Chapter 2: The High-Speed Airfoil Program
[36] The emergence of the transonic research airplanes in the mid-forties (Chapter III) greatly heightened our interest in the supercritical behavior of airfoils and in developing testing techniques for exploring the supercritical and transonic regions. But an even stronger motivation had developed concurrently with the advent of the turbojet engine. In 1944, the Army had sent an XP-59A, the first U.S. jet-powered airplane, for flight demonstrations at Langley. Standing beside the main runway and watching this airplane fly by at nearly 400 mph, we sensed for the first time that here was the key to transonic and supersonic flight, a practical new propulsion concept capable of the enormous power required to penetrate the transonic region. The so-called "sound barrier," which had been almost universally thought of as a set of adverse aerodynamic problems, in reality also involved a fundamental limitation of the piston engine due to its fixed "displacement," or capacity to inhale air for combustion. Since the displacement was independent of airspeed, no significant increase in peak power could occur as the flight speed increased. Thus, there had been no realistic hope that piston engines could be developed in the sizes that would have been needed for transonic flight; the transonic "barrier" was actually as much a piston engine barrier as an aerodynamic barrier. The jet engine on the other hand ingests a volume flow of air that increases as the flight speed [37] increases, permitting a continuous increase in power in contrast to the fixed power of the piston engine. This power increase is significantly augmented at high speeds by the "ram" pressures of the air which provides supercharging and improves the cycle efficiency.
We understood the principles and enormous potential of the turbojet only vaguely at the time of the XP-59A demonstration. Very little data were available to us on the details and performance of the G.E. I-16 engine. Several of us spent the next few days in exciting speculations of possible jet-engine thermodynamic cycles, airflow characteristics, and crude performance estimates, which gave us a better understanding. K. F. Rubert, who had taught internal combustion engines at Cornell, undertook a more careful systematic analysis, published in 1945 in a paper which I reviewed as chairman of the Langley Editorial Committee (ref. 44). (Periodic editorial duties of this kind were of great value as a means of education and stimulation of all involved-in addition to their obvious direct benefit to the quality and accuracy of Langley reports.)
By now our limited goal of the 550-mph subcritical airplane of the mid-thirties had become meaningless and we could foresee the imminent achievement of supersonic flight. Few doubted that operational supersonic military aircraft would soon follow the research airplanes. The need to acquire accurate supercritical and transonic aerodynamic data had become acute, and Langley researchers responded to the challenge with considerable inventiveness. Eight innovative techniques were eventually devised and explored in various forms by NACA, ending with general acceptance of the semi-open tunnel for two-dimensioned airfoil testing up to Mach 1, and the slotted transonic tunnel for wing and aircraft configuration testing as the most satisfactory devices. (These developments are listed in the Appendix and discussed in detail in Chapter III.)
Unknown to us, the Italians had already succeeded in obtaining airfoil force data through the supercritical range up to about Mach 0.94, and the Germans to about 0.92. We first learned of this in 1944-before any of the new Langley schemes materialized-upon the arrival of Antonio Ferri, formerly of the Italian aeronautical laboratory at Guidonia, and recently an Italian Partisan in the war. Ferri brought with him extensive airfoil data from their tests in a semi-open high-speed tunnel in the early forties. He completed analysis of the data at Langley [38] and we published the results in a NACA wartime report (ref. 45). His English at that time was negligible, and I wrote the final text after much heated consultation with Ferri and help from Lou Nucci who acted as interpreter. (Major confusions arose from Ferri's pronounciation of "subsonic" and "supersonic," both of which sounded to me like "soupsonic.") The proportions of Ferri's tunnel (1.31 feet across the open top and bottom and 1.74 feet on the closed sides) corresponded to 43 percent of the perimeter being open. This closely approached the value of 46 percent suggested by a theoretical analysis of Wieselberger in Germany (ref. 46) as the correct proportion for zero "blockage" (zero axial-velocity correction) applicable to a three-dimensional test model. However, this large degree of "openness" had a serious drawback in the large pulsations which occurred at high speeds. None of us was quite sure of the validity of the semi-open tunnel technique at that time.
