The paper sets out to show that by considering the requirements of a motorcycle crash helmet from first principles short comings in existing designs become apparent. Results were obtained from 150 drop tests of instrumented head forms in helmets under controlled conditions. The performance-of practical helmets is compared with the protection that would be provided if the available space could be fully utilised for energy absorption as predicted by applying equations of motion to an accepted injury criterion. H.I.C. is used as the initial criterion but alternative methods of predicting brain injury are discussed and their underlying principles examined. Choice of present helmet materials and the current British Standard test procedure are examined.