Yield potential of historical IRRI rice (Oryza sativa L.) varieties

Rice yield potential has remained constant since IR8 was released 30 years ago. Despite this apparent yield ceiling, breeding to further increase yield potential continue. This study aims to determine the trend in yield potential of IRRI varieties released in different years and compared it with those of traditional varieties, new elite lines, and tropical hybrids. Field experiments were conducted at IRRI and PhilRice during the dry season of 1996. Fifteen genotypes consisting of IRRI-bred rice varieties were evaluated in these studies. Fourteen-day-old rice seedlings were transplanted at a hill spacing of 20 x 20 cm with 4 seedlings per hill. The design used was randomized complete block with four replications. Plants received a total of 200 kg N/ha. Phenological and morphological measurements were done to evaluate any improvement in the IRRI-bred varieties through the years. Grain yield was taken from a 5 sq m plot and expressed at 14 percent moisture content. Average across locations, yield of IRRI-released varieties increased by 75 kg/ha or 1 percent per year. The increase in yields of recently bred varieties was due to larger sink size coupled with heavier grain weight. This resulted to increase sink capacity (spikelet number per sq m x mean grain weight) or potential grain yield. Despite the high biomass and larger sink size of the early bred varieties such as IR26, yields were low due to poor grain-filling percentage and low grain weight. Varieties released from 1984 to 1985 generally had short growth duration, shoot stature, and low biomass production. Although grain filling was improved in these varieties, the sink capacity limited their grain yield. These findings indicate that the yield increase of IRRI-bred varieties was a result of the improvement in sink capacity