Despite the questions of technique, Ferri's data revealed a most important new finding: the loss in lift associated with the compressibility burble did not persist indefinitely. At about Mach 0.9 a marked recovery in lift occurred, suggesting that the separated ("shock-stalled") flow tended to disappear as Mach 1 was approached. Later that year support for this result was indicated in tests of small wings by means of the "wing-flow" technique (ref. 47). In 1946 we obtained German airfoil data from their large 2.7-meter closed-throat tunnel (ref. 48) which provided further verification at speeds up to Mach 0.92. And early in 1947 the first airfoil pressure distributions ever obtained at Mach 1 were successfully measured in our rotating-disc annular transonic tunnel (ref. 49). These showed conclusively that at Mach 1 the shocks had moved to the trailing edge and the flow was supersonic about the entire section except for a small region at the blunt leading edge. The German tests had included a systematic study of the effects of airfoil camber at high speeds which clearly showed that conventional positive camber was undesirable for Mach numbers greater than 0.75, and in fact best lift-drag (1/d) ratio was obtained with negative camber at supercritical speeds, a result with which Ferri's data agreed.
I had become involved in a study of all available transonic data in 1947 in connection with writing a chapter on "Transonic Aerodynamics" for a prospective aeronautical handbook (ref. 50). At Stack's suggestion [39] I discussed my airfoil material at a meeting of the Langley General Aerodynamics Committee on January 16, 1948. This was the first time that many of the members had seen the German results and the general agreement of all of the new data as regards flow phenomena and trends of airfoil performance at supercritical speeds approaching Mach 1 (ref. 51). Dick Whitcomb was an interested participant at this meeting. In commenting on the effects of camber at supercritical speeds Whitcomb suggested that upper surface curvature might be the important parameter and that the use of "proper" curvature might reduce the upper-surface shock strength and tendency of the flow to separate (ref. 51). Some 16 years later he would resurrect this idea and apply it successfully in the "supercritical" airfoil (see pp. 55ff.).
A few months after this meeting I presented the unclassified parts of my summary material at a NACA University Conference (ref. 52). Airfoil shock and separation patterns inferred from the available force and pressure data (refs. 47, 49) throughout the transonic zone were illustrated (fig. 3). The points brought out in the discussion included:
Ferri's successful use of the semi-open tunnel, together with encouraging results of Langley studies of this configuration by Donaldson and Wright (ref. 53) and Lindsey and Bates (ref. 54) led to our decision in the fall of 1947 to convert the 4 x 18-inch high-speed tunnel to the semi-open arrangement with the object of systematic airfoil testing at Mach numbers up to 0.95, and higher if possible. The first results, obtained in 1948, showed that the tunnel (now 4 x 19 inches in size) [40] could be operated with 4-inch chord models at a nominal Mach number (if 1.0, but it was not immediately certain that the sonic results were valid. This tunnel was ideally proportioned for schlieren photography, and from the start impressive photographs were obtained which provided the first visual proof that our speculations about the flows at Mach 1 based on force and pressure data were correct. Figure 4, constructed from photographs taken in 1949, contains typical results from this program. The top row of photographs, for Mach 1, are of particular interest, showing that the shocks lie downstream behind the trailing edge. The flow on the airfoil is virtually separation-free and entirely supersonic in character except for a small subsonic region near the leading edge. Extensive systematic pictures of this kind for other sections were obtained in the 4 x 19-inch tunnel by B. N. Daley and R. S. Dick and published later (ref. 55). Similar flow pictures were also obtained....

drawing indicating subsonic and supersonic regions of an airfoil
FIGURE 3-Airfoil flow patterns at transonic speeds discussed at NACA University Conference, 1948.

41] FIGURE 4.-Transonic flows and pressure distributions, Mach. 0.79 to 1.00. Angle of attack, 3.2 deg. From the 4 x 19-Inch Semi-Open High-Speed Tunnel, 1949.

