Microsoft Word - 6. BAJ-291_Revised Bangladesh Agron. J. 2018, 21(1): 61-70 SENSITIVITY OF ANNUAL WEEDS AGAINST METOLACHLOR+ BENSULFURON-METHYL HERBICIDE IN TRANSPLANTED RICE M. K. A. Bhuiyan*, M. M. Mahbub and M. Z. I. Baki Agronomy Division, Bangladesh Rice Research Institute (BRRI), Gazipur, Bangladesh *Corresponding author, E-mail: bhuiyan072003@yahoo.com (Received: 12 December 2017, Accepted: 16 September 2018) Keywords: Metolachlor + bensulfuron methyl, transplanted rice, annual weed, weed control Abstract Field trials were conducted at Bangladesh Rice Research Institute (BRRI), Gazipur during aman, 2014 and boro, 2014-15 to evaluate the efficacy of metolachlor + bensulfuron methyl 20% WP on weed suppression of transplanted rice. Metolachlor + bensulfuron methyl 20% WP @ 150, 190 and 230 g ha-1 were applied and pyrazosulfuran ethyl @ 125 g ha-1, weed free and unweeded control were the treatment variables. Visual observations indicated that this herbicide possesses high selectivity and not toxic to rice plants. The results revealed that the major weed flora associated with the transplanted rice was mainly comprised of two grasses, two sedges and three broadleaf in aman and two grasses, two sedges and two broadleaf in boro season. The most dominant weeds were Cyperus difformis, Echinochloa crus-galli, Scirpus maritimus and Monochoria vaginalis in both the growing seasons. Application of metolachlor + bensulfuron methyl 20% WP @ 190 g ha-1 was most effective to suppress weeds in both the seasons resulting in increased grain yield more than 40% as compared to unweeded control. Therefore, metolachlor + bensulfuron methyl 20% WP@ 190 g ha-1applied at one to two leaf stage of weed effectively controls the weed infestation in transplanted rice. Introduction Rice is a staple food of Bangladesh. About 160 million peoples in Bangladesh depend on rice as main food and about 75 percent of agricultural land use to grow rice. The average yield of rice in Bangladesh is 4.5 t ha-1 (BRRI, 2016). Bangladesh needs to feed 215.4 million people in 2050 (Kabir et al., 2015). In Bangladesh weed infestation reduces rice grain yield by 70-80% in aus rice, 30-40% in transplanted aman rice, 22-36% for modern boro rice cultivars and 80-82% in direct wet seeded rice (Bhuiyan, 2016; BRRI, 2006; Mamun, 1990). According to Willocquet et al., (1998) and Bari, (2010) that Yield losses due to the infestation of weeds are greater than the combined losses caused by insect, and diseases in rice. Weeds not only cause huge reductions in rice yields but also increase the cost of cultivation, reduce input efficiency, interfere with agricultural operations, impair quality, act as alternate hosts for several insect pests, diseases, they affect the aesthetic look of the ecosystem as well as native biodiversity, affect human and cattle health (Bhuiyan et al., 2017). Weeds compete for nutrient, space, sunlight and moisture with rice resulting in crop yield loss (Sunyob et al., 2015). Weed management in rice production is a major constraint and expensive. Since hand weeding and other weed control methods are labour intensive and complicated, chemicals are the obvious and cost-efficient weed control practices. Chemical weed control has become popular in Bangladesh mainly due to the scarcity of labour during peak growing season, and lower weeding cost. Now a days the use of herbicides is gaining popularity in rice weed control due to their rapid effects and less cost involvement compared to traditional methods of weeding. Pre and post emergence herbicides such as butachlor, pretilachlor, oxadiazone, pyrazosulfuron ethyl, ethoxysulfuron 62 Bhuiyan et al. alone or supplemented with one hand weeding have