Upland rice has been left without intensive studies on the inheritance of agronomic traits unlike lowland rice and there were very few reports on the genetic analysis of heading time. Since the Lm locus having multiple alleles controls primarily the varietal earliness and lateness of lowland rice (YOKOO et al. 1980), this report examined the role of the Lm locus in heading time of six native upland rice varieties. Under the natural day-length, the lowland rice tester lines ER with the early-heading gene Lm(e) and LR with the late-heading allele Lm(u) headed 84 and 105 days after sowing, respectively. Four upland rice varieties, Kurumi-wase, Hideri-shirazu, Kyushu and Wase-dango-mochi, headed at almost the same time as ER. Kuroka-mochi headed intermediately between ER and LR, and Kirishima did two days later than LR. Under the long day (14-hour) condition, LR headed 103 days after sowing and later than ER by 28 days, whereas under the short day (9-hour) condition LR headed 46 days after sowing and earlier than ER by 11 days. These differential responses of the two lines were due to the characteristics of the Lm(e) and Lm(u) alleles (YOKOO and KIKUCHI 1982). The heading order of the upland rice varieties under the long day condition followed the trend under the natural day-length (Table 1). Under the short day condition, all the six upland rice varieties headed four to 15 days later than ER. These photoperiodic reactions revealed that the four early-maturing varieties had the relatively larger basic vegetative growth period than ER or LR, and the medium- and late-maturing varieties were photoperiod-sensitive (Table 2). The larger basic vegetative growth period of upland rice showed the incomplete dominance in the crosses with ER or LR (Fig. 1). The weaker photoperiod-sensitivity was completely or partially dominant over the stronger one in the crosses with ER, but was incompletely dominant or partially recessive in the crosses with LR (Fig. 1). These differences in dominance reaction for photoperiod-sensitivity were possibly due to the distinctive reactions of ER and LR to the two day-lengths. Under the natural day-length condition in the lowland field at 36-degrees 4' North Latitude, the F2 populations of crosses between the four early-maturing varieties and ER showed the narrow heading variations, whereas the distributions in the crosses with LR were wide and bimodal in a 1 early : 3 late ratio with some transgressive segregations (Fig. 2, Table 3). Since ER and LR are isogenic except for the early allele Lm(e) of ER and the late allele Lm(u) of LR, the difference in heading variations of the F2 populations can be attributed to the different actions of Lm(e) and Lm(u). The earliness of Kurumi-wase, Hideri-shirazu, Kyushu and Wase-dango-mochi was controlled by Lm(e) or such an early Lm allele as Lm(e) and at least one more gene. The F2 populations of the crosses of Kuroka-mochi or Kirishima showed the continuous variations for heading time with the comparatively large transgressive segregations (Fig. 2). The mode of variation was around the values of ER in the crosses with ER and of LR in the crosses with LR. There were a very few earlier-heading plants like ER in the crosses with LR. These heading patterns suggested the heading time of Kuroka-mochi and Kirishima was controlled by a late Lm allele and at least two more genes. It was concluded from the above results that the earliness and lateness of the six native upland rice varieties also were mainly due to the Lm alleles, which were originally found in lowland rice.