[42] the British some years Langley work and by were used by W. S. Farren in his Wilbur Wright Memorial Lecture of 1955 (ref. 14).
It is most important to note that our large burst of understanding about airfoil behavior beyond shock stall was acquired in the 1945-1947 time period, several years before any data on this problem were obtained from the research airplanes. My summary airfoil paper (ref. 52) was prepared in the spring of 1948, before the X-1 pressure data had been obtained. When I first saw the X-1 pressure data for the 10 percent thick wing about a year later the fact that it confirmed what was already known from the wind tunnels was satisfying but not at all surprising. Nevertheless the wing pressure distributions obtained in flight on the X-1 were of very great value because they provided the ultimate indisputable basis for judging the relative merits of the various ground facilities. The basic airfoil used on the number 2 airplane was the NACA 65-110, and both the Annular Transonic Tunnel and the 4 x 19-Inch Semi-Open High-Speed Tunnel programs had scheduled this section for their initial tests in anticipation of critically important comparisons. (A minor flaw in the plan was discovered after the tests had been made; in building the airplane the slight cusp in the basic 65-110 section had been removed for structural reasons, and this caused a minor change in the flight pressures just ahead of the trailing edge.) Figure 5 compares the X-1 flight data with the results from the two transonic facilities at Mach 1. Agreement with the 4 x 19-inch tunnel was considered excellent. The annular transonic tunnel data, although showing the generally correct shape, indicate pressures consistently too high. This same type of discrepancy was noted in subsequent tests of other sections and was never satisfactorily explained. Of the several transonic techniques only the 4 x 19-inch semi-open tunnel remained active in airfoil testing throughout the concluding years of the NACA program.
An early airing of our new knowledge of airfoil behavior near Mach 1 was made by Daley and Habel at the NACA Transonic Airplane Design Conference of September 1949 (ref. 56). During preparations for this meeting both the 4 x 19-inch tunnel data and the X-1 data were so new that Daley balked at presenting them without more time for analysis, but he finally yielded to management pressure. No conclusions were drawn, however, concerning the relative merits of the test techniques.

graphed wind tunel results for the X-1
[43] FIGURE 5-The X-1 pressure distributions compared with those of the Annular Transonic and the 4 x 19-Inch tunnels. Mach 1 .0, NACA 65-110 airfoil, c. = 0.41.

The discovery that airfoil flows beyond about Mach 0.95 did not suffer from significant viscous separation effects lent new encouragement to the theorists. It had been previously believed almost universally that sonic flows in real gases would be characterized by large viscous separation effects, so that any theoretical treatment, based on the usual ideal gas assumptions, would have little realism. Thus the main theoretical concentration up to the mid-forties was on refinement of subcritical compressible-flow calculations. While this was consistent with the original belief that practical aircraft would not be able to operate much above the critical speed, in retrospect it is apparent that these efforts were not [44] very profitable. The simple approximations developed early in this work were adequate for most engineering purposes, although minor refinements were laboriously attained later.
By 1947, however, an increasing number of theorists, encouraged by the new experimental findings, started to tackle the transonic problem in a variety of new ways. The development of the transonic similarity laws was a useful aid in data correlations, although these laws, of course, provided no solutions for any flow problems. Major progress came when the special case of the wedge airfoil at zero angle of attack at Mach 1 was solved by Guderley and Yoshihara in 1948 (ref. 57). I was privileged to see this accomplishment before its publication when Guderley visited Langley to discuss the work with A. Busemann, who had been assigned to the Compressibility Research Division after he had been brought to this country under the auspices of a Navy postwar program. For some years previous Busemann had consulted with Guderley on this problem and had contributed suggestions for its solution. The initial solution was for a cusped wedge shape, but this was followed shortly by similar results for a symmetrical double wedge (ref. 58). These results were very important to us; at long last we had theoretical sonic pressure distributions against which the experimental data from our new test techniques could be evaluated. These assessments would constitute a valuable supplement to the X-1 data as a means of insuring the validity of the experiments. I enthusiastically arranged for tests of the wedges in both the Annular Transonic Tunnel and the 4 x 19-Inch Semi-Open High-Speed Tunnel (ref. 59). The gratifying results (fig. 6) were presented for the first time at the September 1949 conference (ref. 56). A photograph of the flow about the wedge at Mach 1 confirmed the absence of any significant viscous separation effects except for a very small bubble just downstream of the sharp crest.