been found to be useful for weed management in transplanted paddy. Use of single herbicide might be effective for only sedges or only grass or broadleaf weeds. Metolachlor + bensulfuran methyl has been recently developed for post emergence control of weeds in the rice field. So, there is a need to develop environment friendly molecules of newer chemistries with a different mode of action. The present study was, therefore, planned to evaluate the efficacy of metolachlor + bensulfuran methyl for rice weed suppression and find out an appropriate dose of the herbicide and its impacts on transplanted rice. Materials and Methods The experiments were conducted at the Bangladesh Rice Research Institute, Gazipur, situated at 23059´33´´ N and 90024´19´´ E at an elevation of 8.4 m from the mean sea level, and is characterized by sub-tropical climate during aman, 2014 and boro, 2014-2015 seasons. The soil of the experimental site was non-calcareous dark grey floodplain with pH around 6.2 and low in organic matter (1.2%). The experiment was carried out with six (6) treatments viz. T1= Metolachlor + bensulfuran methyl 20% WP@ 150gha-1 (7.5 g a.i. ha-1), T2= Metolachlor + bensulfuran methyl 20% WP @ 190 g ha-1(9.5 g a.i. ha-1), T3= Metolachlor + bensulfuran methyl 20% WP @ 230 gha-1(11.5 g a.i. ha-1), T4= Pyrazosulfuran ethyl @ 125 g ha-1(12.5 g a.i. ha-1), T5= Weed free by hand weeding, and T6=Control (Unweeded). All treatments were laid out in a randomized complete block design with three replications. Twenty-five days of BRRI dhan49 for aman, and thirty-five days old seedlings of BRRI dhan28 for boro were transplanted at 20 x 20 cm spacing with 2 seedlings hill-1. Fertilizers were applied as aman: N:P:K:S= 69:10:41:11 kg ha-1 and boro; N:P:K:S= 120:19:60:24 kg ha-1 (BRRI, 2013). All herbicides were sprayed at 1 to 2 leaf stage of weed (7 DAT in aman and 12DAT in boro) with the help of a knapsack sprayer. In weed free treatment, the plots were kept weed free up to 50 DAT by hand weeding. Metolachlor + bensulfuran methyl herbicide is innovative in Bangladesh and its phytotoxicity needs to be evaluated on rice crop. The phytotoxicity of the herbicide to rice plants was determined by visual observations (yellowing leaves, burring leaf tips, stunting growth, etc.). The degree of toxicity on rice plant was measured by the following scale used by IRRI (1965). Rating Visual Symptom Effect 1 No effect No toxicity 2 Yellowing leaves Slightly toxicity 3 Yellowing leaves and stunting Moderate toxicity 4 Burning leaves and leaf tips Toxicity 5 Burning plant, plant kill Highly toxicity The rating of toxicity was done within 7 days after application of herbicides. It was observed three times at 3, 5 and 7 days after application of herbicide and the mean rate was calculated from 10 sample plants of a unit plot. Data on weed density and dry weight were taken from each plot on 40 DAT. The weeds were identified species-wise. Dry weights of weeds were taken by drying them in electric oven at 600 C for 72 hours followed by weighing by digital balance. Relative weed density (RWD), relative weed biomass (RWB), The sum dominance ratio (SDR) and weed control efficiency (WCE) of different weed control treatments were calculated with the following formulas (Janiya and Moody, 1989; Rao ,1985): RWD = Density of individual weed species in the community Total density of all weed species in the community × 100 Sensitivity of Annual Weeds Against Metolachlor 63 RWB = Dry weight of a given oven dried weed species Dry weight of all oven dried weed species × 100 SDR = RWD(%) + RWB(%) 2 WCE(%) = ୈ୰୷ ୵ୣ୧୥୦୲ ୭୤ ୵ୣୣୢୱ ୧୬ ୵ୣୣୢ୷ ୡ୦ୣୡ୩ ୮୪୭୲ୱିୈ୰୷ ୵ୣ୧୥୦୲ ୭୤ ୵ୣୣୢୱ ୧୬ ୲୰ୣୟ୲ୣୢ ୮୪୭୲ୱ ୈ୰୷ ୵ୣ୧୥୦୲ ୭୤ ୵ୣୣୢୱ ୧୬ ୵ୣୣୢ୷ ୡ୦ୣୡ୩ ୮୪୭୲ୱ ×100 Data on panicle m-2, grains panicle-1, sterility and grain yield were collected. Yield attributes and yield data were analyzed with STAR 2.0.1 software for analysis of variance and graphical presentation. Results and Discussion Phytotoxicity of herbicides on rice plant The degree of toxicity of the herbicide to rice plants and the symptoms produced on the plant are presented in Table 1. It is observed that metolachlor + Bensulfuran methyl 20% WP@ 150 g ha-1 showed no toxicity, and metolachlor + bensulfuran methyl 20% WP@ 190 g ha-1 showed slight yellowing of leaves while metolachlor + bensulfuran methyl 20% WP@ 230 g ha-1 showed temporary yellowing of leaves. Results showed that phytotoxicity symptoms were not more prominent for using this herbicide. Phytotoxicity of rice plant found by combined herbicide is less which is like the findings of Bhuiyan et al. (2010). Table 1. Rating of herbicide toxicity on rice plant under different treatments in field condition Treatments Rating Symptom observed in rice field Aman Boro Metolachlor + bensulfuranmethyl 20% wp @ 150 gha-1 (7.5 g a.i. ha-1) 1.10 1.10 No toxicity Metolachlor + bensulfuran methyl 20% wp @ 190 gha-1 (9.5 g a.i. ha-1) 1.14 1.17 Sometimes slight yellowing of leaves Metolachlor + bensulfuran methyl 20% wp @ 230 gha-1(11.5 g a.i. ha-1) 1.90 1.85 Slight yellowing of leaves which required 5-7 days to recover Pyrazosulfuran ethyl @125 gha-1 (12.5 g a.i. ha-1) 1.12 1.10 No toxicity Weed infestation Aman season The experimental field was infested with eight different types of weed species belonging to different families of Poaceae, Cyperaceae, Pontederiaceae, Marsileaceae, Sphenocleaceae and Asteraceae. The relative density of these weed species was also different (Table 2).Among the infesting weeds, twowere grasses, two sedges, and four broadleaf. The weed species belonging to the families of broadleaf were: Monochoria vaginalis, Marsilea minuta, Sphenoclea zeylanicaand Eclipta alba; grasses were: Echinochloa crus-galli, Cynodon dactylon, and sedges were Cyperus difformis and Scirpus maritimus. Among the weed species maximum relative weed density (RWD) was observed for Cyperus difformis (31.24%) followed by Echinochloa 64 Bhuiyan et al. crus-galli (30.40%) but the highest relative weed biomass (RWB) observed for E. crus-galli (33.75%) followed by C. difformis (33%). Eclipta albawas the minor weed with 3 and 4 RWD and RWB, respectively. It was also observed that broad leaved were less dominating weed species. Priya et al., 2017 revealed that application of herbicide combination, bispyribac sodium + metamifop 14% SE at 70 g ha-1 recorded significantly lower total weed density and and higher weed control efficiency. Boro season The number of infesting weed species was slightly different in boro than aman. These weed flora were ecologically categorized into two broad leaved species, two sedge and two grasses (Table 3). The major weed was C. difformis which (RWD) and (RWB) was 32.50% and 34.07%, respectively. The second top weed was E. crus-galli which RWD was 31.48% and (RWB) was 33.76%. So, in boroseason broadleaf weeds were less dominating than aman season. Combination of Trisulfuron + Pretilachlor effectively control Echinochola sp. and Cyperus sp. which found by Puniya et al., 2007. Table 2. Weed composition, relative weed density (RWD), and relative weed biomass (RWB) in aman, at Gazipur Name Family Type RWD (%) RWB (%) Cynodon dactylon Poaceae Grass 7.65 9.63 Echinochloa crus-galli Poaceae Grass 30.40 32.75 Cyperus difformis Cyperaceae Sedge 31.24 32.45 Scirpus maritimus Cyperaceae Sedge 24.76 23.92 Monochoria vaginalis Pontederiaceae Broadleaf 22.28 27.29 Marsilea minuta Marsileaceae