Vous êtes sur la page 1sur 200

Research programs Irrigated rice ecosystem

BREEDING TO BREAK YIELD CEILINGS: A SYSTEMS APPROACH 4 Radiation use efficiency in rice (APPA) 4 Peculiar growth patterns on route to high yields (APPA) 6 Progress in increasing grain yield by breeding a new plant type (APPA, PBGB) 7 Hybrid rice 9 Hybrid rice network (PBGB) 9 Release of IRRI hybrids (PBGB) 9 Elite rice hybrids (PBGB) 9 New cytoplasmic male sterile (CMS) lines (PBGB) 11 Development of thermosensitive genic male sterile (TGMS) lines (PBGB) 11 IR lines named as varieties (PBGB) 11 Importance of nitrate vs ammonium for potential yield (SWS) 12 Identifying new genes with durable resistance to tungro viruses (EPP) 13 Development of rice blast-resistant monogenic lines (PBGB) 13 SUSTAINING SOIL QUALITY IN INTENSIVE RICE SYSTEMS 15 Variability of optimal nitrogen applications for rice 15 Optimal N rates (SWS, SS) 15 Spatial variability at the same site (SWS, SS) 16 Internal nutrient efficiencies in irrigated lowland rice 17 Current situation in farmers fields (SWS) 17 Modeling the nutritional balance (SWS) 17 Case study: Maligaya, Central Luzon, Philippines (SWS) 19

INCREASING WATER USE EFFICIENCY IN RICE CULTURE 20 Management of cracked soils for water saving during land preparation (SWS) 20 Population dynamics of Echinochloa crus-galli in wetseed rice (APPA) 22 Semidry rice production for irrigation water savings (SWS) 24 IMPROVING PEST MANAGEMENT 26 Motivating farmers to test pest management changes through use of printed materials and radio (EPP) 26 Testing the validity of surrogate taxa for canopy and floodwater invertebrates (EPP) 26 Directional movement of predators between the irrigated rice field and its surroundings (EPP) 27 Seed health evaluation for farmer rice crop management (EPP) 28 COPING WITH GLOBAL CLIMATE CHANGE: REDUCING METHANE EMISSION FROM RICE FIELDS (SWS) 28 IRRIGATED RICE RESEARCH CONSORTIUM 29 PROGRESS OF UNREPORTED PROJECT 30 Improving the productivity and sustainability of ricewheat systems (SWS, APPA, EPP) 30 PROGRAM OUTLOOK 30

Irrigated rice ecosystem

Irrigated rice ecosystem

Rice production must be increased by 70% over the next three decades to meet growing demand for food and to sustain the global food security. For the foreseeable future, the bulk of that demand will continue to be met by intensive irrigated rice systems. Because of limited opportunities for expansion of ricegrowing areas plus increased urbanization to further reduce rice areas, reduced labor available for agriculture due to its shifting to nonagricultural sectors, and increased scarcity of water resources, it is essential to increase rice productivity per unit of area, per unit of labor, and per unit of input to meet the projected growth in demand. The irrigated rice ecosystem research is premised on increasing rice production through 1) shifting the yield frontier upward, 2) increasing the efficiency of crop production inputs, 3) sustaining the rice base for irrigated lowlands, and 4) minimizing the effects of the global climate change on the external environment.

Breeding to break yield ceilings: a systems approach

Radiation use efficiency in rice J. Sheehy, J. Dionora, A. Ferrer, and R. Torres The physical limits to rice yield are ultimately set by the absorption and conversion of light and the maximum diffusion rate of CO2. Two approaches were taken to answer the question whether a thermodynamic barrier prevents increases in yield beyond 10 t ha-1. q The concept of radiation conversion to biomass in rice was examined theoretically and experimentally. q Estimates of the conversion obtained throughout a growth cycle in irrigated rice growing in the dry season (DS) with abundant fertilizer were compared. A general equation for a rice crop can be developed over the growing season from transplanting (tr) to maturity (tm) so that Y = H tm (Q(t) f(t))dt tr (1)

where Y is the yield of required parts of the plant (g m-2); H is the fraction of crop dry matter (DM) that becomes the yield, i.e., the harvest index; is the radiation conversion factor (g DM MJ-1), t is time, tr is date of transplanting, and tm is time of maturity; Q(t) is photosynthetically active radiation (PAR) incident on the crop as a total (MJ m-2); and f(t) is the fraction of radiation intercepted at time t (dimensionless). In high-yielding rice, H is about 0.5. The instantaneous value of radiation use efficiency, (E(t)) expressed as a percentage at time t, can be written

IRRI program report for 1998

E(t) = 100

dW/dt Ia(t)


Table 1. Values of crop and weather variables used in radiation use efficiency calculations. IRRI, 1998. Variable or parameter Value Units m2 leaf m-2 ground t ha-1 g CH2O g-1 DM d-1 (21-26) oC dimensionless h mol CO2 m-2 s-1

where is the energy content of the biomass (J g-1); W is the weight of the total crop biomass (g m-2), t is time, and Ia(t) is the solar energy (MJ m-2 PAR) absorbed by the canopy and is the difference between incident and transmitted plus reflected PAR. It should be noted that when the factors 100 and are omitted, the radiation conversion factor (RCF)((t)) is estimated and has units such as g MJ-1. After much simplification, the RCF can be estimated for the crop when it is vegetative, and losses from death and detachment are negligible, and written as 0.75Pg maWT dmean = (3) Itotal where Pg is the daily total gross photosynthesis (g CH2O m-2 d-1), ma is the value of the maintenance coefficient at the average temperature for the day, WT is the total biomass, and Itotal is the daily total of intercepted PAR (MJ m-2 d-1). These values can all be calculated from a combination of published and experimental data. The variation in the RCF caused by changes in incident PAR (Fig. 1) was examined using equation (3) for the conditions shown in Table 1. The relationship highlighted two interesting features. The first was that RCF varied markedly with PAR. It increased from a negative value to a maximum, 3.84
Radiation conversion factor (g DM MJ-1) 4

Leaf area index 1.12 Total biomass including roots 1.2 Maintenance respiration coefficient at 20 oC 0.015 Av daily temperature 23.5 Extinction coefficient 0.45 Daylength 11.7 Leaf photosynthesis 36


-4 0 5 10 Incident PAR (MJ m-2 d-1) 15

1. The variation in the radiation conversion factor (shoot+root/ PAR) with a range of incident PAR recorded at IRRI. Calculated for a young rice crop from the values recorded in Table 1. IRRI, 1998.

g total DM MJ-1 and then declined again. It was negative at low values of PAR, when maintenance respiration was greater than the difference between gross photosynthesis and synthetic respiration. The second decline was caused by the decrease in the efficiency of canopy photosynthesis with increasing irradiation. An experiment with IR72 was conducted at IRRI in 1997 DS. The crop was transplanted at the standard spacing of 20 20 cm and irrigated. It had high applications of N (480 kg N ha-1), and insecticides applied to control insect pests. Panicle initiation occurred at 43 d after transplanting (DAT). Anthesis started at 61 DAT and finished between 75 and 82 DAT. The final grain harvest was taken at 102 DAT. Measurements were made immediately before each harvest of instantaneous incident PAR above the canopy and transmitted PAR below the canopy. The fraction intercepted was computed from these results and an average for the period was obtained by interpolation. The amount of PAR (daily total) incident on the site was taken from records at the IRRI meteorological station. PAR was assumed to be 50% of solar radiation. There are two sets of values of DM with which to compute the RCF: the actual values measured on the harvest dates and the values obtained for the harvest dates after fitting a polynomial curve (r2 > 0.95). The RCF was computed using the mean intercepted PAR for the interval up to the harvest date. Figure 2 shows that variation in consecutive measures of the biomass gave rise to considerable variation in the estimated values of the RCF. Values ranged from -17 to 21 g DM MJ-1 and the overall average was 2.5 g DM MJ-1. The value of the RCF was highly sensitive to errors in both biomass and intercepted PAR. The RCF calculated for the fitted values of biomass (Fig. 2) shows a much smoother

Irrigated rice ecosystem

Radiation conversion factor (g DM MJ-1) 25 20 15 10 5 0 0 -5 -10 -15 -20

2. The RCF (shoot/PAR) calculated at intervals during the growth of a rice crop using the measured shoot biomass (--o--) and for the same periods the shoot biomass obtained from a curve fitted to measured biomass ( ). IRRI, 1998.






Days after transplanting

pattern but still varies during the season. The overall average using this approach was 3.7 g DM MJ-1. Both of the estimates of the RCF suggest a decline during anthesis and an increase during grain filling. The grain yield in this experiment was 11.63 0.67 t ha-1, suggesting a value for the RCF close to 2.6 g DM MJ -1. The traditional method of plotting biomass at each harvest against accumulated intercepted PAR and fitting regression through the origin gave a value of 2.5 DM MJ-1. The results suggest that the value of the RCF varies because of changes in daily irradiation. The RCF declines at high irradiation even though the canopy is increasing its rate of carbon capture and potential growth rate. Even the method used to calculate the RCF can influence its estimated mean value. Errors of measurement in particular have a strong influence (Fig. 2). It seems likely that the high N nutrition of this experimental crop enabled it to make better use of intercepted PAR. We suggest a value for the RCF close to 2.6 g DM MJ-1 (shoot/PAR) for rice grown with high levels of N when yields beyond 12 t ha-1 are attainable. Peculiar growth patterns on route to high yields For 30 yr, 10 t ha-1 has remained the maximum value regarded as a thermodynamic yield barrier in the tropics. We demonstrated that the yield barrier can be exceeded when sufficient N was given to maintain the critical concentration for metabolic

activity, and that the highly productive crop showed an unusual growth pattern. Crops of the elite indica-type rice cultivar IR72 and current lines of the new plant type (NPT) were transplanted and irrigated with standard treatment during 1997 DS. Weekly applications of N fertilizer totaled 420 kg N ha-1. Strings across some plots prevented lodging in IR72. Yields of both IR72 and the NPT were 11.6 t ha-1. The NPT was damaged during grain filling by the striped stem borer. The harvest index (grain dry matter as a fraction of aboveground DM) was 0.55 for IR72 and 0.40 for NPT. In contrast to the usual logistic curve exhibited by annual crops, we found an apparent plateau around the time of flowering (Fig. 3a and 3b). Harvest of aboveground biomass was every 2 d from the start of flowering and it was that frequency of sampling which revealed the plateau. The growth anomaly could be the slowing of growth during flowering, deviation from expected curve A, or very rapid growth afterwards, deviation from curve B (Fig. 3c). If harvests had been sampled every 1 or 2 wk, as commonly practiced, a standard logistic-type curve would have appeared acceptable (Fig. 3d). Changes in incident solar radiation and the fraction intercepted by the crops from the start of flowering were too small and inconsistent to account for the slowing of growth. Other possible explanations for slow growth, such as movement of DM between roots and shoots, loss of weight due to pollen shed,

IRRI program report for 1998

Aboveground biomass (t ha-1) 25 a 20 15 10 5 0 0 30 25 20 15 10 5 0 0 20 40 60 80 100 b 20 40 Fstart 60 Fend 80 100

25 c 20 15 10 5 0 0 d 20 15 10 Fstart Fend 5 120 0 0 20 40 60 80 100 Sigmoid Gaussian 20 40 60 80 100


Days after transplanting

3. The growth of aboveground biomass (points) of productive rice crops grown in DS at IRRI: (a) the cultivar IR72, (b) the new plant type, (c) as in (a) with logistic curves A and B, and (d) a logistic through data representing weekly sampling. The growth curves for 1997 are backtransformed from cubic polynomials, chosen for parsimony and interpretability of the coefficients, fitted to the natural logarithm of aboveground dry weight, to make variances more homogeneous. The period of flowering is shown for each crop. In (c), the pecked lines are the expected near logistic curves with early exponential growth then linear growth before leveling off toward maturity. IRRI, 1998.

and loss of weight from decay of dead plant matter, did not seem to account for the size of the phenomenon. If the results are interpreted as a marked upswing in growth after flowering, an alternative explanation is required. The imposed critical N management enabled these crops to sustain two or three live leaves per tiller through to final harvest, a time when most leaves are usually dead. Consequently, we calculated (with the model Oryza1) that grain filling declined by only about 50% during photosynthesis, whereas it is reported that respiration decreased ninefold over the same period. The difference between the rates of the two processes would provide the resources for the late surge in growth. Our inability to explain the anomalous growth pattern suggests that we do not yet have a complete understanding of the physiology of high-yielding crops.

Progress in increasing grain yield by breeding a new plant type S. Peng, G. Khush, R. Visperas, and A. Evangelista Breeding of the NPT with large panicles and low tillering aimed to break the yield barrier of 10 t ha-1 in the tropics. However, field tests at several sites between 1994 and 1997 indicated low yields from NPT lines due to poor grain filling and low biomass production. The design of the NPT was slightly modified by moderately increasing tillering capacity, plant height, and growth duration, and slightly decreasing panicle size. In 1998 DS and WS, we evaluated recently developed NPT lines at IRRI to determine if potential grain had increased. Forty-three lines were grown in DS and 51 lines were grown in WS. IR72 was the check variety. In 1998 DS, 22-d-old seedlings were transplanted on

Irrigated rice ecosystem

Table 2. Grain yield and yield components of 43 new plant type (NPT) lines and IR72. IRRI, 1998 DS. Grain yield (t ha-1) 8.84 8.59 8.45 8.38 8.27 8.19 8.18 8.07 8.04 8.00 7.90 7.90 7.89 7.86 7.83 7.80 7.58 7.58 7.48 7.47 7.46 7.45 7.44 7.43 7.40 7.36 7.36 7.28 7.27 7.26 7.25 7.25 7.24 7.22 7.08 7.01 7.00 6.97 6.91 6.88 6.75 6.67 6.64 6.58 0.83 Panicles m-2 (no.) 229 216 222 234 219 259 232 532 280 250 239 263 220 279 224 239 313 279 215 309 250 263 207 341 326 221 242 273 226 287 276 329 237 239 219 248 275 281 215 229 274 200 245 191 30 Spikelets panicle-1 (no.) 161 186 186 166 195 199 183 90 136 177 211 142 170 153 150 167 117 129 146 98 126 167 170 114 95 177 210 121 143 135 177 96 140 154 184 130 132 154 197 164 134 147 97 154 20 Spikelets m-2 (no. 103) 36.9 40.1 41.2 38.9 42.7 51.5 42.6 47.8 38.0 44.2 50.4 37.1 37.3 42.9 33.7 39.8 36.7 35.8 31.4 30.3 31.3 43.6 34.8 38.9 31.0 39.2 50.8 33.1 32.3 38.9 48.8 31.6 33.1 36.7 40.0 32.2 36.2 43.2 42.1 37.6 36.6 29.4 23.8 29.3 5.5 Filled spikelets (%) 71.6 72.7 77.8 84.2 65.7 57.2 59.1 80.9 71.8 60.6 52.0 72.3 78.7 84.0 77.9 76.8 78.6 57.0 80.2 90.4 83.1 59.4 38.3 77.7 89.4 57.3 45.3 83.1 67.7 77.4 56.6 92.5 79.0 83.6 50.6 88.4 78.9 61.4 54.9 53.1 82.0 82.7 84.0 82.6 5.8 1,000grain weight (g) 29.2 25.1 23.3 24.9 24.8 22.3 25.1 21.4 25.4 23.3 25.1 24.7 25.5 21.4 24.5 26.4 24.0 32.2 29.6 25.9 26.0 26.6 23.9 23.7 25.5 26.1 24.4 25.8 26.8 23.1 26.1 25.0 28.9 22.0 24.5 24.9 24.1 23.9 24.2 24.5 23.1 27.4 33.7 25.1 0.5

NPT line

IR70554-48-1-2 IR65564-44-2-2 IR69093-41-2-3-2 IR68552-100-1-2-2 IR69853-70-3-1-1 IR70479-45-2-3 IR66738-118-1-2 IR72a IR69137-34-1-3-1 IR67962-84-2-2 IR67966-188-2-2-1 IR68011-15-1-1 IR69116-67-3-2-3 IR68552-55-3-2 IR65564-44-5-1 IR68544-29-2-1-3-1-2 IR69923-31-3-2-3 IR65600-87-2-2-3 IR66158-38-3-2-1 IR66160-121-4-4-2 IR65600-54-6-3 IR65564-22-2-3 IR69432-54-1-1-2-2 IR68019-60-3-3-2 IR66160-121-4-1-1 IR70491-33-2-2 IR67962-40-6-3-3 IR65600-42-5-2 IR65600-77-4-2-1 IR66160-5-2-3-2 IR69092-57-3 IR66160-121-4-5-3 IR65600-96-1-2-2 IR65600-38-1-2-1 IR65598-112-2 IR69800-5-3-1-2 IR69132-17-2-2-2 IR67962-84-2-2-2 IR67966-44-2-3-2 IR66750-6-2-1 IR66159-189-5-5-3 IR65600-129-1-1-2 IR65600-27-1-2-2 IR65600-127-6-2-3 LSD (0.05)

Check variety.

7 Jan. Fertilizer N was applied at 120 kg ha-1. In 1998 WS, transplanting was on 25 Jun with 22-dold seedlings. Fertilizer N input was 100 kg ha-1. Hill spacing was 0.10 0.15 m, with one seedling per hill in both seasons. Samples were taken from a 0.5-m 2 area at maturity to determine panicle number, spikelets per panicle, grain-filling percentage, and 1,000-grain weight. Grain yield was deter-

mined from a 5-m2 area and adjusted to a moisture content of 0.14 g H2O g-1 fresh weight. Seven NPT lines produced greater or the same yield as IR72 in 1998 DS (Table 2). All of those had a panicle size of 150-200 spikelets panicle-1, which may be the optimal range of panicle size for the NPT. High-yielding NPT lines had large spikelet number m-2, or high grain-filling percentage, or

IRRI program report for 1998

both. Four NPT lines had grain-filling percentage of more than 85%, significantly higher than that of IR72. However, the high grain-filling percentage was associated with their small number of spikelets m-2, and grain yield of those lines was limited by sink size. In 1998 WS, 30 NPT lines had equal or greater yield than IR72 (Table 3). IR65600-42-5-2 produced 7.7 t ha-12.5 t ha-1 more than IR72. Many NPT lines produced significantly more spikelets m-2 because of their larger panicle size compared with that of IR72. Nine NPT lines had grain-filling percentage higher than 70% compared with 65% from IR72. Most NPT lines had greater grain weight than IR72 in both seasons. These results indicate significant progress in improving the NPT. The improvement was mainly reflected by the increase in grain-filling percentage. Hybrid rice S. Virmani, R. Toledo, C. Casal, R. Ona, and D. Sanchez The goal of hybrid rice research at IRRI is to increase rice yields beyond the level of high-yielding semidwarf rice varieties by exploiting the phenomenon of hybrid vigor. During 1998, 386 elite inbred lines were tested for their ability to maintain sterility or restore fertility of three cytoplasmic male sterility (CMS) systemsCMS WA, CMS-ARC, CMS-mutagenized IR62829Bused in the breeding program. In all 1,001 testcrosses were evaluated and 192 new backcrosses were initiated to develop new CMS lines in BC1 to BC6 generations. Seeds of 645 hybrids were produced and 780 hybrids were evaluated in observation yield trials (495), preliminary yield trials (214), and advanced yield trials (71). Fifteen hybrids were nominated for national trials by the Philippine Rice Research Institute (PhilRice). Thirty-one hybrids were nominated for inclusion in International Rice Hybrid Observation Nursery (IRHON). Nucleus and breeder seed of 32 CMS lines and 21 restorer lines were also produced for sharing with public- and private-sector institutions working on hybrid rice in national programs.


IRRI launched a network project titled, Development and Use of Hybrid Rice Technology in Asia. Funding was provided by the Asian Development Bank (ADB) and collaboration established between IRRI, the Food and Agriculture Organization of the United Nations (FAO), Asia Pacific Seed Association (APSA), and seven Asian countriesBangladesh, China, India, Indonesia, Philippines, Sri Lanka, and Vietnam. The project will expedite the development and use of hybrid rice technology in the six countries. Meetings of the project during 1998 established project goals, objectives and expected outputs, and specified the roles of IRRI, FAO, APSA, and the NARS. Action plans and budget of the activities to be implemented by member countries and their participating institutions during 1998 and 1999 were prepared and approved.

The IRRI hybrid IR69690H (IR58025A/BR827-353-1-1-1R) was released by Maharashtra state government in India as Karjat Rice Hybrid-1 or Sahyadri. It has intermediate height (110-120 cm), medium growth duration (125-130 d), long slender grains with slight aroma and intermediate amylose, and outyielded variety Jaya by about 30% in onfarm trials. Four other rice hybrids released in India CORH2 and ADRH1 in Tamil Nadu, and Narendra Sankar Dhan 2 and Pant Sankar Dhan 1 in Uttar Pradeshwere derived from the IRRI-bred CMS line IR58025A, which indicated the direct utility of IRRI-bred parental lines for the national programs.

Sixteen rice hybrids showed at least 1 t ha-1 yield advantage in advanced yield trials conducted at IRRI during 1998. Five of those hybrids yielded consistently higher in both seasons (Table 4).

Irrigated rice ecosystem

Table 3. Grain yield and yield components of 51 NPT lines and IR72. IRRI, 1998 WS. Grain yield (t ha-1) 7.67 6.90 6.40 6.39 6.26 6.24 6.00 5.75 5.58 5.55 5.54 5.50 5.48 5.43 5.41 5.37 5.35 5.34 5.34 5.26 5.25 5.23 5.21 5.19 5.19 5.13 5.12 5.11 5.02 5.02 5.01 4.98 4.95 4.94 4.92 4.88 4.87 4.86 4.82 4.76 4.75 4.74 4.68 4.59 4.57 4.50 4.45 4.44 4.27 4.12 4.10 3.66 0.54 Panicles m-2 (no.) 327 340 383 273 326 328 280 259 293 280 276 284 264 295 302 209 356 288 275 276 260 271 269 237 270 243 185 260 218 302 469 219 230 310 229 311 266 247 331 280 230 260 253 286 228 248 237 251 237 245 268 247 29 Spikelets panicle-1 (no.) 96 110 86 139 112 91 142 102 137 81 119 129 131 123 145 192 91 107 85 172 123 120 133 140 141 125 122 95 133 115 72 136 116 135 179 105 179 128 87 108 150 152 151 139 132 201 157 129 132 133 119 140 16 Spikelets m-2 (no. 103) 31.4 37.5 32.5 37.9 36.9 29.6 39.8 26.5 40.1 22.8 32.9 36.5 34.6 36.3 43.8 40.0 32.3 30.8 23.6 47.3 31.8 32.5 35.6 33.3 38.0 30.5 22.4 24.7 28.9 34.8 33.7 29.8 26.7 41.7 40.7 32.5 47.5 31.6 28.7 30.4 34.4 39.3 38.4 39.8 30.0 50.0 37.1 32.6 31.2 32.5 31.8 34.7 5.2 Filled spikelets (%) 74.4 58.5 78.3 52.0 73.3 80.6 66.3 74.0 49.6 71.7 61.1 51.5 26.1 43.9 38.5 37.5 53.4 68.8 53.1 40.0 70.9 60.9 22.4 61.4 50.0 55.2 58.3 73.9 45.3 57.1 65.0 62.5 70.7 42.9 32.5 35.2 20.2 57.3 55.4 58.8 35.2 35.0 41.0 32.2 12.0 38.0 35.2 66.8 47.9 15.3 33.6 47.8 8.2 1,000grain weight (g) 22.7 25.6 22.2 29.2 22.5 24.5 21.6 25.7 22.4 31.0 24.1 23.6 25.6 27.2 24.3 22.9 25.5 24.0 33.1 24.1 22.9 21.8 24.0 23.0 23.6 28.0 31.7 25.9 27.9 25.2 20.1 22.4 27.2 23.5 25.8 24.7 25.5 27.2 25.6 23.7 25.5 26.2 25.4 24.5 23.3 21.9 26.3 27.8 25.4 23.8 27.7 20.9 0.6

NPT line

IR65600-42-5-2 IR65564-44-2-3 IR66160-121-4-1-1 IR66158-38-3-2-1 IR66160-121-4-5-3 IR66160-121-4-4-2 IR68552-55-3-2 IR69132-17-2-2-2 IR66159-189-5-5-3 IR65600-27-1-2-2 IR65564-44-5-1 IR66160-5-2-3-2 IR65600-77-4-2-1 IR65600-96-1-2-2 IR65564-22-2-3 IR69853-70-3-1-1 IR65564-22-2-3 IR65600-54-6-3 IR65600-87-2-2-3 IR67962-84-2-2 IR69800-5-3-1-2 IR69923-3-1-3-2-3 IR66750-6-2-1 IR70559-AC5 IR65600-127-6-2-3 IR65600-129-1-1-2 IR67966-188-2-2-1 IR68544-29-2-1-3-1-2 IR71204-78-3-3 IR69137-34-1-3-1 IR72a IR69093-41-2-3-2 IR68552-100-1-2-2 IR66160-5-2-3-2 IR72926-AC1 IR65564-22-2-3 IR66738-118-1-2 IR69092-57-3 IR68019-60-3-3-2 IR68011-15-1-1 IR70491-33-2-2 IR71218-5-2-1 IR69116-67-3-2-3 IR67962-40-6-3-3 IR69432-54-1-1-2-2 IR70479-45-2-3 IR70554-10-3-1-3 IR67966-44-2-3-2 IR67962-84-2-2-2 IR65598-112-2 IR70554-48-1-2 IR65600-38-1-2-1 LSD (0.05)

Check variety.


IRRI program report for 1998

Table 4. Yield of five elite rice hybrids identified in advanced yield trials. IRRI, 1998. Yield (t ha-1) Hybrid or check variety DS IR69622A/IR29723-143-3-2-1R IR68888A/IR63875-196-2-2-1-3R IR68897A/IR65620-58-2-3-2-3R IR68897A/IR60819-34-2-1R IR68897A/IR63874-187-2-2-1-2R PSBRC 28 PSBRC 18 LSD (5%) LSD (1%) 9.4 9.3 9.0 8.5 8.4 7.4 7.4 1.0 1.3 WS 5.8 5.7 5.4 5.4 4.2 3.5 3.6 0.8 1.0

servation, preliminary, and advanced yield trials. Those were derived from IR68897A, which possesses long slender grains.

Table 5. New cytoplasmic male sterile lines developed in 1998 at IRRI. Designation Parentage Source of cytoplasm WA WA Dissi Dissi Dissi Dissi Dissi Gambiaca Gambiaca Gambiaca Gambiaca Gambiaca WA WA Kalinga Kalinga

IR74604A IR74605A IR75595A IR75596A IR75597A IR75598A IR75599A IR75600A IR75601A IR75602A IR75603A IR75604A IR75605A IR75606A IR75607A IR75608A

IR69624A/7*BW306-2 IR69628A///7*BPI76/ N22//Taichung 65 D297A/7*IR68885B D297A/7*IR68897B D297A/7*IR68898B D297A/7*IR69619B D297A/7*IR69626B G46A/7*IR68889B G46A/7*IR68897B G46A/7*IR69619B G46A/7*IR69626B G46A/7*IR69624B IR68280A/7*IR62898-2-8-6-2 IR68897A/7*IR68271-127-2-2 Krishna A/7*IR69619B Krishna A/7*IR69627B

We evaluated 2,107 F3-F5 progenies in pedigree nursery. Among F5 progenies, 112 good looking male sterile plants were selected and transferred to the phytotron to induce fertility. Fifty of those plants reverted to partial fertility in the phytotron. Seeds produced from them were used for evaluation of male sterility in the field at high temperatures (maximum temperature above 30 C). Based on their good phenotypic acceptability score (1-3), good fertility reversion (score 1-5), and complete male sterility in the field (using the natural temperature variation), four TGMS lines were identified for sharing with national programs: IR68492-1-6-13-B4-8-B, IR71018-13-73-3-6-B, IR70977-16-5-4-112-B, and IR70978-8-22-5-14-13-B.

Thirteen IRRI breeding lines from the irrigated breeding program were named as varieties in three countries (Table 6). This brought the number of IRRI breeding lines named as varieties by national programs to 308.
Table 6. Breeding lines from the irrigated breeding program named as varieties in 1998.


Breeding line

Name given

Country where named Madagascar Mali Madagascar Madagascar Madagascar Madagascar Madagascar Madagascar Cambodia Cambodia Cambodia Cambodia Cambodia

Sixteen new CMS lines were bred possessing WA, Gambiaca, Dissi, and Kalinga cytoplasms (Table 5). In addition, we also have elite CMS lines possessing ARC and mutagenized IR62829B cytoplasms. Thus, the cytoplasmic base of CMS lines developed at IRRI has been reasonably diversified. The CMS line IR68897A developed in 1992 is stable for complete pollen sterility, has good outcrossing, and has adaptability to tropical conditions. During 1998, many heterotic combinations were identified in ob-

IR2061-213-2-17 (IR34) IR2070-747-6-3-2 IR2071-625-1-252(IR36) IR8866-82-1-3-1-3 IR20913-B-160 IR28128-45-2 IR21363-13-2-2 IR25579-135-3 IR48525-100-1-2-1 IR49830-7-1-2-1-3 IR56383-35-3-2-1 IR57529-9-2-1-3 IR62037-71-3-1-1-3

Mamokatra IR32 Tsy Milofika Mazana (X1228) Rojovo (X1289) Mananoro Mahadignirano Rojomena Rohat Popoul Chulsa Baray Rumpe

Irrigated rice ecosystem


Importance of nitrate vs ammonium for potential yield H. Kronzucker and G. Kirk Growth and yield in most plant species are superior under mixed NO3-NH4+ nutrition compared with either N source alone. However, only small concentrations of NO3- are generally present in flooded rice soils, arising as a result of nitrification of NH 4+ in aerobic zones at the soil surface and in the rhizosphere. Nitrate so formed may diffuse into the soil bulk and be lost through denitrification. It has, therefore, been assumed that NO3- uptake is not important. But although NO 3- concentrations may be small, fluxes and rates of uptake may be large if the plant has the necessary capacity for processing NO3-. In collaboration with the University of British Columbia, we used the short-lived tracer 13N to study root transmembrane fluxes and cytoplasmic pool sizes of NH4+ and NO3- in IR72. This showed that the capacity of rice roots for NO3- acquisition is greater than that for NH4+. Influx of both N species in the relevant concentration range followed Michaelis-Menten kinetics typical of high-affinity transport systems (Fig. 4). Km values were 26 5.6 mM for NO3- and 51 18.4 mM for NH4+, indicating a higher affinity for NO3- than for NH4+. Influx of NO3- exceeded that of NH4+ by about 50% at all

concentrations examined; Vmax values were 8.1 for NO3+ and 5.7 mmol g-1 h-1 for NH4+. Using a technique based on rates of efflux of labeled N from roots, we quantified cytoplasmic pool sizes, halflives of cellular exchange, and flux partitioning. Cytoplasmic NH4+ concentrations were large (20 3.21 mM), but cytoplasmic NO3- concentrations were even larger (37 4.18 mM), surpassing values reported for barley. Notwithstanding the differences in cytoplasmic concentrations, half-lives of cellular exchange were similar (14 0.9 min for NH4+ and 16 2.3 min for NO3-). These indicate the magnitude of influx that can be sustained at a given cytoplasmic concentration, i.e. the intensity of negative feedback of internal N accumulation on uptake. The values for NO3are substantially larger than have been observed in other species, including highly efficient NO3- users; the values for NH4+ are more typical. This indicates an unusually high capacity for NO3- capture and retention. There were pronounced differences in subcellular N flux partitioning between seedlings provided with NH4+ and those provided with NO3- (Fig. 5). While similar proportions of incoming N were channeled into assimilation and to the vacuole, the
N flux components (mol g-1 h-1) 8 7 6 5 4 3
0.95 24.4% 51.7% 2.27 37.43% 0.93 23.9% 0.59 9.65%
assimilation/vacuole xylem co

N influx (mol g-1 h-1) 9 8 7 6 5 4 3 2 1 0 0 100 200 300 400 External N concentration (M) 500 NH4+ NO3

2 1 0


52.92% 2.01



4. Concentration-dependence of steady-state NO3 influx (open symbols) and NH4+ influx (closed symbols) in roots of intact 4wk-old IR72 rice. All plants were grown at 100 M [NO3]0 or [NH4+]0. Error bars indicate SE (n 16). IRRI, 1998.

