In this paper an occurrence of restitution-nuclei in the formation of the embryosacs in Balanophora japonica has been reported. The number of chromosomal elements found on the equatorial plate of the heterotype division varies from 94 to 112 so far as the present investigations are concerned. There is a correlation between the number of these elements and their sizes. These facts seem to show that at least in some of the chromosomes there is a tendency to form gemini, but the number of such chromosomes varies in different embryosac mother cells. The chromosome number in somatic Cells seems to be not far from the largest number counted in the heterotype division, 112. The present investigations were undertaken at Prof. FUJII'S suggestion and under his supervision more than ten years ago. In publishing this little note, the writer wishes to express his sincerest thanks to Prof. FUJII for his advice and criticism throughout the work, and for his kindness in giving the writer permission to use his material.
It is of great interest to know the correlation between chromosome numbers and the intensity of the function of gametes which are produced from a pentaploid wheat hybrid. In order to study this problem I used mainly 41-chromosome plants (20II+1I) that were (D-2g×T. spelta) F1, (D-2f×T. spelta) F1 and reciprocal crosses. If one univalent chromosome is not lost in the meiosis of spore mother-cells, 20-and 21-chromosome gametes are produced in a 1:1 ratio. But by the “Äquationskreuzung” we have found 20-and 21-chromosome megaspores in a ratio 73:27; then the chance of one univalent chromosome loss in the meiosis of E. M. C. is 46 per cent in (D-2g×T. spelta) F1, and 42 per cent in (D-2g×T. spelta) F1 (Table II). This is in close agreement with KIHARA'S (1924) and WATKINS'(1924, 1925) results that were studied chiefly with P. M. C. We may conclude, therefore, that loss of chromosomes occurs with about the same frequency in the microspore formation as in the megaspore formation. On the other hand it was found that among the progenies of the “Zertationskreuzung” D-2g×(D-2g×T. spelta) F1 and T. spelta×(D-2g×T. spelta) F1, in average only 11 per cent was fertilized by 20-chromosome microspores, and among D-2g×(D-2g×T. spelta) F1 and T. spelta×(D-2g×T. spelta) F1 37 percent (Table III). We can assume that the discrepancy in the microspore ratio is probably largely accounted for by a slower growth rate of pollen-tubes from 20-chromosome microspores. If the ratios of effective 20-and 21-chromosome gametes be 73:27 in the female and 11:89 in the male, we should expect the 40-, 41-and 42-chromosome progenies of a self-fertilized 41-chromosome plant to be in the ratio 8:68:24 in (D-2g×T. spelta) F2; and 27:56:17 in (D-2g×T. spelta) F1. As shown in Table I the observed ratios are roughly in agreement with the calculated ones, i. e, 11:69:20 in the former, and 14:69:17 in the latter. A state of affairs similar to that mentioned above was found in Oenothera lata, the “globe” mutant of Datura, and the “enlarged” mutant of Nicotiana. But in these cases the extra chromosöme is probably duplicated with one of those normally present in the gamete (n+1), and in the case under consideration, on the contrary, one chromosome is wanting (n-1). In F2 progenies we have found some plants with unexpected chromosome combinations, and various chromosome numbers (Table VI), and among them a plant with a non-viable combination, 19II+1I, was found to be most sterile. Although the fertility in a selfed D-2g is as low as 56 per cent, male and female gametes, are effective more than 83 per cent even in artificial pollination (Table IV). On the sterility in wheat hybrids there are now two hypotheses, i. e. gametic and zygotic sterility hypotheses. My results clearly give strong support to the hypothesis of zygotic terility. The allelomorphic characters for awns, and hairs on glumes are transmitted by two different simple Mendelian factors which do not lie in the g-or f-chromosome.