To understand the effectiveness of tip-clipping in the production of perilla (Perilla frutescens (L.) Briton var. frutescens) grown in the Tokachi region, we compared the growth and yield of plants receiving tip-clipping treatment at different growth stages. Tip-clipping was performed on a local variety, Hokkaido-zairai, at the eight-leaf stage (V8), ten-leaf stage (V10), or flower bud formation stage (R). Plants without tip-clipping treatment were grown in the control plot. The plants in the V8 and V10 plots had fewer primary branches, but more inflorescences than those in the control plot. The total number of florets in the V10 plot was similar to that in the control plot. This may be due to the disruption of apical dominance or increase in allocation of resources to the branches. The plants in the V8 and V10 plots were shorter, resulting in a lower degree of lodging than those in the control plot. On the other hand, the plant form and degree of lodging in the R plot were similar to those in the control plot. Tip-clipping did not show a significant effect on the yields of grain and oil. However, the yield may be reduced due to harvest loss in the control and R plot that showed severe lodging. Our results indicated the effectiveness of tip-clipping for stabilizing perilla production in the Tokachi region, and the appropriate growth stage for tip-clipping was suggested to be the ten-leaf stage.

The duration of the maturity phase, time from heading to the optimum harvest time after grain growth ceased, is important for rice productivity. However, little is known about the determinants for cultivar differences in the duration of the maturity phase due to difficulty of determining the optimum harvest time non-subjectively. The objective of this study was to develop a method for measuring the spikelet yellowing ratio, which is an index of optimum harvest time, by image analysis, and to elucidate the determinants for cultivar differences in the duration of the maturity phase. The relationship between grain number, grain weight, and ripening period was analyzed using ‘Koshihikari’ and its near isogenic lines (NILs) with different numbers of spikelets; and ‘Yamadawara’ and ‘Emidawara’, which have similar genetic backgrounds but differ in duration of maturity phase. The results showed that the number of grains was not always the determining factor in the duration of maturity phase. On the other hand, the analysis of each panicle revealed that the increase in grain weight almost completely ceased when spikelet yellowing began. Furthermore, the duration of the maturity phase was more strongly correlated with the lag time from the end of grain weight increase to maturity, than the duration of grain weight increase. These results suggested that this lag time may also be a determinant for the duration of the maturity phase.

Direct planting—in which storage roots are planted instead of transplanting stem cuttings—has been studied as a labor-saving system in sweet potato. However, weeding is necessary until the shoots emerge. Therefore, it is important to estimate the days to emergence (DTE) for planning weeding practice. In this study, we estimated DTE based on growing degree days (GDD) calculated from soil temperature in direct planting of sweet potato. Seven genotypes were directly planted on four dates from 2018 to 2019, and the average daily soil temperature and date of emergence were investigated. Using these data, we calculated GDD for emergence of seven genotypes with lower limits of soil temperature of 8, 10, 12 and 15°C and upper limits of 20, 22, 24, 26 and 28°C. Then we decided the best combination of lower and upper limits of soil temperature that minimized the root mean squared error (RMSE) of the difference between the observed and predicted DTE. Consequently, the combination with a lower limit of 10°C and an upper limit of 24°C had the smallest RMSE, i.e., 2.48 days. This prediction model was also validated using data obtained by planting on three dates in 2020. The average RMSE in 2020 was 3.01 days, indicating that DTE in direct planting could be predicted with an error of about 3 days.