5. Component fluxes of NO 3 and NH 4+ as estimated by compartmental analysis. IR72 rice seedlings were grown and maintained for 4 wk at 100 m [NO3]0 or [NH4+]0. Total bar graph heights indicate influx. Component fluxes are efflux from the cytoplasm (co), combined flux to assimilation and the vacuole (assimilation/vacuole), and flux to the shoot (xylem). Absolute flux contributions are indicated to the left of respective bar segments, percentages of influx are indicated to the right. Data are the means of eight to nine experiments (n8). SE was less than 15%. IRRI, 1998.


IRRI program report for 1998

proportions translocated to the shoot and lost through efflux were quite different. Translocation to the shoot comprised 37% of incoming 13NO3- but only 24% of incoming 13NH4+. In absolute terms, more than twice as much N was made available to the shoot with NO3- provision. Given that more than 70% of N entering the rice caryopsis and more than 50% of N in growing leaves during the grain-filling stage is derived from remobilization of N storage compounds accumulated in the shoot during the vegetative stage, this difference is potentially of great significance. Moreover, N loss from the roots through efflux was minimized when NO3- was provided. Efflux was less than 10% for NO3- and about 24% for NH4+. Measurements in field experiments with rice in flooded soil have confirmed high concentrations of NO3- in roots and shoots under some circumstances (collaboration with Zhejiang Agricultural University). Hybrids appear to be particularly good in generating NO3- in the soil, or capturing it, or both. Although some NO3- may be lost through denitrification, it may be that the benefits to the plant of mixed NH4+-NO3nutrition outweigh this. Thus manipulation of the interchange between N species (for example by managing water flow through the field to maximize NO3formation in the rhizosphere) may increase the efficiency of N utilization. Identifying new genes with durable resistance to tungro viruses O. Azzam and E. Coloquio Screening the rice germplasm collection for sources of resistance for tungro viruses includes four evaluation tests: q Initial mass screening of the germplasm accessions using the Standard evaluation system for rice and measuring the symptom severity and percent infection. This eliminates lines that are susceptible to the vector and the viruses. q Forced test-tube inoculation and serological indexing for resistance to rice tungro bacilliform virus (RTBV) and rice tungro spherical virus (RTSV) using polyclonal antisera raised against each of the viruses. This identifies those lines that are resistant to the vector and to either of the viruses.

Screening for the vector Nephotettix virescens resistance using the seedbox test. The mortality of the plants based on 1 to 9 score system is used as an indicator of susceptibility or resistance to the green leafhoppers. This identifies which lines have vector (and possibly virus) resistance and which ones have virus resistance. q Further screening of selected lines for resistance to virus strains that are kept on a differential host. This identifies lines that have resistance to more than one strain of the virus (using the IRRI greenhouse collection of strains). By 1996, 1,277 lines from Bangladesh, Pakistan, India, Sri Lanka, and Indonesia were screened using the forced test-tube inoculation and serological indexing against the RTSV and RTBV polyclonal antisera. Results were that 336 lines had 0% infection with RTSV and 264 lines had <10% infection. For RTBV, only 3 lines showed 0% infection but 109 lines had <30% infection. In 1997 and 1998, the 336 selected lines with 0% infection for RTSV and the 3 RTBV lines were screened against tungro virus strains. Three previously identified RTBV-resistant lines were not resistant when tested with the mixture of RTBV strains in the greenhouse. Fortyfive out of the 109 RTBV-tolerant lines (<30% infection) are still being evaluated for RTBV resistance by insect and Agrobacterium-mediated inoculations. Evaluation results for the 336 lines against RTSV-A and RTSV-Vt6, the more virulent and resistance-breaking strain on TKM6, are shown in Table 7. Most of the lines have already been screened for vector resistance and 17 out of the 52 selected lines (0% infection) could be identified as potential resistance donors for broad RTSV resistance. Most of those lines are from Bangladesh and India.

Development of rice blast-resistant monogenic lines H. Kato and M. Yanoria For developing durable resistance against rice blast, pyramiding of major genes, accumulation of minor genes, a combination of major and minor genes, and multilines of major genes are considered the effec-

Irrigated rice ecosystem


Table 7. Number of accession lines with their resistance reaction to rice tungro spherical virus (RTSV) strains (RTSV-A and RTSV-Vt6) using forced test-tube inoculation and serological indexing using enzyme-linked immunosorbent assay at 21 d post inoculation. IRRI, 1998. RTSV-Aa Origin 0% infection Bangladesh Pakistan India Sri Lanka Indonesia Total 52 10 261 2 11 336 0% 10(19) 0 (0) 41(16) 0 (0) 1 (9) 52(15) 1- 10% 8(15) 0 (0) 51(19) 1(50) 1 (9) 61(18) 11-20% 7(13) 0 (0) 36(14) 0 (0) 0 (0) 44(13) 21-30% 5 (10) 2 (20) 24 (9) 1 (50) 4 (36) 36 (11) 31- 100% 17 (33) 8 (20) 70 (27) 0 (0) 9 (82) 98 (29) Untested 5 (10) 0 (0) 39 (15) 0 (0) 1 (9) 45 (13) RTSV Vt6 strainb

a Part of the data for RTSV-A were taken from Cabunagan and Koganezawa, unpubl. data. bNumbers in parenthesis reflect the percentage of lines out of the total lines tested. To ensure that Vt6 strain was used, TKM6 was included in all the experiments as a positive check. Percent infection of TKM6 ranged between 40 and 86% for the different batches tested.

Table 8. Rice blast-resistant monogenic lines. Designation IRBLa-A IRBLa-C IRBLi-F5 IRBLks-F5 IRBLks-S IRBLk-ka Tsuyuake IRBLkp-K60 IRBLkh-K3 IRBLz-Fu IRBLz5-CA IRBLzt-T IRBLta-K1 IRBLta-CT2 Reiho IRBLb-B IRBLt-K59 IRBLsh-S IRBLsh-B IRBL1-CL IRBL3-CP4 IRBL5-M IRBL7-M IRBL9-W IRBL12-M IRBL19-A ARL20

Resistance gene

Donor Aichi Asahi CO 39 Fujisaka 5 Fujisaka 5 Shin 2 Kanto 51 K60 K3 C101A51 Toride 1 K1 C105TTP2L9 BL1 K59 Shin 2 BL1 C101LAC C104PKT RIL249 (Moro.) RIL29 (Moro.) WHD-IS-75-1-127 RIL10 (Moro.) Aichi Asahi Asominori/IR24RIL

Combinationa LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//3*LTH LTH/Donor//LTH LTH/Donor//2*LTH LTH/Donor//3*LTH LTH/Donor//LTH LTH/Donor//2*LTH LTH/Donor//LTH LTH/Donor//LTH LTH/Donor//3*LTH LTH/Donor//2*LTH LTH/Donor//3*LTH LTH/Donor//3*LTH LTH/Donor//3*LTH LTH/Donor//2*LTH LTH/Donor//LTH

Generationb BC1F8 BC1F8 BC1F8 BC1F8 BC1F8 BC1F6 BC1F6 BC1F6 BC1F8 BC3F6 BC1F8 BC2F6 BC3F6 BC1F6 BC2F6 BC1F8 BC1F6 BC3F6 BC2F6 BC3F6 BC3F6 BC3F6 BC2F6 BC1F8


Pi a Pi a1 Pi i Pi k-s Pi k-s Pi k-s Pi k-m Pi k-p Pi k-h Pi z(?) Pi z5 Pi zt Pi ta Pi ta Pi ta-2 Pi b Pi t Pi sh Pi sh Pi 1 Pi 3 Pi 5(t) Pi 7(t) Pi 9(t) Pi 12(t) Pi 19(t) Pi 20

= Pi 2(t) = Pi 4(t) with Pi a with Pi a

V86010-R with Pi a

LTH= Lijiangxintuanheigu. bGenerations = second crop, 1998.

tive methods. Knowledge of the genetic constitution of major genes in rice cultivars and the pathogenicity of blast isolates is a prerequisite for all of those strategies. We are establishing simple differential systems for the identification of pathogenicity of blast isolates of each country. We selected lines with a single resistance gene from crosses with Lijiangxintuanheigu (LTH) and

resistance gene donor parents (Table 8). LTH is a japonica variety with no major genes for blast resistance. So far, all IRRI blast isolates were compatible to LTH. Most of the lines probably possess a gene for photoperiod sensitivity from LTH, and therefore showed poor growth in the tropics. Harvested seeds were distributed as a set of standard differentials to scientists working with blast and collaborating with


IRRI program report for 1998

the International Network for Genetic Evaluation of Rice (INGER).

tion where the slope of the function equaled the ratio of the N price to price of rice: dY PN = dN PP where Y represents yield per hectare, N represents N applied per hectare, PN represents the price per kg of N, and PP represents the price per kg of rough rice. Calculated optimal N rates are occasionally large (or negative). In such cases, the optimal N rate was truncated at either 0 kg N ha-1 or 120 kg N ha-1 in WS and 0 kg N ha-1 or 200 kg N ha-1 in DS.

Sustaining soil quality in intensive rice systems

Variability of optimal nitrogen applications for rice D. Dawe and P. Moya Philippine rice farmers vary their nitrogen (N) application rates substantially from year to year, even for the same climatic season. A sample of 32 irrigated rice farmers in Laguna Province in WS showed the average difference between maximum and minimum application rates for individual farmers during a period of five WS to be 66 kg N ha-1. Socioeconomic factors such as availability of fertilizer or biophysical constraints to growth are reasons for this variability. To investigate whether such variability makes sense, more than 20 yr of data were analyzed to determine the variability of theoretical profit-maximizing N rates under experimental conditions at research stations. The data used were from a series of N response experiments at four Philippine experiment stations during 1965-88. The experiment stations were IRRI, the Maligaya Rice Research and Training Center in Muoz, Nueva Ecija (now known as PhilRice), the Bicol Rice and Corn Experiment Station in Camarines Sur, and the Visayas Rice Experiment Station on Panay Island. Each experiment typically used five or six different rates of applied N, usually 0, 30, 60, 90, and 120 kg N ha-1 in the WS, and 0, 60, 90, 120, 150, and 180 kg N ha-1 in the DS. The varieties used were all modern semidwarf varieties that originated from breeding programs at IRRI. Other data were from the Longterm Continuous Cropping Experiment (LTCCE) at IRRI during the same period in which the N response experiments were conducted. The LTCCE used four different rates of applied N, typically 0, 30, 50 (or 60), and 90 (or 100) kg N ha-1 in the WS and 0, 50, 100, and 150 kg N ha-1 in the DS. Nitrogen response functions were estimated by ordinary least squares regression of yield for three variables: 1) a constant term, 2) the level of applied N, and 3) the square of the level of applied N. This generated a quadratic response function, and the optimal N rate was the point on the response func-

The average optimal WS N rates varied across the four sites from 52 kg N ha-1 to 72 kg N ha-1, and the standard deviation varied from 32 kg N ha-1 to 37 kg N ha-1. In about 15% of the cases (average across stations), the optimal N rate was less than 15 kg N ha-1 (Table 9). The variability of optimal N rates was similar in the DS, with the mean varying from 108 kg N ha-1 to 133 kg N ha-1 and the standard deviation ranging from 34 kg N ha-1 to 42 kg N ha-1. Multiple regression analysis was conducted to determine factors that explain the year-to-year variability at a given site. The most obvious explanation is changing climatic conditions from year to year. This variation was captured by use of dummy variables for each year. Another likely possibility is that optimal N rates are different for different IR varieties, so dummy variables were also specified for each variety used. A third possibility is variations in the indigenous N supply (INS), assuming that the optimal N rate should vary inversely with INS. A high INS implies it is not necessary to apply large quantities of N, while a low INS implies large quantities of N should be added. INS can be defined as N uptake in a zero N plot, but this measure is also highly correlated with grain yield in a zero N plot. Because no data on N uptake were available for the N response experiments, grain yield in the zero N plot was used as an index of INS. Regression of the optimal N rate on INS and dummy variables for year and variety at each of the four sites explains 55-70% of the variation in the dependent variable (as measured by the R2) in the WS, and 38-76% in the DS (Table 10). The magnitudes of the various coefficients (not shown) indicate that year-to-year variations (presumably

Irrigated rice ecosystem


Table 9. Frequency distribution (in percent) of optimal N rates in experiments at IRRI and three Philippine research stations. Based on 1965-88 data. Optimal N rate (kg N ha-1) Sitea 0-15 1530 3045 4560 6075 7590 90105 105- >120 135- 150- 165120 150 165 180 180- >195 Obser- Av 195 vations (no.) SD

Wet season Bicol 20 IRRI 14 PhilRice 20 VRES 6 All 15 Dry season Bicol 1 IRRI 2 PhilRice 1 VRES 2 All 1

7 3 3 5 5

10 6 13 3 8

18 16 15 14 16

20 17 16 21 18

17 16 18 20 18

4 10 8 15 9

0 6 2 7 4

4 12 5 8 8

100 125 100 95 420

52 67 56 72 62

33 37 35 32 35

0 0 0 0 0

0 0 0 0 0

2 1 0 0 1

7 3 4 9 5

14 5 3 15 8

28 17 15 18 19

25 17 15 29 20

9 14 24 13 15

2 11 9 2 7

4 9 9 4 7

1 3 7 2 4

1 3 1 2 2

5 15 12 5 10

81 121 94 55 351

108 131 133 111 123

34 42 38 36 40

Bicol = Bicol Rice and Corn Experiment Station, Camarines Sur. VRES = Visayas Rice Experiment Station, Panay Island.

Table 10. Optimal N rate regressed on year dummy variables, variety dummy variables, and indigenous N supply (INS) estimated from grain yield measured in a 0 N control plot. For regressions including data from all sites, year dummy variables are site-year interactions. IRRI, 1998. Sitea Coefficient INS P value R2 Observations (no.)

Wet season IRRI 7.5 PhilRice -23.6 Bicol -12.7 VRES 3.9 All -3.7 Dry season IRRI PhilRice Bicol VRES All
6.0 11.4 12.5 0.3 7.6

0.17 0.00 0.03 0.48 0.18

0.55 0.70 0.67 0.69 0.61

125 100 100 95 420

0.43 0.15 0.06 0.96 0.03

0.38 0.68 0.52 0.76 0.55

121 94 81 55 351

Bicol in the WS, the effect of INS is statistically significant at the 5% level and negative, as would be expected. In all other cases, however, the effect is positive (typically, the positive coefficients are not significantly different from zero). The analysis suggests that optimal farmer behavior may not necessarily result in a negative correlation between INS and applied N. A possible explanation for this result is that yield in a zero N plot may be a poor measure of INS. More precise measurements of INS would be preferable, but such measurements would also be more costly and of less ultimate value to farmers in formulating cost-effective N management strategies.

a Bicol = Bicol Rice and Corn Experiment Station, Camarines Sur. VRES = Visayas Rice Experiment Station, Panay Island.

due to climate) have larger effects on the optimal N rate than either variety or INS. This suggests that real-time N management, which is able to account for changing year-to-year conditions, is the most important factor for optimal crop management. The coefficient on INS indicates the effect of INS on the optimal N rate, after controlling for the influence of climate and variety. The effect of INS on the optimal N rate is mixed. At PhilRice and

By using data from different experiments at IRRI, we estimated the spatial variability of optimal N rates over a relatively small area. For this analysis, data from both the N response experiment at IRRI and LTCCE were used. Experiments were selected only if the same variety was used in the same season of the same year. Optimal N rates were calculated for both experiments, and the difference between the two rates was compared. On average, regardless of season, the absolute value of the difference in optimal N rates for the same variety and the same season, at two different sites at IRRI, was about 30 kg N ha-1. Part of this difference may be


IRRI program report for 1998

due to different planting dates, but this spatial variability has implications for the design of yield gap studies. Yield gap studies in farmers fields typically calculate a response function for an entire area or village, and then compare the behavior of individual farmers relative to this area-wide response function. The analysis here suggests, however, that response functions may vary significantly over a small area, casting doubt on the use of a single function for many different farmers. Practical implications. There are several practical implications of this analysis. First, because optimal N rates vary substantially from year to year, and because INS appears to help little in explaining this variation, optimal N management strategies must be done in real time, i.e., while the crop is in the ground. This will most likely require technology such as the chlorophyll meter or the leaf color chart, or strategies based on the number of tillers. We note, however, that knowledge of the INS is important to 1) estimate the total crop N demand as a basis for the amount of N to be applied in various splits; and 2) decide the need and amount of N applied basally or during early vegetative growth, when plants are too small to be used as an indicator of N status. Second, as noted earlier, yield gap studies will provide more information if response functions are estimated at the level of individual farms instead of large areas. Internal nutrient efficiencies in irrigated lowland rice C. Witt, A. Dobermann, G. Simbahan, S. Abdulrachman,1 G. Gines,2 W. Guanghuo,3 R. Nagarajan,4 S. Satawatananont,5 T. Son,6 P. Tan,7 and L. Tiem8 In a situation where crop growth is not limited by water supply, weed problems, or pest infestation, biomass production is mainly driven by the supply of N. Thus, the demand of the rice plant for other macronutrients mainly depends on the N supply. There are, however, considerable uncertainties about crop N, P, and K requirements because the internal efficiencies vary greatly, depending on nutrient supply, crop management, and climate. We present the derivation of an empirical model for estimating crop NPK requirements in rice, which will be part of a general nutrient decision

support system for improving nutrient management in intensive rice systems of Asia.

Within the Mega Project on Reversing Trends of Declining Productivity in Intensive Irrigated Rice Systems (RTDP), we have collected a database on nutrient efficiencies in more than 200 farmers fields. Straw yield, grain yield, components of yield (unfilled spikelets, 1,000-grain weight, number of panicles m-2, spikelets m-2), and nutrient concentrations of modern, high-yielding varieties were measured using the same methodology at seven sites in China, India, Indonesia, Philippines, Thailand, and Vietnam between 1994 and 1997 (n>800). The data set covers a wide range of soil nutrient supply, varieties, crop management, and climates. Grain yields ranged from 1.4 to 9.9 t ha-1 with an average of 5.1 t ha-1. The average harvest index was 0.47. The method of crop establishment greatly affected components of yield but differences in grain yield and internal efficiencies among areas with TPR vs WSR were mainly attributed to season- and site-specific conditions. The total nutrient removal per ton of grain yield in farmers fields ranged from 10 to 44 kg N, from 0.9 to 9.9 kg P, and from 6 to 42 kg K. Across all sites and seasons, average nutrient uptake requirements were 17.1 kg N, 2.9 kg P, and 16.1 kg K t-1 grain yield. Internal efficiencies ranged from 23 to 100 kg grain kg-1 N (mean 58), from 101 to 1069 kg grain kg-1 P (mean 350), and from 24 to 179 kg-1 K (mean 62). These are obviously extremely wide ranges representing situations where nutrients are either limiting or available in surplus in the plant.

The nutrient requirements of irrigated lowland rice can be estimated for specified target yields using a modeling approach based on Quantitative Evaluation of the Fertility of Tropical Soils (QUEFTS) developed by Janssen and coworkers in Wageningen, The Netherlands. The model requires the estimation of two boundary lines describing the maximum accumulation and dilution of N, P, and K in the plant (Fig. 6). These parameters were derived by exploiting the RDTP database on grain yield and plant nutrient accumulation in aboveground plant

Irrigated rice ecosystem


Grain yield (t ha-1) 10 8 6 4 2 0 0 YNA YKD YKA a YND b YPD c YKD

50 100 150 Plant N (kg ha-1)


10 20 30 Plant P (kg ha-1)


50 100 150 Plant K (kg ha-1)


6. Relationship between grain yield and accumulation of N, P, and K in total aboveground plant dry matter at maturity of rice, based on data with a harvest index greater than 0.40 kg kg1. The lines on the left of each figure represent the boundary of maximum dilution of N (YND), P (YPD), and K (YKD), and the lines on the right indicate the boundary of maximum accumulation of N (YNA), P (YPA), and K (YKA). Slopes of the boundary lines were calculated by excluding the upper and lower 2.5 percentiles of all internal efficiency data (IE, kg grain produced per kg nutrient in total aboveground dry matter, n 2,200). The respective slopes for the maximum and minimum IEs are 96 and 42 kg kg1 for N, 622 and 206 kg kg1 for P, and 115 and 36 kg kg-1 for K. IRRI, 1998.

Grain yield (t ha-1) 10 8 6 YNA








4 2 0 0 50 100

11 t ha-1 10 t ha-1 9 t ha-1 8 t ha-1 7 t ha-1 6 t ha-1











Plant N (kg ha-1)

Plant P (kg ha-1)

Plant K (kg ha-1)

7. The balanced N, P, and K uptake requirements (YN, YP, and YK) for targeted grain yields depending on the yield potential (Ymax) as calculated by QUEFTS. YND, YPD, YKD, YNA, YPA, and YKA are defined in Figure 6. IRRI, 1998.

dry matter (DM) at physiological maturity of rice. In addition to the farmers data, the database also contains data from Long-term Fertility Experiments at about 15 sites and on-farm trials including nutrient omission plots. Therefore, the data used here cover the whole range of possible internal efficiencies of N, P, and K in modern rice varieties (Fig. 6) across a yield range of 6-11 t ha-1. In QUEFTS, the relation between grain yield and nutrient uptake is assumed to be linear at lower uptake levels because the nutrient uptake would be at its maximum under conditions of limited nutrient supply. The actual plant nutrient accumulation un-

der such conditions should theoretically be close to the line of maximum dilution of the respective nutrient in the plant (e.g., YND in Fig. 6a), but it is not likely that all major nutrients can be maximally diluted. Instead, internal efficiencies of all three nutrients would be between their maximum and minimum values in an ideal situation of balanced N, P, and K nutrition so that neither nutrient is limiting or available in surplus. This situation is reflected by YN, YP, and YK (Fig. 7). At elevated yield levels, however, internal nutrient efficiencies decrease in a nonlinear fashion when actual yields approach the potential yield. The


IRRI program report for 1998

nonlinear part of the relationship between grain yield and nutrient accumulation, as predicted by the QUEFTS model, depends on both the definition of the boundary lines describing the envelope of maximum and minimum accumulation and on the maximum potential yield that is genetically determined (Y max, Fig. 7). Assuming that the internal efficiencies of modern high-yielding varieties are relatively similar, the standard boundary lines of maximum and minimum nutrient accumulation are applicable for all tropical and subtropical sites with irrigated lowland rice in Asia (Fig. 6, 7). On condition that plant growth is limited only by nutrient supply, QUEFTS predicted a linear increase in grain yield if nutrients were taken up in balanced ratios of 14.7 kg N, 2.6 kg P, and 14.5 kg K 1000 kg-1 of grain (internal efficiencies of 68 kg grain kg-1 N, 385 kg grain kg-1 P, and 69 kg grain kg-1 K) until yield targets reached about 70-80% of the climate-adjusted yield potential (Table 11). With yields approaching the maximum yield, internal efficiency of N, P, and K decreased. Regardless of the selected yield potential, the N:P:K in plant DM as recommended by QUEFTS is about 5.7:1:5.6 (Fig. 7), which is similar to the average plant N:P:K of 5.9:1:5.5 that was derived from plant measurements at project sites (Fig. 6).


The concept of nutritional balance is further explained using data from farmers fields in Central Luzon and the experimental station at PhilRice, Maligaya, Philippines. The maximum potential yield is greater in the DS (10 t ha-1) than in the WS (6 t ha-1), mainly due to lower solar radiation in the latter. Thus, YN, YP, and YK differ in DS (straight lines) and WS (broken lines) when yields exceed 5 t ha-1 (Fig. 8). Grain yields were usually greater in the DS than in the WS and there were even clear indications that plants were N-limited in the DS because most data points were closer to the line of maximum N dilution (YND, Fig. 8a). In contrast, most WS data points were below the optimum line of N accumulation as predicted by QUEFTS (YN, Fig. 8a) and, independent of fertilizer N application, closer to the line of maximum accumulation. The WS results were probably caused by unfavorable weather, especially during the grain-filling period. This was well reflected by the lower harvest index of the WS than the DS crops (0.42 vs 0.48). Thus, it is economically reasonable to apply more fertilizer N to the DS than the WS crops.

Table 11. Balanced uptake requirements, internal efficiencies (kg grain kg-1 nutrient), and reciprocal internal efficiencies (kg nutrient 1,000 kg-1 grain) of N, P, and K for irrigated lowland rice as calculated by QUEFTS to achieve certain grain yield targets. Grain yield potential was set to 10 t ha-1. IRRI, 1998. Yield (t ha-1) Required nutrient uptake (kg ha-1) N 1 2 3 4 5 6 7 7.5 8 8.5 9.0 9.5 9.8 9.9 10.0 15 29 44 59 73 88 104 115 127 142 159 182 205 217 243 P 2.6 5.2 7.8 10.4 13.0 15.6 18.4 20.4 22.6 25.1 28.2 32.2 36.3 38.6 43.1 K 15 29 43 58 72 87 103 114 126 140 157 180 203 215 240 N 68 68 68 68 68 68 67 65 63 60 57 52 48 46 41 Internal efficiency (kg grain kg NPK-1) P 385 385 385 385 385 385 380 368 354 339 319 295 270 256 232 K 69 69 69 69 69 69 68 66 64 61 57 53 48 46 42 Reciprocal internal efficiency (kg NPK 1,000 kg grain-1) N 14.7 14.7 14.7 14.7 14.7 14.7 14.8 15.3 15.9 16.7 17.6 19.2 20.9 21.9 24.3 P 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.7 2.8 3.0 3.1 3.4 3.7 3.9 4.3 K 14.5 14.5 14.5 14.5 14.5 14.5 14.7 15.1 15.7 16.5 17.4 19.0 20.7 21.7 24.1

Irrigated rice ecosystem


Grain yield (t ha-1) 10 a YN b YPD YP



0 0 50 100 Plant N (kg ha1) 150 200 0 10 20 Plant P (kg 30 ha1) 40

8. Relationship between grain yield and accumulation of N and P in total aboveground plant DM at maturity of rice in Maligaya, Central Luzon, Philippines, 1995-97. Data shown include farmers' fields (27-48 farms) and various field experiments conducted at PhilRice. IRRI, 1998.

Plant P accumulation was generally close to the line of maximum dilution in both DS and WS crops, which indicates insufficient P supply to the rice crops (Fig. 8b). Practical implications. The derived standard model (Fig. 7) appears to be valid for all indica rice varieties with a harvest index of about 0.50 and can be used regardless of the method of crop establishment. Therefore, estimating crop nutrient requirements within a site-specific nutrient management approach in irrigated rice requires no site- or season-specific information other than the climate-adjusted potential yield, which can be obtained from crop simulation models, long-term experiments, or expertise (Fig. 7). It is more profitable for farmers to maximize nutrient efficiencies by a more balanced nutrition than to aim for yield targets that are approaching the maximum potential yield. The approach using QUEFTS allows not only the estimation of nutrient requirements to achieve a certain yield target; it also provides a useful tool for identifying economic yield targets and yield levels at which nutrient use becomes too inefficient. The nutrient requirements of rice as defined by using our large data set and the

QUEFTS model were noticeably much lower than values commonly used in the literature.

