Experiments were conducted in 1980 and 1983 to evaluate the response of yam varieties to staking and fertilizer application on a sandy-loam soil located in the rain-forest zone of SE Nigeria. In 1980, only one variety, Nwopoko (Dioscorea rotundata Poir.), was used; in 1983, two varieties, Obioaturugo (D. rotundata) and Um 580 (D. alata L.), were used. Two levels of staking (staking v. none) and two levels of fertilizer treatment (NPK fertilizer v. none) were applied. The appropriate factorial combinations (four in 1980 and eight in 1983) were arranged in a randomized complete block and replicated three times in 1980 and four times in 1983. Plot size was 70 m2 in 1980 and 28 m2 in 1983; in both years the plant population was 10 000 plants ha-1. Staking increased the tuber yield of the varieties in both seasons. Leaf area index (LAI) was similar in both staked and unstaked yams up to maximum LAI but thereafter unstaked plants lost their leaves and vines faster than staked plants, especially with varieties of D. rotundata. Fertilizer application increased the tuber yield of Um 680 (D. alata) in both staked and unstaked crops by increasing the LAI. Fertilizer application had no significant effect on the staked plants of Nwopoko and Obiaoturugo (D. rotundata), and reduced the yield of unstaked plants of these varieties. Varietal differences in response to staking/no staking may arise from a difference in canopy structure which, if it allows vines and leaves to contact flood water, leads to their early senescence.