| Meiotic studies of 3 generations of backcrosses to
the amphidiploid hybrid Triticum durum DESF. x Agropyron intermedium
(HOST) BEAUV.1) B. A. YOUNG and J. SCHULZ-SCHAEFFER Department of Plant and Soil Science and Genetics Institute, Montana State University, Bozeman. Montana 59715, U.S.A. Progress has been made towards the development of a meiotically stable male sterile derivative from the cross Triticum durum DESF. x Agropyron intermedium (HOST) BEAUV. (SCHULZ-SCHAEFFER et al. 1971, 1973). Material and methods Perennial amphidiploids (AD) in the F7 and F8 generations were used as parent material in these investigations. T. durum served as the female parent. A. intermedium was used as the recurrent backcross parent in all 3 backcross generations (SB1, SB2 and SB3). Spikes were collected in the field, fixed in a 3:1 mixture of absolute ethanol and acetic acid (Farmer's formula), and stored in 70% ethanol. Staining was done in acetocarmine according to the smear technique (SMITH 1947). Part of the study of univalents in th SB3 was done in metaphase I (MI) of meiosis. Since MI is a relatively long stage, it enables a favorable chance for observing a high number of cells. We analyzed 2915 cells of 52 plants in MI. Only 234 cells could be analyzed where bivalents and univalents could be identified in the same cell. These observations were done in stages of shorter duration such as diakinesis and anaphase I. Results and discussion In Table 1 the observed chromosome pairing and total chromosome numbers of all 3 backcross generations (SB1, SB2 and SB3) are compared with the pairing in A. intermedium and with the expected values. The observed average number of bivalents increased from 11.7 in the SB1 to 19.1 in the SB2 and to 20.2 in the SB3. Stabilization towards the expected number of 21 bivalents has proceeded rapidly, yet has still not been completed as reflected in the range (14-21 bivalents) and the pooled variance of the mean (Table 3, Sp2/n=0.14) of the SB3. The observed average number of univalents decreased from 21.0 in the SB1 to 7.9 in the SB2 to 2.4 in the SB3. This is rapidly approaching the zero value for univalents in A. intermedium. The observed average total chromosome number decreased from 47.0 in the SB1 to 45.0 in the SB2 to 43.0 in the SB3. This is very close to 2n=42 of A. intermedium. Observed and expected values were compared by a heterogeneity chi-square test (Table 2). In the SB1, heterogeneity was high among both the bivalents (chi2=106.51) and the univalents (chi2=221.93), and the goodness of fit to the expected values of 21 bivalents and 14 univalents was poor (p<0.005). In contrast heterogeneity for the total chromosome number in the SB1 was not very high (chi2=18.82) and the goodness of fit to the expected chromosome number of 2n=56 was acceptable but not exceptional (p>0.25). Heterogeneity for the bivalents decreased in the SB2 (chi2=11.32) with a good fit to the expected (p>0.95). However, the SB2 univalent fit was still poor, though heterogenity decreased (chi2=120.71). Also, the heterogeneity for the total chromosome number decreased in the SB2 with a very good fit to the expected (p>0.95). In the third backcross heterogeneity of bivalents, univalents, and total chromosome numbers decreased giving a good fit for bivalents and total chromosome number (p>0.95). Heterogeneity for the SB3 univalents was still high (chi2=56.45), with a poor fit (p<0.005) caused by a lower average (2.4I) than expected (3.5I) indicating that the elimination of univalents is proceeding at a faster rate than expected. Meiotic instability in this material is expressed in a high discrepancy between the observed bivalent number and the expected of 21 bivalents as well as in a high number of univalents. A high variance of chromosome numbers from cell to cell within plants and among plants of the same generation is another measure of meiotic instability. The pooled variances of the mean (Sp2/n) for univalent, bivalent and total chromosome numbers over all three generations are reported in Table 3. Table 4 shows the F values calculated from the pooled variances of the mean of Table 3 and their p-values are indicated. The variability of univalents, bivalents and total chromosome numbers decreased from the SB1 to the SB3. In the SB1 generation the variance of the univalents was much higher than that of the bivalents (F=20.45) and this difference decreased in the SB2 (F=2.07). However, the difference between variances of univalents and bivalents increased slightly after the third substitution backcross (F=2.94). This increase indicates that after an initial very rapid stabilization of univalents from the SB1 to the SB2 generation (F=42.21) the rate of stabilization has slowed down to such an extent that in the SB3 generation the univalents are still much less stable than the bivalents (Table 3). One would expect a larger p-value in the SB3. The differences between the variances of the univalent and the total chromosome numbers (2n) are smaller than the differences between the variances of the univalent and the bivalent numbers in all 3 generations. This indicates that there is a compensation between the numbers of univalents and bivalents which results in the appearance of a more stable total chromosome number. |
| 1) Contribution of the Montana Agricultural Experiment Station, Bozeman Montana, U.S.A. Paper No. 530, Journal Series, Montana Agr. Exp. Sta., published with approval of the Director. |
| --> Next |