Increasing water use efficiency in rice culture

Management of cracked soils for water saving during land preparation T.P. Tuong and R. Cabangon Drying of a puddled soil during the fallow period between rice crops usually results in soil shrinkage and cracking. Irrigation for land preparation of the next rice crop, thus, involves water application to cracked rice soils and results in bypass flow through the cracks. We hypothesized that measures that minimize crack formation or impede the flow of water through the cracks may reduce these bypass flows and decrease the water requirement. Field experiments in the Philippines assessed straw mulching and dry shallow tillage as possible measures to reduce bypass flows during land preparation. The study was in IRRI fields during the 1993 and 1994 DS; in the Angat River Irrigation System, Bulacan, during the 1993 WS; and in Muoz,


IRRI program report for 1998

Nueva Ecija, during the 1995 DS. Plots (6 m 11 m) at IRRI were sun-dried during the fallow period until irrigation for land soaking for the next rice crop. Three soil management treatments were imposed: q Straw mulching. Rice straw (5 t ha -1) was broadcast over the plot 1 d after drainage (DAD). The straw mulch remained on the soil surface during the fallow period. This treatment was not included in 1994. q Shallow surface tillage. Plots were rototilled to 5-10 cm depth by two passes of rototiller at 18 DAD in 1993 and 14 DAD in 1994. Shallow tillage resulted in a topsoil consisting of clods with an average diameter of about 20 mm. q Control. No treatment was applied during the fallow period. In Bulacan, six farmers fields were selected. In Nueva Ecija, the experiment was in eight farmers fields. Two tillage treatments (control and shallow surface tillage) were imposed prior to the land preparation period. Crack depth and width were monitored at 1- to 3-d intervals during the fallow period in one 1- 1m subplot in each plot (IRRI Program Report 1995). We also measured irrigation application rates and monitored the dynamics of the water table and the distance traveled by the surface waterfront along the central longitudinal transect of the control and the shallow tillage field. The land preparation period was divided into two stagesland soaking (from first water application to harrowing) and harrowing stage (from harrowing to leveling). The water flow components in each phase, expressed in mm of water over each field, can be quantified with the following equation (IRRI Program Report 1995): I + R = Ss + Sc + A + E + L where I = irrigation water; R = rainfall; Ss = surface water storage; Sc = crack storage, i.e., the amount of water that fills the cracks in the field; A = water absorbed in the topsoil matrix (0-0.2 m depth); E = evaporation from the field; and L = losses. I, R, Ss, and E were measured directly, Sc from volume of cracks, and A from the changes in soil moisture contents. Losses were derived from the difference between (I + R) and (Ss + Sc + A + E).

Both crack width and depth increased more rapidly in the control plots than in the mulched plots (Fig. 9). In 1993, at the end of the fallow period, the crack width in the mulch treatment was significantly lower than that in the control treatment. This corresponded to wide differences in the final moisture content of the soil surface layer in the two treatments (data not shown). The final mean crack depth in the mulch treatment was also less, but not significantly, than that in the control treatment. During land soaking of the control treatment in Nueva Ecija, water in the cracks advanced faster than on the soil surface. In the control treatment, the water table rose close to the soil surface (less than 30 cm) at a position 10 m ahead of the advancing surface waterfront. No significant rise of the water table was observed ahead of the surface waterfront in the shallow tillage treatment (Fig. 10). The surface waterfront advanced faster in the tilled treatment than in the control. Small soil aggregates in the shallow tillage treatment blocked the cracks, made them discontinuous, impeded the water flow in the cracks, and reduced the recharge to the groundwater. Less bypass flow meant higher flow
Crack width (mm) 140 120 100 80 60 40 20 0 Crack depth (mm) 140 120 100 80 60 40 20 0 0 10 20 30 40 Days after drainage 50
Control 1993 Mulched 1993 Rainfall Control 1993 Mulched 1993

Rainfall (mm) 70 b 60 50 40 30 20 10 0 60

9. Mean crack widths (a) and depths (b) in different treatments after field drainage at the IRRI fields, 1993 DS. Bars indicate standard errors of the mean values for 55-161 measurements.

Irrigated rice ecosystem


Water table depth (cm) 2 0 -20 -40 -60 -80 -100 -120 -140 1 2 3 0 Distance from surface waterfront (m) 4
Initial water table depth Control Shallow tillage Surface water

Water savings during land preparation may increase the service area of an irrigation system. In rainfed areas, shallow surface tillage may also lead to earlier crop establishment and, thus, reduce the risk of late-season drought. Population dynamics of Echinochloa crus-galli in wet-seeded rice M. Mortimer, R. Lubigan, A. Man,9 P. Vongsaroj,10 and D. Chin7 The Echinochloa species constitute a serious threat to rice production in all rice ecosystems. In irrigated rice, they are one of the most noxious weed groups worldwide. They mimic rice in a number of biological and ecological characteristics and require integrated management practices for successful control. An assessment of Echinochloa crus-galli in wetseeded rice (WSR) was conducted to determine the effect of early flooding on weed seedling recruitment, yield loss in relation to infestation density, and long-term population change. A field experiment was conducted at sites in Malaysia, Thailand, Philippines, and Vietnam. At each site, a representative ecotype of E. crus-galli was seeded to give a range of infestation densities in replicated WSR plots. Treatments varied the onset of flooding and compared early flooding (4 DAS) with late flooding (12 DAS). Seedling counts of E. crusgalli and rice were taken during 14 d following sowing. The number of surviving E. crus-galli plants at maturity was recorded together with seed production of survivors and rice yields. The experiment was then continued into the following season. Studies of seedbank longevity were conducted concurrently with the field experiment. Seeds were buried in open mesh bags at different depths, samples recovered at monthly intervals for 12 mo, and the viability of surviving seed determined. Figure 11 illustrates the representative flux in population size (log scale) of E. crus-galli from an initial infestation of 1,000 seeds m-2 over two cropping seasons. This underlying pattern in the population dynamics of the weed was common to all sites. Early flooding, as opposed to late flooding, resulted in about 50% seedling mortality and prohibited the successive recruitment of seedling cohorts. Surviving plants increased the seed population size by about two orders of magnitude by the time of weed seed dissemination at rice harvest

10. Groundwater profiles with distance from surface waterfront in control and shallow tillage plots during flood irrigation for land soaking at Nueva Ecija, 1993. Bars indicate standard deviation for 25 (in the control) and 23 (shallow tillage) measurements.

rate was available for the surface water flow, and resulted in a faster rate of advance of the surface waterfront in the tilled plots. Shallow tillage can thus help reduce time for land soaking. Straw mulching, compared with the control plots, did not significantly reduce the amount of irrigation water for land soaking at IRRI. The shallow-tilled plots used significantly less water for land soaking than the control plots. This resulted from much lower bypass flow in the shallow-tilled plots. The amounts of irrigation water for land soaking and water loss in shallow-tilled plots were significantly less than in the mulched plots, despite similar crack depths in shallow tillage and in the mulched plots. This highlights the importance of small soil aggregates in blocking and impeding water flow through the cracks. Shallow surface tillage in farmers fields reduced the total water input for land preparation by 31-34% of the amount needed in the control plots. Much of the differences in water use between the two treatments occurred before the first harrowing was performed. After harrowing, the loss rates were not significantly different among treatments in both Bulacan and Nueva Ecija. Harrowing broke soils into aggregates, which sealed the cracks and produced puddling, which reduced soil permeability. Shortening the duration between land soaking and the first harrowing may be an important measure to reduce water loss during land preparation.


IRRI program report for 1998

Numbers m-2 (log scale)

95% Cl

Seeds at harvest

Stage 2

Seeds at harvest

100,000 10,000
Stage 1


Stage 3

100 10 1


Stage 4
Early flooding Late flooding Seedlings Seedlings

14 1997 DS

112 DAS Fallow

19 1997 WS

109 DAS

11. The population dynamics of Echinochloa crus-galli in wet-seeded IR72, during 1997 WS and DS. Note interrupted time line. IRRI, 1998.

(stage 2). In the Philippines, the population suffered a 97% loss by stage 3 through seed removal with rice during harvesting and seed germination in the fallow period. However, surviving plants compensated by increased seed production to return the total seed population to high densities, by the end of the second season. Rice yield losses due to E. crus-galli were severe. Figure 12 illustrates typical yield loss curves at the Malaysian Agricultural Research and Development Institute in relation to weed density and flooding time. Yield declined at all sites sharply with increasing weed density. No statistically significant differences were detected in yield due to alteration of flooding time within sites. Intersite comparisons suggested that densities of 10 E. crusgalli plants m-2 resulted in rice yield loss of between 0.8 and 1 t ha-1. The persistence of viable E. crus-galli seed populations in the soil was quantified in terms of the time taken to decline by 50% (half-lives). Noticeable differences were evident among sites. Buried seed populations in Vietnam had a half-life of 190 d and was independent of burial depth (5 or 15 cm), whereas longevity increased with depth at other sites and half-lives in excess of 1000 d were recorded at 15 cm burial.

Yield (t ha-1) 4 Early flooding Late flooding

3 2

1 0 25 50 75 100 125

Weed density (plants m-2)

12. Yield-density relationships for Echinochloa crus-galli in wet-seeded rice. MARDI, 1996-97.

This study exposed noticeable differences among E. crus-galli populations in terms of yield loss and demographic traits, but the significant result was the importance of early flooding in regulating population increase. Advancing the time of flooding was beneficial at all four sites in terms of reducing seedling recruitment (by up to 50%) and did not adversely affect final rice yield. This practice, however, was insufficient as a sole long-term measure of weed control due to strong compensatory processes in the weed species.

Irrigated rice ecosystem


Semidry rice production for irrigation water savings S.I. Bhuiyan, D. Tabbal, and E.B. Sibayan The increasing scarcity of water in irrigated rice areas and competition from industrial, domestic, and nonrice agricultural sectors mean that rice cultivation in the future must be more water-efficient. Increased water use efficiency was examined by collaborative research by IRRI and PhilRice on semidry rice production system in farmers fields at San Jose City, Central Luzon, Philippines during 1995-97 WS. Semidry rice describes a crop established as dry-seeded rice (DSR) during the premonsoon period, nurtured as rainfed crop for most of the vegetative growth stage, and then irrigated during the remaining growth periods. Our research objectives were to q establish the feasibility and management requirements of semidry production system within irrigated areas; q determine opportunities for saving irrigation water and more effective use of rainfall and to quantify them; and q identify the requirements for, and assess the potential impact of, widespread adoption of semidry rice system. Four treatment combinations consisting of two water regimes during the irrigation period and two weed control methods in four replications were laid out in farmers fields. The two water regimes were continuous shallow (2-5 cm depth) standing water and continuously saturated soil, while the two weed control methods were recommended chemical weed control and chemical weed control plus hand weeding. The data from the field experiments were supplemented by data from a survey of 21 randomly selected farmers to assess their perceptions on the feasibility and practical applicability of semidry rice cultivation. In 1998 WS, four farmer collaborators volunteered to adopt the semidry rice production technique. q The 3-yr field experiments showed high potential for water savings. q Yields obtained in 1995 WS from plots with weed control by both chemical and hand weeding were significantly higher than those

obtained in plots with chemical control (Table 12) only. In general, yields in the 1995 experiment were relatively lower because of typhoon damage. In 1996 WS, when weed growth was less, hand weeding did not significantly add yield to that obtained from chemical weed control only (Table 12). In 1997 WS, the highest yield was 4.7 t ha-1 for the treatment combinations in which the plots have either shallow standing water or saturated soil condition plus chemical weed control and spot hand weeding. In contrast, average yields in neighboring farms were 4.1 t ha-1 for transplanted (TPR) and 3.9 t ha-1 for wet-seeded rice (WSR) (Table 12). Irrigation water use for land preparation for all treatments was basically eliminated. Fields were plowed, harrowed, leveled, and seeded as dry or moist soil. Rainfall supported seed germination. Previous water management studies in the area had reported average water use for land preparation amounting to 740 mm for the WSR and 895 mm for TPR (IRRI Program Report for 1993). Those amounts are equivalent to about 40% of total water used in WSR and TPR. The potential amount of water that can be saved by a semidry rice system for land preparation is 7,400-8,950 m3 ha-1.

Table 12. Average yield of dry-seeded IR64 in different treatment combinations of two water regimes and two weed control methods at different locations and seasons. Upper Pampanga River Integrated Irrigation System (UPRIIS), Central Luzon, Philippines. 1995, 1996, and 1997 WS. Av yield (kg ha-1) Treatment combinationa 1995 WS (5 sites, 2 reps) 1404.8 c 2320.9 b 1249.7 c 2433.3 b 1852.2 320.3 23.0 1996 WS (3 sites, 4 reps) 4292.4 a 4383.1 a 4142.5 a 4201.2 a 4254.8 b 944.4 8.42 1997 WS (2 sites, 4 reps) 4663.7 a 4728.0 a 4424.0 a 4667.2 a 4620.7 b 501.0 6.35

TC1 (chemical) TC2 (chemical + hand weeding) TC3 (chemical) TC4 (chemical + hand weeding) Mean LSD CV (%)

a TC1 and TC2 = continuous standing water (2-5 cm depth), TC3 and TC4 = continuous saturated soil condition up to field capacity.


IRRI program report for 1998

The irrigation water used during the crop growth stage was low as rainfall was more effectively used and irrigation inputs were supplemental to rainfall. The average water use (irrigation plus rainfall) during the 1995, 1996, and 1997 WS ranged from about 1,487 mm for the plots with continuous standing water and 1,231 mm for the continuous saturated soil conditions (Table 13). Twenty to 30% of those water uses were in terms of irrigation water applied for maintaining continuous standing water and continuously saturated soil. The 21 selected farmers who were interviewed after the 1997 WS were unanimous in saying that the semidry rice production technique is a good alternative to TPR. They emphasized that the DSR technique is appropriate, especially when irrigation water is scarce. However, some farmers (15%) expressed the concern that weed control could become a serious problem, and 5% of farmers perceived that the yield may be lower. About 5% of farmers were apprehensive that it would be difficult to adopt the method in areas with clayey soils because of tillage problems.

The management of the National Irrigation Administration in Central Luzon, Philippines, advanced the 1998 cropping schedule by 1 mo and conducted a massive campaign for the adoption of semidry rice technology within the service area of the Upper Pampanga River Integrated Irrigation System (UPRIIS). About 2,000 farmers, covering an area of 5,000 ha, adopted the technology. The average crop cut yields of four farmer cooperators who volunteered to adopt the technology in 1998 WS ranged from 3 to 3.7 t ha-1. A survey of 17 randomly selected farmers within the vicinity of the experimental area, and who also adopted the semidry rice technology, indicated that about 53% obtained an average yield of 4 t ha-1 which was statistically the same as the yields from their WSR or TPR in the previous 1997 WS (Table 14). About 47% of farmers obtained an average yield of 3.4 t ha-1, which was significantly lower than in the previous WS. Despite the lower yield, the farmers were satisfied because they were able to reduce the labor for transplanting and chemicals and to grow a rice crop with scarce water supply.

Table 13. Average water usea (mm) of dry-seeded IR64 rice in different treatment combinations of two water regimes and two weed control methods at different locations and seasons. Upper Pampanga River Integrated Irrigation System, Central Luzon, Philippines, 1995, 1996, and 1997 WS. Treatment combination 1995 WS TC1 (chemical) TC2 (chemical + hand weeding) TC3 (chemical) TC4 (chemical + hand weeding) 1131.2 a (8.8%) 1118.4 a (8.0%) 1100.8 a (6.8%) 1088.3 a (5.7%) 1109.7 44.3 3.47 Seasonb Mean 1997 WS 1403.9 a (38%) 1430.2 a (36.6%) 1340.8 a (32.7%) 1318.1 a (31.7%) 1389.7 209.0 5.48 1997 WS 1976.4 a (50.2%) 1863.4 ab (47.9%) 1279.2 ab (28.5%) 1257.7 b (27.4%) 1594.2 501.2 16.08 1503.8 (32.3) 1470.7 (30.8) 1240.3 (22.7) 1221.4 (21.6) 933.1 1356.8

Mean LSD (.05) CV (%)


Total water use (mm) is irrigation water applied + rainfall. bNumbers in parenthesis are percent of irrigation applied.

Irrigated rice ecosystem


Table 14. Average yields (t ha-1) obtained by two groups of farmer adopters of semidry rice in 1998 WS compared with their yields obtained from either WSR or TPR in 1997 WS. UPRIIS, Central Luzon, Philippines. Yield (t ha-1) Farmers' group DSR (1998 WS) I (high yield, 53%) (n=9) II (low yield, 47%) (n=8) Total (100%) (n=17) 4.0 a 3.4 a 3.7 a WSR or TPR (1997 WS) 4.2 a 5.3 b 4.7 b

Farmers who sprayed (%) 100 80 60 40 20 0 Seedling

August 1994 February 1996 March 1997

Tillering Booting Heading Maturing Crop stages

Improving pest management

Motivating farmers to test pest management changes through use of printed materials and radio K. Heong Insecticide use decisions of many rice farmers are based on perceptions of potential damages and losses caused by pests. Farmers generally overestimate the seriousness of the rice leaffolder from its highly visible damages and apply insecticides in the early crop stages. Because perceptions, rather than economic rationale, determine most farmers pest management decisions, changing their perceptions may help reduce unnecessary spraying. Farmers in Long An Province, Vietnam, were motivated to test a heuristic, Insecticide spraying for leaffolder control in the first 40 d after sowing is not needed. The communication media reached 97% of the farmers in the study sites with 21,000 households. The leaflet, radio drama, and poster had the most effective reach. During 31 mo after the media campaign, farmers mean insecticide sprays dropped significantly from 3.35 to 1.56 sprays farmer-1 season-1. The proportion of farmers who did not use any insecticides increased from 1 to 32%. There were also significant reductions in number of farmers spraying at the early crop stages of seedling, tillering, and booting (Fig. 13). Farmers perceptions of leaffolder damage were favorably changed. The proportion of farmers who believed that leaffolders could cause losses was reduced from 70 to 25%. The proportion who believed that early-season spraying was required was reduced from 77 to 23%. Farmers insecticide spray frequencies and the belief index were significantly correlated. Cost saving (insecticide and labor) as the most important in-

13. Rice farmers' insecticide applications in different crop stages in Tan Tru and Tan Thanh districts, Long An Province, Vietnam, before (Aug 1994) and after (Feb 1996 and Mar 1997) introduction of the communication media materials.

centive to stop early-season spraying was cited by 89% of the farmers. A survey of the other districts in the province showed that the message reached 82% of the provinces 210,000 households. About 77% stopped early-season spraying, about 20% had not applied any insecticides, and the average insecticide sprays was 1.6. That compared with 1.56 in study sites. The belief index was 8.11, which was not significantly different from the study sites. The media approach stimulated 15 provincial governments to launch their own program and extended it to the whole Mekong Delta. That potentially reached a farm household population of 2 million. Testing the validity of surrogate taxa for canopy and floodwater invertebrates K. Schoenly, H. Justo, Jr., A. Barrion, M. Harris, and D. Bottrell An issue confronting plant protection specialists is how to characterize tropical rice ecosystems with its staggering biodiversity, interconnectedness, and spatial-temporal flux. One approach, advocated by conservation biologists, is the use of single taxa as surrogates for all taxa. This approach assumes that the characteristics of single taxa represent biodiversity over wide ranges and disturbance gradients. Surrogate-based methods of biodiversity assessment were applied to invertebrate time-series data collected during DS from a farmers irrigated rice field in Calauan, Laguna Province, Philippines.


IRRI program report for 1998

Canopy and floodwater invertebrates were vacuumand strainer-sampled, respectively, at roughly weekly intervals from seedling to harvest (eight sampling dates). The cumulative samples included 202 taxa and 9,570 individuals for the plant canopy and 180 taxa and 84,905 individuals for the floodwater. Because ideal surrogate taxa should be abundant, we analyzed the most common taxa sampled from the canopy and floodwater faunas. In each test, time-series abundance of each taxon was plotted against time-series abundance of all canopy or all floodwater taxa, depending on the origin of the sample. To determine the suitability of a surrogate taxon in predicting total invertebrate abundance, we used the coefficient of correlation (r), judged at P = 0.05, to test the linear fit of the eight date-specific abundance of the independent (surrogate taxa) variable against the dependent (all taxa) variable. This process was repeated for each taxon that fell within the 95% abundance threshold, established from rankabundance curves for canopy (82) and floodwater (24) samples. Because of the small sample size of each correlation (n = 8), the likelihood was high that one observation could bias the regression toward statistical significance. Therefore, statistical significance and visual inspection of each correlation were used as companion tests to make final suggestions for surrogate taxa. Of the 82 taxa that made up 95% of the total abundance of canopy invertebrates, 5 taxa were significantly positively correlated with all canopy taxa: veliid bugs (Microvelia douglasi atrolineata), corixid bugs (Micronecta quadristrigata Bredin), mymarid wasps (Anagrus spp.), brown planthoppers (Nilaparvata lugens Stl) and notonectid bugs (Anisops sp.). Of these, mymarids gave the best linear fit to the data. However, for all five taxa, each correlation was driven by a single outlying data point (41 DT) whose subsequent removal made each correlation nonsignificant. Owing to their sensitivity to single sampling dates, we judged these five taxa as unacceptable surrogates of the canopy fauna. Of the 24 floodwater taxa that made up 95% of the total abundance of floodwater invertebrates, 5 were significantly positively correlated with all floodwater taxa. After removing single outlying taxa, our analysis revealed that all five were acceptable floodwater surrogates: ostracods (Heterocypris

luzonensis), chironomid flies, corixid bugs (Micronecta quadristrigata), hydrophilid beetles, and gastropods (Pila sp.). Hydrophilid beetles and four other taxa (ostracods, chironomids, corixids, and gastropods) were judged as reasonable surrogates of total floodwater biodiversity (abundance) in the dry season. Conversely, none of the 82 canopy taxa tested were found to be suitable surrogate taxa of the canopy fauna. The floodwater surrogates satisfied criteria of ease of sampling and continuous presence over the crop cycle, but whether they also exhibit wide geographical distribution in Asian rice ecosystems requires further investigation. Directional movement of predators between the irrigated rice field and its surroundings L. Sigsgaard and S. Villareal Unsprayed irrigated rice fields have proven to be almost pest-free, largely due to the existing predator fauna. Early establishment of a predator complex requires that predators either remain in the field during fallow or that they move into the field from the outside. We assessed the possible impact of nearby surroundings supplying predators to a newly established irrigated rice field or vice versa. Predators were trapped in directional pitfall traps. Preliminary results of catches around an irrigated rice field at IRRI during the first half of the rainy season of 1998 showed that the most common predators caught were spiders, ants, and carabids. The most caught spider was P. pseudoannulata (Fig. 14). The highest relative abundance of P. pseudoannulata was found in the bund, stressing the importance of this habitat for this important predator of the irrigated rice agroecosystem. P. pseudoannulata was present early and brought its young into the newly established field. Later the directional movement from the field was reversed and the field may have changed to become a source of P. pseudoannulata to other fields. After harrowing, the number of P. pseudoannulata adults increased, while the number of 2-10 instar spiderlings was reduced to one-tenth. Mortality inflicted in the less mobile younger instars is the most probable explanation. Farm operations also affected directional movement. Thus, immediately after rice planting, there was a significant directional movement of females out of the

Irrigated rice ecosystem


P. pseudoa nnulata S. geminata

Habitat and trap type Bund-pitfall Bund-bv Uncropped- bv Field-bv 0

C. formosan a Pheidole sp.

Other spid ers Other ants

P. javanus E. stali

Other carabids






Percent of total number caught

14. Relative abundance of the most common predators caught in pitfall traps in the bund (n = 90), and with blower-vac (bv) in the bund (n = 72), the field (n = 480), and in uncropped area next to the field (n = 24).

field. Results indicate that directional movement between the field and its surroundings, as well as total catches of predators, depend on farm operations and crop age. The bund may be an important source of predators for the young rice field. Seed health evaluation for farmer rice crop management S.D. Merca, L. Diaz, M. Hossain, and T. Mew A study of seed quality of 590 lots of farmer rice seed from Cavite, Laguna and Quezon provinces, Philippines, revealed pure seed fraction ranging from 88 to 95%, rice mixtures of 4-12%, and a weed seed contamination of 22-35 weed seeds 40 g-1 of rice seed. Most Nueva Ecija farmer rice seeds were below the seed qualities set for certified seeds. Different seed health practices were evaluated to improve farmer seed health management. Chemical seed treatment gave an additional yield of 351 kg ha-1. Flotation increased yield by 255 kg ha-1 and manual sorting gave additional yield of 742 kg ha-1. Manual seed sorting consistently gave a significant 8-20% increase in grain yield. Use of high-quality IR64 seeds showed that physical quality, seedling vigor, and seed infection were significantly different from farmers IR64 seeds. In the three trials involving 30 farmers season-1 in Nueva Ecija, an increase of 8-14% yield of high-quality IR64 seed over farmers IR64 seeds was demonstrated, which indicated a degeneration of potential yield when farmers handle their own seed production. Farmer seed had varietal mixture,

weed seed contamination, and lower physical qualities. Our findings suggest that the use of high-quality seeds or certified seeds is imperative to improve yield and realize the genetic potential of improved rice varieties. An estimated 85-90% of farmers cannot obtain certified seeds. For them, seed health management is vital to maintain physical and genetic purity of improved rice varieties and maintain their yield potential.

Coping with global climate change: reducing methane emission from rice fields
R. Wassmann and R. Lantin Activities on methane emissions from rice fields were conducted at IRRI and at eight collaborating research stations in Asia. Methane emissions were generally high in irrigated rice, but could effectively be reduced in various environmental settings through mid-season drainage. The seasonal courses of methane emissions in rainfed rice were determined by precipitation patterns. In Jakenan, Central Java, Indonesia, the shift from dry to wet conditions resulted in a gradual increase of methane emission rates during the wetseason crop (Fig. 15). The reverse shift from wet to dry conditions caused high emissions in the early stage and low emissions in the late stage of the ensuing dry-season crop. In deepwater rice, average fluxes were relatively low while cumulative values were high due to long


IRRI program report for 1998

Methane emission (mg CH4 m-2 d-1) Emission Rainfall

Rainfall (mm)







0 15 Oct

14 Nov

14 Dec

13 Jan

12 Feb

14 Mar

13 Apr

0 13 May

1996-97 WS

1997 DS

15. Methane emission rates (area) and rainfall (bars) during two consecutive seasons (1996-97) in Jakenan, Central Java, Indonesia.

duration of flooding in a given season. Irrespective of the ecosystem, methane emission rates were extremely sensitive to quantity and quality of organic amendments. The mode of fertilizer application seemed to have only a minor effect on emission rates. Direct seeding yielded substantially lower emission rates throughout the season. The impact of rice varieties on methane emissions was detectable in field and greenhouse experiments, but was in a smaller range than the impact of water regime and organic inputs. Furthermore, the methane emission potential of a given cultivar showed a pronounced variability, depending on the nutritional status of the plant. Phosphorus deficiency increased root exudation and, in the next step, enhanced methane production and emission. In all cultivars, the plants methane transport capacity was linearly correlated to tiller numbers, indicating that methane transport is determined by outlets rather than biomass. Results in 1998, however, did not allow a distinct classification of rice cultivars according to emission potentials. The experimental findings of the project were incorporated in a model to simulate methane emissions from rice fields. The CERES-Rice crop simulation model was used as a basis, with existing routines simulating soil organic matter decomposition to predict the amount of substrate available for methanogenesis. This was linked to a submodel cal-

culating steady-state fluxes and concentrations of methane and oxygen in flooded soils. Extra routines were also incorporated to simulate the influence of the pool of alternative electron acceptors in the soil. The model was able to explain the seasonal variations in methane emissions in an experiment involving mid- and end-season drainage and additions of organic material. Because the model also predicts grain yields, it may be further applied to identify situations for modifying crop management for higher yields and lower methane emissions.

Irrigated Rice Research Consortium

R. Zeigler, P. Sta. Cruz, and M.A. Quilloy The Irrigated Rice Research Consortium (IRRC) aims to develop multi-institutional and multidisciplinary research partnerships to solve critical productivity and sustainability problems in the irrigated rice ecosystem by formally linking highly successful, but separate, research initiatives. The Consortiums anticipated impacts are q Creation of a regionwide, multi-institutional, multidisciplinary, and integrated research mechanism; q Increased irrigated rice system productivity and efficiency through balanced and optimum combination of inputs at high production levels;

Irrigated rice ecosystem


Creation of environmentally and ecologically sound rice production technology through improved resource use efficiency and input management.

Progress of unreported project

Improving the productivity and sustainability of rice-wheat systems J.K. Ladha

During 1998, IRRC activities included q general coordination of IRRC with regard to its membership, agenda, resource allocation, and related matters; q conduct of the joint experimentation on nutrient-pest interaction; q training-planning workshops for the joint experimentation, and second joint meetings of the IRRC, INMNet, and IPMNet steering committees; q support of Consortium participants in IRRIconducted training courses; and q establishment of the IRRC web page. Under an IRRC-led nutrient-pest interaction activity, work plans, research protocols, and implementing guidelines were developed; participating institutions and scientists were identified; and field researcher capability was calibrated. Field experiments were established on pest impact assessment in the RTDP Project monitoring farms and crop residue management. Their first cropping cycle was monitored. Agreement was reached to include an ADB-funded Hybrid Rice Network under the IRRC. The governance and management of Integrated Pest Management Network (IPMNet), Integrated Nutrient Management (INMNet), (RTDP), and Hybrid Rice Network (HRNet) in the IRRC will continue to q explore opportunities for research collaboration, direction, and broad resource allocation; q identify and lay out mechanics for integration of other networks, projects, and other research interfaces in the Consortium; and q strengthen interdisciplinary research on soil and crop management, professional development, and interaction among rice scientists in the Consortium.

Soil N supply capacity was found to be higher in rice than in wheat. Dynamic nutrient tests for P and K relate well to yield and P and K uptake. Nitrogen use efficiency is lower in rice than in wheat in silt loams but equal in loamy sand. Nitrification-denitrification and leaching are the main loss pathways. Surveys conducted to assess the importance of pests at Pantnagar and Faizabad in India allowed further documentation of the injury profiles of rice (which are dominated by brown spot disease and weed infestation). Soilborne and seedborne rice pathogens associated with typical crop rotations in the system have been characterized in a small sample of farmers fields at Pantnagar, India. This has led to the development of methods that will be used on a larger scale. Continued collection of on-farm data from the rice-wheat system in Pantnagar to analyze trends in productivity and the profitability of site-specific nutrient management.

Program outlook
The program will maintain its efforts toward raising the yield plateau by developing the NPT and hybrid rices backed up by a systems approach that integrates understanding of yield determinants. Radiation use efficiency, high yield determining growth patterns, N source nutrition, new genes for tungro viruses, and blast resistance are some of the research considerations that are integrated in the breeding of the NPT. The program will work on sustaining soil quality in intensive rice systems. Emphasis will be on optimal N applications, internal nutrient efficien-


IRRI program report for 1998

cies, and modeling the nutritional balance, all of which are directed toward sustainable nutrient management in intensively cropped irrigated lowlands. Identification and quantification of key biotic and soil organic determinants of sustainability and ecological resilience will be addressed. Studies on the constraints to nutrient supply and the development of practical approaches through site-specific nutrient management technologies in partnership with NARS and farmers are under way. Efforts are under way to improve water use at farm and irrigation system levels. That includes saving water during land preparation and within the cultivation period, weed population dynamics in DSR, and crop establishment methods under semidry production conditions. Understanding of the processes of water quality degradation from agrochemicals and development of feasible options for its alleviation will be a continuing effort. Characterization of pest problems and generation of practical pest management strategies will be sustained.