Broadleaf 11.47 12.64 Sphenoclea zeylanica Sphenocleaceae Broadleaf 3.81 2.58 Eclipta alba Asteraceae Broadleaf 2.84 3.45 Table 3. Weed composition, relative weed density (RWD), and relative weed biomass (RWB) in boro, at BRRI, Gazipur Name Family Type RWD (%) RWB (%) Cynodon dactylon Poaceae Grass 10.76 9.21 Echinochloa crus-galli Poaceae Grass 31.48 33.76 Cyperus difformis Cyperaceae Sedge 32.50 34.07 Scirpus maritimus Cyperaceae Sedge 28.16 27.63 Monochoria vaginalis Pontederiaceae Broadleaf 21.42 26.74 Marsilea minuta Marsileaceae Broadleaf 10.90 12.65 Weed ranking The summed dominance ratio (SDR) is more informative than any single measure in reflecting the contribution of a species in the community. The most dominant weeds in aman seasonwere C. difformis, E. crus-galli, S. maritimus and M. vaginalis (Fig. 1). Cyperus difformis, Echinochloa crus-galli, Scirpus maritimus and Monochoria vaginalis were also the most dominant weeds in boro season. Mamun et al., 2011 showed that SDR of a weed against the same herbicide is similar in different seasons. Sensitivity of Annual Weeds Against Metolachlor 65 Fig.1. Summed dominance ratio (SDR) of weed species in transplanted rice Weed control efficiency (WCE) Lower weed biomass as well as higher weed control efficiency in all the growing seasons exhibited by Metolachlor + bensulfuran methyl. Weed control efficiency improved with increases of herbicide dose irrespective of weed species. Treatment, T1 controls all the weeds less than 80% due to a lower dose of application, whereas T2, T3 and T4 (check) control E. crus- galli, C. difformis, S. maritimus and M. vaginalis more than 80% in aman season (Table 4). The trend of weed control efficiency in Boro, 2014-15 was almost similar to Aman, 2014. Treatment T2, T3 and T4controls E. crus-galli, C. difformis, S. maritimus and M. vaginalis more than 80% (Table 5) in boro season. It was evident from the study that the post emergence herbicide Metolachlor + Bensulfuran Methyl @ 190 gha-1 and 230 g ha-1 was effective for controlling weed than other doses of that herbicide. Lodhi et al., 2017 reported that application of Bensulfuron methyl + Pretilachlor Control weeds more than 80% in transplanted rice in India. Table 4. Effect of Metolachlor + Bensulfuran methylon weed control efficiency in transplanted rice in aman at BRRI, Gazipur Name of weeds Weed control efficiency (%) T1 T2 T3 T4 Cynodon dactylon 40.54 52.30 51.26 45.80 Echinochloa crus-galli 62.74 81.36 83.25 84.61 Cyperus difformis 68.52 83.20 85.67 81.54 Scripus maritimus 72.49 82.36 86.32 83.72 Monochoria vaginalis 68.43 80.71 80.87 80.59 Marsilea minuta 51.30 62.80 64.38 69.26 Sphenoclea zeylanica 62.14 75.33 68.49 72.65 66 Bhuiyan et al. Eclipta alba 52.62 60.45 62.92 62.75 T1= Metolachlor + bensulfuran methyl 20% WP @ 150 g ha-1, T2= Metolachlor + bensulfuran methyl 20% WP @ 190 g ha-1, T3= Metolachlor + bensulfuran methyl 20% WP @ 230 g ha-1 and T4= Pyrazosulfuran ethyl @ 125 g ha-1 Table 5. Effect of Metolachlor + Bensulfuran Methylon weed control efficiency in transplanted rice in Boro, 2014-15 at BRRI, Gazipur Name of weeds Weed control efficiency (%) T1 T2 T3 T4 Cynodon dactylon 37.94 46.39 48.36 52.45 Echinochloa crus-galli 66.34 83.68 84.09 83.35 Cyperus difformis 70.45 81.60 85.20 85.60 Scripus maritimus 68.25 83.35 84.25 83.68 Monochoria vaginalis 63.16 80.65 81.85 83.40 Marsilea minuta 45.10 57.70 65.50 59.30 T1= Metolachlor + bensulfuran Methyl 20% wp @ 150 g ha-1, T2= Metolachlor + bensulfuran Methyl 20% wp @ 190 g ha-1, T3= Metolachlor + bensulfuran Methyl 20% wp @ 230 g ha-1 and T4= Pyrazosulfuran ethyl @ 125 g ha-1 Yield and yield attributes From Table 6 it was found that all the treatments