Share This Article


‘Engineers sometimes suffer from an excess of zeal and crazed imagination that trumps common sense and conservative design. But who would have it any other way?’

An old aviation truism holds that if an airplane looks good, it’ll fly well, which is one reason you won’t find a handsome machine in this entire selection. These are our nominations for the 10 most difficult to fly aircraft ever built, but it’s a list that can be endlessly expanded. Gee­Bee R1/R2 hybrid, Pou du Ciel, Grumman XF-10F variable-geometry jet, Curtiss XP-55 “Ass-Ender,” the Lockheed Have Blue prototype that preceded the F-117 stealth fighter…more candidates are waiting in the wings, as it were, making it clear that aeronautical engineers sometimes suffer from an excess of zeal and crazed imagination that trumps common sense and conservative design. But who would have it any other way?


The stunningly advanced 1917 DE BRUYÈRE C 1 canard biplane was not difficult to fly; it was impossible to fly. The C 1 made one short flight, really just a takeoff, and crashed—inverted. Instead of going back to the drawing board, the French designer went so far away that de Bruyère’s first name seems to have been lost to history.

A shame, for the airplane was in fact amazingly ahead of its time. The pusher propeller was at the extreme rear of the advanced metal fuselage, turned through a long driveshaft by what may have been—nobody knows for sure—a 150-hp Hispano-Suiza V8 mounted midship, behind the pilot, P-39 Airacobra–fashion. The prop had clean air to work with, since there was no horizontal tail—just sleekly curved dorsal and ventral vertical fins. Pitch was controlled by the nose-mounted canard—hardly revolutionary, since the Wright Flyer was a tail-firster, but a rig like the de Bruyère’s wouldn’t come back into favor until the 1970s. Roll control was not by ailerons or even wing-warping but through pivoting wingtips, which may well have been the reason for the airplane’s aversion to upright flight. The windscreen was not the already traditional flat pane of glass but a curved, wind-cheating assembly that wouldn’t have been out of place on a 1930s Thompson Trophy racer. And the de Bruyère also had tricycle landing gear, with a pair of mains under the wings and an even bigger nosewheel half-recessed into the airplane’s aardvarky nose. With a few rare exceptions, nobody would see that configuration in serious use until the 1940s.

The de Bruyère’s nose was intended to be the heart of the airplane. It was configured to carry a big-bore 37mm Hotchkiss cannon, which required a propeller-free field of fire. Had it worked, the C 1 could have been the A-10 tank-killer of its day.


The 1919 TARRANT TABOR, a six-engine triplane that crashed fatally during its sole, never-to-be-repeated takeoff run, was a “What were they thinking?” design. It was originally a relatively conventional, albeit enormous, biplane bomber with four engines in two tractor/pusher nacelles between the wings, but its designer, Walter Barling (later of U.S. Barling Bomber fame), decided the Tabor needed more power. Rather than putting an additional engine outboard of the existing pair in each wing bay, he mounted the extra two atop the upper wing and added a third wing above them. What had been a relatively sleek and fairly conventional huge biplane became a mass of tree trunk–size struts, uncowled engines, box-kite tail surfaces and massive landing gear. It looked like a lumbering lumberyard. Which was perhaps appropriate, since manufacturer W.G. Tarrant had made his reputation not in aviation—the Tabor was his first and only airplane—but in building one of the world’s first housing developments and then prefab wooden structures for the British army.

Engines high above an airplane’s fuselage centerline create a strong downward pitch when throttled up, as pilots of contemporary Lake Amphibians and the like have been quick to learn, but the Tabor’s test pilots were plowing new ground. Literally, as it turned out, for when they brought the uppermost engines online after starting the takeoff run on the lower four, the Tabor went straight onto its nose and dug a dreadful furrow with it as the far-forward cockpit and its occupants were ground away. The Tabor ended up like a lawn dart, with three of the six aboard killed.

As is too often the case with much-mocked obvious failures, the Tabor was in some ways an admirable airplane. The sleek fuselage was as aerodynamic and advanced as that of a World War II de Havilland Mosquito. It was a monocoque wood cross-ply structure devoid of then-conventional interior stiffening, obviously intended to become the cabin of a spacious airliner once it had finished bombing Berlin.


