RJ. ZHANG1, S. MING1, C.W. XU1, L.S. YANG1, Y.S. BAI1, C.Q. SUN2 and X.K. WANG2
1) Anhui Academy of Agricultural Sciences, Hefei, 230031 China
2) China Agricultural University, Beijing, 100094 China
Utilization of heterosis of F1 hybrid between Japonica and Indica has been one of the main directions of hybrid rice breeding in the recent years. But several problems of F1 hybrid between Japonica and Indica have not been solved; for example, tall plant height. long growth period, low seed set and low degree of grain filling, etc. Thus, direct utilization of F1 hybrid between Japonica and Indica has been restricted. Under such situation, use of Japonica-like rice (J') which is not typical but Japonica-inclined type is expected to be one of the effective ways to improve grain yield of hybrid rice.
Twelve parents (Table 1) were crossed according to NC II design and thirty-two F hybrids were obtained. The F1 s, their parents, and CK Shanyou 63 were grown in Hefei in midseason in 1995, in three replications of a randomized block design. Twelve parents were classified by Cheng's index (classification index of Indica and Japonica rice. Cheng 1993). Eight characters, i.e., days to heading, plant height, panicle length, number of effective tillers, number of grains per panicle, seed set, 1000-grain weight and grain yield per plant were observed in F1 s and parents.
The effects of female, male and their interaction on hybrids were estimated in terms of % variance. As shown in Table 2, the male effect was greater than female and interaction effects in all characters except for 1000-grain weight in which female effect was greater than the sum of male and interaction effect. Therefore, characters carried by R-line (male parent) was extremely important in breeding of Japonica Two-line hybrid rice.
Relative heterosis (RH) and standard heterosis (SH) were also estimated (Table 3). Positive crosses outnumbered negative crosses in all characters for RH, while only in three characters for SH. It was found that degree of heterosis remarkably diferred among crosses. In number of grains per panicle, heterosis ranged from -1.1% to 107.9% with a mean 33.5% in RH, and from -17.1% to 82.7% with a mean 23.1% in SH. In grain yield, SH showed a range of -36.8% to 33.6% with a mean -2.4%. SH of seed setting rate ranged from -23.5% to 3.4% with a mean -7.4%.
Next, heterosis was compared among three cross-combination groups; J x J, J x J', and J x I. In general, heterosis showed an order of J x J < J x J' < J x I, indicating that remotely related parents showed high degree of heterosis as expected. The differences in Cheng's index between the two parents were significantly correlated with respective RH (r = 0.52-0.72). The cases with number of effective tillers and number of grains per panicle are shown in Fig. 1. In SH, however, number of grains per panicle showed J x J < J x J' > J x I, suggesting that a large number of grains per panicle which is one of the important yielding components is expected in J x J' crosses. But seed set and 1000-grain weight showed a trend of J x J > J x J ' > J x I. The above results suggest that J x J' cross may contribute to increase number of spikelets per panicle but all spikelets are not always fully filled.
Japonica PGMS line 4008s, Indica R-line PA64 and Japonica-like WCV YH101 had higher G.C.A. (general combining ability) value of grain yield. Hybrids with high S.C.A. (special combining ability) value of grain yield were 80-4B/YH101, 80-4B/02428 and 4008s/HP214. The difference in Cheng's index between the two parents had very little effect on S.C.A. values of seven characters, correlation coefficients ranging from -0.02 to 0.10 except for effective tillers (-0.36).
Cheng Kan-Sheng, 1993. Discrimination of Indica and Japonica subspecies in Asian cultivated rice. Yunnan Science and Technology Press. Kunming, China, pp.45.