1. Allo-octoploid plants (2n=96, CCCCDDDD) were induced by treating the germinating seeds of allo-tetraploid species, Oryza latifolia DEsv. (2n=48, CCDD) with 0.05% colchicine solution. 2. In the octoploid-latifolia, anther development is delayed and the meiosis in PMCs begins only when the top floret of panicle is just coming out of the flag-1eaf sheath. 3. At MI in PMCs, tetravalent chromosomes ranging from O to 10, on the average 4.11 per PMC are formed, the mode being 2. Trivalent chromosomes are rarely met with, while univalent chromosomes are observed rather frequently, 8 in maximum. Thus, the average number of bivalents per PMC was calculated as 39.19. 4. Octoploid-latifolia gives gigas-type in most characters, showing thick culm, thick and coarse leaves, thick and large auricles. The length of ligules, panicles and awns, and the size of stomata, pollen grains and spikelets increased. Especially the increase of spikelet size is conspicuous, but the number of spikelets per panicle showed a reduction. Under short-day condition it headed about 5 days later than its original species. 5. The seed fertility was low, recording 33 % in selfed condition. The crossability with tetraploid-sativa, however, increased remarkably in comparison with that between undoubled species

View full abstract

In view of distinct differences of pH values noticed in anther tissue, especially at pollen tetrad and pollen grain stages, studies have been conducted on the application of rejection ellipse to breeding of complete male sterile strain and determination the various types of male sterility. The rejection ellipses were obtained on the basis of the critical region of bivariate statistics of two developmental stages of pollen. The long axis and short axis of the ellipse correspond to the pollen tetrad and the pollen grain stages respectively. The rejection ellipses of normal and complete male sterile types which are diagrammed at 5% level of significance distinctly separated from each other, while that of semi-sterile type overlapped partially with two above mentioned ellipses. Such rejection ellipses are seemingly useful for breeding of the male sterile strain from a given population.

View full abstract

The main purpose of this study is to discuss the possibility of combining the favourable characteristics which exist independently in three types, i. e., Virginia, Spanish and Valencia. In the previous paper (SYAKUDO and KAWABATA 1963 a), the authors showed that crossings between varieties are successful both within each type and among the three types ; and the seed fertility of F₁ plants is similar to that of the parents. The present investigation was undertaken to study the interrelationship of some main characters in the F₂ generations. Phenotypic and genotypic correlations among the characters studied are presented in Table 3. In general, genotypic correlations are higher than the phenotypic correlations but the values themselves are very low, with few exceptions. The highest correlation is found between pod length and kernel length (O.935∼O.959 in genotypic and O.853∼0.899 in phenotypic, respectively). Comparatively high values are found between leaflet size and plant type index, and between length of branches and top weights. Both flowering date and oil content do not correlate with other characters studied. These results suggest that the combination of some favourable characters to form a variety may be successful. In addition, Table 2 shows the heritabilities in the broad sense for all characters in the three crosses. Plant type index and height of main axis showed relatively high heritabilities (O.89 and 0.82, respecitively). Kernel yield and air-dried top weights showed lower heritabilities (O.66 and O.64 on the average for the three crosses)

View full abstract

There are two types of growth habit in the tomato, i.e. normal and self pruning. Several forms of the self pruning type of tomato are known. The author attempted to analyse quantitatively the varietal differences in the self pruning type. The branches (main stem and lateral branches) of the tomato are sympodially formed. Sympodial axillary shoots terminate after a few leaves in an apical inflorescence. There are basically four types of sympodial axillary shoots (Table 1) . The branches of the normal type usually consist of Type III alone (Fig. 1A), while those of the self pruning type consist of various combinations of four types (Figs. IB-J) . Frequency distributions for types of sympodial axillary shoots in five successive shoots of the main stem showed that the eight varieties differed markedly from each other in degree of self pruning (Table 2) . Scores from O to 3, corresponding to the four types of sympodial axillary shoots, were given :for quantitative analysis (Table 1). From the results shown in Figs. 2 and 3, it was accepted that the average mean score of the first and second sympodial axillary shoots of the main stem was good enough to represent the relative degree of self pruning among varieties and was designated as the self pruning index (SPI) of the variety. The SPI of a variety may be obtained easily by the following formula :

View full abstract

According to BRIQUET (1897), Mentha viridis L. (Syn. M. spicata L., M.spicata L. var. viridis L.) and M. longifolia(L.) HUDs. (Syn. M. spicata L. var. longifolia L., M. silvestris L.) belong to the Spicatae section in the genus Meletha. The Spicatae section is characterized by flowers in spikes. In the Spicatae 'sectiole, M. longifolia is distinguished from M. viridis by the former being more or less hairy while the latter is glabrous. This difference seems to be due to a few Mendelian genes. Either species, however, does, not always breed true to its type as is seen in Table 3. When we propagate by seed, from downy clone (M. longifolia) we cannot always expect only downy plants in the progeny, or only glabrous progenies from glabrous clone (M. viridis). From the above mentioned fact, we suppose the names of the species were given to the phenotypes of these clones, not to their genotypes. Hence, it seems insufficient to separate these two species according to such an unfixed character as that of hairiness when no other definite difference can be found. By another experiment of ours, it was proved that from the interspecific crossings between the two species mentioned above, no difference could bc fcund in genome constitutions. (This will be published later.) So we propose to unite these two species into one, M. spicata L., in accordance with LlNNAEI (1753). If it is necessary to separate one from the other by the phenotype of the clone, the glabrous spicata should be called M. spicata L. var. viridis L., and the hairy one M. spicata L. var. longifolia L

View full abstract

Morphological and histochemical comparisons of pollen were made among the following rice strains (a) Fujisaka No. 5, an improved Japanese variety; (b) a strain of Chinese wild rice, O. sativa f. spontanea ; (c) a male-sterile, Fujisaka No. 5-type strain with spontanea cytoplasm ; (d) a normal spontaneatype strain with Fujisaka No. 5 cytoplasm. The last two of these were made by Katsuo and Mizushima (1958) through nuclear substitution. In microcytes at the tetrad stage polysaccharide content was observed to be less in (c) than in (d), (a), and (b). Localization of starch grains to the polar site of germpore of pollen was recognized in (d), (a), and (b), but not in (c). Translocation of starch grains to the site was confirmed to take place at the time of anther dehiscence in the former three. The pollen of (d), (a), and (b) germinated on artificial culture media, whereas those of (c) did not and puffed out cytoplasm. Majority of the pollen of (d), (a), and (b) puffed out cytoplasm soon after being soaked in distilled water. Occurrence of this plasmoptyse was not so remarkable in the case of (c) . Pollen of all the strains showed similar amylase activity as well as equal maltose and glucose content.

View full abstract

In Part 3, it described the problem of the partition of the genetic variance and the genetic covariance into genic random variables and other random ones in case where the linkage exists. The results, such as formulas (15) and (40), are derived and show that genic random variables and random variables of the deviation due to dominance are not correlated in any cases. In any part of this report, the problems on the estimation and more complicated stochastic models have not been considered, but these problems will be dealt with afterwards, considering practical breedings. Then this Part 3, is the last part in the problems of handling genetic structures of quantitative characteristics with stochastic strictness

View full abstract