Tailless airplanes were all the rage in Europe in the 1920s and ’30s, in part because designers believed that eliminating tail surfaces would lessen drag and increase speed. Aero engineer Geoffrey Hill, however, was more concerned with eliminating stalls when he designed his Pterodactyls and had them built by Westland. Hill got his Pterodactyl concept from watching birds fly, concluding that they maneuvered with their wingtips, not their tails, and birds never stalled. The 1931 WESTLAND-HILL PTERODACTYL MK.IV was typical of the breed, and Westland test pilot Harald Penrose later wrote of the single-seater that after being introduced to it by “a few soothing words from Hill about the straightforwardness of the controls,” he had the misfortune to actually fly it.

“For a hundred yards we bucked, lurched and yawed, the [wingtip] controllers being totally ineffectual at low speeds until a sudden surface roughness bounced her airborne…then she touched, only to be thrown into the air again, my instinctive control movements momentarily building up an increasing longitudinal oscillation combined with lateral yawing and lurching that was almost beyond control. If this was Pterodactyl flying, it seemed beyond me.”

Penrose later found that the Pterodactyl could undergo tip stall just after liftoff, which instantly spun the airplane through 90 degrees and sent it flying directly across the runway. A later version, the Mk.V, with a 600-hp Rolls-Royce V12 in the nose, was fast enough (190 mph) that displacing its aileron-like tip controllers flexed the wing in the opposite direction—called aeroelasticity—and thus canceled out all roll control.

Westland tried five different Ptero­dactyl configurations—some with tractor engines and a gun turret at the rear, others the reverse. In each case, the gunner had an entirely unobstructed field of fire, though the Boulton-Paul Defiant would prove that the turreted fighter was an unworkable concept.


The 1939 CURTISS SO3C SEAMEW was built to be catapulted off battleships and cruisers as a gunnery-spotter and observation airplane, but it had such unfortunate handling characteristics that most were quickly replaced by Curtiss SOC biplanes—the very airplane the Sea­mew had been tasked to replace. In an attempt to correct its longitudinal instability, the SO3C was given an enormous, ungainly vertical fin and rudder, with a forward fillet riveted to the top of the observer’s sliding canopy. It was hard to do much observing from inside the multi-paned greenhouse, and nobody wanted to be enclosed during risky catapult launches and even riskier open-water landings, so the canopy usually stayed slid as far forward as possible, thus negating much of the fin’s effect.

The Seamew—known to its handlers as the Sea Cow—was laterally unstable as well, since nobody had thought to give the wings any dihedral. The “solution” was awkward, cranked-up wingtips that looked broken, but they did little to help. One of the most frequently depicted Seamews is a bird that briefly served aboard the light cruiser USS Biloxi in 1944; it bears the uncomplimentary name “War Junk.” Others called the Seamew the “Reluctant Dragon,” since its Fairchild Ranger V12 engine was so underpowered that the airplane couldn’t take off at gross weight even from calm water. Fairchild was good at making inverted straight sixes, but trying to put two of them on a common crankcase nearly bankrupted the company. In rough seas, Seamew engine mounts flexed so much that the propeller occasionally cut through the front of the main float.

Ultimately, a number of Seamews were lend-leased to the Royal Navy, which knew exactly what to do with them. They were turned into radio-controlled gunnery targets.


You’d think it would be easy to design a glider, since all it needs to do is descend in a relatively straight line and land once released from its towplane, but apparently not when it has a wingspan almost eight feet greater than a 747’s. That would be the 1941 JUNKERS JU-322, an assault glider intended to carry a Panzer IV medium tank and troops to in­vade England. When the Battle of Britain put an end to that plan, the Germans figured they’d use the enormous semi–flying wing to help conquer the Soviet Union—also a fool’s errand.

Hugo Junkers had proposed payload-carrying flying-wing designs throughout the 1920s and ’30s, though none ever flew. Some postulated eight engines, others assumed engines of 3,800 hp, a power level not even reached in the 1950s. In any case, the appropriately named Mammut (Mammoth) was towed aloft by a four-engine Junkers Ju-90 airliner and turned out to be uncontrollable in both yaw and pitch, climbing so defiantly that it pulled the big towplane’s tail upward. Nose down and looking at the ground a short distance below, the tow pilot cut the Junkers loose, and the glider landed on a fortuitously open field.

