This report describes the use of simple mathematical models to examine (i) the size of bus which, operated on the central London service network, would minimise total traveller and transport costs (time and money), and (ii) the likely result of operating small buses in competition with existing London Transport services. Naturally, the optimum size of bus depends sensitively on the assumptions made about the costs of operating different sizes of bus: taking current average costs in the bus industry the optimum size is around 60 seats, but smaller buses, possibly as small as 30 seats, might prove optimum if flexibility of operation could reduce their costs relative to those of big buses. Small buses operated in competition with existing services, and using the same routes as the big buses so that most passengers could catch either type of bus, depending on which arrived first, would easily attract sufficient revenue to cover their costs, and would provide a net social benefit. By contrast, small buses operated on services which are physically separate from those of LT would in most circumstances find it difficult to break even. (Author/TRRL)