With regard to the paper by Professor Sheikh and his colleagues (June 2004 JRSM1) and the subsequent critical correspondence, no argument on effectiveness can be valid unless related to a particular helmet. Helmets have a centuries-old history and have been designed to protect against specific sources and sites of injury. In 1952, during the Korean War, the Admiralty raised a requirement for a protective helmet for Naval aircrew. It stemmed from reports of ditchings in which pilot or observer failed to abandon the sinking aircraft as a result of concussion following an apparently minor head injury. It was triggered by an accident in which a pilot was scalped when his aircraft went into a deck-landing barrier. I was serving at the RAF Institute of Aviation Medicine at the time and was tasked with developing the helmet. The RAF were not interested. Cade,2 following a study of 1545 cases of head injury in RAF aircrew, had pronounced that only a heavy helmet would give adequate protection and this would interfere with operational efficiency. I started by analysing reports of head injuries in Naval aircrew resulting from flying accidents, to determine the frequency of various types of injury and, where possible, to relate them to protrusions within the cockpit (e.g. the gyrogunsight). I then studied the mechanics of skull and brain injury, from which were derived the constructional requirements of the helmet. The final product was constructed of layers of woven nylon bonded with phenol formaldehyde resin, with an extra layer coronally across the forehead, the commonest site of injury. Additional protection for the forehead was provided by padding of expanded polyvinylchloride, and the helmet harness was made of nylon, which had the property of stretching under tension, thereby tending to absorb an impact as well as distributing it. The helmet was polished to give a low coefficient of friction, so that it would skid over obstructions and minimize the negative g force in an impact. After extensive trials the helmet was accepted by the Navy and 350 were ordered.3 The RAF, without reference to the Institute, ordered 3000. A member of the Institute’s staff, a neurologist, had contacts within the motor-racing fraternity. Drivers used a variety of helmets, none of which met the criteria of consonance with the mechanics of skull and brain injury. They were soon replaced by helmets based upon the Mk 1 Aircrew Protective Helmet. The only motorcyclist’s helmet was the Don R (despatch rider’s) helmet, consisting of a dome of asbestos fibre with leather cheek-pieces, on which Cairns and Holbourn reported in 1943.4 Although the helmet evidently had a substantial effect in reducing the incidence of skull fractures and concussion it was far from ideal. Maximum protection was provided to the roof of the skull, which is very seldom the site of injury in motorcycle accidents. There was little protection to the frontal area and no protection to the temporal area. However, it had a British Standards number, and in 1956 I wrote to The Lancet suggesting that the BS was unwarranted. The British Standards protested but at the same time invited me to join them in the development of a new BS for an improved motorcycle helmet. I presented the aircrew helmet at a meeting of the Department of Transport’s Road-user Sub-committee where a helmet resembling a deerstalker hat, with fore and aft peaks and made of cork, was favoured by the Police Federation and supported by a consultant neurologist from Oxford. The helmet gave no protection against penetration and its high coefficient of friction was enhanced by a covering of grained leather. The aircrew helmet prevailed and, in cooperation with Road Research at Hayes, Middlesex, the Royal Aircraft Establishment and the Snell Memorial Foundation Inc, a new BS was created that was accepted in 32 countries. Today’s motorcycle helmets are derived from the new BS version, to which has been added improved coverage with protection to the face. Today’s bicycle helmets come in several very different forms, obviously differing in stability, protective coverage and resistance to penetration. Before considering the question of legislation it is essential to conduct a study of bicycling head injuries in relation to individual helmets, and if necessary, to develop a helmet combining the best features of the existing helmets, following the procedures outlined above for developing protective helmets for other activities. Bicycle-racing enthusiasts might consider stream-lining to be an important requirement; leisure cyclists might attach more importance to appearance. These features would have to be taken into account.