||Wheat yields rose rapidly in the Punjab of Pakistan during the Green Revolution decade, 1966-76, but the rate of growth has been significantly slower in the decade that followed. Official statistics indicate that the yield of semidwarf varieties in irrigated areas has hardly changed since 1970. Slower growth in yields contrasts strongly with the rapid growth of inputs, especially irrigation water from tubewells and fertilizer use, which reached 120 kg nutrient/ha for irrigated wheat in the mid-1980s. Farm survey data from different periods also indicate that between 1970 and 1985 late planting of wheat has become much more common (because of increased cropping intensity), tractor use has replaced animal power, and the use of organic manure seems to have fallen sharply. A simple model is used to disaggregate the effects on yields of three factors: 1) the conversion of rainfed land to irrigated land, 2) the adoption of high-yielding varieties (HYVs), and 3) the increasing yields of HYVs resulting from the release of newer HYVs and increased fertilizer application. Applying the model to the period 1971-73 to 1984-86 indicates that, given the change in inputs, irrigated wheat yields should have increased by at least 725 kg/ha. Yields in fact increased by only 375 kg/ha, reflecting the presence of long-term negative influences on yields. The stagnation in yields of HYVs in the 1970s and 1980s is contrary to earlier projections based on extensive on-farm experimentation. The results suggest that important sustainability issues must be resolved if wheat yields in Pakistan's Punjab are to be maintained. These issues appear to relate to 1) increased cropping intensity, which leads to delayed planting of wheat, 2) use of poor quality tubewell water, 3) increased weed and disease problems, and 4) low efficiency of fertilizer use. Many of these factors are not well understood or quantified. Projections of wheat supply and demand to the year 2000 suggest that Pakistan needs to reformulate its strategy of increasing yields by adding more inputs and focus instead on increasing the efficiency of use of inputs and arresting the tendency for yields to decline over the long term. This strategy will require well-coordinated, long-term research combined with efforts to improve extension services and farmers' technical knowledge.