An important feature of the Guderley flow field was a region of smooth shockless deceleration of the local flow downstream of the crest of the wedge, caused by reflections of the expansion waves from the curved sonic line extending upward from the crest. The reflections from this free boundary were compression waves which decelerated the flow in smooth reversible fashion. Previously, for conventional airfoil shapes at low supercritical speeds, no such shockless compressions had been [45] identified and it was thought that shocks were the inevitable device employed by nature to return the flow of stream velocity.
We now know that a considerable degree of smooth recompression prior to the terminal shock can occur for a wide variety of airfoil shapes near Mach 1. Actually this could be seen in the conventional airfoil pressure distributions obtained in the late forties at Mach 1 (see fig. 4, top row, for example). A more direct indication of the effect can be seen in fig. 7 which shows a Mach 1 pressure distribution obtained in 1948 in the Annular Transonic Tunnel. By comparison with supersonic expansion theory the measured pressures over the rearward portion of the airfoil were unaccountably high, and not understanding the possibility of the recompression effect we speculated that boundary layer growth might be the cause. The effect is actually primarily recompression due to reflections from the sonic fine and secondarily the boundary layer contribution. Theoretical treatment revealing that all conventional....

chart illustrating the Guderley field flow
FIGURE 6. Guderley theory for Mach 1 compared with Langley transonictunnel data.

graph of pressure distribution on upper and lower  air foil surfaces
[46] FIGURE 7.-Pressure distribution obtained in the Langley Annular Transonic Tunnel at Mach 1. NACA 66-006 airfoil at zero angle of attack.
.....sections experience this effect near Mach 1 came about 10 years later in 1959 (ref. 60). Of great interest here is the implication that smooth recompression can in some circumstances also play a major role at speeds well below Mach 1 in the achievement of improved supercritical airfoils, accomplishing the benefit suggested by von Karman in 1941 (ref. 61).
One of Busemann's first projects after his arrival was to summarize the theoretical possibilities for treating transonic flows, starting at Mach 1 and extending upward in speed through the detached shock region (ref. 62). Applying these methods, Vincenti and Wagoner extended the flow field calculations for the wedge to low supersonic speeds with detached bow waves, showing that the transition to pure supersonic flow with attached shock was a stable, orderly process (ref. 63). These results tended to support the conclusion we had already come to from [47] the experimental work, that there was little need for systematic experimental airfoil research in the supersonic part of the transonic region. We believed that such wing investigations as would routinely be made in the course of configuration development in the slotted tunnels and supersonic tunnels would be sufficient, and later experience proved this assessment correct.
The development of airfoils with improved supercritical characteristics was a major thrust of the 1945-1955 decade. Nearly everyone working in this field naturally thought of the possibilities of achieving a "delayed compressibility burble" Stack in 1938 (ref. 28) and von Karman in 1941 (ref. 61) specifically discussed this possibility. The term "supercritical" in its broadest sense means any speed beyond the critical Mach number, but as used by most of us in that period it meant speeds greater than the force-break speeds and extending upward into the sonic or low supersonic region. In recent years Whitcomb has introduced a more restrictive meaning: his "supercritical" airfoil is designed to delay the drag rise and thus the term refers to airfoil operation in the speed region between critical Mach number and drag-rise Mach number.