The IRRC will provide the institutional framework for integration of three networks: IPMNet, INMNet, and HRNet, which are involved in helping the NARS in developing and using these technologies. This integration of research efforts will link and better benefit from each others experiences and outlook. The IRRC will eventually include other aspects of irrigated rice research, such as the Rice-Wheat Consortium, farm mechanization, and systemwide program on water management and global climate change. Two SDC-supported projects, the IPM Project (within IPMNet) and RTDP Project (within INMNet), and the ADB-supported Hybrid Rice Network (HRNet) have been formally linked in a consortium mode. The nutrient-pest interaction research in intensive irrigated rice systems is under way. Initial research work within IPMNet, INMNet, and HRNet interfaces is expected to start by the year 2000. IRRC eventually will include other aspects of irrigated rice research on water management, farm mechanization, and global climate change.

Irrigated rice ecosystem


Research programs Rainfed lowland rice ecosystem

MANAGING CROP, SOIL, AND WATER RESOURCES FOR ENHANCED PRODUCTIVITY AND SUSTAINABILITY 34 Rainfed rice and risk-coping strategies (SS) 34 Improved water conservation and nutrient use efficiency via subsoil compaction and mineral fertilization (APPA) 36 Analyzing weed community dynamics in rainfed lowland rice (APPA) 38 GERMPLASM IMPROVEMENT FOR RAINFED LOWLAND RICE 39 Germplasm for rainfed lowland ecosystems (PBGB) 40 Breeding materials adapted to drought (PBGB) 40 Screening for drought resistance at vegetative stage (PBGB) 41 Mechanisms of submergence tolerance and recovery (APPA) 41 Phosphorus efficiency (SWS) 43 Organic anion release into nutrient solutions 44 Organic anion release into soil 44 Phosphate solubilization by organic anions 45 Mapping genes for root traits (APPA) 46 Water extraction and recovery ability following drought (APPA) 47 PROGRESS OF UNREPORTED PROJECTS 49 Characterizing and analyzing rainfed rice environments 49 Addressing gender concerns in rice research and technology development 50 Rainfed Lowland Rice Research Consortium 50 PROGRAM OUTLOOK 51

Rainfed lowland rice ecosystem

Rainfed lowland rice is grown on about 25% of the worlds rice area (nearly 40 million ha) and contributes 18% of the global rice supply. Average yields are low and farmers grow mainly traditional varieties. The ecosystem has high potential for increasing production to meet the increasing demand for high-quality rice. The objectives of research in the rainfed lowland rice ecosystem include q better understanding of anthropological, socioeconomic, and biotic components to allow identification of major constraints and opportunities for improvement; q better technology for managing soil, water, and crop to achieve high yields; and q high-yielding germplasm adapted to overcome the constraints of poor soils, drought, and submergence. These objectives are pursued in collaboration with national agricultural research systems (NARS) through the Rainfed Lowland Rice Research Consortium (RLRRC).

Managing crop, soil, and water resources for enhanced productivity and sustainability
This project develops systems for improved crop, soil, and water management for sustainable increases in productivity. It also identifies and analyzes the processes governing productivity and sustainability of crop production systems in the rainfed lowlands. Research covers nutrient and water management, crop establishment and weed management, and risk management. Rainfed rice and risk-coping strategies S. Pandey, R. Villano, F. Lagasca, H.N. Singh,11 D. Naik,12 and D. Behura12 A good understanding of farmer risk management strategies is needed to assess responses of farmers to new technologies and policies. Collection of farm-level data was initiated in 1994 in Mungeshpur and Itgaon villages of Faizabad, Uttar Pradesh, India. Thirty farmers from each village were selected randomly for in-depth monitoring of cropping pattern, input use, and income levels. The two villages differed mainly in risk of rice production due to differential access to irrigation. The analyses reported here used panel data for 1994-97. The major characteristics of the study villages are summarized in Table 1. The rainfall data from a monitoring station for 1994-97 are presented in Figure 1. Based on rainfall distribution, 1994 is classified as a normal year and 1995 and 1996 as drought years, with drought more severe in 1995.


IRRI program report for 1998

Table 1. Characteristics of two villages in Faizabad, Uttar Pradesh, India, where farm-level data were collected, 199497. Mungeshpur Characteristic Small farmers Irrigated area (%) Av operational holding (ha) Proportion of land type (%) Upland Medium land Lowland Av length of schooling of household head (yr) Av household size (members) 93 0.5 Medium farmers 74 1.5 Large farmers 75 2.2 Overall Small farmers 34 0.5 Medium farmers 35 1.4 Large farmers 39 2.9 Overall Itgaon

85 0.7

37 1.4

38 6 57 5

44 16 40 8

24 43 33 12

34 14 52 5

59 15 26 3

60 21 20 8

48 34 18 9

55 24 21 6

Rainfall (mm) 600 500 400 300 200 100 0 Jun Jul Aug Sep Oct
1996 1994 1997 1995 Long term average

1. Kharif season rainfall for Mungeshpur, Faizabad, Uttar Pradesh, India, 1994-97.

The adoption of modern varieties in Mungeshpur was over 90% of the area. The adoption rate in Itgaon was less than 70%. Short-duration varieties were more popular in Itgaon (55-66% of the area) than in Mungeshpur (25-31%). Transplanting was the dominant method of rice establishment in Mungeshpur, whereas direct seeding dominated in Itgaon. Overall, the variability of yield and net returns was higher in Itgaon, where average level of fertilizer use was lower. The differences in productivity and variability of rice as well

as crop management practices are reflections of the differential levels of risk in the two villages. The major risk adjustment mechanisms used by the rice growers in response to rainfall variability were a) changes in rice area, b) changes in methods of rice establishment, and c) changes in rice varieties. Rice area was reduced in 1995 and 1996 in both villages in response to delayed start of the rainy season. While some farmers planted less-moisture-sensitive crops, such as pulses and maize, others left the land fallow. Changes in rice establishment method from direct seeding to transplanting in years with favorable early-season rainfall, and vice versa, were practiced. Farmers were also found to increase the area growing traditional varieties in years with delayed rains and expand the area planted to modern varieties when rains were more favorable. A major risk adjustment mechanism, when the rainy-season crops were adversely affected, was to rely on the dry-season (DS) crops such as wheat, oilseeds, and vegetables. Nonrice crops and nonfarm activities were the major sources of income that helped cushion the effect of shortage in rice production in the study area. These sources of income were more important in Itgaon than in Mungeshpur (Table 2). The extent of adoption of modern varieties was found to be positively correlated with farm size, es-

Rainfed lowland rice ecosystem


Table 2. Average percentage share of different sources of household income for two study villages in Faizabad, Uttar Pradesh, India, 1994-97. Source Rice Wheat Other cropa Livestock Farm labor Nonfarm activitiesb Othersc Total income Total incomed ($ household-1)
a b

Mungeshpur 11 14 29 5 3 37 1 100 460

Itgaon 4 13 28 3 1 50 1 100 642

Includes pulses, mustard, maize, and other rabi and zaid crops. Includes services, laborer, business, blacksmith, barber, and pensioner. cIncludes sale of fruits and timber products. dExchange rate: US$1 = Rs 38.00.

pecially in Itgaon, suggesting the possibility that lower risk-bearing capacity of small farmers may be a factor constraining the adoption of modern varieties. The share of rice in the total income of small farmers was also lower than that of the larger farmers. Thus small farmers appeared to have developed income strategies less reliant on rice production. Given the low share (4-11%) of rice in total income (due to small farm sizes, the importance of post-rainy season crops, and nonfarm income in the study area), the economic benefit from a reduction in variability of income from rice was found to be relatively small (1-3% of total mean income). Nevertheless, improvements in the productivity of rice should be important in increasing farmers incomes. Improved water conservation and nutrient use efficiency via subsoil compaction and mineral fertilization G. Trebuil, D. Harnpitchitvitaya,13 G. Pantuwan,13 T.P. Tuong, and L.J. Wade Rainfed lowland rice, with an average yield of 1.5 t ha-1, is grown on some 4.5 million ha of coarse-textured, low-fertility soils in northeastern Thailand. Water stress ensues whenever rains are interrupted for about a week. Efficient management of rainwater and nutrients through soil management practices is a key to improving the productivity and stability of the rice crop on such highly permeable soils. Two-factor

field experiments in collaboration with the Ubon Rice Research Center were conducted during 1993 and 1994 wet seasons (WS). The objectives were to study the combined effects of soil compaction and fertilization, to quantify the magnitude of the interaction with different rates and types of fertilizer application, and to assess the relative impact of subsoil compaction and mineral fertilization on rice productivity. The main plots compared subsoil compaction with shallow dry tillage (C1), subsoil compaction with deep dry tillage (C2), and shallow dry tillage without compaction (C0). Five mineral fertilization techniques were used as subplot treatments: no fertilizer (F0), 40-13-25 kg NPK ha-1 in two applications of conventional fertilizer (F1), 40-13-25 kg NPK ha-1 using slow-release fertilizer (F2), 80-2650 kg NPK ha-1 in two applications of conventional fertilizer (F3), and 80-26-50 kg NPK ha-1 using slow-release fertilizer (F4). Soil was compacted with 10 passes of a vibrating road roller on 15 May 1993. Rice seedlings (RD6) were transplanted in all plots. Subsoil compaction significantly increased penetration resistance to at least 75 cm depth, with the 15-40 cm layer being the most compacted (Fig. 2). This degree of subsoil compaction obtained on a very coarse-textured soil without removing the topsoil was satisfactory. Subsoil compaction significantly increased depth and duration of floodwater in rice fields (Fig. 3 a,b). Compaction increased total weeks with
Soil depth (cm) 0 20 40 60 80 100 120 0
C0 = (Uncompacted + shallow tillage) C1 = (Compacted + shallow tillage C2 = (Compacted + deep tillage)

100 200 300 400 Soil penetration resistance (kPa)


2. Soil penetration resistance profiles with and without subsoil compaction (meanSE values of three repli-cations), Ubon Rice Research Center, Thailand, 1993 WS.


IRRI program report for 1998

Floodwater depth (cm) 25


20 15 10 5 0 25 20 15 10 5 0

C0 = (Uncompacted + shallow tillage) C1 = (Compacted + shallow tillage C2 = (Compacted + deep tillage)

Table 3. Effects of subsoil compaction and fertilization on grain yields (t ha-1) of rainfed lowland rice at Ubon Rice Research Center, Thailand, 1993 and 1994 WS.



Grain yield (t ha-1) Subsoil compaction treatmenta F0 Fertilizer treatmentsb F1 F2 F3 F4 Mean




C0 C1 C2 Av

1993 WS: CV = 19.8% 0.9 1.7 1.8 1.5 1.4 2.0 2.3 2.6 1.8 1.9 2.7 2.7 1.4 a 1.9 b 2.3 b 2.3 b 1994 WS: CV = 32.9% 0.8 0.4 1.2 0.9 1.9 2.0 1.6 1.6 2.0 2.3 1.7 2.2 1.5 a 1.6 a 1.5 a 1.6 a

2.4 1.7 a 3.6 2.4 a 3.5 2.5 a 3.2 c

C0 C1 C2 Mean

0.6 0.8 a 1.8 1.8 b 2.8 2.2 c 1.7 a

0 10 20 30 40 50 60 70 80 90 100 110 120 130

3. (a) Floodwater depth (meanSE values of three replications), 1993 WS; and (b) floodwater depth (positive values) and perched water table depth (negative values), 1994 WS, Ubon Rice Research Center. Harv = harvest; IR2, IR3 = 2nd and 3rd irrigations; PEb, PEe = beginning and end of panicle emergence.

a C0 = shallow dry tillage with no subsoil compaction, C1 = shallow dry tillage with subsoil compaction, C2 = deep dry tillage with subsoil compaction. bF0 = no fertilizer; F1 = 40-13-25 kg NPK ha-1 in two applications; F2 = 40-13-25 kg NPK ha-1 as slow-release fertilizer; F3 = 80-26-50 kg NPK ha-1 in two applications; F4 = 80-26-52 kg NPK ha-1 as slow-release fertilizer.

surface water accumulation from 3.7 and 2.4 wk for C0 to 11.0 and 14.3 wk for C1, and 11.7 and 14.9 wk for C2 for 1993 and 1994 WS, respectively. Because three irrigations were applied during the 1993 WS, drought stress was far more severe during the non-irrigated 1994 WS. But following the disappearance of surface water, no lasting and gradually receding perched water table was observed in the compacted plots in the later part of the 1994 cropping season. Subsoil compaction did not permit any increase in cropping intensity through cultivation of a short-duration upland crop in the post-rainy season. The benefits of subsoil compaction are thus limited to effects on the performance of the rice crop. Although highest grain production was harvested in compacted plots in 1993 WS, grain yield increase due to subsoil compaction was significant (P=0.05) in 1994 WS only (Table 3). Grain yield was 2.2 t ha-1 in C2 plots, 1.8 t ha-1 in C1 plots, and 0.8 t ha-1 in C0 plots. This confirmed the superiority of deeper tillage as observed in previous experiments at the same site. Effects of mineral fertiliza-

tion were significant in the 1993 WS only. The F4 slow-release fertilizer treatment gave a grain yield of 3.2 t ha-1 compared with 1.4 t ha-1 in the F0 plots. No significant yield differences between fertilizer treatments were measured in the 1994 WS partly because of the severity of drought stress. With the exception of the effect on rice grain weight in 1994 WS, no significant interaction was observed between compaction and fertilization treatments. The combination of subsoil compaction and controlled-release fertilizer constituted a way to increase soil productivity and stabilize rice yields, and improve food security and incomes at the farm level. Among the three factors studiedsubsoil compaction, tillage depth, and fertilization techniquesthe subsoil compaction had the highest impact on rice productivity. Although the compaction technique helped mitigate climatic risk, it did not improve soil water status after the rainy season, thus excluding the possibility of its use to increase cropping intensity at Ubon. Further investigations are needed. Persistence of the subsoil compacted layer should be assessed to determine how long the effect of subsoil compaction will last.

Rainfed lowland rice ecosystem


Analyzing weed community dynamics in rainfed lowland rice M. Mortimer, C. Piggin, and R. Lubigan Dry, direct seeding (DSR) in rainfed lowlands is an alternative to TPR and suited to areas with uncertain seasonal rainfall. DSR is sown at the onset of the rains into a seedbed that is often rapidly prepared. Crop growth is threatened by rapidly growing weed communities, which frequently have more than 10 species. Farmers typically repeat patterns of cultural and chemical weed management practices each season. Such repeated use of the same weed management over cropping cycles is known to result in mixtures of weed species that escape the combined effects of weeding practices. Moreover, this process of interspecific selection often results in particularly intransigent weeds remaining a threat to crop yield. A key part of the process of the design of integrated weed management systems involves understanding the factors that drive the shifts in weed species. A factorial combination of three land preparation (main plot) and five weed management (subplot) treatments was imposed in the 1996 WS on a farmers field at Tarlac, Philippines. The field had been regularly cropped with rainfed rice. A survey in the previous cropping season indicated relative homogeneity in the weed flora. Land preparation treatments in main plots contrasted conventional farmer practice (rototillage, furrowing, and immediate seeding) with the use of a stale-seedbed technique designed to remove weeds with glyphosate (applied at two rates) prior to crop establishment. Preemergence herbicides, butachlor and oxadiazon, were applied immediately after crop seeding. Butachlor is recommended for control of annual grass and sedges and oxadiazon for control of grass and broad-leaved weeds. No manual weeding followed the herbicide treatments. The imposition of the stale seedbed delayed crop planting by 22 d to allow weed emergence. In consequence, crops received different precipitation patterns. Fertilizer was applied both as a basal (6060-60 kg NPK ha-1) and topdressing (40 kg N ha-1). Details of the treatments are in Figure 4. Weed communities were destructively harvested, after canopy closure, from a 50- 50-cm quadrat in each plot, and the dry biomass measured for each species.

The average yield of IR72 in weed-free plots was 1.7 t ha-1. Redundancy analysis, the canonical form of principal component analysis, was used to examine the relationship between weed community composition and experimental treatments. This form of ordination graphically represented similarity among plots on the basis of species composition, and concurrently, with similarity among species on the basis of occurrence within plots. Ordination axes were restricted to linear combinations of experimental treatments. Figure 4 presents the biplot describing the relationships between selected individual species and treatments. A qualitative assessment of the strength of correlation may be gained by noting the arrows: those pointing in similar directions indicate high positive correlation, and those opposite to one another, a negative one. The cosine of the angle between the arrows of a species and an experimental treatment approximates the correlation coefficient between that species and the experimental treatment. The relative length of an arrow in the biplot is a reflection of the importance of that species or treatment in describing interrelationships within the ordination. The relative proximity of circles in Figure 4 indicates the degree of similarity in the weed community composition of individual plots. The distance of the circles from the origin reflects increasing diversity in composition. Plots that received herbicide and hand weeding (lower left quadrant) exhibited no association with any weed species, and contrasted strongly with unweeded plots (upper right quadrant), which were dominated by Ludwigia perennis, Fimbristylis miliacea, and Echinochloa colona. The absence of a noticeable association of E. colona with any practice reaffirms the hypothesis that this is a general weed that will persist under all management conditions except manual weeding. The analysis also indicates that two pernicious weeds, Cynodon dactylon and Cyperus rotundus, were strongly associated with farmers practice of land preparation and the sole use of butachlor and oxadiazon, and negatively correlated with the use of glyphosate in preplanting land preparation. Both species survive extensively by vegetative means from season to season and rototillage alone promotes population increase from fragmented stolons


IRRI program report for 1998

Land preparation treatments G1. Land rototilled and left for 22 d for weed emergence, then sprayed with glyphosate at 0.96 kg ai ha-1. Rice 1.0 seed subsequently sown in rows to a depth of 1-2 cm -1 at 70 kg ha Unweeded WS G2. As in G1, but with glyphosate sprayed at 1.2 kg ai ha-1. G3. Land was rototilled, furrowed, and seeded on the same day. Rice was dry-seeded in rows at a rate of Ludwigia 110 kg ha-1. Weed control treatments perennis W1. Butachlor (1 DAS) at 1.5 kg ai ha-1 W2. Butachlor (1 DAS) at 1.5 kg ai ha-1 + 1 Fimbristylis miliacea manual weeding (28 DAS) 15 W3. Oxadiazon (1 DAS) at 0.625 kg ai ha-1 5 -1 W4. Oxadiazon (1 DAS) at 0.625 kg ai ha + 1 manual Ischaemum weeeding (28 DAE). rugosum 7 W5. Not weeded.

17 Preplanting seedbed preparation (G2)

Echinochloa colona

11 Eclipta prostrata Preplanting seedbed preparation (G2) 14 24 13 19 1.0 10


anua lor + m Butach (W2) eeding w

12 20 Butachlor W1 Conventional farmer seedbed preparation (G3) 9

ual n + man Oxadiazo (W4) eeding w

Cyperus rotundus Cynodon dactylon Oxadiazon W3 -1.0

4. Biplot redundancy analysis of the relationships among weed species and weed management treatments in rainfed DSR. The ordination shows species locations with solid lines and experimental treatments with dotted ones. Circles represent the weed communities of different treatments. Numbers indicate unnamed weed species of lesser importance. (See text for explanation.)

and tuber chains that escaped earlier butachlor and oxadiazon application. Conversely, rototillage combined with glyphosate results in high mortality through within-plant translocation of the herbicide. Glyphosate use in land preparation was partially correlated with the abundance of Ischaemum rugosum and Eclipta prostrata. Both of those species can exhibit delayed germination in rainfed systems, escape being killed by glyphosate during seedbed preparation, and emerge into a community dominated by rice but with fewer other weed species. They may, therefore, increase with the use of glyphosate in land preparation and require additional postemergence weed control.

Our analyses gave first insights into possible shifts that may be expected to occur under differing weed management treatments. While not providing underlying mechanistic explanations and being seasonally dependent, the analyses are valuable in that they can be conducted in farmers fields by experimentally overlaying treatments on the indigenous weed flora.

Germplasm improvement for rainfed lowland rice

Germplasm improvement for the rainfed lowlands is complicated due to the heterogeneity and com-

Rainfed lowland rice ecosystem


plexity of the environment. A dispersed breeding program based on physiological understanding of plant adaptation and environmental characterization has been established at key representative sites. It operates through a shuttle breeding partnership in collaboration with NARS. A methodology for field evaluation of GxE interactions was developed, and preliminary analysis identified principal factors influencing GE interaction. An important constraint to adoption of improved germplasm that tolerates drought and submergence is the preservation of traits found in traditional cultivars that are highly valued by farmers and consumers. This problem is approached through farmer participatory breeding, and selection of advanced lines to draw on indigenous knowledge of farmers. Germplasm for rainfed lowland ecosystems S. Sarkarung Shuttle breeding programs were strengthened at the rainfed lowland key sites. Early-generation materials, mainly F2, were evaluated in 1997 and 1998 WS in a wide range of submergence-prone environments in Assam, Bihar, Madhya Pradesh (MP), Uttar Pradesh (UP), Orissa, and West Bengal. Among these sites, only India Gandhi Agricultural University (IGAU) at Raipur, MP, represents drought-prone conditions. The eastern states have more than 14 million ha of rainfed lowland rice areas. Most of the breeding materials for drought-prone environments in Southeast Asia were initially evaluated at the Ubon, Thailand breeding site. That site represents intermittent drought during the vegetative stage. The selected materials, particularly the early generations starting from F3, were systematically sent to Lao PDR, the Philippines, and Myanmar with emphasis on favorable rainfed lowland in the latter two countries. Breeding materials adapted to drought S. Sarkarung and S. Pushpavesa13 Breeding materials (F2 onward) including nonglutinous and glutinous lines were dry seeded in 1998 WS at the Ubon Rice Research Center (URRC). Table 4 shows the number of breeding lines and crosses for each generation.

Table 4. Breeding materials evaluated in drought-prone conditions at Ubon Rice Research Center, Thailand, 1998 WS. Lines evaluateda (no.) Generation NG G Total crosses F2 F4 F5 F6 F7 F8

Plants selected (no.)b NG G Total crosses 49 53 52 39 26 8



49 75 74 72 38 10

4129(22) 1173 (24) 570(35) 1517(125) 146(47) 67(34) 234(59) 294(36) 700(477) 331(129) 795(177) 3

455 1051 398 246 272 149 949 405 337 1

NG = nonglutinous type, G = glutinous type. bNumber in parentheses is number of lines harvested in bulk.

In the F2, F4, and F5 generations, in which segregation within the lines still prevails, single individual plants with good agronomic traits were selected. A modified bulk method was employed in the later generations (F6 to F8). In addition to the breeding activities at URRC, 22 advanced breeding lines and 2 check varieties, KDML 105 and RD23, were evaluated on farms at Lamplaimat, Buriram, in collaboration with the Population and Community Development Association. Lamplaimat represents a favorable rainfed lowland area of northeast Thailand in the rice-growing areas that produce the best grain quality. The experimental plots were dry-seeded on 24 Jun 1998 with 25 cm between rows and a 60 kg ha-1 seeding rate. Table 5 shows grain yields, days to 50% flowering, and plant height of the breeding lines and check varieties. Many lines yielded significantly higher than KDML 105. However, IR68835-99-6-7-B, which combines good grain quality with blast resistance produced a yield about 15% higher than KDML 105. On the other hand, IR69550-2-KKN2-UBN-4-3, even though not producing the highest grain yields, has grain quality similar to the popular KDML 105. It also has good resistance to blast.


IRRI program report for 1998

Table 5. Grain yields, days to 50% flowering, plant height, and blast reaction of advanced breeding lines evaluated in a farmers field at Lamplaimat, Buriram. Thailand, 1998 WS. Line Grain yield (t ha-1)a 4.7 4.6 4.6 4.5 4.4 4.4 4.3 4.3 4.3 4.3 4.2 4.2 4.0 4.0 4.0 3.7 3.6 3.6 3.3 3.2 3.1 2.9 3.7 4.0 Rank Days to 50% flowering 108 114 101 108 104 109 104 103 108 98 112 109 105 110 106 102 104 101 98 101 101 99 94 119 Height (cm) 138 125 116 128 125 110 116 118 128 107 160 127 131 217 114 106 122 113 124 108 104 105 105 140 Blast scoreb

IR68835-99-6-7-B IR71505-43-2-B IR69515-27-KKN-3-UBN-6-1 IR69515-26-KKN-3-UBN-5-1 IR69513-14-SRN-1-UBN-2-B IR69550-3-KKN-1-UBN-12-1 IR69550-2-KKN-2-UBN-4-3 IR69515-2-KKN-3-UBN-1-3 IR69515-10-KKN-3-UBN-B406-2 IR69550-2-KKN-1-UBN-5-2 IR70220-26-1-B IR69515-26-KKN-3-UBN-4-3 IR69515-26-KKN-3-UBN-10-1 IR68835-105-1-1-B IR68801-46-1-B-2-B IR69550-2-KKN-1-UBN-4-4 IR68815-51-PMI-3-UBN-1-B IR71506-27-2-B IR68801-6-2-B1-2-B IR68796-27-1-B-1-B IR68796-27-1-B-2-B IR68796-27-1-B-5-B Check varieties RD23 KDML 105 CV (%) 13.3 LSD (5%) 120.24

1 2 3 4 5 6 7 8 9 10 11 12 14 15 16 17 19 20 21 22 23 24 18 13

4 3 1 2 4 2 2 4 3 2 3 3 2 3 3 4 1 2 3 3 2 3 9 9

Av of four replications. bStandard evaluation scale for rice (SES) 1 to 9: 1 = highly resistant, 9 = highly susceptible.

Screening for drought resistance at vegetative stage S. Sarkarung and G. Pantuwan13 More than 200 advanced breeding lines (F5 onward) were screened for drought tolerance and recovery ability at URRC during 1997-98. Thirty-day-old seedlings were subjected to different levels of stress after irrigation stopped. Drought resistance was scored at 1-wk intervals for 4 wk and recovery scored 4 d after rewatering. Table 6 presents the drought resistance score at different stages of stress development, and their ability to recover after rewatering. The check varieties were KDML 105 (high recovery), NSG19 (drought-resistant), and IR20 (drought-susceptible). The line IR68796-362-B-1-1-B gave the lowest drought score and was the fastest to resume growth after stress. IR6879627-3-B-5-2-B and its sister line had low drought score and good recovery ability.

Mechanisms of submergence tolerance and recovery O. Ito, E. Ella, and N. Kawano The restoration of normal oxygen conditions during recovery of submerged organisms is damaging to plant and animal tissues. An active oxygen-scavenging system plays an important role in minimizing induction of oxidative stress that may cause lipid peroxidation, protein denaturation, and DNA mutation. Such a system may include 1) antioxidants known as free radical scavengers such as reduced glutathione and ascorbate, and 2) active oxygen-scavenging enzymes such as superoxide dismutase (SOD) and glutathione reductase (GR). Two cultivars, FR13A (submergence-tolerant) and IR42 (submergence-intolerant), were used to confirm the role of the active oxygen-scavenging system in protecting the plant from oxidative stress during recovery from submergence.

Rainfed lowland rice ecosystem


Table 6. Drought resistance and recovery scores of selected breeding lines evaluated at Ubon, Thailand, 1997 DS. Drought resistance scorea 20 Feb IR70828-18-1-B-B IR70849-4-3-B-B IR70849-29-B-B-B IR68821-70-1-B-5-8-B IR68823-9-5-B-1-2-B IR69502-28-SKN-1UBN-1-B-1 IR69513-1-SKN-2UBN-1-3-2 IR69514-51-KKN-3UBN-1-4-3 IR69514-51-KKN-3UBN-1-5-1 IR69515-9-KKN-4UBN-2-1-1 IR69515-27-KKN-1UBN-1-1-1 IR68815-51-PMI-2UBN-7-1-2 IR68815-69-PMI-6UBN-B-2-1 IR68796-27-3-B-5-1-B IR68796-27-B-B-6-1-B IR68796-36-2-B-1-1-1 Checks KDML 105 IR20 NSG19 IR54081-CPA-3-B-1-3/ Sabita//SLK3-1-2-2 CT6241-17-1-5-1/ KDML 86G2 4-//IR46329-SKN CT6241-17-1-5-1/ KDML 86G2 4-//IR46329-SKN CT9899-32-1/ IR55829-B-B//IR450-SKN-516 CT9981-23-5-M-2/ IR55776-16-2//IR54977-UBN IR57514-PMI-5-B-1-2/ IR57515-PMI-8-1-SKN// IR43524-55-1-3-2 IR57514-SKN-299-3-2/ IR57515-PMI-8-1-SKN// IR43524-55-1-3-2 IR57514-SKN-299-3-2/ IR57515-PMI-8-1-SRN// IR54119-4-B-1-1 IR57514-SKN-299-3-2/ IR57515-PMI-8-1-SRN// IR54119-4-B-1-1 IR57514-SKN-299-3-2/ IR57515-PMI-8-1-SRN// IR54119-4-B-1-1 IR57514-SKN-299-3-2/ IR57515-PMI-8-1-SRN// IR54119-4-B-1-1 CT9993-5-10-1-M/ IR41431/68-1-2-3// IR57514-PMI-5-B-1-2 CT9993-5-10-1-M/ IR41431/68-1-2-3// IR57514-PMI-5-B-1-2 CT9992-22-2-4/ IR56592-21-1-3-1//KDML 105 CT9992-22-2-4/ IR56592-21-1-3-1//KDML 105 CT9992-22-2-4/ IR56592-21-1-3-1//KDML 105 2.33 3.33 3.00 1.67 1.33 2.67 2.33 2.00 1.67 27 Feb 3.00 2.67 3.33 3.33 2.67 3.00 6 Mar 4.00 5.00 4.67 4.67 4.33 4.67 13 Mar 5.33 5.33 5.67 5.67 5.33 5.67 Drought recovery scoreb 20 Mar 23 Mar 5.33 5.33 5.67 5.67 5.33 5.33 1.67 1.67 3.67 1.67 2.33 2.33

Breeding line












































2.33 1.00 1.00

3.33 2.33 2.00

5.67 3.67 4.67

6.67 5.67 5.00

5.67 5.67 5.00

3.67 1.67 1.00

3.33 4.00 4.00

5.67 6.00 6.47

6.37 6.67 7.27

7.00 7.00 7.50

4.33 5.00 4.73

a Av of four replications. Standard evaluation scale for rice (SES) 1 = highly resistant/very good recovery, 9 = highly susceptible/very poor recovery. bRecovery score was taken 4 d after rewatering.

