It is suggested that there is a need for research into the problem of collision avoidance between ships with particular emphasis on human factors. Work already done is of interest, but is largely theoretical and it seems likely that the usefulness of this approach is nearly exhausted. Collision avoidance at sea is achieved mainly by the rules for preventing collisions, by routing schemes for separating traffic in certain areas and, within certain port limits, by local rules, advisory services or occasionally by direct control of traffic movements. Similar methods are adopted to prevent collisions in the air and on the roads, but with differences which reflect the respective environments. It is suggested that these are differences of emphasis rather than differences of kind. A great deal of research has gone into methods by which collisions between road vehicles and collisions between aircraft can be avoided. Analysis of this work to establish its application to marine collision avoidance should form a prelude to a study of particular problems in the marine field.