In a sense, the "dive-recovery" flaps developed for the P-38 were the first attempt to obtain an airfoil with improved supercritical performance (ref. 20). Throughout the forties, the tendency of diving aircraft to lock into a severe nose-heavy condition from which recovery was often difficult remained the principal problem for supercritical research. The buffeting which accompanied the lift loss in shock-stalled flows was a parallel concern. It had become generally accepted by the mid-forties that high critical Mach number was no index of good supercritical performance. There is little correlation between critical Mach number and force-break Mach number for a wide variety of sections. It was generally agreed that new criteria would have to be found for the design of airfoils with good supercritical performance. H. J. Allen came up with a fresh idea for minimizing the lift loss and moment changes at shock stall, which was tested with some success (ref. 64). He reasoned that if both upper and lower surface flows reached local sonic velocity at the same flight speed, a more equal separation would occur on each surface, leaving the net lift relatively unchanged. He and D. Graham developed an airfoil having an "M-shaped" camber line which achieved [48] a reasonable approximation of this type of flow. Unfortunately, it had high subcritical drag and was never used as far as I have been able to learn. Nevertheless, it was the first attempt to tailor a specific fixed-geometry airfoil for alleviation of the shock-stall lift loss; no mention was made of improvement in supercritical 1/d.
In the early forties when the P-38 was in trouble, I recall a conversation with Allen and Stack in which we agreed that conventional cambered airfoils showed improved supercritical lift and moment performance if operated inverted in the negative-lift attitude (i.e., with negative camber in the positive lifting sense). Negative camber meant a less curved upper surface which had reduced separation losses at shock stall. Allen dismissed this approach as being rather unthinkable and remarked facetiously that pilots would hardly accept inverted flight as a technique for pulling out of supercritical dives. None of us gave much thought to the supercritical lift-drag ratio at that time; I was certainly unaware that negative camber in addition to the lift-loss benefit resulted in better supercritical 1/d until I noticed that this was so in editing Ferri's airfoil report in 1945. I looked back at our own data and some 1945 Ames data (ref. 65) obtained in systematic airfoil tests in their 1 x 3.5-foot tunnel at speeds up to about Mach 0.85, and noted with interest that the supercritical 1/d was significantly better for negative camber. Figure 8 taken from the Ames data shows this result.
My past upbringing to the effect that positive camber was inherently beneficial and essential to conventional lifting airfoils at normal speeds was so deeply ingrained that I dismissed these results as an impractical aerodynamic curiosity. Three years later, in 1948, I included in my summary NACA Conference paper on high-speed airfoils (ref. 52) a plot based on the German airfoil data (ref. 66) which showed in detail how the camber for best l/d quickly diminished to negative values as the Mach number advanced beyond about 0.75 (fig. 9). Actually the data clearly showed that negative camber (dashed lines on fig. 9) gave best l/d at the higher speeds. But still believing that negative camber was unthinkable for practical applications, I terminated the plots at zero camber and suggested as a major conclusion that zero camber (symmetrical) airfoils were the best compromise for transonic applications (ref. 52). This interpretation was shared by the other airfoil.....

graph of lift drag vs mach
[49] FIGURE 8.-Improvement in supercritical lift-drag ratio (1/d) with negative camber. Ames data,1945.

....specialists at Langley at that time and also by Allen and Graham (ref. 64).
The physical mechanism by which the improvements due to negative camber came about was thought at that time to be related to the location of the peak suction pressure and shock near the leading edge for the "peaky" distribution of pressure that occurred over the relatively flat upper surface with negative camber. For this forward shock position, the boundary layer was thin and not as prone to separation as it was for the positively-cambered case where the shock occurred far aft on the curved afterbody where it triggered separation. We did not realize then that an additional mechanism was at work for the "peaky" case, namely some degree of shockless recompression due to reflections from [50] the sonic line. ("Peaky" is a term coined a decade later by Pearcey, who called attention to the recompression effect (ref. 67).)