The activities of some active oxygen-scavenging enzymes and the levels of antioxidants (reduced glutathione and reduced ascorbate) in the leaves of seedlings submerged for 6 d were measured. FR13A had significantly higher SOD and GR activities in the submerged seedlings than in the nonsubmerged control. IR42 had more or less the same levels of activity under the two conditions

(Fig. 5, 6). Increases in leaf SOD and GR activities in FR13A were observed during the first 3 d of recovery. However, the levels of reduced glutathione and reduced ascorbate decreased in submerged seedlings in both cultivars. The level of antioxidants during the recovery of submerged seedlings increased in FR13A at day 5 of recovery for reduced ascorbate and at day 7 for reduced glutathione. In


IRRI program report for 1998

Leaf SOD activity (units mg protein-1) 50 40 30

FR13A, submerged (6 d) FR13A, not submerged lR42, submerged (6 d) lR42, not submerged

Lipid peroxidation in leaf (nmole MDA g leaf fresh wt-1) 6

FR13A, submerged (6 d) FR13A, not submerged lR42, submerged (6 d) lR42, not submerged

5 4

20 10 0

2 3 4 Days of recovery

5. Activities of superoxide dismutase (SOD) in leaves during recovery from submergence for FR13A (submergence-tolerant) and IR42 (submergence-sus-ceptible). Vertical line indicates LSD (P = 0.05). IRRI, 1998.

2 3 4 5 Days of recovery

7. Formation of malonylaldehyde (MDA) as an indicator of lipid peroxidation during recovery from submergence for FR13A and IR42. Vertical line indicates LSD (P = 0.05). IRRI, 1998.

Leaf GR activity (nmole glutathione mg protein-1 min-1) 50 40 30 20 10 0 0 1 2 3 4 5 6 7 Days of recovery 6. Activities of glutathione reductase (GR) in leaves during recovery from submergence for FR13A and IR42. Vertical line indicates LSD (P = 0.05). IRRI, 1998.
FR13A, submerged (6 d) FR13A, not submerged lR42, submerged (6 d) lR42, not submerged

have protected itself from oxidative stress better than IR42. Phosphorus efficiency G. Kirk, E. Santos, M. Santos, and G. Findenegg The mechanisms by which rice roots extract soil P was studied to develop a mechanism-based technique for germplasm screening. Earlier research (Program Report for 1992) found that rice plants growing in an aerobic, highly weathered, P-deficient soil, with and without added P, were able to solubilize P and thereby increase their P uptake. The solubilization could not be explained by root-induced pH changes nor by the release of organic Pmobilizing phosphatase enzymes. It was speculated that it might be explained by release of organic acids from the roots. Other researchers had reported release of citric acid from rice roots into nutrient solution cultures, and we found that the addition of citrate to experimental soil released large amounts of phosphate into solution. To confirm that hypothesis, we needed to show that the amounts of citrate released were sufficient to explain the observed solubilization. Organic acids released from roots in soil are quickly decomposed by microbes. Also, any P solubilized diffuses into the soil bulk, as well as toward the root, and not all the P solubilized is taken up.

IR42, the levels remained more or less constant. These findings indicate that high SOD and GR activities and high levels of antioxidants during recovery may explain why FR13A is better at tolerating submergence (and thus protect itself from oxidative stress) and has better survival than IR42. The extent of lipid peroxidation during the first week of recovery of submerged seedlings was less in FR13A than in IR42 (Fig. 7). IR42 may have suffered more from oxidative stress than FR13A and this is consistent with the earlier observation that FR13A, because of high SOD and GR activities and high levels of antioxidant during recovery, may

Rainfed lowland rice ecosystem


Table 7. Rates of organic anion excretion and concentrations in roots for different rice cultivars grown in nutrient solution. Rates and concentrations were measured 17 d after planting. Data are means of two replicates. IRRI, 1998. Rate of excretion (pmol g-1 root FW s-1) Cultivar +P Salumpikit Nohrin Mochi 76 Aus 257 WAB56-50 IR55411-50 IAC47 Toyohata Mochi Azucena Oryzica Sabana 6 IR47686-09-2-4 11.9 36.4 17.8 56.7 52.2 24.7 55.6 33.6 40.0 0.0 Citrate -P 57.8 7.5 63.1 29.4 51.7 6.1 124.4 66.4 27.2 21.4 +P 0 0 0 0 0 0 0 0 0 0 Malate -P 0 0 0 0 0 0 0 0 0 0 +P 0.60 0.34 0.34 0.36 0.43 0.33 0.36 0.27 0.32 0.15 Concentration in root (pmol g-1 root FW) Citrate -P 0.24 0.31 0.46 0.30 0.33 0.34 0.35 0.24 0.40 0.17 +P 1.17 1.66 1.77 1.23 1.08 1.96 2.71 1.42 1.41 0.63 Malate -P 0.94 0.70 1.83 1.14 0.73 0.85 0.99 0.68 1.06 0.32



Rates of release of organic anions from rice roots grown in nutrient solution cultures with and without P were measured by transferring the roots of 1to 3-wk-old intact plants to fresh nutrient solutions for 30 min and assaying the organic acids released using high-performance liquid chromatography. In all the rice cultivars tested, the main organic anion identified in the exudates was citrate (Table 7). Two, as yet unidentified, organic anions were also present and their excretion consistently increased about sixfold under P deficiency. Rates of citrate excretion also tended to increase under P deficiency, though to a lesser extent. Peak rates were of the order of 100 pmol g-1 (root fresh weight) s-1. Rates of excretion differed among cultivars and there were significant interactions between cultivars and P status for the two unidentified anions. Malate was present in the roots at a greater concentration than citrate, but rates of malate excretion were negligible (Table 7). The concentration of malate in the roots under P deficiency tended to decrease but those of the two unidentified anions increased while that of citrate was little affected. This indicates that the increased excretion of citrate and unidentified anions was associated with increases in their rates of synthesis under P deficiency.

Rates of organic anion release into soil are likely to be different from those into nutrient solution cultures. Measuring release into soil is complicated by a) decomposition by soil microbes; b) adsorption onto soil particles, making it necessary to extract the soil with a reagent to recover the organic anions; and c) interference in the assay of organic anions by other soil constituents extracted. Also, as a result of adsorption and decomposition, organic anions will diffuse only a short distance from root surfaces. To enable recovery of the organic anions released, we grew plants in 3-mm layers of soil. The roots were arranged in a 2-mm thick, 13- 8-cm layer, sandwiched between two 3-mm thick layers of soil connected to a nutrient solution reservoir via glass-fiber filter paper wicks. Roots were separated from soil by 24-mm-pore-diameter nylon mesh sheeting. Two soil P levels (100 and 1000 mg P g-1) and either wholly NH4+ or NO3- nutrient solution were used. Organic anions were recovered from the soil by shaking in 0.01 N H2SO4. The main organic anion recovered from the rhizosphere soil in the thin-layer experiments was citrate (Fig. 8). Smaller quantities of oxalate, malate, lactate, and fumarate were also detected, and moderate quantities of two unidentified carboxylates and unidentified amino acids. Only trace amounts of the organic anions, detected but not identified in the nutrient solution, were found.


IRRI program report for 1998

Citrate in thin layers (mol system-1) a P100-NH4 500 400 300 200 100 0 500 400 300 215 200 100 0 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 Weeks after transplanting
8. Amounts of citrate recovered from soil in experiments in which rice plants were grown in thin layers of soil at low and high rates of P addition. Data are means SE of three replications. Dotted lines indicate means over time. IRRI, 1998.

c P1000-NH4

250 122

b P100-NO3

d P1000-NO3


In the P1000 treatments, the citrate concentration reached steady state after 1-2 wk in which the rate of excretion from the roots was equal to the rate of decomposition in the soil. We calculated the apparent flux, FC, of citrate across the roots from the measured steady-state concentration in the soil and independently measured soil diffusion characteristics and rate constant for citrate decomposition. These gave FC = 33 nmol m-2 s-1. The corresponding rates of excretion per unit root weight ranged from 94 to 43 pmol g-1 (root fresh weight) s-1 over the course of plant growth. These compared with 34 and 66 pmol g-1 (root fresh weight) s-1 under P-sufficient and P-starved conditions for the appropriate cultivar (Azucena) in the nutrient solution experiments (Table 7). We assumed that citrate recovered from the thin layers was all excreted by roots. But it is possible that part of it was generated by microbes subsisting on root-derived carbon in the soil. In this regard, the nature of the unidentified organic anions in the nutrient

solution experiments, whose synthesis and excretion were stimulated under P deficiency, is interesting. They may act as substrates for citrate synthesis or otherwise influence its longevity in the rhizosphere.

We used the rates of citrate release and decomposition to calculate how much P might be solubilized. To do that, we defined a P-citrate interaction coefficient, l, such that the quantity of P that must be removed from the soil for a given uniform addition of citrate to leave the P concentration in solution unchanged, equals bP/bC where bP is the soil P buffer power (d[P]/d[PL], where [P] equals the concentration of labile P in the whole soil and [PL] equals the concentration in the soil solution. We measured bP, bC, and by shaking the soil with different concentrations of citrate at a wide range of soil-solution ratiosand a corresponding range of sinks for P and measuring P and citrate left in solution.

Rainfed lowland rice ecosystem


We used the values attained together with the rates of citrate excretion and decomposition measured above to calculate the extent to which P solubilization by rice could be explained by citrate excretion. Figure 9 compares P concentration profiles so calculated with P profiles measured previously (Program Report for 1992). The calculations show that the observed solubilization and increase in P uptake could indeed be accounted for by citrate release. Running the calculations with FC = 0 showed that the quantity of P taken up in the absence of solubilization was negligible. Given that the model parameters were measured independent of the output, and that the output is sensitive to the parameter values, the good agreement between the observed and predicted concentration profiles shows that the model provides a satisfactory description of the system. Further research will seek to explain germplasm differences in P efficiency in terms of the above mechanisms, so as to develop a germplasm screening technique. Mapping genes for root traits A. Kamoshita, L. Wade, S. Sarkarung, and H.T. Nguyen14 Choice of relevant environments for phenotypic screening and appropriate mapping populations is essential for identification of quantitative trait loci (QTLs) responsible for traits conferring drought tolerance, and development of marker-assisted selection technologies. Little attention has been focused on the screening system for phenotyping of constitutive root traits (those conferring a better root system before the onset of drought) that is relevant to rainfed lowland rice. Variation and QTL for deep and thick root systems were compared for two rice mapping populations grown in pots in the greenhouse in anaerobic soil. Comparison was made about 45 d after sowing for 220 doubled haploid lines from a cross between breeding lines from upland japonica and lowland indica (CT9993/IR62266), and 184 recombinant inbred lines from a cross within lowland indica (IR58821/IR52561). Thirteen traits, categorized into three groups (shoot, deep root, or thick root traits) were analyzed in two plantings of contrasting temperature and radiation conditions (dry and rainy seasons).

[P]/mmol kg-1 14 12 10 8 6 [C] 4 7d

[C]/mmol kg-1, [PL*]/M



12 [PL*]

4 2 0 0

14 12 10 8 6 4

21 d [P] 16



8 [C] 4

2 0 14 12 10 8 6 [C] 4 2 0 0 1 2 3 4 5 6 Distance, x mm-1 0 4 [PL


0 35 d [P] 16


9. Observed (points) and predicted (solid lines) concentration profiles of P ([P]) in soil near a planar layer of rice roots after different periods of root-soil contact. Predictions are made with a model of P solubilization by root-released citrate. Dashed lines are predicted citrate concentration ([C]) profiles; dotted lines are predicted profiles of P in the soil solution ([PL]). IRRI, 1998.


IRRI program report for 1998

Count IR62266 50 a 40 30 CT9993

Proportion per bar

Count IR62266 40 0.2 30 b

Proportion per bar CT9993 0.18 0.16 0.14 0.12

20 20 0.1 10 10 0.0 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 IR52561 60 c 50 40 0.2 30 20 20 10 0 0.0 0.25 0.1 10 30 0.3 40 IR58821 50 d 0 0.9 1.0 1.1 1.2 IR52561

0.10 0.08 0.06 0.04 0.02 1.3 1.4 1.5 1.6 0.0









0 1.1






0.0 1.7

Deep root dry weight (g)

Root thickness at 0-10 cm layer (mm)

10. Distribution of deep root dry weight (a,c) and root thickness at 0-10 cm soil depth (b,d) among 220 doubled haploid lines from CT9993/IR62266 (a,b) and among 184 recombinant inbred lines from IR58821/IR52561 (c,d) in the dry season experiment. IRRI, 1998.

Large transgressive variation was observed for deep root traits in CT9993/IR62266 and for thick root traits in IR58821/IR52561 (Fig. 10). Genotype planting interaction was significant but always smaller than genotypic variation with one exception. For deep and thick root traits in IR58821/ IR52561, where genotype planting interaction and genotypic variation were comparable in size, only three out of 22 QTLs identified were common across the two plantings. In CT9993/IR62266, six out of 26 QTLs identified were common for deep or thick root traits, but three of them were also associated with the shoot dry weight and plant height. We conclude that constitutive root traits may be improved by crossing within lowland indicas, or by

introgression with upland japonicas. Even subtle variation in the phenotyping environment influenced the QTLs identified, so QTL environment interaction was significant. There was evidence that difference in shoot growth affected expression of root traits and QTLs identified. Water extraction and recovery ability following drought A. Kamoshita, L.J. Wade, T. Azhiri-Sigari,2 and A. Yamauchi15 Quantitative data are lacking on the dynamics of plant water use and root and shoot growth in response to drought and rewatering in rainfed lowland rice.

Rainfed lowland rice ecosystem


Table 8. Green leaf biomass at the end of drought period, incremental total root length during the later drought period, transpiration (kg d-1), incremental leaf area (cm-2), and crop growth rate (g pot-1 d-1) during 9 d of rewatering period among eight rice genotypes. IRRI, 1998. Drought Genotype Leaf biomass (g) IR20 NSG19 CT9993 IR62266 Mahsuri KDML 105 IR58821 IR52561 LSD (0.05) 4.67 d 6.23 ab 5.08 cd 5.90 abc 6.34 ab 6.60 a 5.69 bc 5.68 bc 0.89 ** Root length (m) 38.1 c 59.9 bc 43.6 c 75.2 abc 108.6 a 86.8 ab 98.9 a 43.4 c 38.8 ** Transpiration (kg d-1) 0.83 c 0.97 b 0.81 c 0.80 c 0.94 bc 1.20 a 0.87 bc 0.86 bc 0.14 ** Leaf area (cm-2) 1783 d 2153 bcd 844 e 2442 abcd 2935 a 2739 ab 2043 bcd 2514 abc 683 ** Crop growth rate (g pot-1 d-1) 1.36 b 1.93 a 1.55 b 1.41 b 2.03 a 2.15 a 1.89 a 1.56 b 0.29 ** Rewatering

Shoot and root growth, transpiration, and water extraction of eight diverse rice lines were quantified in three pot experiments: one under severe water deficit after panicle initiation (average transpiration of 15.0 mm d-1; Experiment 1), and two under slow and progressive water deficit during tillering (average transpiration of 2.1 and 7.6 mm d-1 in Experiments 2 and 3). Higher transpiration was generally associated with higher crop growth rate, both early in drought development, when soil became aerobic but soil water was still readily available, and after rewatering. Early in drought development in Experiments 2 and 3, there was genotypic variation for relative amounts of tiller and leaf area production compared with the well-watered treatment. Genotypes with high vigor before stress imposition and during this transitional phase, such as NSG19, KDML 105, Mahsuri, and IR58821, produced longer root length during the following severe drought period and had larger green leaf biomass at the end of the drought period in Experiment 3. These genotypes sharply increased transpiration and rapidly expanded leaf area after rewatering, which caused superior drought recovery (Table 8). In Experiment 1, plant size before stress imposition was large, and genotypic variation in response to drought and rewatering was small, except in KDML 105, which recovered more rapidly after rewatering. Both constitutive root traits (those present before stress) and adaptive root traits (those developed in response to stress) varied among genotypes. In wellwatered conditions, CT9993 and IR58821 had high root-shoot ratios, deep and thick root systems, and

high root weights per tiller. In response to milder drought in Experiments 2 and 3, the total amount of assimilate distributed to roots was reduced and roots became thinner, but the proportion of total assimilate supply assigned to deeper layers increased, thereby maintaining deep root mass and increasing branching. NSG19, KDML 105, Mahsuri, and IR58821 partitioned a larger proportion of assimilate to deep roots, had more deep root branching, and took a shorter time to extract 4 kg of soil water during drought in Experiment 3 (e.g., 17 d for NSG19 compared with 22 d for IR20). The average root length density below the 30 soil layer was positively correlated with the rate of water extraction at similar depths during the latter half of the drought period, respectively explaining 66 and 58% of the variation around 30 and 40 cm depth (Fig. 11). In Experiment 3, osmotic adjustment tended to be higher in the genotypes that had a slower rate of transpiration and a lower pre-dawn leaf water potential at the end of drought period. Among the genotypes that extracted water rapidly, KDML 105 had a higher osmotic adjustment. IR58821 and CT9993 had the lowest levels of osmotic adjustment. We conclude that the amount of water transpired was more important than water use efficiency in determining the capacity of a line to perform well in the imposed conditions of water deficit and rewatering. Vigor before and during transitional phase from anaerobic to aerobic soil conditions, and a capacity to elongate and branch roots deeply in the profile as the drought progressed, were advanta-


IRRI program report for 1998

Daily extraction (g d-1) 100 a 10-cm depth 80

IR20 NSG19 CT9993

b 20-cm depth


IR62266 Mahsuri KDML 105


IR58821 IR52561


0 100 c 30-cm depth 80 d 40-cm depth




0 0.0






0.4 (mm-3)




Average root length density

11. Relationship between root length density and daily extraction of soil water around 10 (a), 20 (b), 30 (c), and 40 cm depth (d) among eight genotypes during later half of drought period. IRRI, 1998.

geous for quicker water extraction during mediumterm drought and for superior recovery ability after rewatering. Root development, water extraction, and other physiological responses need to be examined under more prolonged and severe drought, where seedling vigor may be less important.

Progress of unreported projects

Characterizing and analyzing rainfed rice environments T.P. Tuong Characterization of the heterogeneous rainfed rice environment for land evaluation in northeastern Thailand

GIS-based and statistical methodologies developed for analyzing spatio-temporal variability of climatic and soil parameters in the rainfed lowland rice environment in northeastern Thailand. Carried out agroclimatic analysis of northeastern Thailand based on long-term monthly rainfall and evapotranspiration data of 16 meteorological stations to determine spatial and temporal variation of rainfall and water balance and duration of humid periods. Generated rainfall probability surfaces using long-term weekly data from an additional 100 stations, to be used for producing weekly moisture availability index surfaces for drought analysis for the northeast region. Gridded interpolated sur-

Rainfed lowland rice ecosystem


faces of various soil fertility parameters were generated using geostatistical techniques and were combined using a mathematical procedure to produce a joint classification of soil fertility for a pilot site in the Ubon land reform area. This is to be compared with a gridded surface of farmers assessments of soil fertility. Set up pilot study for testing new land use evaluation procedure for a land reform area, Ubon, northeastern Thailand.

GIS data layers of infrastructure, land cover, and economic development generated for Giridih and Purulia districts. Household farm survey carried out for eight villages in Giridih District of Bihar State, data encoded and ready for economic analysis. Household listing of another eight villages in Purulia District of West Bengal conducted in preparation for household survey this winter.

Delineation and mapping of intensified rainfed rice areas in the Mekong River Delta q Methodologies developed for using synthetic aperture radar (SAR) to delineate and map rice-based cropping systems in the Mekong Delta. q Completed mapping of rice cropping systems using multidate ERS-2 SAR data for the 199697 crop year for a 100 100-km area, and carried out GIS analysis to determine expansion of intensified double-cropped rainfed rice in relation to the retreat of saline water intrusion. q Multidate ERS-2 SAR and RADARSAT data for the 1997-98 crop year acquired and preprocessed for a 100 200-km strip of two adjacent areas; began preliminary classification and interpretation of the ERS-2 data. q Field survey data for the 1997-98 crop seasons from seven provinces compiled to provide ground truth for validating the map output from radar image processing. Analyzing sustainability criteria of rainfed ricebased cropping systems in the Bihar Plateau q Gridded weekly moisture availability index surfaces generated for Bihar and West Bengal for agroclimatic analysis. q A computer program for determining the start and end of the humid period, as well as the occurrence of mid-season drought, was developed for analyzing the weekly moisture availability index surfaces. q Block-level and village-level mapping of rice production, land use area, and demographic characteristics completed for Purulia District.

Characterizing principal rainfed subecosystems in eastern India q Validated the GIS and remote sensing-based methodology and delineated and characterized the rainfed environments at the meso level in Madhya Pradesh. q Agroecological atlas of eastern India is being compiled, featuring climate, land use, landforms and hydrology, cropping systems, and selected socioeconomic features. Addressing gender concerns in rice research and technology development T. Paris

Began impact evaluation of a rice flour mill and drum seeder in the Philippines. Assessed farmers gender-specific criteria for using traditional and improved rice varieties in three rainfed villages in eastern Uttar Pradesh, India, under the Farmer Participatory Breeding Project. Developed on-farm selection of new rice lines for rainfed lowland conditions in Faizabad District by male and female farmers. Trained NARS partners on data analysis of socioeconomic and gender components of participatory research in plant breeding and varietal selection.

Rainfed Lowland Rice Research Consortium C. Piggin


Workplan meetings for 1998 were held at each site from March to May and attended by 4-5 IRRI scientists and 20-25 NARS scientists. An International Workshop on Nutrient Research in Rainfed Lowlands, which re-


IRRI program report for 1998

viewed the current knowledge and needs for research on nutrients, was held at Ubon, Thailand, 12-15 Oct, and attended by 40 NARS, 8 ARI, and 25 IRRI participants. A 304-page proceedings was published by years end. The annual steering committee meeting was held on 15 Oct, after the Nutrient Workshop, to plan 1999 Consortium research and budgets. Sixty-four collaborative long- and short-term experiments (Bangladesh=18; Cuttack, India=6; Faizabad, India=13; Indonesia=7; Philippines=9; Thailand=11) were conducted between IRRI and NARS scientists. Thirteen research papers were jointly written by IRRI and NARS scientists for publication, including 10 papers that summarized research findings from RLRRC to be published in a special 1999 issue of Experimental Agriculture. Two training courses on problem-based technology generation and biometry were conducted in Indonesia (15 NARS participants) and Bangladesh (6 NARS participants). Six NARS collaborators attended courses on strategic nutrient management and weed research methodologies conducted at IRRI headquarters. Ten NARS collaborators had on-the-job training for 2-4 wk at IRRI headquarters on data analysis and report writing.

Program outlook
Since the Rainfed Lowland Program began in 1990, substantial progress has been made in clearly defining the problems, understanding the key constraints, and developing workable solutions. The clearer vision of the hydrologic constraints has provided a basis for progress in subecosystem definition, germplasm improvement, cultural intervention, and better characterization of target subecosystems. New tools have been developed to better characterize the areas of rainfed lowland and have helped prioritize future research directions. Varieties better adapted to the principal constraints of drought and submergence have been released, and more will follow as integrated efforts among breeders, biotechnologists, and physiologists come to fruition.

Our understanding of the genetics and ecology of blast will result in more stable blast resistance. Better varieties provide a basis for more accurate agronomic interventions in target subecosystems. Integrated efforts on agronomic systems, nutrient cycling, and weed ecology are providing the depth of understanding needed for successful longer term management of these intensified but risk-reduced systems. In concert with these efforts, socioeconomic (risk management, technology appreciation, and adoption constraints) and gender concerns will be monitored in village-level studies, and the findings considered interactively with other research outcomes. Partnerships in collaboration with NARS are now providing clear benefits. Research is conducted where key constraints are clearly expressed and in collaboration with scientists vitally concerned with successful outcomes. Collaboration with IRRI scientists has improved the scientific rigor of research in the NARS while IRRI has benefited enormously from the clear priority-setting and outcome focus of our NARS partners. The next cycle of research will feature a smaller, more sharply focused agenda. It must address how to improve the reliability of crop establishment in direct seeding, while at the same time ensuring that weed competition is effectively managed. Early sowing may permit capture of the early flush of nitrate at the start of WS. Nutrient manipulation will be critical to enhancing seedling vigor and then to ensuring continued nutrient availability in the complex anaerobic-aerobic transitions of the rainfed lowlands. We already have some evidence that nutrients may partially lessen the adverse effects of drought and may enhance survival under submergence. We will focus on these issues because of their likely immediate impact. But longer term studies will continue to ensure the sustainability of increasingly more complex cropping systems. Success will ultimately depend on adoption of improved varieties and technologies. Socioeconomic and gender studies will provide the basis for developing targeted and acceptable packages. During the next few years, the Rainfed Lowland Ecosystem program will make significant advances in understanding and providing technology for more productive and sustainable rice-based cropping systems in the rainfed lowlands.

Rainfed lowland rice ecosystem


Research programs Upland rice ecosystem

IMPROVED PRODUCTIVITY AND SUSTAINABILITY OF FARMING SYSTEMS IN UPLAND RICE AREAS 54 Upland rice root system and nutrient effects on its development (APPA) 54 Tiller-nodal root development 55 Dynamics of soil and applied nitrogen in upland rice soil culture (APPA) 57 A multiscale approach for on-farm erosion research 58 Selected results at field-plot scale (APPA) 58 Catchment scale (APPA) 59 Farming household scale (APPA) 60 On-farm crop diagnosis of upland rice yields (APPA) 60 Cropping patterns and food security in northern Vietnam (SS) 63 A characterization of upland rice biotic constraints in Lao PDR (EPP) 64 UPLAND RICE RESEARCH CONSORTIUM (APPA, EPP, PBGB, SS, SWS) 67 Progress in 1998 67 PROGRESS OF UNREPORTED PROJECT 67 Genetic improvement of upland rice (PBGB, EPP, APPA) 67 PROGRAM OUTLOOK 68

Upland rice ecosystem


Upland rice ecosystem

Upland rice is grown annually on about 17 million ha worldwide10.5 million ha in Asia, 3.7 million ha in Latin America, and 2.8 million ha in Africa. Total upland production is about 20 million t. Rice is a major staple crop for upland farmers in many parts of Asia, Africa, and Latin America. The total area supporting upland rice-based cropping is considerably larger because of rotation with fallow and other crops. The crop is grown alone or in diverse mixtures in shifting or permanent fields under a wide range of conditions of climate, slope, and soil type, often as a subsistence crop receiving few purchased inputs, although it is commonly a commercial crop receiving inputs in favorable areas such as Brazil, Indonesia, and southern Philippines. The major thrusts of the program are developing knowledge and technology to maximize productivity and sustainability of upland rice where it is grown, helping maximize returns for farmers efforts, and reducing the area needed to satisfy demands for upland rice. Many studies are also providing a scientific understanding with broad application outside the upland ecosystem. The Program has several major themes. One involves germplasm improvement to overcome major abiotic (drought, erosion) and biotic (weeds, blast, nematodes) constraints, moving away from traditional breeding and selection and using new technologies to target, characterize, and incorporate desired genes. Two novel projects are on (1) developing a perennial rice for the uplands to help control erosion and improve food security and (2) investigating allelopathy in rice to assist with sustainable weed management. The perennial upland rice and allelopathy projects are also providing valuable information on genetic characterization of rice and its wild relatives and the genetics and physiology of tolerance for such constraints as drought and nematodes.

The second theme is on abiotic constraints, focusing on a strategic understanding of the dynamics of nutrients (P and N), organic matter, and soil acidity in upland soils. The third theme covers biotic constraints, investigating the biology and management of weeds, nematodes, and blast. Underpinning these themes is a study of the socioeconomics of the uplands, which is designed to characterize and understand the production systems and their dynamics and the impact of new technologies and policies. The Program is implemented in close collaboration with national agricultural research systems (NARS) through the Upland Rice Research Consortium (URRC), which includes Brazil, India, Indonesia, Lao PDR, Philippines, Thailand, and Vietnam. Bangladesh, China, and Myanmar are associate members. The Consortium, in operation since 1991 with support from the Asian Development Bank, the German Agency for Technical Cooperation (GTZ), and Japan, provides a framework for NARS-IRRI collaboration. It focuses on strategic issues and themes of importance to the upland ecosystem. Each partner focuses on a theme and share outputs with others, in which all share in the planning, implementing, and reporting of the research agenda.

Improved productivity and sustainability of farming systems in upland rice areas

Upland rice root system and nutrient effects on its development M. Kondo, P.P. Pablico, D. V. Aragones, R. Agbisit, J. Abe,38 S. Morita,38 and T. Winn Increased root-length density in upland rice is important for maintaining plant water status. Characteristics of root development and nutrient effects on root distribution were investigated.


IRRI program report for 1998


Nodal roots (no.) 25.0 Moroberekan 20.0 15.0 10.0 5.0 0.0 1st-3rd 4th UPLRi-5

The structure of the nodal root system in relation to tillers and nodes was examined at flowering of two upland cultivarslow-tillering and deep-rooted Moroberekan and medium-tillering and mediumdeep rooted UPLRi-5. They were grown in a root box in aerobic soil (soil water content at 35% v/v). Nodal roots per tiller and average length of nodal roots were highest for the main tiller, followed by primary tillers, and lowest in secondary tillers. Length of nodal roots from main tillers accounted for 50% of total nodal root length in Moroberekan and 43% in UPLRi-5. Roots from main tillers had a steeper elongation angle than the roots from other tillers, resulting in deeper roots. Root numbers were more in lower internodes than in higher internodes on a main tiller (Fig. 1). In both Moroberekan and UPLRi-5, nodal roots were longer at the 1st to 4th internodes compared with higher internodes (Fig. 2). The results indicate that stimulating the growth of nodal roots from early internodes in early tillers is likely to promote deeper rooting. Root development along growth stages. Root development and shoot growth for UPLRi-5 were monitored at an IRRI upland site during 1995 WS. Vertical root distribution, root length depth index (RLDI), and root mass depth index (RMDI), the depth which contains 50% of roots, were calculated as RLDI(cm)={(RL1 D1) + (RL2 D2) + (RL3 D3) +}/(RL1+RL2+RL3) RMDI(cm)={(RM1 D1) + (RM2 D2) + (RM3 D3) }/(RM1+RM2+RM3) where RLx = root length in the layer x (cm cm-1) RMx = root mass in the layer x (g cm-1) Dx = depth of layer x (cm) Increase in total root length was at a faster pace than increase in shoot dry weight before maximum tillering (Fig. 3). Intrahill root length increased until flowering, then decreased sharply while intrarow root length reached a maximum at panicle initiation. A similar trend was found in root dry weight (data not shown). The rooting zone expanded both vertically and horizontally from crop establishment to maximum tillering stage and slightly more horizon-


6th 7th Internode unit




1. Number of nodal roots in internode unit in main tiller in Moroberekan and UPLRi-5 grown in root box for 112 d. IRRI, 1998.

