For many potential applications of fuel cells, especially road transport, it is highly desirable to use liquid fuels. In this paper it is shown that whether direct fuel cell systems (methanol/air) or indirect, hydrogen reformer to hydrogen are chosen, catalyst performance and cost plays a dominant role in the overall design. Notably, the maximum current densities which can usefully be employed are limited by the conductivity of low cost intercell plates in bipolar designs. Thus to attain respectable volume power densities, thin cell stack designs are essential. Considerations of overall simplicity, cooling and water control together with the need to operate at atmospheric pressure because of the cost, noise and inefficiency of air compressors all favour operating at 100 degrees C or below. Although a number of systems could be constructed with known technology, costs are still too high for all but specialised applications. While useful engineering design work can be done using existing catalysts, present costs are far too high to justify widespread application. The recent discovery in the authors＇ laboratory of a base metal catalyst for the electrolytic oxidation of methanol together with past workers who have :presented data on effective non-platinum air cathodes in acid electrolytes suggest that the quest is worthwhile. Bearing in mind the need to reduce total costs, while some system engineering development is worthwhile to provide solutions to the various physical problems involved, the authors argue that the primary effort on fuel cells should be in the fields of catalysis and electrode structure,