The first attempt to derive "supercritical" airfoils in the restricted Whitcomb sense was made in 1951 by Woersching (ref. 68). He had studied all of the available negative camber data including the negative lift operation of the Allen/Graham "M" cambered NACA airfoil 847B-110 (ref. 64). He noticed that this airfoil in the inverted or negative lift attitude had a drag-rise Mach number of 0.81 at c1 = 42, while in the normal attitude the drag rise occurred at M= 0.73. He also examined the inverted airfoil in the region of shock stall and beyond and found it to be generally as good as or superior to the normal attitude. After further study, Woersching concluded, "Maximum drag rise Mach number is obtained with negative camber over the forward chordwise portion of the airfoil, and positive camber aft to the trailing edge-but at the expense of large negative moment coefficients." This, of course, is a qualitative description of the features of the Whitcomb "supercritical airfoil"-together with one of its special problems. Woersching goes on to advocate inclusion of the last arm of the "W" camber in order to relieve the pitching moment problem at some loss in drag-rise Mach number. He also visualized aircraft incorporating both sweep and the proposed sections, designed "for cruise near Mach 1." This work was undoubtedly the first serious attempt at delaying the drag rise-with a profile that would qualify as a "supercritical airfoil" in the present-day sense. By way of explanation of the action of negative camber Woersching pointed out that it results in a degree of flatness of the suction surface comparable to that of a much thinner symmetrical section. Unfortunately, he did not have the resources to continue development.
Probably inspired by the Woersching paper, Britisher W. F. Hilton published in 1953 a report (ref. 69) which he had written in 1947 on the advantages of negative camber. The original report had apparently been given only restricted circulation in Great Britain, perhaps for reasons of security. It is interesting to note that Hilton had been employed in the United States for several years following the war and had access to and personal interest in the available American, Italian, and German airfoil data. Hilton did not recommend any particular.....

graph shows how the camber for best l/d
   quickly diminished to negative values as the Mach number advanced
[51] FIGURE 9.-Effect of Mach number on camber for best lift-drag ratio. 12-percent-thick airfoils.

....distribution of negative camber. His aim was primarily to reduce the adverse lift and moment changes due to shock stall and secondarily to improve I/d beyond shock stall.
Without doubt the period from 1945 to 1951 was one of the most productive eras in the history of high-speed airfoil research. Several new transonic ground facilities and flight techniques were developed and applied successfully; reliable wind tunnel data at Mach numbers up to 1.0 were obtained, including airfoil flow photographs; new theoretical treatments of the flow were accomplished for wedge airfoils at Mach 1 and throughout the detached shock range; criteria were established for airfoils having delayed drag rise and an inverted NACA airfoil meeting these criteria was specified (the first "supercritical" airfoil in today's [52] parlance). NACA program activities were at the core of this progress, although there were also important outside contributions, especially on the theoretical side. On June 2, 1950, 1 reviewed this satisfying progress in considerable detail for the NACA Executive Committee, concluding, "The principal details of two-dimensional transonic flow are now known as a consequence of recent progress both experimental and theoretical.... Many problems remain for the three-dimensional case of complete wings. . . . Our 8-foot high-speed tunnel with its new slotted throat provides a transonic facility of adequate size for the needed work on complete wings."
John Stack's role in the high-speed airfoil developments of this period was quite different than his intimate personal participation during the first dozen years of his career. From about the time of his Wright Brothers lecture it had seemed likely that he would be moved into a management position in the Langley "front office." His special talents as a tough, persuasive technical salesman were badly needed and, furthermore, it was obvious that life would be much more pleasant for Langley management with Stack as a member of their office rather than as a combative division chief who increasingly was cast in an adversary relationship to higher management in regard to approvals and funding allocations for our projects. Thus in mid-1947 Stack became an assistant to Chief of Research F. L. Thompson, and I succeeded him as Chief of the Compressibility Division.
Although he remained invariably supportive of our projects, my relationship with Stack was inevitably changed. He was now one of "them" rather than a close colleague in research. His principal preoccupation became the promotion and development of major new transonic and supersonic tunnels, and he also became involved with other problems beyond our field of interest. He observed the airfoil developments with interest as they unfolded but had no direct part in them except through related facility developments-such as the Annular Transonic Tunnel, which might never have been successfully promoted without Stack's support. His early experiences with the open-throat tunnels made him rather suspicious of semi-open tunnels and this was reflected in his encouragement of studies of their transient disturbances by Lindsey and Bates (ref. 54). In contrast to Stack's many publications in his earlier [53] airfoil research period, the paper covering his review of facility developments (ref. 54) was his principal publication in the 1945-1951 period.
