This thesis discusses the possibility of applying pattern recognition techniques to the identification of individuals typing at a computer keyboard. The patterns of times between keystrokes and of times each key was held down, recorded when a particular sentence was typed, were used as the data sources for the recognition. A system using the techniques explored in this work could be employed to restrict access to computers for security purposes. A pattern recognition system was developed based on template matching. Templates were compiled from sentence data recorded in a learning phase, and then matched against another set of data in a classification phase. Initially a simple least-squares fit matching algorithm was employed. This was then developed and improved in order to increase the levels of identification. The highest levels were obtained when the match of the templates against the sentence data was completed using a fitting algorithm calculated on only the keystroke times which were found to be repeated consistently. The best data source for identification proved to be the key held time. Identification levels of over ninety per cent for a five per cent level of misrecognition were achieved. The feasibility of designing a pattern recognition system to recognise typists typing at a terminal has been confirmed by the work carried out for this thesis. High levels of recognition could be expected from an on-line system designed using the algorithms that have been developed.