After the airplane was fitted with tanks that held 2,100 gallons of water in an attempt to correct the imbalance, four further flights re­vealed dangerous spiral instability—the glider didn’t want to fly wings-level. Only two Mammuts were completed, though Junkers had optimistically begun building another 98. All 100 were chopped into firewood for locomotives. Firewood? Yes, Junkers, which had pioneered metal fabrication with its very first design, the 1915 J-1, had been ordered to build the Ju-322 from wood, a material with which it had so little experience that the Mammut had to be extensively redesigned when the first Panzer to clank aboard it crashed through the plywood floor.


Much like the original de Havilland Comet, the 1945 AVRO TUDOR airliner was a good example of too much too soon. The postwar British aircraft industry was desperate to build modern airliners, so Avro slapped a cylindrical, passenger-carrying fuselage between the Lincoln bomber’s wings and four Merlin engines. The Lincoln was an uprated version of the Lancaster, which meant the Tudor was basically a taildragger DC-4, albeit pressurized, with high-maintenance engines in an era of tricycle-gear, round-engine airliners.

Eric “Winkle” Brown, the most widely experienced test pilot of all time, told Aviation History that he considers the Tudor one of the three worst aircraft he’s ever flown (the other two: the de Havilland D.H.108 Swallow tailless research jet and the General Aircraft GAL-56 tailless glider). “There wasn’t just one thing wrong with the Tudor,” he once wrote. “Basically it suffered from excessive cruising drag, high engine-failure safety speed, bad stalling characteristics, and control difficulties on takeoff.” Brown was too kind to mention that the Tudor also was unstable in pitch and yaw and buffeted badly at low speeds.

Perhaps the scariest thing Brown had to do with the benighted Tudor was high-speed dives probing the onset of compressibility, since a four-jet-engine version, the Avro Ashton, was to be built. He got it up to Mach 0.7 and quit, since the pullout required the strength of two pilots.

Brown told Avro’s engineers that the airplane needed to be totally redesigned, which they did. But the Tudor Mk.2 prototype crashed on takeoff (the aileron cables had been reversed during servicing the night before), killing its designer, Roy Chadwick. Two Tudor Mk.4s in airline service en route to Bermuda mysteriously disappeared in 1948 and ’49 in good weather with no distress signals sent, thus adding to the lore of the mythical Bermuda Triangle. And in 1950 a Tudor Mk.5 airliner crashed and killed 80 passengers and its crew, at the time the highest fatality toll in British civil aviation history.


The 1953 SHORT SEAMEW SB.6 is a frequent candidate on all the world’s-ugliest-airplanes lists. It was intended to be a cheap, simple, lightweight antisubmarine hunter launched off small escort carriers. But the only person who was ever able to fly the fixed-gear turboprop well and keep its vicious tendencies on a short leash, no pun, was Wally Runciman, the Short Brothers test pilot who made the first flight. (Even so, he landed, uh, short and badly damaged the prototype during that initial hop.) Runciman was also said to be the one brave man who could actually perform aerobatics in the ludicrous-looking Seamew. Not for long, though: He stalled at low altitude during an airshow demo in 1956 and was killed.

The Seamew handled badly during its entire nasty, brutish and short life, despite being festooned with an increasing number of slats, slots, flow correctors and other aerodynamic band-aids. One of its most unusual features was that the main landing gear was designed to be blown free before ditching, to keep the airplane from turning turtle. Apparently no curious Seamew pilot ever pulled the anonymous gear-ejection handle wondering what it was for. One Royal Navy pilot who had flown a Seamew recently posted on an aviation forum: “It must have been designed by an admiral who had never been to sea or had an obsession that one must…design an aircraft to be like a ship. I felt like I was flying standing up.” It was said in the Fleet Air Arm at the time that the Seamew was “half a [Fairey] Gannet at twice the cost.” Only 19 were ever built, and the Seamew never actually went into service. All 19 were scrapped, as though the airplane was too ugly to leave alive.


To some engineers, airplanes are dumb because they need long runways to become airborne. Vertical takeoff and landing is their answer. Obviously helicopters can do it, at the cost of cruise speed, and today the Harrier jump-jet and the V-22 Osprey tilt-rotor hint at the ideal. But in 1954 the CONVAIR XFY-1 POGO was the best that the brightest could come up with: a stubby prototype turboprop fighter that sat on its tail like a milking stool and took off straight up, then transitioned to fast, level flight.