In the final period of the NACA program from 1951 through 1956 a rapid dwindling of the effort took place. This was due partly to the large shift in research emphasis to swept and low-aspect-ratio wings for supersonic aircraft, and partly to the fact that a substantial high-speed airfoil technology base had been established. The demand for two-dimensional airfoil research diminished to low levels except for the special area of helicopter blading. Ames experimental work in the field V4, had been terminated in 1951 when their 1 x 3.5-foot tunnel was phased out as a closed-throat facility. Finally, the abolishment of W. F. Lindsey's section at Langley in 1956 brought to a close the NACA high-speed airfoil program which had started 29 years before. Although several worthwhile projects were left unfinished, they could not compete in priority with the demands of supersonic aircraft and the burgeoning space program.
Curiously, the impressive progress in-high-speed airfoil technology in the last decade of the program is often overlooked. At a recent NASA airfoil conference (ref. 70) several practitioners in the current program seemed to believe that the NACA program had terminated with the 16-series sections and Stack's Wright Brothers Lecture of 1944 which, so to speak, left high-speed airfoils in the depths of the shock staff. The most likely explanation is that the researchers of 1945-1956 did not produce any specific new airfoil families. They produced important new understanding of transonic flows and they extended the accurate data for existing airfoil families to Mach 1, but unfortunately, perhaps, there were no associated clever baptisms or new acronyms to help publicize the progress that was made. This solid but unspectacular airfoil progress was overshadowed by the more dramatic events of that period-the first supersonic flight, the slotted transonic tunnel, and the area rule.
The high-speed airfoil program provides an excellent example of NACA accomplishing its mission in an important problem area of aeronautics. For the first 20 years, from the early twenties to the early [54] forties when the propeller was the primary application, the program provided both fundamental understanding of the flow phenomena and new airfoils to improve propeller performance at high speeds. These solutions were in hand well before they were needed by industry. Only for the brief period from about 1944 to 1947 was the program deficient in meeting the new needs for transonic data beyond Mach 0.85. During this interim, foreign data plus information from the "wing-flow" and "body-drop" techniques were used effectively. Early in 1947 the first airfoil/pressure data at Mach 1 were obtained (ref. 49) and by 1949 an effective semi-open wind tunnel was being used routinely for airfoil testing to Mach 1, the technique verified by X-1 flight data.
The program wound down rapidly in the mid-fifties, partly because there was no obvious need then to expand the technology beyond its already substantial proportions, but mainly because several of its talented researchers had been lured into more urgent and fascinating supersonic and space-related projects. Almost a decade would pass before the renaissance described in the next section would take place, based on the recurrence of an old need, but carried forward with fresh inspiration by a wholly new research team.
The report editing procedure mentioned in this chapter and elsewhere deserves comment. The primary technical editing was accomplished by an inter-divisional committee of the author's peers. This was followed by editing for grammar, availability of references, etc., by a female "English critic" in the editorial office. The generally superior reliability, clarity, and freedom from "governmentese" of the NACA reports produced by this system have been widely acclaimed. Unfortunately, however, most of them are rather dull from a literary point of view. The report-writing manual used to indoctrinate young NACA engineers emphasized accuracy, clarity, and adherence to the standard format, rather than any matters of style or technique to make the report interesting. Language which added humor or sparkle was frowned on and almost always deleted. Imaginative speculation was forbidden unless specifically identified as such. All of this was perhaps appropriate for simple reports intended to present reliable data in a readily usable form. But by the time NACA writers had progressed to more sophisticated subjects such as advanced concepts or state-of-the-art papers for a [55] national audience, most of us were so crippled by habitual adherence to the system that these writings also tended to be stereotyped and less interesting than they might have been.