significantly increased rice grain yields over unweeded condition. In aman season, the highest grain yield (5.18 t ha-1) was recorded in the weed free treatment which was statistically similar to treatments T2 and T4 (5.05 and 4.98 t ha-1, respectively). Minimum grain yield (3.21 t ha-1) was found from BRRI dhan49 in weedy check plots as compared to weed-free treatment due to high weed density which resulted from a smaller number of panicle m2, grains panicle-1 and high sterility. Treatment wise box plot of yield attributes in aman season confirm that most of the yield contributing parameters are showed similar range in T2, T4 (check) and T5 (weed-free) treatments; whereas T6 was outside of the normal range and its data was also in dispersing condition than other treatments due to severe weed infestation (Fig. 2). The similar trend of results was observed during the boro season where unweeded control (T6) produced the minimum number of panicles m-2, grains panicle-1 and high sterility which resulting the lowest (3.37 t ha-1) grain yield of BRRI dhan28. The minimum number of panicles m-2 in the control plot was the result of higher competition for nutrient, space, light and water between crop plants and weeds. Maximum grain yield of 5.71 t ha-1 that was recorded with T2treatment could be due to lower weed-crop competition at crop growth stages. In Boro season, T2, T4 (check) and T5 (weed-free) treatments are in similar range in boxplot of yield attributes (Fig. 3). Metolachlor + bensulfuran methyl @ 150, 190, 230 gha-1gave effective control of grass, sedge and broad leaf weeds lead to increased grain yield.Herbicide treatments contributed to higher yield performance compared to control in all the growing seasons (Bari, 2010). Table 6. Effect of Metolachlor + bensulfuran methyl on the yield attributes and yield of transplanted rice at BRRI, Gazipur Treatments Panicles m-2 Grains panicle-1 Sterility (%) Grain yield (t ha-1) BRRI dhan49 BRRI dhan28 BRRI dhan49 BRRI dhan28 BRRI dhan49 BRRI dhan28 BRRI dhan49 BRRI dhan28 T1 222 264 105 114 17.43 16.03 4.93 5.44 T2 228 281 105 114 16.60 14.87 5.05 5.71 T3 216 258 98 111 18.13 16.33 4.72 5.07 T4 237 283 103 114 17.27 15.27 4.98 5.69 T5 236 287 107 117 16.23 14.60 5.18 5.75 Sensitivity of Annual Weeds Against Metolachlor 67 T6 186 214 78 93 19.77 19.07 3.21 3.37 CV (%) 8.57 3.12 3.08 4.57 4.24 4.20 2.73 2.45 LSD (0.05) 11.08 14.99 5.57 9.21 1.45 1.22 0.23 0.23 T1= Metolachlor + Bensulfuran Methyl 20% wp @ 150 gha-1, T2= Metolachlor + Bensulfuran Methyl 20% wp @ 190 gha-1, T3= Metolachlor + Bensulfuran Methyl 20% wp @ 230 gha-1, T4= Pyrazosulfuran ethyl @ 125gha-1, T5= Weed free and T6= Unweeded (control) Fig. 2. Box plot of yield attributes in BRRI dhan49 at BRRI, Gazipur T1= Metolachlor + Bensulfuran Methyl 20% wp @ 150 g ha-1, T2= Metolachlor + Bensulfuran Methyl 20% wp @ 190 g ha-1, T3= Metolachlor + Bensulfuran Methyl 20% wp @ 230 g ha-1, T4= Pyrazosulfuran ethyl @ 125 g ha-1, T5= Weed free and T6= Unweed (control) 68 Bhuiyan et al. Fig. 3. Box plot of yield attributes in BRRI dhan28 at BRRI, Gazipur T1= Metolachlor + Bensulfuran Methyl 20% wp @ 150 g ha-1, T2= Metolachlor + Bensulfuran Methyl 20% wp @ 190 g ha-1, T3= Metolachlor + Bensulfuran Methyl 20% wp @ 230 g ha-1, T4= Pyrazosulfuran ethyl @ 125 g ha-1, T5= Weed free and T6= Unweed (control) Sensitivity of Annual Weeds Against Metolachlor 69 Conclusion Results reveal that yield and yield attributes, and weed dynamics were greatly influenced by different weed management practices. Metolachlor + bensulfuran methyl 20% WP @ 230 g ha-1 showed better weed control efficiency although it has slight phytotoxicity. On the other hand, Metolachlor + Bensulfuran methyl 20% WP @ 190 g ha-1 applied at 1 to 2 leaf stage would be more effective for controlling annual weeds in transplanted rice. References Bari, M. N. 2010. Effects of herbicides on weed suppression and rice yield in transplanted wetland rice. Pakistan J. Weed Sci. Res. 16(4): 349-361. BBS. 1991. Statistical Yearbook of 1990. Bangladesh Bureau of Statistics, Ministry of Planning, Dhaka. Bhuiyan, M. K. A., M. M. Mahbub, Z. I Baki and L. Nahar. 