STUDIES ON COMPETITION IN RICE I. COMPETITION IN MIXTURES OF VARIETIES.

Abstract

Cultivated rice (Oryza sativa L.) may be loosely differentiated into tropical and temperate variety groups separated geographically, morphologically in vegetative character associations, and yielding potential. The thousands of tropical varieties are distinct in having vigorous seedling and early vegetative growth; large, pale-green, and lax leaves; tall, weak culms; sensitivity to photoperiod and/or late maturity; profuse tillering; strong grain dormancy; appreciable grain shattering; low yield response to nitrogen; and low to, moderate yield potential (Jennings, 1966b). This tropical plant type is well adapted to a monsoonal climate and relatively primitive cultural practices, but fails to respond to improved agronomic practice. Improved spacing procedures, fertilization, and water and weed control do not increase yields of the tropical plant types to levels comparable with those grown in temperate countries. Varietal improvement in tropical Asia has been confined largely to selection within, and hybridization among, local varieties, and little improvement in yielding ability has been realized. A concerted effort has been made in several countries, however, to increase variability by hybridization of local varieties with temperate types, mostly japonica varieties from Japan. But again, despite considerable international effort, little success has been achieved. Although such indica X japonica crosses show considerable hybrid sterility, and there is a progressive increase in indica characters from generation to, generation (Oka, 1957, 1958), a recent review of the subject (Jennings, 1966a) contended that early generation sterility is not a prime cause of breeding failure. Most breeders in the tropics have concluded that the bulk breeding method used in the plant improvement programs has given particularly discouraging results. Bulk breeding, especially with progenies of crosses involving contrasting plant types, results in a progressive loss from generation to generation of desired intermediate recombinants and types resembling the introduced parental plant types. Breeding materials, objectives, and procedures followed at The International Rice Research Institute are determined by the principle that certain morphological and physiological characters determine grain yield potential (Tsunoda, 1965; Tanaka et al., 1966). This concept of plant type involves strict selection for specific combinations of leaf dimensions, leaf color and arrangement, culm height and tillering (vigor and arrangement), early maturity and insensitivity to photoperiod, as traits associated with productivity, nitrogen responsiveness and lodging resistance (Jennings, 1964; Jennings and Beachell, 1965). As the fundamental objective of the program is to, increase yield of tropical varieties by 100 to 300%, it follows that existing tropical types must be drastically modified in gross morphology (Beachell and Jennings, 1965). The breeding program based on these principles has resulted in rapid and dramatic progress. Experimental and farm yields of fixed lines of tropical indicas crossed with temperate indicas and japonicas, grown in several countries under careful management, have exceeded 10 ton/ha and averaged from 5 to 9 tons compared with 2 to, 4 tons for local varieties grown under the same conditions (International Rice Research Institute, 1967). This rapid progress is in contrast to the situation that has prevailed throughout tropical Asia for so long and questions the