In 1938 Messerschmitt started to produce the Bf 109E, popularly called the Emil because of the German phonetic alphabet. The basic change was the fitting of a Daimler-Benz 601A engine. This provided another 400 hp for its additional 400 lbs of weight. But the new engine needed more cooling than the Jumo of the previous models. Just to enlarge the air scoop under the nose would create so much drag that much of the extra power would be lost. The only solution was a whole new cooling system, with proper ducted radiators, that is to say, radiators with controlled airflow to minimize drag. There was only one place they could go — under the wings. Again the wing was changed so that at least some of the cooling equipment could go aft of the centre of gravity, to make up for the extra weight of the new engine.

There were many redesigned Bf 109s still to come after the Battle of Britain. It stayed in production until the end of the war and was produced in greater numbers than any other combat aircraft of the Second World War. Because the Bf 109 was such a fine aircraft, the butchery done to its airframe was not fatal. In 1940, it was still marginally better than the Hurricane and as good as the Spitfire, but changing the initial specification, for a two machine-gun monoplane, crippled it. The problems of the Bf 109 wing were never solved. After the Emil, the designers reverted to wings without armament for all the later variants.

The undercarriage was of an ingenious design. On the ground the weight was carried by the fuselage, rather than by the wings as in other aircraft. The wings were not strong enough to take the Emil's weight. There were obvious advantages: for instance, the wings could be taken off the aircraft while it was still supported by its own undercarriage. But such an undercarriage had to keep the legs, and wheels, very close together. This narrow undercarriage, the extra weight of the wing armament, and the rough airfields of northern France did not go well together. It has been estimated that 5 per cent of all Bf 109s manufactured were written off in landing accidents.

Messerschmitt's enemies said that he was still building the gliders for which he was a keen enthusiast. Certainly the lightweight airframe and high-lift wings were the two basic elements of sailplane design. And the tail was so fragile that it was supported on struts. And of course a glider had no undercarriage, the mysteries of which Messerschmitt found so difficult to master. Even in 1941, when the Messerschmitt Me 210 was being tested, the usually amiable General Udet, now in the Air Ministry's Technical Department, wrote: "One thing, dear Messerschmitt, must be made clear between us, there must be no more aircraft lost in normal landings as a result of faulty undercarriages: this can hardly be described as a technical novelty in aircraft construction."

Willy Messerschmitt was not the only manufacturer who had difficulties with the design of undercarriages. At Heinkel's they actually harnessed an engineer to hang in the wheel bay of an He 70 so that the retraction could be studied in flight.

By the time of the Battle of Britain, the Luftwaffe's High Command was anxiously watching the tests of a new fighter — the Focke-Wulf FW 190 — which had so obviously been designed to avoid the weaknesses of the Messerschmitt Bf 109. Perhaps there was intended irony in making the FW 190's undercarriage wider than had been seen on any other similar aircraft.

In 1928 the Supermarine aircraft company had been bought by the giant Vickers-Armstrong concern. Some said that the Vickers decision had largely depended upon securing the services of the design department.

Born in 1895 Reginald Mitchell, Chief Designer of the Supermarine company, had designed the S.6B seaplane which in 1931 won the Schneider Cup race, and later that year did 407 mph. Now he wanted a chance to use some of this expertise on a new fighter. As a starting point for a revolutionary aircraft, this record-breaking seaplane design was far more liberating than the military biplane specification of the Hawker Fury, or the strict economies of the Bf 108 'tourer'.

It has been written, many times, that the Spitfire was a privately developed machine that impressed the Air Ministry so deeply that they wrote a specification to fit it. This is not true. The true story is less dramatic and more complex. It begins with Air Ministry specification F.7/30, which had been issued in 1930 (and was also the beginning of the Hawker Hurricane).

First Mitchell submitted designs and a report. By March 1932 the Air Ministry agreed to one prototype being manufactured. But Reginald Mitchell was a sick man. In 1933, while the prototype was being built, he had an operation on one lung. He took a continental holiday to convalesce. During that holiday he talked with some young German aviators, and he returned to England convinced that war would come soon. With a prescience sometimes given to the very old or the very sick Mitchell began to believe that his fighter design could influence the outcome of the war, and from that time onwards he refused all medical advice about resting, devoting his life entirely to his aeroplane.

He looked again at his prototype, which began flying by the end of 1933. It was a curious-looking gull-wing monoplane with an open cockpit and a spatted fixed undercarriage. (It looked more like a Junkers Ju 87 than like a Spitfire.) The unsatisfactory Goshawk II engine gave it a top speed of no better than 238 mph. No one had to tell Mitchell that it wasn't good enough. Even while the prototype was under construction Mitchell had been scribbling radically different shapes on the drawing board: closed cockpit, outward-retracting undercarriage, and smaller wings with split flaps.

During 1934 the Air Ministry was responding to intelligence reports about the secret German air force. Demand for better RAF fighters was now more urgent but there were still many Air Ministry officials who insisted upon the superiority of the biplane. In July 1935 the Gloster Gladiator biplane was ordered into production for Fighter Command.

Towards the end of 1934 while Hawker's were building a Hurricane mock-up and still waiting for an official order Reginald Mitchell prepared detailed drawings of his entirely new fighter. This time he ignored the unsuccessful Goshawk engine and designed his fighter around the PV 12. By now the Air Ministry was working more closely with the two designers, and as the new specification was written it deliberately incorporated much of the work already done.

In January 1935 work on Mitchell's new prototype started, and yet the Rolls-Royce engine that was essential to its success had not yet flown. In fact the engine was not ready. Overheating had to be solved by means of ethylene-glycol coolant. Engine and airframe were being developed simultaneously, and it wasn't until the very end of 1935 that the engine got its certificate of airworthiness. By now even the Air Ministry had begun to realize that it was a race against time. The engine went straight into the waiting airframe and on 5 March 1936 the Spitfire prototype flew.

Mitchell still had a long way to go. He had not designed the elliptical wings which were the key to the Spitfire's excellence. And when Mitchell had designed them, it was obvious that the curved shape was going to make for complex problems during the tooling-up, so the first deliveries would certainly be later than those of the Hurricane.

Camm's energies were devoted to manufacturing Hurricanes using as many as possible of the jigs, tools, and skills already available at Hawker's. Mitchell's time was devoted to solving his problems in original ways. The Spitfire fuselage was built in three sections: a tubular case to enclose the engine, a monocoque centre part, and then a detachable aft section (of which the last two formers extended upwards to become the tail fin).