Average nodal root length (cm) 80.0 Moroberekan 60.0 UPLRi-5



0.0 1st-3rd







Internode unit 2. Average length of nodal roots in internodes in main tiller of Moroberekan and UPLRi-5 grown in root box for 112 d. IRRI, 1998.

tally from maximum tillering to panicle initiation. Maximum rooting depth was attained before maximum tillering. The RLDI and RMDI at panicle initiation were slightly lower than at maximum tillering (Fig. 4). Increased intrarow root length after maximum tillering was probably related to an increase in shallow nodal roots emerging from late tillers. Thus, there is a close linkage of tillering and nodal root development in the spatial expansion of the rice rooting zone.

Upland rice ecosystem


Root length (m m-2) 8000 6000 4000 2000 0

Shoot weight (g m-2) 800 600 400 200 0 120

Total root length in plant row Total root length between plant rows Shoot dry weight






Days after seeding 3. Change in root length and shoot dry weight in UPLRi-5 grown at an IRRI upland site, 1995 WS.

Tillers (no. m-2) 500 Tiller number 400 300

Root length depth index, root mass depth index (cm) 18 16

Root length depth index Root mass depth index

14 200 100 0 12


40 60 80 100 Days after seeding

10 120

4. Change in tiller number, root length depth index, and root mass depth index during the growth of UPLRi-5, IRRI, 1995 WS.

Effect of P and N on root growth and distribution in an acidic Ultisol. The effect of P and N on root distribution at flowering of upland rice was investigated in acidic Ultisols in Siniloan (pH 4.4) and Cavinti (pH 4.5), Philippines. Basal applications of P and N were on the soil surface as single superphosphate and controlled-release urea. Application of P increased total root length and total root and shoot dry weights (Table 1). Roots were deeper with P as indicated by higher RMDI and RLDI. Root length density tended to increase from the surface to lower layers, especially below 30 cm (Fig. 5). Similar effects of P on roots were observed in Cavinti. Nitrogen tended to increase total root length and root dry weight but with a shallower root distribution than P. The results of an experiment that grew Moroberekan and UPLRi-5 in a root box indicated that the deeper root distribution with P is mainly associated with stimulated growth of long nodal roots from main and primary tillers. Phosphorus increased the number of nodal roots per tiller and the long nodal roots from lower internodes in early tillers, which would lead to the increase in root density in deep layers (Table 2). Results overall imply that stimulating the development of main and primary rice tillers by P in the early vegetative stage results in greater root distribution in deeper soil layers during the reproductive stage. Nutrient management in the surface layer significantly affects the root length density near the soil surface and in deeper layers.

Table 1. Effect of P and N on shoot and root growth and root depth of an upland rice grown in Siniloan, Philippines. IRRI, 1998. Shoot dry weight (g m-2) Total root length (m m-2) Total root dry weight (g m-2) Root length depth index (cm) Root mass depth index (cm)


Effect of N No N N (90 kg ha-1) Effect of P No P P (50 kg ha-1)

310a 356a

5161a 5670a

143a 178a

23.3a 21.3a

18.9a 15.9b

281b 385a

4075b 6756a

123b 199a

19.9b 24.7a

16.2b 18.5a


IRRI program report for 1998

Depth (cm) 0-15 15-30 30-45 Control 45-60 60-75 75-90 0 0.5 1 1.5 2 2.5 Root length density (cm cm-3) 3 + 90 kg N ha-1 + 50 kg P ha-1 + 90 kg N ha-1 + 50 kg P ha-1

5. Effect of N and P on root length density of an upland rice grown in an acidic Ultisol at Siniloan, Philippines, 1998.

Dynamics of soil and applied nitrogen in upland rice soil culture M. Kondo, P.P. Pablico, D.V. Aragones, and R. Agbisit Improved N use efficiency is needed for upland rice because N recovery is generally low due to NO3leaching. Nitrate in soil solution and soil NO3--N and NH4+-N were monitored during an upland rice crop at Batangas (volcanic ash origin, pH 5.18) and

Cavinti (acidic Ultisol, pH 4.50) in 1995 WS, and at Batangas and Siniloan (acidic Ultisol, pH 4.41) in 1996 WS. Applied N in the soil and N recovery by IRAT216 were traced using 15N. The amount of NO3- and NH4+-N were higher in the Batangas soil in 1996 than in 1995. In 1996, the soil to 80 cm depth contained 170 kg KCl-extractable NO3- and 80 kg NH4+ just after the onset of the rainy season. Thereafter, NO3- decreased sharply until about 60 DAS, possibly due to leaching as rainfall also peaked during this period. Ammonium levels remained more constant than NO3- in all soil layers until crop maturity. A different soil NO3- profile was found in Siniloan and Cavinti. NO3- in the top 40-cm layer decreased slowly from the start of the crop season, while NO3- in the 40-80 cm increased despite high rainfall. A positive correlation between KCl-extractable NO3- and NO3- in soil solution was observed in Batangas, indicating that most of the KCl-extractable NO3- existed in soil solution. In contrast, NO3in soil solution was much lower than KCl-extractable NO3- in Cavinti and Siniloan. Thus, most of the NO3- extracted was NO3- absorbed on soil particles probably by Fe and Al oxides. These results indicate that high NO3- sorption capacity reduces NO3leaching and maintains high NO 3- in subsoil in acidic Ultisol. This is supported by the observation that N uptake by rice was higher in Siniloan than in Batangas after heading, possibly due to deeper roots recovering NO3--N from the subsoil.

Table 2. Effect of P level on number of nodal roots per tiller and number of long nodal roots (>30 cm) in internode units in main tillers in Moroberekan and UPLRi-5 grown in a root box. IRRI, 1998. Nodal roots tiller-1 (no.) Main tiller Primary tillers Long nodal roots (>30 cm) in internode units (no.) 1-4th 5-7th 8th>

Variety and P level

Moroberekan No P P (0.5 g box-1) P (1.0 g box-1) UPLRi-5 No P P (0.5 g box-1) P (1.0 g box-1)

16 50 53

2 19 19

8 21 26

4 13 11

0 4 0

14 38 44

2 7 10

4 10 13

4 6 8

1 0 0

Upland rice ecosystem


In Batangas, applied 15N recovery by rice was greater for late applications than for early applications, possibly a reflection of higher soil N availability early in the season. Early 15N application also resulted in high unrecovered 15N fraction, most of which was considered lost through leaching. In Siniloan, there was less difference in 15N recovery among timings of N application. Siniloan soil retained more 15N than Batangas soil, a likely consequence of NO3- absorption in the subsoil. Fifteen to 26% of applied 15N was found in the top 10-cm soil at crop maturity in both sites, presumably incorporated in the soil organic fraction. In summary, N mineralization upon the onset of the rainy season produces high inorganic N in the surface soil. NO3- then decreases from the surface soil mainly due to leaching of NO3- and incorporation into organic fraction. NO3--absorbing capacity of soil significantly affects the level of soil-available N in the root zone. In soils that do not absorb much NO3-, matching N application with N demand of rice would lead to increased N recovery. A multiscale approach for on-farm erosion research Guy Trbuil and F. Turkelboom41 Highland cropping and farming systems in Southeast Asia are diversifying rapidly due to population growth and market integration. Farming is characterized by an increasing diversity of small farmers practices and strategies. Research in northern Thailand showed that the approach using long-term runoff plots and long-term experiments to develop erosion control techniques is not adapted to current farming circumstances. Alternative approaches should be based on an understanding of erosion processes in farmers circumstances at the scale of fields, farms, and catchments. Three research tools were combined: an on-farm erosion survey, a catchment survey, and a farmer classification (Fig. 6). The research was conducted in Pakha Sukjai, a highland Akha village (800-1100 m) of Chiang Rai Province in northern Thailand. Population density was about 65 inhabitants km-2. The area is characterized by a monsoon climate (1,600-2,200 mm rainfall yr-1), a strong relief (slope angle of 30-70%), phyllite-derived soils, and an extensive range of cropping systems managed on a semipermanent basis.


Least erosive rainy event. Rainfall of at least 11 mm, a minimum erosivity value of 53 MJ mm ha-1 h-1, a minimum 30-min intensity (I30) of 15 mm h-1, and a duration of at least 37 min was necessary to generate visually detectable erosion. New erosion symptoms were always observed for daily rainfall amounts of 20 mm or more. Typology of erosion symptoms. The correspondence found between the diagnosed five stages of erosion severity and calculated soil losses was a useful tool for roughly, but rapidly, assessing soil losses at field level (Table 3). Thresholds for slope angles and lengths. Rills could be observed in nearly every field. Plow layer erosion started in fields with slope angles superior to 47% and lengths superior to 25 m. Field history and the fallow effect. Erosion symptoms differed between fields created after clearing a fallow and older fields. Complementary experiments showed that this was due to a high content of fallow vegetation roots in new fields, which caused higher macroaggregate stability. This fallow effect disappeared in the second year of cultivation. Thresholds for soil cover. With more than 50% of total soil cover and 30% of contact cover, erosion was found to be negligible. These thresholds were applied to delimit the duration of the critical periods for erosion. Delimitation of critical periods for cropping systems. An erosion efficiency score (EE) of 100% for a given storm means that the maximum increase of erosion features (from least severe type to the most severe one) was observed in all the monitored fields. An EE of zero means that no visual changes in erosion symptoms could be detected following that specific rain event. Results were that upland rice displayed a long period (4 mo) of susceptibility to erosion, while maize, soybean, and cabbage had a 1.5-mo maximum duration of the critical period. This relates to differing evolutions of total soil cover between upland rice (with slower growth rate and more intensive weeding) and the other shorter duration crops. Delimitation of the high-erosion-risk domain. The integration of the soil cover thresholds enabled the identification of the field combinations corresponding to high erosion risk (Table 4). This matrix is useful in planning appropriate improvements in


IRRI program report for 1998

= effect of erosion from one scale to another scale = effect of land-use dynamics at different scales


Spatial distribution of erosionsusceptible household groups

Identification of erosionsusceptible household groups

Socioeconomic and physical conditions Catchment scale

Identification of land-use strategy per household group

Identification of erosionsusceptible locations in watershed Gullies and landslides

Runoff generation

Runoff concentration - Landscape features - Linear rural infrastructure

Field scale

Identification of erosion-susceptible fields Inter-rills, rills, (gullies) Constant field properties - Slope properties - Soil properties - Previous land use Dynamic field properties - Soil/crop management (land use) - Soil cover - Soil roughness - Erosivity (rain) - Soil conservation measures

6. A proposed multiscale approach for on-farm erosion research. IRRI, 1998.

land use with farmers, and to provide local stakeholders and decisionmakers with knowledge of the erosion risk measured at field level.

Because of a mosaic of small fields or fallows and rural infrastructure interrupting runoff flows, only

10% of the gullies observed were developed from rills in long fields or from runoff that could flow from one field into another. Runoff concentration took place mainly as the result of contour concavity (36% of cases) or in 49% of the cases because of rural infrastructure (paths, declining ditches, field border ditches, irrigation channels, paddy terraces, etc.).

Upland rice ecosystem


Table 3. Relationship between calculated soil losses and maximum development of observed erosion features at Pakha Sukjai, Thailand, 1994 and 1995 WS. Stagea Year Cropb (t ha-1 yr-1) 5 1995 1994 1994 1994 1994 1994 1994 1995 1994 1994 1994 1995 1994 1995 1995 1995 1994 1995 1994 1995 1995 1995 1995 1995 1995 UR UR Mz UR UR UR Bn UR UR Mz Bn Bn UR Mz UR Mz Bn UR Mz UR Bn Mz UR Mz UR 350 270 229 187 145 108 64 61 60 36 20 18 17 13 10 10 10 7 6 6 5 4 2 2 2 x x x x Calculated soil loss Plow layer erosion x x x x (x) (x) Red rills Rill network Pre-rill network Pre-rills or occasional rills

x x x x x x

x x x x x x x x x x

x x

x x x x x x x x x x x x x x x x

x x

Stage 1: Very mild erosion (0-5 t ha-1 crop-1). Only 1-2 occasional rills (per 10-m contour) and there is low density of pre-rills (<1 per 2-m contour). Stage 2: Mild erosion (5-20 t ha-1 crop-1). Presence of a pre-rill network (>1 pre-rill per 2-m contour) or more than 3 active rills (per 10-m contour). Stage 3: Moderate erosion (20-100 t ha-1 crop-1). Presence of rill network (>1 rill per 2-m) or presence of red rills. Stage 4: Severe erosion (100-150 t ha-1 crop-1). Several active red rills and erosion of tilled topsoil at some local spots. Stage 5: Very severe erosion (150-350 t ha-1 crop-1). Plow layer erosion at the bottom of the field by widening rills or planar slides. bUR = upland rice, Mz = maize, Bn = bean).


Four main types of farming households were identified having different levels of susceptibility to erosion risk. Relatively high, and currently not improving erosion risks characterized the profit maximizer and diversifier categories of farms, which represented more than half of the local farming community. They were seen as the most appropriate target groups for participatory soil and water conservation efforts. The multiscale erosion survey approach is a fast and cheap way to generate new knowledge about erosion in diverse farming situations. Because ob-

servations are made in farmers circumstances and aim at understanding the key processes at work, the outputs are relevant to local issues. On-farm crop diagnosis of upland rice yields K. Van Keer, G. Trbuil, E. Goze, and C. Vejpas Swidden cultivation of upland rice in northern Thailand is still an important component of remote upland farming systems, but yields in farmers fields are generally low (1-1.5 t ha-1) and variable. Local cultivars display yield potentials of 3-4 t ha-1 in noinput cropping systems. But such potentials are seldom reached due to multiple limiting factors. The


IRRI program report for 1998

Table 4. Effects of slope characteristics and field history on the risk of erosion when soil cover was under the critical threshold. Pakha Sukjai, Thailand, 1994 and 1995 WS. Total soil cover Field cropping history <25m Fallow clearing 18 Older field 7 7 25 50 Slope angle and lengtha <47% >25m <25m 47-57% >25m <25m >57% >25m

Under critical threshold







Above critical threshold

Fallow clearing 0 Older field 0 33 50 0






Percentage of field observations in a given situation for which new erosion features were found and measured compared with the previous field visit330 field observations analyzed.

identification and ranking of limiting factors, as well as the understanding of their effects on upland rice crop functioning, is a prerequisite to setting research priorities for improving current practices. A 4-yr (1993-96) on-farm agronomic survey was done in a highland (600-800 m) Lahu village of northern Thailand to q date and quantify upland rice yield differentiation in actual farmers conditions, q characterize upland rice environmental conditions and farmers practices along the whole crop cycle, and q identify, rank, and understand the main environmental and cropping system variables influencing upland rice crop production and causing major yield limitations. Data on crop population status for two contrasting local upland cultivars, crop environmental conditions, and cropping practices were obtained through regular monitoring of 432 squares (1 m2 each) delimited at crop emergence in 63 farmers fields. Phase realization indices, which synthesize the impacts of crop constraints during a given phase of upland rice yield buildup, were calculated and distribution patterns analyzed to assess the relative

importance of successive growth phases in yield determination. Identity and causes of yield differentiation were assessed by a principal component analysis with instrumental variables (PCAIV, logtransformed data for yield components) on 1993-96 pooled data as well as for individual year subdata sets. The study site was characterized by monsoonal climate, a strong relief, deep granitic soils with clayloamy texture and medium chemical fertility, heterogeneous (2- to 10-yr-old) fallow vegetation, and a wide range of weed species, diseases, insects or other pests. Local farmers still use traditional swidden cultivation practices and no-input cropping systems. Only limited weed control practices (hoe tillage and NaCl application) were observed. Frequency distributions of phase realization indices pinpointed panicle formation and spikelet differentiation as key periods of yield differentiation (Fig. 7). During the vegetative period, poor crop biomass accumulation was of greater importance than low plant densities. Vegetative biomass accumulation per plant influenced panicle and spikelet formation, but not spikelet fertilization or grain filling.

Upland rice ecosystem


Percentage of squares 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 Early-maturing varieties (n = 90) Late-maturing varieties (n = 206)





100 80 60 40 20 0 0.0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

7. Frequency distribution of the successive phase realization indices of each growth phase for early and late UR cultivars grown in Mae Haeng. IRRI, 1993-96 pooled data.


IRRI program report for 1998

The multivariate analysis of pooled data (Fig. 7) revealed strongest negative relationships for q plant density vs late weed competition and late-maturing cultivars (lower sowing densities), q panicles plant-1 and spikelets panicle-1 vs rice root aphid infestation, and q percentage of filled spikelets and 1,000-grain weight vs 1995 wet season (dry spells occurred during the wet season). Weak negative relationships were found for panicles plant-1 vs slope angle, erosion, number of successive crops, and early weed stress. The analysis of individual year subdata sets isolated rice root aphid infestation as the single major limiting factor with a strong and consistent effect on final upland rice yield. In Mae Haeng, yield differentiation was found to occur mainly during biomass accumulation in the vegetative phase and the early part of the reproductive phase. The survey results support the validity of several common hypotheses on limiting factors (weed competition, drought stress, soil erosion). No major soil nutrient as a limiting factor was diagnosed in this study. The survey showed that more attention should be paid to soil-borne upland rice pests when prioritizing issues for the improvement of upland rice-based cropping systems in northern Thailand highlands.

Cropping patterns and food security in northern Vietnam S. Pandey, T. Blohm,39 Tran van Dien,40 and L. Velasco A socioeconomic survey of 100 farmers in Bac Yen District, Son La Province, examined the effect of market access on the choice of cropping pattern and food security in the uplands of northern Vietnam. The villages of Ta Xua, Hong Ngai, and Phieng Ban were surveyed to record the cropping pattern, landholding, income sources, and food consumption. The major characteristics of the villages are shown in Table 5. Ta Xua is least accessible to market. Phieng Ban is close to the district headquarters. Both Ta Xua and Hong Ngai have a small proportion of lowland area but in Phieng Ban, lowland area accounted for 25% of the total landholding. Ethnic groups in Ta Xua and Hong Ngai are mostly Hmong. In Phieng Ban, Hmong, Thai, and Muong are in equal proportions. Lowlands are used for growing summer rice, mostly rainfed. Several crops are grown in the uplands with maize and upland rice the major ones. The land use pattern in the uplands is most diversified in Phieng Ban, which had high market access, and least diversified in Ta Xua, which had low market access. Farmers in Phieng Ban have diversified their cropping patterns to include fruit trees. The

Table 5. Major characteristics of three villages in Bac Yen District, Son La Province, Vietnam, covered by a socioeconomic survey, 1998. Ta Xua Population density (no. km ) Market access Family size (no.) Lowland area cultivated (ha household-1) Upland area cultivated (ha household-1) Lowland rice area (ha household-1) Upland rice area (ha household-1) Lowland rice yield (t ha-1) Upland rice yield (t ha-1) Lowland rice output (t household-1) Upland rice output (t household-1) Rice supply (kg capita-1)

Hong Ngai 44 Medium 9.0 0.13 1.45 0.13 0.58 2.8 2.5 0.4 1.4 201

Phieng Ban 103 High 6.1 0.31 0.9 0.22 0.17 4.7 1.8 1 0.3 218

17 Low 12.6 0.59 3.1 0.59 0.5 3.2 1.9 1.9 0.9 227

Upland rice ecosystem


importance of upland rice varies among the villages. Average returns to land and labor in the production of upland rice, lowland rice, and maize are presented in Table 6. Due to their high returns to land, lowland fields are almost invariably planted to rice. Although returns to land are higher for upland rice than for maize, high labor requirement for upland rice, especially for weeding, lowers the returns to labor. There was an increasing trend toward substitution of maize for upland rice in all villages. Discussions with farmers about changes in cropping patterns during 1987-97 indicated three major types of adjustments to increased population pressure. Farmers in Ta Xua responded by increasing the terraced area to grow lowland rice and by expanding upland rice area. A similar expansion of upland rice was made by farmers in Hong Ngai, but terracing of uplands there was limited due to steeper slopes. Upland rice area in Phieng Ban was reduced over time as land use was increasingly diversified and more terraces were constructed in the lower slopes. Phieng Ban farmers took advantage of the more productive terraced fields for growing rice and generated additional incomes through a more diversified land use in the uplands. The total household income per capita in Phieng Ban was almost double that in Ta Xua (Table 7). Rice and livestock are important sources of income in Phieng Ban where markets for small animals, such as pigs and chickens, are readily available. The economy of Phieng Ban is market-oriented and its share of cash income in the total income is highest among the villages. At the same time, the share of upland rice in the total income is lowest in Phieng Ban, indicating that diversification of land use can be an important mechanism for income enhancement in areas with good market access.

The nature of food self-sufficiency was judged by asking farmers the number of years during the past 10 yr when rice production was insufficient to meet their consumption requirement. In Ta Xua and Hong Ngai, 20% of the farmers reported no shortage during all of the 10 yr. The corresponding figure for Phieng Ban was 34%. However, about 36% of the farmers in Phieng Ban reported food shortage in each of the past 10 yr. This indicated a somewhat bimodal pattern of food self-sufficiency in Phieng Ban with the farmers who had smaller area of lowlands being less food self-sufficient. Farmers in Phieng Ban were able to maintain their calorie consumption by purchasing rice in the market place. Farmers in Ta Xua and Hong Ngai bought less rice but increased consumption of maize, cassava, and other tuber crops. The major finding of the study was that although the production of upland rice is less correlated with food security in areas with better market access and diversified land use than in other areas, it is of vital importance where these conditions are not satisfied. In remote areas with limited lowland area, farmers have little choice but to rely on upland rice for their survival. An increase in productivity of upland rice will help discourage such farmers from continuing to expand upland rice area into the fragile environments. A characterization of upland rice biotic constraints in Lao PDR S. Savary and K. Fahrney Rice is an important crop in the uplands of Lao PDR, where 219,000 t are produced in a wide range of production situations on an area of 340,000 ha. Our analysis refers to farmers fields where the

Table 6. Average returns to land and labor in lowland rice, upland rice, and maize production in three villages in Bac Yen District, Son La Province, Vietnam, 1998. Ta Xua Crop Land ($ ha-1) Upland rice Lowland rice Maize 273 474 124 Labor ($ d-1) 0.6 1.5 0.7 Land ($ ha-1) 363 405 160 Labor ($ d-1) 1.0 1.9 1.2 Land ($ ha-1) 268 685 202 Labor ($ d-1) 0.4 1.7 1.0 Hong Ngai Phieng Ban


IRRI program report for 1998

Table 7. Household income per capita and its composition in three villages in Bac Yen District, Son La Province, Vietnam, 1998. Ta Xua Total income ($ capita-1) Share of cash income in total income (%) Income shares (%) Upland rice Lowland rice Other crops Livestock Home gardens Forest products Off-farm 59 32 19 37 16 14 1 13 Hong Ngai 72 50 32 9 17 27 1 14 Phieng Ban 114 63 8 26 14 34 4 14

variation in rice yield was from 200 kg to more than 3 t ha-1. This is probably a reasonable reflection of the range of production situations encountered in the uplands in Lao PDR, and perhaps in many upland rice fields of the region. Alleviation of upland rice pest constraints is seldom considered a priority, for a number of possible reasons: q secondariness of these constraints relative to others, such as erosion or soil fertility; q difficulty of pest management, e.g., soil insects; and q lack of documentation on pest importance. The approach was to monitor ongoing on-farm experiments, quantify crop growth and injuries, and try to link the resulting groups of variables to yield variation. In essence, the objective was not to quantify yield reduction due to upland rice pests (weeds, insects, and pathogens), but to identify circumstances in which those pests may lead to injuries, which in turn may have a potential yield reduction effect. Several statistical techniques were used. The first followed the design used in the on-farm experiments, with analyses of variances based on their original design. Because the data sets are sufficiently large, principal components, multiple regressions (1995 data), and correspondence analysis (combined data sets of 1995 and 1996) were also used. Each of these techniques has advantages and drawbacks. Their combination, however, allowed an overview of complex relationships in the upland system from a plant protection perspective.

A large and significant variation in yield across farmers fields, and a small and insignificant overall variation associated with the tested N treatment were found. In addition, small but significant variation in yield was attributed to the site treatment interaction the N treatment had a small but significant effect in some of the farmers fields. Eigen vectors generated from principal component analysis on crop growth variables were used as new, composite variables to describe yield variation using a stepwise, multiple regression analysis. This allowed a fair description (50.2% of yield variation accounted for by the regression) of yield variation using four of the six eigen vectors available. Introduction of individual injuries in the regression model in a stepwise procedure marginally improved (53.4 %) the fraction of yield variation explained. Only two injuries were retained in the final modelleaf blast and weed infestation. Relationships among variables are summarized in a graphic output of principal component analysis using its two first axes (25.3 and 16.7% of total variance explained, Fig. 8). Not all variables are shown on the graph. Crop density (DTY) in an early stage of the crop (Aug, DTY1) and at a later stage (Sep, DTY2), initial number of tillers (TILL1), and leaf numbers (LEAF1 and LEAF2) are associated, reflecting vigorous crop growth. This cluster of variables is associated with (early) leaf blast (LB1) and opposed to weed infestation (WEED1 and WEED2). It also is strongly opposed to (early) white grub injury (WG1). Early weed infestation (WEED1) is associated with (early) brown spot infection (BS1). Figure 8 primarily highlights the collinearity among variables, especially injuries. Collinearity among variables was one among several other reasons to use correspondence analysis and derive a more synthetic view of the data (Fig. 9), where four broad domains are defined. Domain A corresponds to presence of neck blast, whiteheads, and sheath rot, and to high levels of brown spot. Domain B corresponds to presence of sheath blight and white grubs, medium brown spot, and (late)

Upland rice ecosystem


Factor 2 1.0





-1.0 -1.0


0.0 Factor 1



8. Principal component analysis among some crop growth and injury variables in upland rice. IRRI, 1998.

weed infestation, and absence of deadhearts. Domain C corresponds to absence of whiteheads and neck blast and to low brown spot and (late) weed infestation. Domain D corresponds to low defoliator injury, to medium (early) weed infestation, and to high leaf blast and (late) weed infestation. Domains B, C, and D correspond, in turn, to yield levels. Domain C corresponds to very low yields (YLL), domain B to low-medium yields (YL-YM), and domain D to high yields (YH). This additional analysis was much more powerful than the previous parametric ones in describing yields. The quality of description of yield classes in this analysis (which uses injury levels only to define axes) varies from adequate (53%) to very good (85%). Figure 9 leads to a few points: q The figure shows a path of increasing yields, from very low to high. q No particular injury should be ascribed to major yield-reducing effects. Rather, combinations of injuries should be considered. q Weed infestation appears to be one important factor that determines yield levels. The relationship is not a straightforward one, however. Low weed infestation early in the crop cycle corresponds to medium-low yields and not to high yields. Medium weed infestation corresponds to high yields. A high weed in-

festation in the early crop stages corresponds to a range of possible actual yield levels. This reflects the competition among weeds and the rice crop, and also soil fertility, which enhances weed emergence. This complex relationship can also be phrased using the path of increasing yields. Low weed infestation in the beginning of the cropping season is not associated with high yield (and this could be attributable to exhausted soils, for instance). Rather, a link between medium weed infestation early in the cropping season and high yields is indicated. Weed infestation, its pattern over time, and the succession of weed species (not addressed here) might well prove useful indicators of fertility in the future. These domains may also be used for IPM recommendations. Domain C corresponds to low yields. It also corresponds to injury levels that are usually low. In domain C, therefore, low yields cannot be ascribed to losses due to rice pests. Environmental and agronomic features of the rice crop (poor production situations) seem rather to be the cause of the low yields, and there does not seem to be a need to target any particular pest for IPM under such production situations. In domain B, corresponding to medium-low yields, medium weed infestation toward the end of the crop cycle is the only emerging injury. It may be that this contributes to reducing actual yields. Under production situations that conventionally lead to such medium-low yield, weed control in the middle of the crop cycle might then be one constraint to consider. Under production situations that usually lead to high yields (domain D), weed infestation, both early and late in the crop cycle, seem to be constraints which, when lifted, might lead to higher yields. It may also be that blast would be a constraint under such production situations. It should be noted, however, that correspondence analysis does not specifically link leaf blast with possible yield reduction. Setting a specific priority on blast should, therefore, be done with caution, and would require specific on-farm testing on traditional varieties.


