Hiko-Ichi Oka

VARIATIONS IN THE BREEDING SYSTEMS OF A WILD RICE, ORYZA PERENNIS

An important factor determining the genetic structure and evolutionary dynamics of a population is its breeding system. I t differs among plant species as has been surveyed by Fryxell (1957) but its variation within a species is little known. Oryza perennis Moench is distributed throughout the tropical countries of the world, and comprises many varieties or forms with different propagation habits. They can be divided into four main geographical groups or races, Asian, American, African, and Oceanian, distinguished by characterassociation patterns and F1 sterility relationships, though not by single characters (the writers’ unpublished data). The Asian race tends to differentiate into perennial (perennis or bdunga) and annual (spontanea or fatua) types, though there are many intermediates (Morishima et a]., 1961), and the perennial type may be the progenitor of cultivated rice, 0. sativa L. (Oka and Chang, 1962; Oka, 1964). The Oceanian race also seems to contain perennial and annual types. The African race, called 0. perennis subsp. barthii, is highly rhizomatous, while the American race, comprising a Cuban form called 0. cubensis Ekman and other South American forms, appears to vary from perennial to annual. In contrast, another cultivated species, 0. glaberrima Steud., and its wild progenitor, 0. breviligulata A. Chev. et Roehr., endemic to West Africa, are wholly annuals (Morishima et al., 1963). To investigate how the variation in the breeding system of 0. perennis is related to that in the mode of adaptation, the writers made a survey of breeding characters of various strains. Data for 0. pa-

Genic analysis for the sterility of hybrids between distantly related varieties of cultivated rice

Summary1.The hybrid sterility between distantly related varieties of cultivated rice can be explained by duplicate genes which work in the gamete as development maintainers (Gametic-Development Genes, or g.d. genes). Hybrids between varieties with those genes in different combinations may produce double recessive gametes which deteriorate at a certain stage of development.2.For analysing these g.d. genes, crossing experiments of an ‘(A xB) xC’ type were used, whereA andB are closely related varieties, but differ in fertility in the hybrid with the third varietyC. In one case, it was found that a g.d. gene was linked with the glutinous genegl with a certain recombination value.3.When a gene is linked with a g.d. gene, causing sterility in a hybrid, its segregation ratio changes, due to gametic selection in a definite manner. This phenomenon was investigated theoretically and experimentally.4.When two independent genes are linked respectively with two pairs of g.d. genes belonging to the same set (the two pairs have the same effect but are located on different chromosomes), recombination of the former genes is restricted to some extent. This phenomenon was demonstrated in theF2 of some varietal crosses.5.It was pointed out that the g.d. gene sets segregating in a hybrid between, distantly related varieties of rice are usually numerous, and produce a continuous array of inter-grades in fertility in theF2 and later generations. The g.d. gene sets may be said to act as polygenes.6.In addition to the g.d. genes, other duplicate genes for fertility seem to be present. The origin of gene duplication was discussed. Since it was found that in tetraploid hybrids between distantly related varieties of rice, the chromosomes derived from the same parent did not tend to pair selectively, the gene duplication in rice was considered to be due more probably to the doubling of chromosomes, as suggested by the hypothesis that rice is a secondary balanced polyploid, than to numerous small translocations.7.When varieties with those duplicate genes in different combinations are hybridized, since chromosome recombination tends to bring about unfavourable effects, genotypes similar to the parental strains will increase in the hybrid population. It was pointed out that this tendency might work as an isolating mechanism in mainly self-fertilized plants.

