벼 雄性不稔性을 利用한 集團改良에 관한 硏究 Ⅱ. 自然相互交雜 및 新遺傳子源 浸透交雜에 의한 集團循環改良 育種過程中 몇가지 量的形質의 變異

Studies on population improvement breeding in rice using genic male-sterile factor Ⅱ. Variation of some quantitative characters in process of population improvement by recurrent random mating and new germplasm introduction using male-sterile factor

This study was conducted to clarify the benefit of active recurrent improvement scheme of rice population and to find out the problems occured in process of recurrent random mating and introuduction of new germplasm into the basic populations using genetic male-sterile factor during 1978-1983. Once (NDR₁, NSR₁) and twice random-mating populations(NDR₂, NSR₂, and modified bulk F₄ populations of two crosses between male-sterile mutants of Nihonmasari(Nihonmasarims) and Suweon 223 or Suweon 235, and improved populations by penetrating outcross between male-sterile plants of NDR₂ or NSR₂ populations and Jinjubyeo or Samnambyeo(NDJ, NDS and/or NSJ, NSS) and their F₂ populations selected from male-fertile plants of NDJ and NSJ populations, and parental varieties included two male donors were investigated for phenotypic variations of some quantitative characters during 1984-1985. The modified bulk F₄, populations were derived from male-fertile F₂selections of two basic crosses and were maintained by single seed descent method(SSD). Although the genetic male sterility of Nihonmasariims was controlled by single recessive gene(msms), observed frequencies of male-sterile plants in once or twice random mating populations were remarkably decreased compared to expected ones. This phenomenon might be resulted from pleiotrophic effect of male-sterile gene such as inferiority of competition in vegetative growth and fertilization compared to normal one. The penetration ratio of new germplasm, calculated by the difference in frequencies of male-sterile plants between basic random mating and penerated mating populations under the natural condition of planting with alternative rows of the two populations, was about 11% as average of four different penerated mating populations varied from 2.1% to 16.2%. The male-sterile plants showed significantly shorter culm and panicle length, less tillers per plant, and less spikelets per panicle compared to normal fertile ones. Twice random mating population revealed significantly increased frequencies of tall, late heading, numerous spikelets and severely awned plants compared to once random mating population. Also, the former showed significantly less frequency of intermediate infections but more frequencies of resistant and susceptible plants to blast disease compared to the latter. This indicates that recurrent random-mating populations should be maintained with relatively desirable range of variations by discarding these undesirable recombinants. The NDJ, NDS, NSJ, and NSS populations manifested the significant change of distributions to the respective directions of earlier heading, shorter culm, longer panicle, more spikelets, less awning, more or less tiller, and more or less tolerance to blast or stripe virus diseases compared to the basic random-mating populations according to respective combinations between the basic populations and new donor parents. Also, these recurrently improved populations and random-mating populations presented wider variations of various agronomic characteristics compared with the modified bulk F₄ populations. These facts point out that the active recurrent improvement scheme using genic male-sterile factor as outcrossing carrier may be utilized effectively to accumulate the numerous gene sources together into a population and to increase the opportunity of continuous recombinations. The phenotypic correlations among some quantitative characters, except the association between panicle length and spikelet numbers per panicle, showed similar or considerably different tendency according to the populations due to different parental combinations and different breeding manipulations. Associations between the quantitative characters such as days to heading, culm length, panicle length, and number of spikelets per panicle, whose variations were largely changed by recombined variation, increased significantly according to random-mating times but decreased