IRRI program report for 1998

0.312 0.289 0.267 0.244 0.222 0.199 0.177 0.154 0.132 0.109 0.087 0.064 0.042 0.019 -0.003 -0.026 -0.048 -0.071 -0.093 -0.116 NBP -0.138 -0.161 WHP -0.183 -0.206 -0.228 -0.251 -0.273 -0.296 -0.318 -0.341 -0.363 -0.386 -0.408 -0.431 -0.453 -0.476 -0.498 -0.87800







Medium yields Sheath blight; White grubs; No deadhearts; Medium brown spot; (Late)weeds

Low yields No neck blast; Low brown spot; Low (late) weeds; YLL




Neck blast; Whiteheads; Sheath rot; High brown spot


High yields Leaf blast; YH Low defoliators; Medium brown spot; High (late)weeds Medium (early) weeds
-0.58780 -0.29760 -0.00740

No leaf blast
W1M LBA 0.57300

0.312 0.289 0.267 0.244 0.222 0.199 0.177 0.154 0.132 0.109 0.087 0.064 0.042 0.019 -0.003 -0.026 -0.048 -0.071 -0.093 -0.116 -0.138 -0.161 -0.183 -0.206 -0.228 -0.251 -0.273 -0.296 -0.318 -0.341 -0.363 -0.386 -0.408 -0.431 -0.453 -0.476 -0.498


9. Domains of injuries and of potential impact of IPM on upland rice yield from a correspondence analysis. IRRI, 1998.

Upland Rice Research Consortium

The Upland Rice Research Consortium (URRC) facilitates collaboration and cooperation in upland research and development among NARS and IRRI. Progress in 1998 Steering and technical committee meetings were attended by 50 NARS, 12 IRRI, and 3 agricultural research organization (ARO) scientists. Site planning meetings, attended by IRRI scientists and NARS researchers were: Vietnam, 22 Feb.; Philippines, 26-27 Feb.; Thailand, 4 Mar; Brazil, 14 Mar; India, 26 Mar; Lao PDR, 29-30 Mar; and Indonesia, 22 Sep. Site work plans were consolidated and distributed among partners. Research programs were implemented at sites in the six full partner countriesIndia, Indonesia, Lao PDR, Philippines, Thailand, and Vietnam. There were 41 collaborative IRRI-NARS experiments: India 13, Indonesia 5, Lao PDR 7, Philippines 2, Thailand 7, and Vietnam 7. Each country site produced a research report that was distributed among partners.

Work exchanges, workshops, and on-the-job training activities continued. One IRRI and two NARS (Vietnam, China) scientists undertook onthe-job training. Four top senior officials from Thailand visited Vietnam upland sites. Three NARS (Lao PDR, Philippines, and Vietnam) scientists undertook training courses on Strategic Research on Nutrient Management and one staff member from Thailand attended training on Advanced Experimental Design and Data Analysis at IRRI. Two NARS scientists from Thailand, and one from Lao PDR participated in a Weed Research Methodology Workshop at IRRI.

Progress of unreported project

Genetic improvement of upland rice Brigitte Courtois

Investigated variety soil N management interactions in root growth under field conditions to enhance root development through proper management. Started modeling nodal root development to understand systematically G E interactions for root growth in the soil profile.

Upland rice ecosystem


Developed techniques to apply stress to individual plots at a specific phenological stage. This system uses drip irrigation with valves on each plot. Evaluated upland varieties (48 entries) differing in maturity for performance with a 20-d stress period at the vegetative, flowering, and grain-filling stages. Aus varieties showed the greatest tolerance for flowering stress. Evaluated selected lines of a mapping population under furrow irrigation with two short stress periods applied near flowering. Identified QTLs associated with grain set under stress. Evaluated another population of 120 recombinant inbred lines (RILs) derived from the cross IAC165 Co 39 for leaf rolling, leaf drying, and relative water content and crop growth under water stress. In parallel, mapped the population using RFLP markers and microsatellites and established a pre-liminary molecular map. Evaluated the same population of RILs for its allelopathic potential under laboratory conditions. Located QTLs controlling this trait but did not find major genes. Advanced the mapping populations of RILs specifically designed to study the genetic control of allelopathy up to the F 4 generation using single-seed descent. Established a collaboration with the University of Mississippi to identify allelochemicals and identified extracts with putative allelochemicals. At least three chemicals, yet to be identified, were found in the extracts using HPLC. Completed the inoculation of a set of 120 upland lines from India, Thailand, and the Philippines with 24 blast isolates. Clustered the varieties based on pathotyping and compared this dendrogram with the dendrogram resulting from isozyme analyses, which showed few relationships. Started probing for analogues to resistance genes on the same material.

Obtained BC1, BC2, and BC3 generations of the crosses needed to transfer QTLs for partial resistance to blast from Moroberekan into Vandana, an elite upland rice variety popular in India, using the advanced backcross QTL analysis method. Broke the linkage between alleles governing hindered endosperm development and rhizome expression in third-generation complex hybrids of O. sativa/O. longistaminata, which allowed production of plants with normal fertility and good rhizome expression. Selected 12 individuals from two accessions of O. longistaminata naturally introgressed with O. sativa on the basis of non-invasive rhizomes and production ability for assessing the prospect of domesticating the wild species and improving its seed set. Evaluated resistance to Meloidogyne graminicola (root-knot nematode) for 190 lines of BCLF1 and BCLF2 mapping populations for further gene tagging. Implemented the second year of comparison of breeders and farmers opinions on upland varieties. To test hypotheses on the importance of grain quality, organized a sensorial evaluation of the material showing how farmers appreciations differ when based on grain quality.

Program outlook
Upland rice ecosystem research will continue to address the major constraints in upland systems and develop better adapted germplasm and better basic understanding of socioeconomic, biotic, and abiotic processes. Research results will be aimed at development of more sustainable and productive ricebased cropping systems. Better understanding of the socioeconomic dynamics of the uplands led to priority being given to the dominant rice-based and perennial-based systems, which together cover some 85% of Asian upland rice area.


IRRI program report for 1998

Yield stability in upland rice is being developed by enhancing its drought resistance and weed-suppressing ability for subsistence systems. Its response to moderate inputs and its resistance to blast disease are being enhanced. In resource management, development of better nutrient, weed, and nematode management technology for favorable areas (where slopes are not steep, or rainfall is reliable, or both) continues. The availability of new research technology in ecosystem and genetic characterization, the use of molecular biology in germplasm improvement and pest and disease characterization, and a long-term approach to understanding the dynamics of nutrients and pests are expected to enable significant gains to be made in the different environments of the uplands. Farmer participation in breeding and selection programs is being evaluated. IRRIs breeding program now focuses strongly on using new technology in prebreeding to provide better genetic material to support the breeding programs of NARS.

Resource management research is developing a more basic understanding of upland processes, including nutrient cycling, soil acidity, biology of pests, and erosion. That will underpin the development of upland systems that have a diversity of rice and other crops. Economic and policy analysis research will continue to investigate technological, policy, and institutional interventions that enhance food security of upland farmers and sustainability of upland systems. The URRC provides a strong framework to facilitate collaborative NARS-IRRI research and institution building in Asia and, through Brazil, in Latin America. The Brazilian production system for upland rice now focuses on favorable areas, with the use of inputs such as irrigation, fertilizers, and chemical weed and disease control. Farmers are achieving yields of more than 5 t ha-1 and profitability of around $500 ha-1.

Upland rice ecosystem


Research programs Flood-prone rice ecosystem

EFFICIENT SELECTION TECHNIQUES AND NOVEL GERMPLASM FOR INCREASING PRODUCTIVITY OF FLOOD-PRONE RICELANDS 72 Rices with high iron and zinc in the grain (PBGB) 72 Optimal growing conditions and mineral content 73 Response to nitrogen 73 Antinutritional factors 73 Improved rice with enhanced iron and zinc in the grain 73 Effect of milling on grain iron content 74 Genes/QTLs affecting flood tolerance in rice (PBGB) 74 Screening technique for zinc efficiency (SWS) 74 Distribution of photoassimilates in deepwater rice (APPA) 75 CROP AND RESOURCE MANAGEMENT TO IMPROVE PRODUCTIVITY AND SUSTAINABILITY OF FLOOD-PRONE RICELANDS 76 Kinetics of phosphorus fixation and availability (APPA) 77 Phosphate dynamics in different treatments 78 PROGRAM OUTLOOK 83

Flood-prone rice ecosystem


Flood-prone rice ecosystem

There are about 12 million ha of flood-prone rice worldwide, with almost 95% of it in South and Southeast Asia. The flood-prone ecosystem incorporates many types of rice adapted to different flood conditionsdeepwater, tidal saline, tidal nonsaline, and borogrown in widely different environments. The programs projects, germplasm improvement and crop and resource management, assist national agricultural research systems (NARS) in promoting productive, sustainable, rice-based farming systems for the ecosystem.

Efficient selection techniques and novel germplasm for increasing productivity of flood-prone ricelands
Germplasm improvement research continues to focus on abiotic stresses of excess water and problem soils, and on screening techniques and the genetics and physiological mechanisms of the predominant stresses in the ecosystem. However, an important research component is identifying and improving rices with high Fe and Zn in the grain. The project also assists NARS in developing improved floodprone rice cultivars and promotes inter-NARS collaboration. India and Thailand have assumed leadership for production and distribution of deepwater rice germplasm for South and Southeast Asia. Rices with high iron and zinc in the grain G.B. Gregorio, D. Senadhira, R.D. Mendoza, A.N.R. Monroy, R.D. Graham,16 R.M. Welch,17 and H.E. Bouis18 Major developments in 1998 were q Identification of a high-Fe content, highyielding rice (IR68144-2B-2-2-3) among improved lines being tested at IRRI. Progress in establishing the Fe bioavailability in highdensity and low-density rice using Fedeficient rats and human colon (caco-2) cells. q The discovery of varieties with high Fe content was made after germplasm screening showed that aromatic varieties tended to be high in Fe content. After milling, they have about twice the Fe content of present commercial varieties. The significance of that is that a substantial amount of time is saved by not having to breed the high-Fe trait into elite


IRRI program report for 1998

lines. This option remains open, however, for possibly increasing Fe content by a factor greater than two.

Table 1. Fe and Zn content of 23 rice varieties grown in a greenhouse with optimum fertility and full protection from insects and diseases. IRRI, 1998. Variety Ganjay Roozy Zuchem YR4194 Banjaiman Xua Bue Nuo IR64446 Kinmaze Tsuyake CNA 6187 Miyazaki 7 Ketan Ireng CT7127 Huri 370 Lagrosa Ketan Menah IR10198 Skybonnet IR60864 Heibao Alan IR63877 IR74 IR72 IR36 Fe (mg kg-1) 26.4 a 23.4 b 23.2 b 22.7 bc 22.5 bed 22.2 bcd 21.7 bcd 21.2 bcd 20.7 bcd 20.3 cde 20.2 cde 20.1 cde 19.8 de 18.0 ef 16.1 fg 15.8 fgh 15.3 ghi 15.0 ghi 14.9 ghi 14.0 g-i 13.1 hij 13.0 ij 11.7 jk 10.1 k Zn (mg kg-1) 58.9 a 51.0 b-e 54.0 abc 53.0 a-d 46.6 efg 53.5 abc 51.7 b-f 42.5 f-i 54.5 ab 42.5 f-i 55.3 ab 47.0 d-g 45.7 efg 48.2 c-f 45.4 e-h 37.9 if 41.3 ghi 41.1 ghi 31.6 k 39.2 hi 36.4 ijk 36.4 ijk 32.5 jk 31.4 k

Twenty-three rice varieties, including three commercially grown (IR36, IR72, and IR74), were grown in the greenhouse with optimum nutrients. Fe content of grain (brown rice) ranged from 10.1 mg kg-1 in IR36 to 26.4 mg kg-1 in Ganjay Roozy (Table 1). Fe was low in all commercial varieties. Zn concentration ranged from 31.4 mg kg-1 in IR36 to 58.9 mg kg-1 in Ganjay Roozy. The experiment showed that there is potential for doubling the Fe density and increasing the Zn density by 60%.

Three high-yielding rice varieties were tested with four different levels of added N (0, 46, 90, and 135 kg ha-1) in the field and Fe and Zn content of grain (brown rice) analyzed (Table 2). Differences in Fe obtained at various N levels were highly significant and indicated that Fe increased with the addition of N. No significant interaction between N and varieties was observed. Zn concentration did not increase with the addition of N.

chosen and the correlations between P and Fe and P and Zn content in grain calculated. Highly significant positive correlations were found for P with Fe (r2 = 0.42) and P with Zn (r2 = 0.50) relationships. The analysis suggests that Fe and Zn content is associated with phytate.

In germplasm screening tests for Fe and Zn, a large variation in P content of brown rice was observed. It ranged from 2.2 to 5.9 g kg-1. All varieties with high Fe and Zn also had high P. There were also varieties with average Fe and Zn content and high P. From the samples analyzed, 200 were randomly

High Fe density and high yield traits can be combined. This was demonstrated in the cross of IR72 and a tall traditional variety from India, Zawa Bonday. The line IR68144-3B-2-2-3 from the cross

Table 2. Fe and Zn content (mg kg-1) of three improved rice varieties grown at four levels of added N. IRRI, 1998. N level (kg ha-1) Fe 0 45 90 135 Mean 12.0c 12.9b 13.9a 14.2a 13.3 IR60819 Zn 27.9b 28.5b 29.8ab 30.9a 29.3 Fe 11.4c 12.4b 13.4a 13.6a 12.7 IR59682 Zn 25.7b 27.0ab 28.9a 29.0a 27.6 Fe 12.0b 12.6b 13.6a 13.8a 13.0 IR72 Zn 26.5a 26.5a 28.4a 28.0a 27.3 Fe 11.8c 12.6b 13.6a 13.6a 13.0 Mean Zn 26.7 27.3 29.0 29.3 28.1

Flood-prone rice ecosystem


had a high concentration of Fe (21.4 mg kg-1) in brown rice The elite lines from this cross have good tolerance for rice tungro virus, are aromatic, and have excellent grain types. The yields were about 10% below IR72 but maturity was earlier. One elite line is a mid-amylose type suited for rice consumers in the Philippines. This elite line will be used for bioavailability feeding trials. It is in the advanced yield trials of the National Seed Industry Council of the Philippines.

Grain Fe content (mg kg-1) 20 18 16 14 12 10 8 Xua Bue Nuo 6 IR68144 IR64 4 Jamagna 2 Tong Lan Mo Mi 0 0 15




Polishing time (min)

A comparison at different polishing times of highFe traditional varieties demonstrated a strong genotype time of milling interaction and indicated that much of the Fe is in the outer layers of the grain. IR64, with the lowest Fe content in brown rice, lost more than 33% Fe with 15-min polishing but remained almost unchanged as polishing time increased. A loss of about 33% after 15-min milling was also observed for the high-Fe traditional rices Jalmagna and Tong Lan Mo Mi, but their Fe concentrations drastically decreased as polishing time increased. A traditional variety from China, Xua Bue Nuo, and the high-Fe IR68144-3B-2-2-3 were less affected by polishing time. With a time equivalent to that of commercial polishing (25 min), IR68144-4B-2-2-3 had about 79% more Fe than IR64 (Fig. 1). Genes/QTLs affecting flood tolerance in rice K. Sripongpangkul, G.B.T. Posa, D. Senadhira, and Zhikang Li To facilitate breeding of high-yielding varieties tolerant of flooding, the genes/QTLs controlling rice elongation ability and submergence tolerance under flooding were investigated using AFLP/RFLP markers and 165 recombinant inbred lines (RILs) derived from IR74/Jalmagna (a rapid elongating rice from India). Under slowly increasing water levels, the parents and the RILs were evaluated for elongation ability and submergence tolerance in both the greenhouse (test 1) and field experiments (test 2). Five QTLs affecting plant elongation were identified in the IR74/Jalmagna RILs (Table 3). These included QIne1, QIne2, and QIne4, which affected internode and plant elongation. The Jalmagna allele

1. Grain Fe content of selected rice varieties at different polishing times (consumption rice equivalent to 15-min polishing). IRRI, 1998.

at QIne1 had a large effect of 13.6 cm and 8.7 cm for plant and internode elongation in test 1; the respective effects in test 2 were 23.2 cm and 18.2 cm. QIne4 was detected only in test 2 and the IR74 allele at this locus had a very large effect of 21.8 cm for elongated internodes. QIne2 was detected in test 2 and the Jalmagna allele at this locus caused elongated internodes. Two QTLs affecting leaf elongation, QLe6 and QLe7, were also identified. Seven genomic regions showed significant associations with submergence tolerance. The Jalmagna allele at five of these regions was associated with submergence tolerance, whereas the IR74 allele at the other two regions contributed to submergence tolerance. Four (on chromosomes 1, 3, 4, and 5) of the seven submergence tolerance loci were located in the QTL regions for plant elongation. Of these, the marker showing the strongest association with submergence tolerance was P2M7-4 on chromosome 9, where a major gene for submergence tolerance, Sub1(t), was reported previously. Surprisingly, it was the IR74 (non-submergence tolerant parent) allele at this locus that contributed strongly to submergence tolerance. Screening technique for zinc efficiency P. Thongbai, C.Q. Guerta, and G. Kirk We seek screening techniques for different types of Zn efficiency in rice: q external efficiency, allowing greater uptake from low-Zn soil; and


IRRI program report for 1998

Table 3. Main-effect QTLs affecting plant elongation ability and submergence tolerance detected in the IR74/ Jalmagna recombinant inbred population. Test 1 Traita QTL Chromosome 1 2 4 4 6 7 Marker interval LOD RG109 ~ sd-1 P2M9-8 ~ P2M6-8 P3M1-5 ~ P3M5-1 P3M5-1 ~ P2M5-8 P1M10-7~ P2M5-17 P1M3-9 ~ P1M10-5 Markerb P1M6-2 P1M3-5 P2M5-8 P1M6-9 P3M7-3 P2M7-4, P3M1-3 7.0 a (cm) -8.7 R2 (%) 20.1 LOD 18.9 5.1 10.7 a (cm) -18.2 -9.6 21.8 R2 (%) 33.1 8.6 36.7 Test 2


QIne1 QIne2 QIne4 QLe4 QLe6 QLe7

6.2 2.5 G2

-6.2 4.9

14.2 9.4 3.7 -3.5 12.0 G2 23.2 15.7 19.7 9.4 11.9 58.3 13.0

QSut1 QSut3 QSut4 QSut5 QSut8 Sub1(t) QSut10


Chromosome 1 3 4 5 8 9 10

Tolerance allele

7.2 6.9 5.1

Jalmagna Jalmagna IR74

Tolerance allele Jalmagna IR74 Jalmagna Jalmagna Jalmagna IR74 Jalmagna

INI and LLI are increments of internode and leaf length. The gene effect a is the phenotypic effect due to substitution of the Jalmagna allele by the IR74 allele. bUnderlined markers are associated with detected QTLs affecting internode or plant elongation. G2 is the likelihood ratio chi-square statistic. The significant values of G2 at P = 0.05, 0.01, and 0.001 are 3.79, 6.63, and 11.60, respectively.

internal efficiency, allowing greater growth per unit Zn uptake. Previous work found that rice plants in low-Zn soil were able to solubilize Zn in the rhizosphere and increase their Zn uptake. They did that by acidifying the rhizosphere through release of H+ from the roots to balance excess intake of cations over anions, and production of H+ in Fe2+ oxidation by rootreleased O2. The latter process suggests it might be possible to screen cultivars for external Zn efficiency based on their ability to oxidize the rhizosphere. We compared two cultivars contrasting in external Zn efficienciesMadhukar which is efficient and IR26 which is notfor their ability to oxidize the rhizosphere. The plants were grown in sand that was continuously perfused with deoxygenated nutrient solution. The system was gas-tight so that the only means of O2 entry was via the roots. We calculated the rates of O2 release from the roots by measuring changes in O2 concentration in the solution bathing the roots. Figure 2 shows plant growth and root morphology in the two cultivars. Madhukar produced four times the dry weight of IR26 by the early tillering stage. Root mass and surface area were larger in Madhukar but not root surface area per unit root mass. Root porosity and rates of O2 release from the

roots (Fig. 3) were also larger. Oxygen release per pot increased as the plants grew, but rates of release per unit root mass decreased, presumably because as the root mass increased, the rate of O2 supply from the shoot was increasingly limiting. However, for a given shoot mass, the rate of O2 release per unit root mass was far greater in Madhukar than in IR26. This is consistent with the greater external Zn efficiency in Madhukar due to solubilization of soil Zn by root-released O2. This method should provide a useful screening technique for external Zn efficiency. Distribution of photoassimilates in deepwater rice O. Ito, T. Setter, and A.M. Mazaredo Deepwater rice (DWR) is grown where maximum floodwater depth can range from 50 cm to about 400 cm. Research has shown that yields of modern DWR were similar to those of high-yielding irrigated rice grown with the same fertilizer management (80 kg N ha-1) and agroclimatic environment. This implied that any above-water plant characters associated with high-yielding irrigated rice should be applicable for DWR and that the grain-filling pattern of DWR should not differ from that of highyielding irrigated rice. An experiment to prove this

Flood-prone rice ecosystem


Shoot dry weight (g pot-1) 2.5 2.0 1.5

Root surface area (dm2 pot-1) 6 5

IR26 Madhukar 4 3

1.0 2 0.5 0.0 Root dry weight (g pot-1) 0.5 0.4 25 0.3 0.2 0.1 0.0 20 15 10 5 0 2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16 Days after planting 1 0 Root porosity (%) 35 30

2. Plant growth and changes in root morphology over time in Zn-efficient Madhukar and Zn-inefficient IR26 grown in sand continuously perfused with deoxygenated nutrient solution. Means of three replicates; error bars indicate 1 SE. IRRI, 1998.

point used 14C as a tracer. Flag leaves were fed with 14CO to compare the photoassimilate partitioning 2 during the reproductive stage of a modern DWR (HTA60) with that of high-yielding irrigated cultivars (IR8 and IR72). The pattern of translocation in irrigated rice was similar to that in DWR. The flag leaf assimilates were translocated quickly to the leaf sheath and eventually stored in the panicle before and after flowering. At booting, 10 d before flowering, radioactive carbon was localized in the flag leaf, leaf sheath, and developing panicle in both DWR and irrigated rice (Fig. 4). No assimilates from the flag leaf were translocated to the lower part of the stems and adjacent tillers. During grain filling, 10 d after flowering, 85-93% of assimilates were translocated immediately into the panicle (Fig. 5) in all cultivars

48 h after 14C feeding. There was no assimilate translocated into the stem in HTA60 and IR8 unlike in IR72 where there were some traces in the stem even at 24 h after feeding. The research on DWR is complete, but the 14C technique developed will be used in research on the sink-source relationship of the new plant type (NPT) in the irrigated rice ecosystem program.

Crop and resource management to improve productivity and sustainability of flood-prone ricelands
Research to alleviate constraints to productivity of flood-prone rice, particularly in acidic soils of coastal areas, includes analyzing the roles of mineral elements in flooding tolerance and productiv-


IRRI program report for 1998

O2 release (mol pot-1 d-1) 30 IR26 25 20 15 10 5 0 O2 release (mol g-1 root d-1) 160 140 120 100 80 60 40 0.0 Madhukar

ity of rice cultivars, and developing environmentally sound soil and water management. Partnerships with NARS for formulating and testing new technologies to alleviate poverty through promotion of sustainable, higher, and more stable production systems are an important part of the flood-prone rice ecosystem program. Kinetics of phosphorus fixation and availability V. P. Singh Phosphorus is a key nutrient in governing the productivity potential of soils. Fertilizer P applied to soil undergoes irreversible sorption and transformations to different forms that strongly influence P availability. It is unclear whether P availability is controlled strictly by solubility or by sorption relationships. Many soil factors such as pH, quantities of Fe, Al, and Ca; type and quantity of clay and organic matter, and level of redox potential are known to influence P fixation. Acid sulfate and inland Fe-toxic soils are extensively used for crop production in South and Southeast Asia, but with poor response to P fertilization because of low pH and high Al and Fe concentrations. Leaching and liming are often recommended to increase P availability. The extent of P fixation and interrelationships of different forms in these soils have not been precisely examined. We investigated the kinetics of P fixation and availability in untreated, limed, and leached acid sulfate soils, and


1.0 1.5 Shoot dry weight (g pot-1)


3. Rates of oxygen release from roots of Zn-efficient Madhukar and Zn-inefficient IR26 into sand cultures continuously perfused with deoxygenated nutrient solution. Means of three replicates. IRRI, 1998.

Radioactive carbon (% uCi) HTA60 IR8 100 80 60 40 20 0


Adjacent tiller Main culm Main panicle Main flag leaf sheath Main flag leaf

O 6 24 48

O 6 24 48 Time (h)

O 6 24 48

4. Distribution of 14C in HTA60, IR8, and IR72 at 0, 6, 24, and 48 h after assimilation of main flag leaf at 10 d before flowering.

Flood-prone rice ecosystem


Radioactive carbon (% uCi) HTA60 IR8 100 80 60 40 20 0


Adjacent tiller Main culm Main panicle Main flag leaf sheath Main flag leaf

O 6 24 48

O 6 24 48 Time (h)

O 6 24 48

5. Distribution of 14C in HTA60, IR8, and IR72 at 0, 6, 24, and 48 h after assimilation of main flag leaf at 10 d after flowering.
Table 4. Properties of soils used in a study of kinetics of P fractions. IRRI, 1998. Soil Acid sulfate Neutral Inland Fe toxic 6.5 1.6 9.8 0.7 7,880 8,000 94 3.8 1.0 6.5 190 1,160 2,300 1,500

Dry pH 3.7 Organic matter (%) 2.7 Available P (mg kg-1) 6.8 Exchangeable Al (mg kg-1) 120 Exchangeable Ca (mg kg-1) 2,400 Active Fe (mg kg-1) 12,000 Acetate soluble sulfate (mg kg-1) 2,570

in inland Fe-toxic and neutral soils. All soils (Table 4) were sampled in bulk from undisturbed areas. The experiment had six treatments: A) untreated acid sulfate soil without P application, B) untreated acid sulfate, C) limed acid sulfate, D) leached acid sulfate, E) neutral, and F) inland Fe-toxic soils, all with P application. For treatments A, B, and C, the soil was put in five tanks (one tank per replicate) to 10 cm thickness. For treatment D, a portion of the same soil was leached seven times with brackish water and dried, pulverized, and filled in five tanks to 10 cm thickness. For treatments E and F, the neutral soil and inland Fe-toxic soil were separately prepared and filled in five tanks in the same manner as treatments A, B, and C. All soils were flooded to 5 cm depth with brackish water, and then leveled.

After water subsidence, 2.5 t CaCO 3 ha -1 was evenly spread on the soil surface of treatment C. All treatments then had 1 t chicken manure ha-1 applied and were again flooded with brackish water to 3 cm depth and maintained for 2 wk. Phosphorus was then applied as single superphosphate in solution form at 5 mg P kg-1 soil, followed by a second application after 2 wk. Soil samples from each treatment were separately analyzed before starting the experiment (Table 4), just before P application, and daily for 6 d in the first week and once in the second week after P applications. Soil pH, exchangeable Al and Ca, active Fe, and organic matter were determined once a week.

Total phosphate. The trends of total P concentration and its mean values are shown in Table 5 and Figure 6. After P applications, total P content increased in all treatments but the increase was significant only in Fe-toxic soil. The increase was partly contributed by the addition of brackish water, which contained 0.3 mg total P liter-1. Iron phosphate. The Fe-P content represented the most dominant fraction of total P in Fe-toxic soil (23% initial and 28% final mean) and was the lowest fraction in neutral soil (8% and 9%) respectively (Table 5, Fig. 6).


IRRI program report for 1998

Table 5. Levels (mg kg-1) of total P, iron-P, and aluminum-P in different treatments. IRRI, 1998.a Total P Treatment Initial Untreated acid sulfate without P Untreated acid sulfate with P Limed acid sulfate with P Leached acid sulfate with P Neutral soil with P Iron toxic soil with P

Fe-P Change 2.2 ns 2.6 ns 4.4 ns 3.8 ns 4.2 ns 3.6 * Initial


Al-P Change 0.2 ns -2.0 ** -2.4 ** 1.6 ** 0.8 ns 2.8 ** Initial




Finalb 38.6 c 40.0 b 42.0 a 41.8 a 0.4 e 20.4 d

Change 0.4 ns 2.8 ** 2.4 ** 3.6 ** 0.2 * 3.2 **

129.4 b 129.0 b 139.8 a 129.6 b 137.6 ab 39.2 c

131.6 b 131.6 b 144.2 a 133.4 b 141.8 a 42.8 c

19.8 b 22.0 a 16.4 c 11.2 d 11.4 d 9.0 e

20.0 a 20.0 a 14.0 b 12.8 c 12.2 d 11.8 d

38.2 a 37.2 a 38.6 a 38.2 a 0.2 c 17.2 b

a * and ** significant at P = 0.01 and 0.05, respectively. bInitial is before P application. Final is the mean of 15 successive samples after P application.