Intermediate perennial-annual populations ofOryza perennis found in Thailand and their evolutionary significance

To reexamine the mode of variations in adaptive mechanisms in the Asian form ofOryza perennis, the breeding behavior of natural populations was observed at eight sites selected in Thailand. One population consisted of seedlings only (annual), and five consisted of ratooned plants only (perennial), while two others were mixtures of both. The annual population had a larger number of buried seeds at the habitat than had the others. In the habitat of the annual population, annual species were predominant, while perennial populations tended to coexist with perennial species. The annual population was in a temporary swamp, while perennial populations were found at sites more deeply inundated in the rainy season and retaining moisture in the dry season. This indicated the role of water stress in conditioning the reproductive strategy.Observations of character variations among plants from those populations in a uniform condition proved that the seedling population was of an annual type while those performing mixed sexual and asexual propagations were of an intermediate perennial-annual type. Isoenzymic variations controlled by alleles at two known loci indicated that the two intermediate populations were highly polymorphic and heterozygous. One of them consisted of nine separate clumps which differed in their characteristics. On the basis of these observations, the dynamics of differentiation of this species in adaptive strategy is discussed. It is further suggested that the intermediate perennial-annual type is probably the wild progenitor ofO. sativa.

COMPARISON OF MODES OF EVOLUTION OF CULTIVATED FORMS FROM TWO WILD RICE SPECIES, ORYZA BREVILIGULATA AND O. PERENNIS

Oryza breviligulata A. Cheval. et Roehr., growing wild in West Africa, is genetically closely related to 0. glaberrima Steud., which is cultivated by the native people of the same region. 0. stapfii Roschev., also found in West Africa, can be regarded as synonymous with 0. breviligulata, as will be pointed out later. These species form an endemic group distinct from the common cultivated rice, 0. sativa L. and the wild species closely related to it, 0. perennis Moench. They can be distinguished from 0. sativa and 0. perennis by short, roundish ligules and other characters, and their F1 hybrids with the latter two species, though they show no disturbance in chromosome pairing, are highly sterile (Morinaga and Kuriyama, 1957; Nezu et al., 1960; etc.). Another wild form occurring in Africa, 0. barthii A. Cheval., is considered by us as an African variety of 0. perennis which can be distinguished from Asian and American forms of the same species by the remarkable rhizome development (Chatterjee, 1948; Sampath and Rao, 1951; Sampath, 1961) . The above rice species may for the sake of simplicity be divided into two series, Series Glaberrima comprising 0. glaberrima and 0. breviligulata, and Series Sativa comprising 0. sativa and 0. perennis. This paper deals mainly with the comparison of the two series regarding the mode of evolution of cultivated forms. Regarding the phylogenetic relationships in Series Sativa, a number of systematic and cytogenetic studies have been made (Chatterjee, 1951; Sampath and Rao,

THE DYNAMICS OF PLANT DOMESTICATION: CULTIVATION EXPERIMENTS WITH ORYZA PERENNIS AND ITS HYBRID WITH O. SATIVA

Experiments were carried out on the farm of Chung-Hsing University, Taichung (24°N), Taiwan, in collaboration with the late Dr. W. T. Chang. Three-week seedlings raised in pots were transplanted with a single plant per hill into irrigated paddy plots (3m X 10m) surrounded by dikes, or presoaked seeds were broadcast in the same plots at a rate of about 150 grains per square meter. To reduce inter-matings, the experimental populations were arranged so that the adjacent ones differed in heading time. No chemical fertilizers were applied since the paddy soil contained enough nutrients. The experiments were carried out as follows: First Experiment (Response oj Natural Population to Cultivation; Transplanted). -Using the original seeds from natural When wild plants are grown by man, cultivation may in itself work as a selective agent. However, to the knowledge of the present authors, the response of a population of wild plants to cultivation does not seem to have been investigated. The Asian form of Oryza perennis Moench may be the wild progenitor of cultivated rice, O. sativa L. (Oka, 1964). It is 5 to 50% cross-pollinated while sativa is predominantly self-pollinated (d. Oka and Morishima, 1967). It usually grows in marshes in the proximity of rice fields, and sometimes produces hybrid swarms with sativa (Oka and Chang, 1961). The populations in natural habitats also seem to contain differing numbers of the sativa genes due to introgression (Morishima and Oka, 1970). To look into the dynamics of plant domestication, population samples of Asian perennis collected in natural habitats, and peretmis X sativa hybrids were grown for several generations in an experimental field at Taichung, Taiwan, and were observed for characters distinguishing wild from cultivated types. The examined populations indicated a change toward the cultivated type at varying rate according to the propagation method. Seeding was found to be an important agent causing natural selection for the cultivated type.