In organic paddy-rice cultivation, weed control is important for preventing the reduction of rice yield but is a labor-intensive task. Therefore, easier and more stable weed-control technology is required. In this study, an automatic weed suppression robot for paddy fields ("weed suppression robot") powered by solar energy and GPS-navigated, was tested in 36 organic paddy rice fields for two years. The weed suppression robot reduced the number of mechanical weeding operations by an average of 58%. The average estimated weed dry-matter weight at the panicle formation stage was 16.6 g m^-2 and the weeds were controlled enough to prevent the reduction of rice yield. The weed suppression robot tended to be more effective on annual weeds than on perennial weeds. This robot was considered to have little effect against perennial weeds, which have a long emergence period, because it could be used for only about three weeks after transplanting paddy rice. Therefore, additional physical control such as mechanical weeding would be necessary for further control of perennial weeds. The average yield was 424 g m^-2, a 10% increase compared to practical organic cultivation. These results suggest that the weed suppression robot is an effective new weeding tool for reducing weeding labor while ensuring yield in organic paddy rice cultivation.

The intra-field variation of soybean growth and its relations to the yield were evaluated using the normalized difference vegetation index (NDVI) measured with a handheld crop sensor ‘GreenSeeker’. NDVI of the soybean variety ‘Satonohohoemi’ cultivated in 2021 and 2022 was evaluated. In both years, the plots with a large number of seedlings per area showed high NDVI at the early growth stage. The plots with a high NDVI at the early growth stage indicated heavy aboveground dry weight at the flowering stage and at the early seed growth stage in both years. However, NDVI at the early growth stage was not related with yield in both years. At the late maturing stage, NDVI showed error values in the two plots among 30 plots because of early leaf withering. Although the yields of those two plots were not always low, the hundred-seed weights were clearly low. Those results suggest that NDVI at the early growth stage can be used to estimate the intra-field variations of soybean growth until the early seed growth stages to some extent and NDVA at the late growth stage can detect the poor maturing plots.

A method was developed to estimate individual dry weight of soybean cultivar ‘Satonohohoemi’ using the main stem diameter. Specifically, a quadratic regression model was developed for each of six developmental phases (1: VC to V3, 2: V4 to V6, 3: vegetative phase after V7, 4: R1 to R3, 5: R4 to R6, 6: R7 to R8), using main stem diameter as an explanatory variable and individual aboveground dry weight as a response variable. The estimation accuracy was examined using the serial numbers of the plots in which the cropping season and planting density were varied (from April to August, about 6 to 38/m²) as the random effect. The estimation accuracy examined with the random effect was not significantly different from that examined without it, suggesting that the difference in the relationship between the main stem diameter and aboveground dry weight with the cropping season was small. Therefore, the estimation equation was considered to be applicable in a wide range of conditions. When the random effect was removed and analyzed by dividing the developmental phases in detail, a prediction could be made with a weight error of 11% to 37%, although there was some difference with the developmental phase. The change in the slope of the estimation equation by developmental phase was considered to represent the ratio of the growth of other organs to the enlargement growth of the stem, which increased after V7, peaked during R6, and decreased in R7 to R8.

To predict the jointing stage of wheat,we compared the accuracy of a new three-parameter model equation that considers the vernalization and photoperiod response(modified Weir’s model)and an existing five-parameter model equation(Horie-Nakagawa’s model) that applies the development prediction for rice to wheat.Validation was conducted using the growth data of‘Norin61’,’Shirogane-Komugi’(spring growth habit),’Satonosora’ and ’Iwainodaichi’(winter growth habit)cultivars grown between 2006 and 2019 in Tsukuba,Ibaraki Prefecture Tsu,Mie Prefecture Fukuyama,Hiroshima Prefecture and Zentsuji,Kagawa Prefecture.The parameter representing the degree of low temperature requirement in the modified Weir’s model equation reflected the winter habit of the cultivar. Both models achieved a good fit for the jointing dates with a root mean square error of 4.8 days averaged over the four cultivars for the two models, with a slightly better fit for the modified Weir’s model in the case of ‘Satonosora’.The minimum number of data points required to determine the parameters was approximately 15.The accuracy improved as the number of data points increased, but little change was observed when it exceeded 40.To improve the accuracy of the model, it is important to determine the parameters not only by increasing the number of data points but also by including data obtained in a wide range of temperatures and photoperiods at various sowing dates and locations.