1942 年 12 巻 4 号 p. 397-417
1) The bivalent in Trillium takes usually a cruciferous form, the pair of paired chromatids being associated only by the kinetochores themselves. This condition of pairing has been regarded by the chiasmatheorists as representing one chiasma or two localized at the vicinity of the kinetochore. In the present study, the paired kinetochores were strictly distinguished from chiasmata and were not counted as the latter in recording. It was pointed out then that earlier workers of the chiasma shool disregarded this point making their data on chiasmata seriously unreliable.
2) Based on this new method, the present study was made on the comparison of the frequency of metaphase chiasmata in one control plant (t0) with those in four other plants (t1-t4) which had been subjected to high temperature prior to metaphase.
3) The mean chiasma frequencies per cell in these plants were found to be: t0, 0.45, t1 1.49, t2, 2.20, t3 4.31, t4 6.53. Such great increases in chiasma frequency have not been reported before. In terms of the chiasma school, this change in the situation is from “localized chiasmata” to “random chiasmata”
4) There are two types of bivalents distinguishable as to the pairing condition of the kinetochores that is, (i) “K-bivalents” where the kinetochores of the homologues remain paired until metaphase and chiasmata may or not be formed, (ii) “k-bivalents” where the kinetochores are unpaired at metaphase and chiasmata are alone responsible for the maintenance as bivalents. In t0-t3, the bivalents are nearly almost of the former type, while in t4 they are of the latter type in a majority of cases.
5) In K-bivalents, the frequency of interstitial chiasmata is a function of arm length, while in k-bivalents it is a function of entire chromosome length (see Graphs 1 & 2). It was also found that the k condition of the kinetochores is correlated to certain extent with an increase in the frequency of interstitial chiasmata and furthermore greatly with the occurrence of univalents.
6) The above principles about interstitial chiasmata were found not to hold true as they stand to terminal ones. In the present material, terminal chiasmata are negligibly low in frequency. In general, however, they occur more frequently in plalnts in which interstitial chiasmata are higher in frequency. It is also significant that terminal chiasmata occur in relatively higher frequencies at the nucleolar forming ends of the chromosomes, that is, at both the ends of A and the end of the short arm of E. It was inferred from these findings that the chromosome matrix acts as the prime factor for the maintenance of terminal chiasmata and the nucleolus plays a subsidiary rôle for it.
7) The present findings are at variance with the view of the chiasma school that the formation of random chiasmata is the sign of complete pairing. With respect to the paired or non-paired condition of the kinetochores (K and k respectively) and the formation or non-formation of chiasmata (C and c respectively), it was concluded that the typical or standard bivalent is represented by the Kc condition, and KC and kC types of bivalents are its derivatives (see Fig. 12).
8) The above view favors the two-plane theory of chiasma formation, that is, chiasmata are formed by the alternate opening-out of sister and non-sister chromatids at diplotene and therefore not related in their origin to crossing-over. It was then assumed that such alternate opening-out is due to the development of two or more points of repulsion in an arm, a condition of rather unusual nature in contrast with the normal case, where the repulsion force initiates at its one point (probably the distal end);