The impact of copper pollution on water foxtail (Alopecurus aequalis sobol.) populations and winter weed communities in rice fields

Water foxtail (Alopecurus aequalis Sobol.) is a winter-annual grass dominant in rice fields in Japan. The seeds were collected from copper-polluted (30–300 ppm) and non-polluted control fields in Ohta (Gunma-ken) and Tsuru (Yamanashi-ken), and were sown in pots with varying levels of copper (0–400 ppm) in the soil. Percent survival on copper-containing (toxic) soils depended mainly on the ability to recover from frost heaving of the roots affected by copper. It varied from 30 to 90% among strains and showed a strong correlation with soil copper content at the collection site. Percent heading of plants on toxic soil also appeared to be an indicator of copper tolerance. Copper-tolerant strains accumulated more copper and manganese in their roots than non-tolerant ones, although the copper content of the tops showed no such relation. When the strains were mix-planted with milk vetch (Astragalus sinicus L.), their resistance to the latter species showed a negative correlation with copper tolerance indicating a trade-off relation between tolerance and competitive ability. A counterbalance of trade-off to the selection for tolerance may account for the correlation found between tolerance and soil copper content. The percentage of water foxtail plants to total plants emerging from buried seeds was correlated with soil copper content at the sampling site. The dominance of water foxtail in polluted fields was also recognized from a direct observation of the fields. The tendency of species diversity and equitability in the community to decline with increasing copper level in the soil was observed with plants from buried seeds as well as in the fields.

Genetic diversity in rice populations of Nigeria: Influence of community structure☆

Two rice species, Oryza glaberrima and O. sativa, are often mix-planted in West Africa, and their populations, particulary those of O. glaberrima, are highly polymorphic in seed type. It was found that the seed-type diversity within glaberrima populations, as shown by the “amount of information”, H = −Σpilogepi, was correlated with species diversity in the field (also shown by H), and gave a curvillinear regression on the proportion of glaberrima plants to the total number of rice plants. The highest diversity occurred when the proportion was 50–60%. A similar trend was also found for sativa populations. The background and significance of this phenomenon are briefly discussed.

The impact of copper pollution on weed communities in Japanese rice fields

In certain localities in Japan, rice fields were polluted by copper after irrigation with river water polluted by drainage from copper mines. Soil samples were taken from two such localities, Ohta and Tsuru, and a survey of the weed community was made using seed reserves in the soil. The soil samples contained 50–310 ppm copper. The rice field was found to have a characteristics composition of weed species. Copper pollution generally reduced species diversity but increased the total number of plants arising from a soil sample. Major species whose plant numbers increased were Vandellia angustifolia, Lindernia procumbens, Rotala indica and Fimbristylis miliacea, while the number of plants of Mazus japonicus decreased in polluted fields. The same patterns of change were observed when very polluted water-inflow parts were compared with slightly polluted water-outflow parts of the same fields. The relative copper tolerance of the species, estimated by tests on copper contaminated and normal soils, did not account for the changes in their relative abundance. The evolution of copper-tolerant genotypes was recognized in Vandellia angustifolia, Mazus japonicus and Cyperus difformis. The tolerant strains were generally inferior to non-tolerant ones in their performance on normal soil.

An experiment on the interaction between cultivated rice and barnyard grass at different planting densities

Abstract To examine the relation between the effects of density and interaction between species, a rice and a barnyard grass strain were tested in pure and mixed (1 : 1) plantings at five different densities. In pure stands, the data for dry weight per plant and other characters gave a good fit to the logarithmic regression equation on plant density (derived from WNa = K). For the mixed plantings, a parameter showing neighbor effect was added to the equation, and its values were estimated. Rice and barnyard grass had different signs for this parameter. The trend of competition effect to increase with density was not observed. Both species tended to reduce reproductive effort at high densities, but at different rates. Rice showed logarithmically linear density responses in both size and number of stems, while barnyard grass showed it only in stem number.