So far, so good. The hard part was transitioning back to vertical flight and landing like, well, a milking stool. It was hard enough that only one pilot, James “Skeets” Coleman, ever successfully flew the Pogo. (Coleman’s backup pilot nearly crashed the airplane on his one aborted flight.) In vertical flight, the Pogo’s controls didn’t work intui­tively but in a confusing pat-your-head/rub-your-belly fashion. It took Coleman more than 50 hours of tethered practice inside a huge blimp hangar to learn how to use them; Skeets was already an experienced helicopter pilot before he tried to fly it, but even that didn’t help.

Coleman once told me the challenge on landing was that “you were faced with three different configuration changes at the same time. You had to rotate your seat more toward the vertical, trim the airplane for vertical flight and make a power change so you didn’t zoom up too high, and your attention goes from your instrument panel to looking back down over your shoulder. Jeez, I still have a stiff neck from that,” he said, laughing. Better a stiff neck than trying to eject from a tumbling, uncontrollable airplane mere feet above the ground.


Called by some the most dangerous aircraft ever tested, the 1954 ROLLS-ROYCE FLYING BEDSTEAD had no aerodynamic stability or lifting surfaces of any sort. The world’s first jet-lift vehicle, it was simply two back-to-back Rolls-Royce Nene jet engines in a light, rectangular tube frame, with the engine’s exhaust nozzles pointing straight down. Since the Bedstead weighed 7,600 pounds and the engines put out 8,100 pounds of thrust, it indeed “flew,” after a fashion, climbing to 50 feet during one test. The pilot sat on top with a throttle like a helicopter’s collective that controlled the rig’s rate of climb and descent, a joystick that vectored bleed air from the engines to nozzles fore and aft for pitch control and on each side for roll control, and rudder pedals that slewed the pitch-control ducts from side to side to provide yaw.

The biggest problem was height control. Because the Nenes spooled up slowly, a pilot sometimes had to add power and then take it off while the engines were still responding to the initial command. And when bleed air was need­ed for attitude control, the slight diminution in thrust avail­able had to be anticipated and the throttle opened precisely to compensate. Fortunately, the Thrust Meas­uring Rig, as it was officially called, had a good auto-stabilization system, without which it was barely controllable. In fact the Flying Bedstead was used to test stabili­ty augmentation for VTOL airplanes as much as it was to venture into the realm of jump-jet thrust technology.

Only two Flying Bedsteads were built, and both crashed—the second one fatally in November 1957, at which point the Brits decided they’d learned quite enough, thank you. Ten years later, NASA built its own Flying Bedstead—the Lunar Lander Research Vehicle—to give Neil Armstrong some moon-landing practice. That one crashed too, seconds after Armstrong ejected from it.


The 1958 TUPOLEV TU-22 BLINDER was the Soviet Union’s first supersonic bomber and one of the worst-handling airplanes the Tupolev design bureau ever released. Crews hated it, and some even refused to fly it—strong medicine in a Communist country. It had, in fact, failed its airworthiness testing but was put into service in 1962 for political reasons. “It was dreaded by its crews, and some regarded it as unflyable,” wrote Jim Win­chester in his fascinating book The World’s Worst Aircraft. One of the airplane’s faults was that skin heating during Mach 1.4 flight caused control-system rods to expand and distort, altering the airplane’s already-shaky handling. The Tu-22’s relatively low-tech wing, swept back 45 degrees, necessitated takeoff and ap­proach speeds far higher than anything Soviet bomber pilots had ever seen, “and the Blinder had a tendency to pitch up and strike its tail on landing” as a result of its strongly aft center of gravity. “The undercarriage was very bouncy and sometimes collapsed—with serious consequences particularly when carrying a fueled-up missile.”

Part of the problem, admittedly, was that the Tu-22 was an enormously complex single-pilot airplane, since Tupolev wanted the narrowest possible fuselage cross-section. Soviet bomber captains considered themselves so elite that many made copilots handle everything but takeoffs and landings, so when these “experts” were upgraded to Blinders and were at the controls from engine start to shutdown, they found they were in over their heads.

Some 20 percent of all Soviet Blinders were lost in accidents, and doubtless more were destroyed by Iraqi and Libyan pilots in the only two other countries to succumb to its charms. Soviets called the Tu-22 the “Awl,” because of its slim, pointy fuselage. The NATO code name was initially to be Bullshot, which might have been more appropriate.

Frequent contributor Stephan Wilkinson suggests for further reading: The World’s Worst Aircraft, by Jim Winchester; The World’s Worst Aircraft, by James Gilbert; and Wings On My Sleeve, by Captain Eric “Winkle” Brown.