In 1989 and 1990, small, medium and large seed tubers, cv. Désirée, were planted 38 cm apart in experiments designed to measure the effects of different combinations of neighbouring plants on the yield and size of tubers from individual plants. Total yield, ware yield (> 150 g), numbers of tubers and numbers of stems increased as seed tuber size increased. Also, decreasing the size of seed tubers decreased the numbers of tubers in all sizes except those > 200 g. Competition from the two neighbouring plants on either side in the same row (first neighbours) increased as the size of seed tubers increased and with all seed sizes yields decreased. For example, plants from large seed benefited from small seed planted as neighbours whereas, with plants from small seed, yield was decreased with neighbouring plants from large seed. Numbers of stems and tubers were not affected by neighbouring plants. Also yields were not affected by the two plants adjacent to the first neighbours (second neighbours) or by the size of seed tubers planted in adjacent rows. Competition from first neighbours also affected tuber size distributions and, with small seed, increasing the competition decreased the number of tubers > 150 g and increased numbers < 50 g. Medium and small seed planted as neighbours of large seed increased numbers of tubers > 150 g. When yields from plants produced by seed of each size with similarly sized seed planted as first neighbours (uniform populations) were compared with mixed populations, yields were always greater from the mixed populations of different seed sizes planted alternately than from the means of uniform populations. Therefore in the mixed populations, there were larger gains from the larger seed than losses from the smaller seed.