluded that “safety in flight may well depend more upon ease of control than upon stability. The almost universal prejudice among accomplished fliers against so-called ‘stable aeroplanes’ appears to have a rational foundation.”
Considerable confusion remained at the end of the war. In the overall aeronautical community—as against the smaller circle of research workers and academics—the jury on stability, though deliberating, was still out.
The verdict was returned unanimously between 1920 and 1935. The final shift is hard to trace. In the early twenties the issue was still unsettled. Typically, D. R. Husted, an engineer at the Curtiss company, advised in Aviation magazine in 1920 that an airplane should have “sufficient pitching stability”—but didn’t specify what sufficient was. The following year the research engineer Frederick Norton and the test pilot Eddie Alien, both of the new NACA lab at Langley Field, Virginia, voiced a different opinion. They had been engaged for two years in the most extensive flight tests of stability and control to that time. From this considerable experience they stated that “generally speaking, a pilot does not know a stable from an unstable machine, and if the forces on the controls are small he is just as well satisfied with the unstable one as the other. … The beginner learns to fly as quickly in an unstable machine as in a stable one.” Boris Korvin-Kroukovsky, a noted design engineer, wrote similarly in 1925 that while “it is true that an unstable machine will not fly very long with hands off … even in stable machines the pilots do not leave the controls for long, simply because the stick is the best place to keep the hands on if not for any other reason.”
After the mid-1920s, statements doubting the need for stability disappeared, and almost all published opinions agreed that planes must be at least somewhat stable. Aeronautical engineering textbooks from the end of the 1920s contain statements saying flatly that “instability in any sense is useless and unsafe” and that “with the technical knowledge now available … there is no excuse for designing a machine of longitudinal instability.”
By the mid-1930s instability had become almost unthinkable. Kelly Johnson, who would become one of the most famous American airplane designers, put the matter clearly in a 1935 article reporting on wind-tunnel tests of the projected Lockheed Electra: “The reasons why an airplane must be stable are more or less obvious. A moderately unstable airplane may be flown safely if the pilot continually manipulates the controls. However, this procedure is very nerve-wracking and tiring. Landing or taking off is dangerous in an unstable ship. In flying blind, it is a great comfort to know that if the controls are released, the ship will continue on its course safely even in rough air.”
Thus what in 1910 had been confused, and in 1920 at least arguable, had by 1935 become “obvious.”
The historical reason for this shift is implicit in Johnson’s statement: piloting tasks had changed drastically with the passing years. Throughout the 1920s the increasing range and duration of flights had increased pilots’ exposure to fatigue when they were required continually to manipulate the controls. Higher landing speeds, especially with the introduction of more streamlined aircraft after 1930, made landing more dangerous, and improvements in cockpit instrumentation made blind flying more common. At the same time, pilots were being required to give greater attention to peripheral duties, such as radio operation, that required them to remove their hands from the controls. And the experiences of pilots like Jimmy Doolittle in the air races of the early 1930s showed the disastrous effects instabilities could have in the high-performance racing planes of the time.
In general, in more ways than one and from a variety of causes, the art of flying had grown up and become more demanding since World War I. How the various influences became translated into a status quo reflected in books and articles is not the kind of thing that appears in any record, but it had to involve pilots, designers, research engineers, and academics all interacting. However it took place, it reflected a complex learning process by a considerable engineering community.
Thus there was a true consensus by the 1930s. Some inherent aerodynamic stability was essential, but not too much, and the degree of it would be different for airplanes with different requirements of maneuverability, such as for fighters as against transports. This was the accepted canon taught to me as a graduate student.
The foregoing history reveals a lot about how concepts in engineering evolve. To begin with, it shows that even ideas that we think are obvious had to be learned somehow and have a history. Second, it shows that the lengthy learning process that generated these engineering ideas was a community activity. Nobody dominated the story; the protagonist is the entire sphere of people actively concerned with the problems of stability and control. One can imagine the story with a different cast of characters; one cannot imagine it without a stability-and-control community.
That stability-and-control community subsumed at least four identifiable groups: airplane designers, research aerodynamicists, pilots, and flight-instrumentation specialists. Knowledge was generated—as it had to be—simultaneously and interactively among them. The aeronautical historian Edward W. Constant II has described the “community structure of technological practice.” The growth of a consensus about stability took place within just such a structure.
As the social scientist Herbert Simon observes in his book The Sciences of the Artificial , whereas the natural sciences are concerned with how things are, engineering design, like all design, is concerned with how things ought to be —that is, with devising artifacts to meet required goals. The knowledge that, all things considered, an airplane is best made stable but not too much so can hardly be classified as scientific. It is in no sense knowledge about how an airplane innately is; rather, it is knowledge about how an airplane ought to be to enable the pilot to perform his flying tasks with ease, confidence, and precision. It is an instance par excellence of engineering, as opposed to scientific, knowledge.
Similarly, the process by which the knowledge was achieved was characteristically an engineering activity. The conclusion that airplanes ought to be inherently stable was reached more or less instinctively by an engineering community as a whole. It was arrived at gradually by relating the growing objective knowledge of stability and control to the likewise growing body of subjective piloting experience. The outcome was a balance, a trade-off between the conflicting requirements of control and stability, a kind of tradeoff engineers often find necessary. The need for a trade-off here, however, did not arise from economic constraints; neither did it derive from purely theoretical requirements. It came into being because of the practical needs and limitations of the human pilot. The balance, therefore, could not have been achieved on purely intellectual grounds or without extensive flight expenence. It summarized a practical judgment (based largely on subjective opinion) of a sort that cannot be avoided in engineering. Achieving it did involve some activity that can properly be regarded as science—for example, the analysis of dynamic response by applied mathematicians. But it clearly, predominantly, required a great deal more. That “great deal more” provides evidence, if any were still needed, of the fundamental difference between applied science and engineering.
There is a postscript to this story. Once the need for inherent stability in aircraft became obvious, it remained so for decades. Today it no longer is generally obvious. Fighter planes are being designed and built to be deliberately unstable aerodynamically in order to obtain the most rapid possible response to the movement of their controls. However, modern control technology makes possible automatic motion of the control surfaces of these planes, and that makes the craft artificially stable. So the planes, while unstable, still feel stable to their pilots. Perhaps that is the final compromise.