Thirteen rotation treatments consisting of 'Bedford', a soyabean cultivar resistant to Heterodera glycines (SCN) races 3 and 14, 'Forrest' resistant to races 1 and 3, susceptible 'Essex', and a nonhost crop (cotton [Gossypium hirsutum]) were evaluated in an SCN infested field for 4 yr in Portageville, Missouri. Yields of soyabean cultivars in continuous production systems were compared with those in rotation with the nonhost, resistant, and susceptible cultivars. Bedford and Forrest yields, over the 4 yr, were less variable than Essex yields. Forrest and Essex produced similar yields in 3 out of 4 yr, and yields of Essex were significantly lower than Bedford yields in 2 out of 4 yr. Beneficial effects of different rotations on cultivar performance were not very conspicuous, however, yields of Essex were increased with rotation in 5 of the 6 comparisons compared with continuous Essex, whereas Bedford in rotation was significantly higher than continuous Bedford in 2 of the 5 comparisons. Grain yields of Forrest were not affected by rotation systems. The nematode density, after 4 yr, was lowest in cotton-Forrest-Bedford-cotton system and greatest in Bedford-cotton-Essex-Forrest system. The preceding crop had a greater influence on the nematode population density than the rotation system per se. Treatments with cotton as the preceding crop had lower nematode population densities than resistant or susceptible soyabean. Rotating cultivars caused changes in SCN race structure. SCN race 14 was present in almost all rotations including Essex and Forrest. Bedford tended to increase the population with genes for parasitism on PI 88788, resulting in a shift to race 4. The shift to highly pathogenic race was slower in rotations that included susceptible Essex and Forrest as a component crop. Inclusion of susceptible soyabean lines in rotations may delay, but not prevent, the development of highly virulent races of SCN. Rotation of soyabean cultivars was found to be a useful practice in the management of SCN-caused losses.