2017. Sensitivity of Annual Weeds against Sulfentrazone 48 SC herbicide in Rice Cultivation. Bangladesh Rice J. 21 (1) : 67- 76 Bhuiyan, M. K. A. 2016.Effect of resource conservation technologies and weed management on nitrogen use efficiency, growth and yield of rice. (PhD thesis), BAU, Mymensingh. Bhuiyan, M. K. A. and G. J. U. Ahmad. 2010. Performance of mefenaset + bensulfuron methyl 53%WP against weed suppression in transplanted paddy. Pakistan J. Weed Sci. Res. 16 (2):181-187. BRRI (Bangladesh Rice Research Institute). 2006. Weed identification and management in rice. Bangladesh Rice Research Institute, Joydebpur, Gazipur, Bangladesh. BRRI (Bangladesh Rice Research Institute). 2013. Adhunik Dhaner Chas (in Bengali). Bangladesh Rice Research Institute, Joydebpur, Gazipur, Bangladesh. BRRI (Bangladesh Rice Research Institute). 2016. Adhunik Dhaner Chas (in Bengali). Bangladesh Rice Research Institute, Joydebpur, Gazipur, Bangladesh. IRRI (International Rice Research Institute). 1965. Annual Report for 1963. IRRI, Los Banos, Philippines, pp.224-231. Janiya, J. D. and K. Moody. 1989. Weed populations in transplanted and wet-seeded rice as affected by weed control methods. Trop. Pest Manag.35(1) 8 – 11. Jayadeva, H. M. 2010. Bioefficacy of post emergent herbicides in weed management of transplanted rice (Oryza sativa L.). J. Crop Weed. 6(2): 63-66. Kabir, M.S., M.U. Salam, A. Chowdhury, N. M. F. Rahman, K. M. Iftekharuddaula, M. S Rahman, M. H Rashid, S. S Dipti, A. Islam, M. A Latif, A. K. M. S. Islam, M. M. Hossain, B. Nessa, T. H. Ansari, M. A. Ali, J. K. Biswas. 2015. Rice Vision for Bangladesh: 2050 and Beyond. Bangladesh Rice J.19(2): 1-18 Lodhi, R., A. rawat, A. K. Jha, J. dubey.2017. Efficacy of bensulfuron methyl + pretilachlor against weeds in transplanted rice. Global J. Bio-sci. Biotechnol. 6 (2): 285-288. Mamun, M. A. A., A. J. Mridha, A. Akter, A. Parvez. 2011. Bio-efficacy of acetochlor 50% EC against weed suppression in transplanted rice ecosystem. J. Environ. Sci. Nat. Resour. 4(2):73 –77. Mamun, A.A. 1990. Weeds and their control: A review of weed research in Bangladesh. Agricultural and Rural Development in Bangladesh. Japan Intl. Co-operation agency. Dhaka, Bangladesh. JSARD. Pub. No. 19. pp. 45-72 70 Bhuiyan et al. Priya, R. S., C. Chinnusamy, P. Murali Arthanari and P. Janaki. 2017. Weed control in direct seeded rice using new herbicide combination under Indian tropical condition. British J. Appl. Sci. Technol. 21(5): 1-14. Puniya, R., P.C. Pandey and P. S. Bisht. 2007. Performance of trisulfuron, trisulfuron+ pretilachlor and bensulfuron-methyl in transplanted rice (Oryza sativa L.). Indian J. Weed Sci.39 (1 & 2): 120-122. Rao, A. N. 1985: Weed vegetation analysis. Lecture presented to participants attending the Weed Science Workshop, IRRI, Los Baños, Philippines. Sunyob, N. B., A. S. Juraimi, M. A. Hakim, A. Man, A. Selamat and M. A. Alam. 2015. Competitive ability of some selected rice varieties against weed under aerobic condition. Intl. J. Agric. Biol. 17: 61-70. Willocquet L, Savary 5, Fernandel L, Elazegui F, Teng P. 1998. Simulation of yield losses caused by rice diseases, insects, and weeds in tropical Asia IRRI Discussion Paper Series No. 34. Manila (Philippines): Int. Rice Res. Inst. p.62.