Before P application, the Fe-P concentrations were significantly higher in both untreated acid sulfate soils and were lowest in Fe-toxic soil. The leached and the limed acid sulfate soils had significantly lower Fe-P concentrations than the untreated acid sulfate soils. This is attributed to the reduction of ferric Fe due to the removal of acidity and Fe through leaching, and to the neutralization of acids and precipitation of Fe compounds with lime. The levels of pH and Ca in the soil, and in the supernatant of both treatments, and pH and Fe content in the leachate of the leached soil support these explanations. The decrease in Fe-P was about two times more pronounced with leaching than with liming, possibly due to greater soil reduction in the leached soil. Iron-P concentration increased after P applications for 2 d and then decreased, in all treatments. However, the increase was not significant in the untreated acid sulfate soil without P and the neutral soil. These changes were gradual in Fe-toxic and leached acid sulfate soils. The final mean values (Table 5) were identical and significantly higher in both untreated acid sulfate soils (with and without P), and were lower in neutral and Fe-toxic soils. Comparison of final and initial Fe-P values showed different trends in different treatments. There was no change in untreated acid sulfate soil without P application because of the impaired progressive soil reduction due to low soil pH and high active Fe concentration, which is a more effective oxidant at low pH. The iron oxides in untreated acid sulfate soil without P were X-ray amorphous. No significant change was noticed in neutral soil, which also had inherently low Fe content. Iron-P decreased in limed acid sulfate soil (Table 5), with an overall decrease of 33% due to

soil reduction and consequent precipitation of iron compounds with lime. The leached acid sulfate soil had an increase, possibly due to the diffusion of reduced iron from underneath and its reoxidation at the soil-water interface. The active Fe retained in the leached soil has been reported to produce FeSO4 and soluble-free Fe upon soil reduction. The oxidation of FeSO4 produces Fe(OH)3 precipitate and soluble H2SO4 and the oxidation of Fe may fix the P present in the system. In this treatment, a continuous decrease in water pH and increased levels of soluble Fe was also noted. Because of that, the initial decrease of 47% in Fe-P reverted to 39% with a 14% increase after P application. The increase of Fe-P in Fe-toxic soil was attributed to higher levels of active Fe and lower degree of soil reduction due to low pH, organic matter, and Ca contents. These conditions are reported to diminish soil reduction and kinetically favor the formation of Fe-P. Aluminum phosphate. The Al-P content, as percentage of total P, was highest in Fe-toxic soil (44% initial and 47% final mean) but lowest (<0.3%) in the neutral soil (Table 5 and Fig. 6). All acid sulfate soils had almost identical values both at the initial (29%) and the final stage (30%). Before P application, all acid sulfate soils had similar, but significantly higher, Al-P concentrations than other treatments. The neutral soil had the lowest Al-P concentration. Fe-toxic soil had significantly lower Al-P content (17.2 mg kg-1) than all acid sulfate soil treatments. The leaching and liming of acid sulfate soil resulted in a significant decrease in exchangeable Al but had no effect on Al-P, apparently due to no hydrolysis of Al-bound P. Al-P concentrations increased after P application in all treatments. The increase was consistent in Fe-

Flood-prone rice ecosystem



Total P (mg kg-1) Fe-P (mg kg-1) Al-P (mg kg-1)

60 30


140 25 50

120 20 40

100 15 30

IRRI program report for 1998

10 20 5 10 0 0 0 0 1 2 3 4 5 6 11 15 16 17 18 19 20 21 25 1 2 3 4 5 6 11 15 16 17 18 19 20 21 25





1 2 3 4

5 6 11 15 16 17 18 19 20 21 25

Sampling period (d) Available P (mg kg-1)


Sampling period (d)

Sampling period (d)

Ca-P (mg kg-1)


120 20

100 15


60 10

Acid sulfate without P Acid sulfate with P Limed acid sulfate with P Leached acid sulfate with P Neutral with P Iron toxic with P

40 5

20 0 0 1 2 3 4 5 6 11 15 16 17 18 19 20 21 25

1 2 3 4

5 6 11 15 16 17 18 19 20 21 25

Sampling period (d)

6. Concentrations of phosphate in different forms in the soil of different treatments. IRRI, 1998.

toxic soil, while other soils showed some fluctuations. The final mean values were significantly higher in limed and leached acid sulfate soils. The increase in Al-P over initial values was significant in all treatments except the untreated acid sulfate soil without P. Higher concentration and the increase in Al-P in leached and limed acid sulfate soils may have been caused by the release of Al from the exchange sites and also from the lattice dissociation of the clay mineral. The amorphous and surface active aluminum hydrous oxides (Al (OH)3), formed from the neutralization of exchangeable Al, also could have contributed to P adsorption. In untreated acid sulfate and in Fe-toxic soils, the Al-P fixation was essentially due to the Al from exchange sites. These treatments also had significantly higher levels of exchangeable as well as soluble Al. The magnitude of change in Al-P, Ca-P, and FeP indicates that the presence of Al also enhances the refixation of P released upon the dissolution of CaP and the reduction of Fe-P. The increase in Al-P in untreated acid sulfate soil was closely associated with the decrease in Fe-P and Ca-P. In limed acid sulfate soil, it was with applied P and with the decrease in Fe-P; whereas, in leached acid sulfate soil, it was only with the decrease in Ca-P. In Fe-toxic soil, the increase in Al-P was mostly from applied P. In that treatment, the levels of exchangeable Ca and organic matter, as well as soil pH, were too low to cause any significant soil reduction. Hence, decrease in Fe-P, increase in pH, or any consequent decrease in Al solubility did not take

place. This treatment had not only the highest concentration of Al-P and exchangeable-Al but also of soluble Al, which could chemically immobilize the soluble P. Calcium phosphate. Calcium-P content was the dominant fraction of total P in neutral soil and was lowest in Fe-toxic soil (Table 6 and Fig. 6). Both untreated acid sulfate soils also had 35% and 36% initial and 30% and 23% final total P as Ca-P. In the limed and leached acid sulfate soils, the Ca-P fraction was 35% and 46% initially, and 48% and 34% at the end. Before P application, the Ca-P concentration was highest in neutral soil and was lowest in Fe-toxic soil. The leached acid sulfate soil had significantly higher concentration than other acid sulfate soil treatments. This was contributed by the brackish water used for leaching, which contained 0.26 mg liter-1 total P, mostly as Ca-P, and 30 mg liter-1 soluble Ca. While Al was being removed through leaching, the soluble Ca apparently precipitated P released by the reduction of Fe-P. This is reflected by the significantly lower initial concentrations of exchangeable Al and Fe-P without any change in AlP in this treatment than in untreated acid sulfate soils. The Ca-P content in the limed acid sulfate soil treatment was identical to untreated acid sulfate soil, but was significantly lower than the leached acid sulfate soil. That was because of the formation of insoluble Ca complexes with Al and Fe at the soil surface as a result of their accumulation due to the flooding of soil and evaporation of water prior to

Table 6. Levels (mg kg-1) of calcium P, available P, and supernatant P in different treatments. IRRI, 1998.a Calcium P Treatment Initialc Untreated acid sulfate 46.2 c without P Untreated acid sulfate 45.2 c with P Limed acid sulfate 46.4 c with P Leached acid sulfate 60.0 b with P Neutral soil with P 105.0 a Iron toxic soil with P 1.5 d

Available P Change -15.8 ** -5.8 ** 22.4 ** -14.8 ** 8.0 ** -0.5 * Initialc 13.6 b 13.6 b 14.6 a 16.0 a 14.0 b 5.8 c Final c 13.0 c 17.4 b 18.4 a 18.4 a 17.8 b 7.8 d Change -0.7 ns 3.8 ** 3.8 ** 2.4 ** 3.8 ** 2.0 **

Supernatant Pb Initialc 0.06 b 0.06 b 0.06 b 0.15 a 0.13 ab 0.05 b Finalc 0.04 d 0.08 bc 0.07 c 0.09 b 0.12 a 0.07 c Change 0.02 ns 0.02 ns 0.01 ns 0.06 ns 0.01 ns 0.02 ns

Finalc 30.4 e 49.4 d 68.8 b 45.2 c 113.0 a 1.0 f

* and ** significant at P = 0.01 and 0.05, respectively. bSum of total and soluble P in water. bInitial is before P application; final is the mean of 15 successive samples after P application.

Flood-prone rice ecosystem


lime application. That was probably because of lime used in inducing soil reduction, which produced ferrous Fe and neutralized the acidity produced by the oxidation of ferrous Fe. Otherwise the increase in exchangeable Ca and the decrease in Fe-P should have shown an increase in Ca-P, as there was no change in other P forms, especially the Al-P. This explanation is supported by the subsequent increase in Ca-P and soluble Ca together with the consistent decrease in soluble Fe and Al. These changes were greater in limed acid sulfate soil than in leached acid sulfate soil. The highest initial Ca-P in neutral soil is a characteristic of its inherent propertiesneutral pH, high Ca, and low Fe and Al content. Likewise, the lowest initial Ca-P in Fe-toxic soil is also due to its inherent properties: low pH and Ca and high Al and Fe levels. After P applications, different treatments showed different trends of Ca-P. It increased consistently in limed acid sulfate soil, whereas it decreased in both untreated acid sulfate soils. The leached acid sulfate and the neutral soils showed a gradual increase for 5 d and a decrease thereafter, while there was practically no change in Fe-toxic soil. The final mean Ca-P level was highest in neutral soil, followed by limed acid sulfate, leached acid sulfate, untreated acid sulfate with P, untreated acid sulfate without P, and Fe-toxic soil. The increase in Ca-P in limed acid sulfate soil was higher than the total amount of applied P, indicating also the fixation of P released from the organic fraction. This observation was further supported by the increase in Al-P, which accounted more than the decrease in Fe-P. The phosphate fixation in this treatment may have been caused by Ca from lime and from the exchange sites of the soil, and due to P precipitation on the surface of CaCO3 particles. The increase in Ca-P in neutral soil may also be traced to the fixation of P from the organic fraction because of the increase in all P forms, which totaled more than the applied P. This soil contained twice as much exchangeable and soluble Ca as the limed acid sulfate soil, yet it had lesser increase in Ca-P. The Ca-P increase, amounting to 80% of the applied P, and a significant increase in available P, indicate Ca-P dissolution and nonreactiveness of exchange-

able Ca due to the highly buffered nature of brackish water. The decrease in Ca-P in the untreated and leached acid sulfate soils was possibly caused by its dissolution due to increased pCO2 during anaerobic decomposition of organic matter. In these treatments, the sum of applied P and the decrease in CaP was more than the sum of increase in Al-P and available-P, indicating a gain in organic P. Available phosphate. The trends of available P concentrations are shown in Table 6 and Figure 6. Before P application, the leached acid sulfate soil had the highest level, whereas the Fe-toxic soil had the lowest level. In limed acid sulfate soil, it was significantly higher than other treatments but similar to the leached acid sulfate soil. Both untreated acid sulfate soils and the neutral soil had nonsignificant differences. The higher available P in leached acid sulfate soil was due to the enhanced soil reduction leading to a decrease in Fe-P and lowering of Fe and Al concentrations. Available P, as percentage of total P, followed the trend of its initial content in all cases except the Fe-toxic soil, which showed the highest percentage but had the lowest concentration. Available P increased in all P-applied treatments on the first day and gradually declined after the 4th day. The final mean available P values were highest and identical in both the limed and leached acid sulfate soils. This was followed by neutral soil, untreated acid sulfate soil with P, and by untreated acid sulfate soil without P. It was lowest in Fe-toxic soil (7.8 mg kg-1). Comparison of final available P values with initial ones showed an increase of 2.0-3.8 mg kg-1 lowest in Fe-toxic soil and highest in the untreated acid sulfate with P, limed acid sulfate, and the neutral soils. In terms of applied P, the increase in the latter treatments was also largest and identical (38%), whereas in the Fe-toxic soil, it was the lowest (20%). The increase in available P was generally contributed by the applied P and by the decrease in FeP and Ca-P levels in all treatments except limed acid sulfate and neutral soils. In the latter treatments, it was largely contributed by the decrease in organic P fraction, as shown by the higher extent of increase in Al-P and Ca-P than the decrease in Fe-P. Moreo-


IRRI program report for 1998

ver, there was also an increase in total P in these treatments, as noted earlier. Total fixation of P was high in all the soils studied. Phosphorus application without treating the soil is not rational. With regard to the wide variations in the intensity of different P-immobilizing pools in soils, different soils warrant different P management strategies. It is apparent that P fixation cannot be effectively mitigated in an acid sulfate soil by the conventional rates of lime application. Leaching with brackish water seems to be highly effective in improving P availability in this soil as it reduces P fixation through the removal of acidity, Al and Fe, and enhances soil reduction as well as the dissolution of Ca-P. A lower rate of liming is likely to be effective as well as economical, if used after leaching, which could help soil reduction. Leaching followed by liming may also be effective in reducing P fixation in Fe-toxic soil.

Program outlook
In flood-prone environments, germplasm improvement research will emphasize developing MAS techniques that can increase selection efficiency and permit simultaneous selection for several traits to deal with abiotic stresses. Mechanism-based screening techniques for P and Zn uptake will be established and breeding procedures for increasing Fe and Zn in grain will likewise be developed. Research will be continued to determine the relationship between soil properties and flooding tolerance and to assess the environmental impact of land, water, and crop management practices on coastal wetlands. Efforts will be made to project changes in water resources and land use in coastal areas. Strategies for enhanced sustainability of flood-prone rice lands will be formulated and implemented through collaborative partnership with NARS.

Flood-prone rice ecosystem


Research programs Cross-ecosystems research

BIOTECHNOLOGY TOOLS FOR RICE BREEDING 86 Molecular markers 86 Use of conserved motifs of disease resistance genes to characterize elite genetic stocks and germplasm (PBGB, EPP) 86 Physical mapping of bacterial blight resistance gene xa13 (PBGB) 88 Improved resistance of new plant type lines to bacterial blight by marker-aided transfer of resistance genes (PBGB) 88 New models and computer software for mapping QTLs involved in epistasis and GE interactions (PBGB) 90 Physical mapping of the rice genome using IR64 library (PBGB) 91 DNA markers and salinity tolerance in rice (PBGB) 91 Tagging of gall midge resistance genes (PBGB) 93 Wide hybridization 94 Production of interspecific hybrids and backcross progenies (PBGB) 94 Characterization of wide-cross derivatives through in situ hybridization (PBGB) 94 AFLP analysis of Oryza species (PBGB) 95 Rice transformation 96 Effectiveness of PEPC promoter in Bt rice (EPP) 96 Transgenic approaches to tungro virus (EPP) 96 Organization of the International Molecular Breeding Program (IMBP) 97 EXPLOITING BIODIVERSITY FOR SUSTAINABLE PEST MANAGEMENT 97 DNA sequence variation in rice tungro viruses 97 Resistance-breaking mechanism in RTSV-Vt6-TKM6 (EPP) 97 Differential reactions of RTBV variants (EPP) 98 IR64 deletion mutants for functional genomics (EPP) 98 Bacteria with disease antagonism and plant growthpromoting properties (EPP) 100

Heritability of Bt tolerance in the striped stem borer (EPP) 100 Invertebrate biodiversity in a Philippine farmer's field (EPP) 102 Effect of water management on sheath blight spread (EPP) 102 ASSESSING OPPORTUNITIES FOR N2 FIXATION IN RICE 105 Homologues of legume ENOD genes in rice (SWS, PBGB) 105 Characterization of the homologues of legume ENOD93 from rice (SWS, PBGB) 105 Ability of rice to perceive rhizobial nod factors (SWS, PBGB) 107 IMPLEMENTING ECOREGIONAL APPROACHES TO IMPROVE NATURAL RESOURCE MANAGEMENT IN ASIA 109 Characterization of tropical lowland rice production situations and injury profiles 109 Characterization: production situations and injury profiles (EPP) 109 A multiple-pest, production situation-specific model to simulate yield losses of rice in tropical Asia 113 Model structure (EPP) 113 Model calibration and testing (EPP) 114 The Systems Research Network (SysNet) for ecoregional land use planning 116 The SYSNET methodology (APPA, EPP) 116 PROGRESS OF UNREPORTED PROJECTS 119 Rice a way of life for the next generation of rice farmers (APPA, AE, SS, CREMNET) 119 Socioeconomic studies for technology impact, gender, and policy analysis (SS, EPP, APPA) 120 PROGRAM OUTLOOK 120

Cross-ecosystems research

The goal of the Cross-ecosystems Research Program is to acquire and develop knowledge and tools for use by ecosystem-based research programs at IRRI and in national agricultural research systems (NARS), and to address current or anticipated problems common to all ecosystems. The research conducted under the Program cuts across disciplines and ecosystems. The program builds on rapid advances in molecular biology, knowledge of evolving rice pests, alternative approaches to natural resource management for relatively homogeneous agroecoregions, understanding farmers current knowledge and factors affecting the absorption of scientific knowledge, assessing farmers need of technologies in the context of emerging trends in the rice sector of the economy, and evaluating the impact of technologies on the wellbeing of the people. The Program consists of the following six projects: 1. Applying biotechnology to accelerate rice breeding and broaden the rice genepool 2. Exploiting biodiversity for sustainable pest management 3. Biological nitrogen fixation 4. Rice: a way of life for the next generation of rice farmers 5. Socioeconomic studies for technology impact, gender, and policy analysis 6. Implementing ecoregional approaches to improve natural resource management in Asia The Program is the home of the Asian Rice Biotechnology Network (ARBN) and continues to provide technical support to NARS scientists involved in the network on Systems Analysis and Simulations in Rice Production Systems (SysNet).

Biotechnology tools for rice breeding

Biotechnology tools can increase the efficiency and scope of rice breeding and broaden the genepool of valuable traits. The tools include DNA marker technology, DNA fingerprinting, anther culture, wide hybridization, and transformation. Significant advances in the development and use of biotechnology since 1988 have solved difficult problems in rice improvement. Since 1993, the ARBN has promoted collaborative research between NARS and presented training courses in biotechnology techniques and their applications. The application of biotechnology tools at IRRI and at NARS institutes continues, with new approaches and new standards of performance helping to increase the power of the biotechnology approach. Molecular markers

H. Leung, M. Bernardo, L. Ebron, H. Tsunematsu, H. Kato, T. Imbe, L.T. Nghia,28 and V.D. Quang28 A new series of blast-resistant, near-isogenic lines in IR24 background (IR24 NILs) were established for evaluating the performance of individual genes, characterizing pathogen populations, map-based cloning, and construction of gene pyramids. Resistance gene analog (RGA) markers were applied to characterize those elite genetic stocks and to tag resistance genes by cosegregation analysis.


IRRI program report for 1998

Table 1. Blast resistance genes in IR24 NILs available for tagging with resistance gene analog markers. IRRI, 1998. Donor Resistance gene Pi-i Pi-k Pi ta-2 Pi-1 Pi-ta Pi-3 Pi-z5 Pi-9(t) Pi-7(t) Diagnostic fungal isolate PO6-6 PO6-6 IK81-3 PO6-6 IK81-3 PO6-6/PO3-82-51 PO6-6/IK81-3 PO6-6 PO6-6

Fujisaka 5 Kasabue Pi no. 4 C101LAC C101PKT C104PKT C101A51 WHD-IS-75-1-127 RIL 29

We used polymerase chain reaction (PCR) primers corresponding to the leucine-rich repeats (LRR), nucleotide binding sites (NBS), and kinase domains of disease-resistance genes to fingerprint IR24 NILs carrying nine blast resistance genes (Pi) (Table 1). Because the PCR markers correspond to conserved motifs of disease-resistance genes, any unique differences between the lines serve not only as diagnostic markers for IR24 NILs, but represent potential markers linked to specific resistance genes. Those markers can be used for marker-aided selection (MAS) in the development of elite lines and cultivars. DNA was extracted from homozygous resistant (RR) and homozygous susceptible (ss) lines from BC6F3 populations from different cross combinations carrying specific Pi genes. Four primer pairs detecting resistance-gene analogs were used to am-

plify polymorphic DNA fragments of IR24 NILs. DNA amplification was performed using 45 C annealing temperature and amplified products were subjected to electrophoresis using 4.0% polyacrylamide gels. The products were detected using silver staining. The background of all IR24 NILs was nearly identical (>90% similarity in banding patterns) indicating genetic homogeneity of the backcross populations. Diagnostic RGA markers for Pi-3, Pi-i and Pi-7 genes were detected providing unique DNA fingerprints for individual NILs (Fig. 1). RGA polymorphic markers were also detected among resistant and susceptible families in advanced backcross populations. Figure 2 illustrates the cosegregation between a RGA marker produced by NBS primers and Pi-i gene. Two potentially linked markers were also detected in Pi-i and Pi-z5 genes using NBS primers. The markers cosegregating with Pi genes could be candidate resistance genes or part of the gene clusters containing the conserved motifs. Experiments are under way to tag additional resistance genes in IR24 NILs using RGA markers. We are cloning the RGA markers linked to Pi-i and Pi-z5 resistance genes. Sequence-tagged site (STS) primers will be designed from these RGA markers to implement MAS. In addition to gene tagging in NILs, the RGA approach is widely used to assess functional diversity of disease-resistance germplasm. Donor parents and






Pi-9 (t)

IR24 Pi-k


Pi-ta IR24

1. DNA fingerprint of IR24 NILs showing uniform background. A diagnostic band for Pi-3 gene was detected by XLRR RGA primer. IRRI, 1998.

Cross-ecosystems research


Fujisaka 5 IR24



Fujisaka 5 IR24

2. A RGA marker S1/As3 that cosegregates with Pi-i gene. IRRI, 1998.

commercial varieties from Vietnam, Nepal, and central and southern China were fingerprinted using RGA markers (Fig. 3). Our results show that with three representative primers corresponding to the conserved motifs of LRR, NBS, and kinase domains, we can reliably determine the genetic relationships among donor and commercial varieties. This provides an efficient tool to track the pedigrees of elite germplasm and may eventually lead to accurate prediction of functional resistance based on RGA polymorphic markers.

(44F8, 41O2, 12A16, and 12F20), the G1149 contig of two clones 23D11 and 21H18, and the R2027 contig consisting of four overlap clones 42C23, 30B5, 6B7, and 21H14. Genetic mapping indicated that the xa13 locus was contained in the R2027 contig. Chromosomal walking on the R2027 contig resulted in clones 33C7 and 14L3. DNA fingerprinting showed that the six clones of the R2027 contig were overlapping. Clone 14L3 hybridized with a single fragment from the clone 44F8, integrating the R2027 and S14003 contigs into a single contig consisting of 10 BAC clones with a total size of ~330 kb. The physical presence of the xa13 locus in the contig was determined by mapping the ends of the BAC inserts generated by TAIL-PCR. In an F 2 population of 230 plants, the BAC-end markers 42C23R and 6B7F flank the xa13 locus. The probes 21H14F and 21H14R derived from BAC clone 21H14 were found to flank xa13 at a distance of 0.5 cM on either side (Fig. 4). Thus, genetic mapping indicated that the contig and the 96-kb clone 21H14 contained the xa13 locus. Further efforts to isolate xa13, including construction of the xa13 cDNA library, identification of the resistant allele at xa13, subcloning of clone 21H14, and genetic complementation, are under way.

A.C. Sanchez, L. Ilag,29 D. Yang, D.S. Brar, F. Ausubel,30 G.S. Khush, M.Yano,27 T. Sasaki,27 N. Huang, and Z. Li The recessive gene, xa13, confers resistance to Philippine race 6 (PXO 99) of the bacterial blight (BB) pathogen Xanthomonas oryzae pv. oryzae (Xoo). Fine genetic mapping and physical mapping using the map-based cloning approach were conducted as initial steps in an effort to isolate the gene. Using nine selected DNA markers and F2 populations of 132 and 230 plants, xa13 was fine-mapped in a genomic region <4 cM on the long arm of rice chromosome 8, flanked by two restriction fragment length polymorphism (RFLP) markers, RG136 and R2027. Four DNA markers (RG136, R2027, S14003, and G1149) in the target region were used to identify bacterial artificial chromosome (BAC) clones potentially harboring the xa13 locus from a rice BAC library. Eleven BAC were identified, forming four separate contigs including a single clone contig, 29I3 associated with the RG136 STS marker, the S14003 contig consisting of four clones

A.C. Sanchez, D.S. Brar, N. Huang, Z. Li, and G.S. Khush IR65598-112-1 and the two sister lines, IR65600-42 and IR65600-96-28, are promising new plant type (NPT) lines with high yield potential. However, they are susceptible to BB caused by Xoo. To improve the resistance of the NPT lines to Xoo, three BB resistance genes (xa5, xa13, and Xa21) were transferred to them via marker-aided backcrossing. STS markers for the three target resistance genes were developed based on DNA sequences of the target gene (Xa21) or of their linked RFLP markers (RG556 and RG207 for xa5 and RG136 for xa13). Marker polymorphism for xa13 was detected after digestion of the PCR products with HinfI enzyme. Fifty-nine BC3F3 NILs in the three NPT backgrounds containing one to three BB resistance genes in various combinations were developed through


IRRI program report for 1998

40 50 60 70 80 90 100
IR36 CN2 CL-88-66 LC-88-66 K13 Heo ran MTL61 Chiem Chiem rau Bat Nep ran Nep sao vang Nep sao vang Nep da nang MT58 C61MT Song hy Lua thom Heo ran Ba trang CN2 CN2 Nep thom Nep thom DH7 Nep CR203 IR66 MT61 Chiem Nep chiem Nep chiem DC1 Nep ma 25 Cu ba

3. Central Vietnam rice varieties fingerprinted by RGA markers. The dendrogram is generated using GelCompar software. The scale indicates percent similarity between groups. The varietal groupings with different spectra of blast resistance are consistent with the genetic relationships expected from known pedigrees. Duplicated names are accessions. IRRI, 1998.

MAS for the resistance genes and phenotypic selection for the NPT phenotype. While showing a wider resistance spectrum, the BC3F3 NILs having more than one BB resistance gene manifested increased levels of resistance to the Xoo races, compared with those having a single BB resistance gene (Table 2). Our results, based on two F2 populations and the progeny testing of their F3 lines, showed that MAS respectively reached the accuracy of 95% and

95.8% of identifying homozygous resistant plants for xa5 and xa13. These NPT NILs for BB resistance genes provided valuable breeding materials for breeding and genetic study of the relationship between rice plants and Xoo. Our results demonstrated the usefulness and efficiency of MAS in gene pyramiding and the strategy of developing PCRbased polymorphic markers for MAS in rice improvement.

Cross-ecosystems research


R 42C23 (68) 30B5 (50) 6B7 (64) F F 21H14 (96) 33C7 (54) R 42C23R 21H14R 1.2 1.0 0.8 F xa13 - R2027 6B7F 21H14F 3.3 a S14003 14L3 (94) 41O2 (53) 12F20 (51) 12A16 (46) 44F8 (138)





4. Alignment of genetic and physical maps of xa13. The physical map is shown as overlapping BAC clones represented as solid line length equal to the insert size as estimated through pulse field gel electrophoresis. Numbers in parenthesis after clone names are insert size in kb. The genetic map is shown as genetic distance in cM. a) Genetic map derived from the NPT population; b) Genetic map using the NIL population; c) Genetic map from the Nipponbare/Kasalath mapping population. IRRI, 1998.

Table 2. Reactions of NPT BC3F3 lines to six Philippine races of bacterial blight pathogen, Xanthomonas oryzae pv. oryzae. IRRI, 1998. Recurrent parent Target gene Race 1 PXO 611 R S R R R R R R S R R R R R Race 2 PXO 85 R S R MR R R R R S R R R R R Race 3 PXO 79 R S R MR R R R R S R R R R R Race 4 PXO 71 MR S R R R R R MR S R R R R R Race 5 PXO 112 R S R MR R R R MR S R R R R R Race 6 PXO 99 S R R R R R R S R R MR R R R

IR65598-112-1 -2 -3 -4 -5 -7 -10 IR65600-96-28 -17 -18 -19 -20 -22 -24

xa5 xa13 Xa21 xa5/xa13 xa5/Xa21 xa13/Xa21 xa5/xa13/Xa21 xa5 xa13 Xa21 xa5/xa13 xa5/Xa21 xa13/Xa21 xa5/xa13/Xa21


D.L. Wang,3 J. Zhu,3 Z. Li, and A.H. Paterson24 A new methodology based on mixed linear models was developed for mapping quantitative trait loci (QTL) with digenic epistasis and QTL environment (QE) interactions. A two-dimensional search and the control of background genetic variation were applied in the mixed model approaches. Reliable estimates of QTL effects (additive effects and

epistasis effects) can be obtained by maximum likelihood estimation, while QE interaction effects (additive environment interactions and epistasis environment interactions) can be predicted by the best linear unbiased prediction method. Likelihood ratio and t statistics were combined for testing hypotheses about QTL effects and QE interactions. Monte Carlo simulations were conducted for evaluating the unbiasedness, accuracy, and power for parameter estimation in QTL mapping. The results indicated that the mixed model approaches could provide unbiased estimates for positions and effects of QTLs,


IRRI program report for 1998

as well as unbiased predicted values for QE interactions. The mixed model approach also showed high accuracy and power in mapping QTLs with epistatic effects and QE interactions. Based on the models and methodology, computer software (QTLMapper version 1.0) was developed, which is suitable for interval mapping QTLs with epistatic effects and QE interactions in DH, RI, and BC populations. This work was accomplished through collaboration with Zhejiang University, China, and Texas A&M University, USA.

A.C. Sanchez, B. Fu, J. Domingo, J. Talag, G. Posa, D. Yang, R. Maghirang, L.T. Nguyen, D.S. Brar, J. Bennett, G.S. Khush, and Z. Li Construction of a physical map of the rice genome using the IR64 BAC library is in progress. We established the backbone of the IR64 physical map with 378 anchor islands (contigs) consisting of 1,100 BAC clones identified by anchor DNA markers. We also established 350 random BAC islands consisting of more than 1000 BAC clones. The current physical contigs covered about 20% of the rice genome with more than 50% coverage for chromosome 11 and more than 30% for chromosomes 12 and 3.

G. Gregorio, B. Ghareyazie, Z. Li, J. Bennett, and D. Senadhira Marker-aided selection development for salinity tolerance in rice. MAS for salinity tolerance would accelerate breeding progress by increasing selection efficiency. Tagging these genes for salinity tolerance is the first step in developing the PCR-based molecular markers. Amplified fragment length polymorphism (AFLP) mapping generated rapid, numerous, reproducible, and dominant markers. Although an F8 recombinant inbred population (IR66946) developed from indica parents (IR29, a salt-sensitive improved variety/ Pokkali, a tolerant traditional variety) was used, 206 AFLP markers were generated using 32 primer combinations. The AFLP markers were able to construct a linkage map of all 12 chromosomes and were used in tagging salinity tolerance.

Salinity tolerance gene tagged. With the use of the phenotypic marker for salinity tolerance (saltol), marker analysis mapped the gene between flanking AFLP markers P3/M9-8 and P1/M9-3. Distances were respectively 14.7 cM and 18.6 cM. The saltol gene was in chromosome 1 with LOD 5.7. The location of the saltol gene in this study may be the same as that of the SalT gene in the Cornell map, which contains an open reading frame coding for a protein of 140 amino acid residues. SalT mRNA accumulates rapidly in sheaths and roots of mature plants and seedlings after treatment with Murashige and Skoog salts. The organ-specific response of salt is correlated with the pattern of Na accumulation during salt stress. QTLs for salinity tolerance mapped. A quantitative measure of the Na and K concentrations, and Na-K ratio absorption in the shoot at seedling stage in salinized conditions, was taken from each line in the population. Marker analysis was performed only on RILs in the extreme tails of the distribution, i.e., those with the lowest and highest value for Na and K concentrations, and Na-K ratio. Any allele frequency at any AFLP marker locus differing significantly at P<0.001 between the two extreme subpopulations was inferred as the location of a salinity tolerance QTL. Based on interval analysis where sets of linked markers were analyzed simultaneously with regard to their effect on the quantitative traits (Na, K, and Na-K ratio), three QTLs were detected for K absorption, four for Na absorption, and three for NaK ratio. These include the major gene saltol in chromosome 1. The QTLs detected for K absorption, in addition to the major gene in chromosome 1, were located in chromosomes 4 and 12 with 83.5% and 21.2% of the total phenotypic variation explained (Table 3). For Na absorption, QTLs were detected in chromosomes 1 and 10 and two in chromosome 3,. and respectively explained 64.6%, 35.6%, 17.1%, and 16.0% of the phenotypic variation (Table 3). For Na-K ratio, the three QTLs were located in chromosomes 1, 10, and 12 which respectively explained phenotypic variations of 64.3%, 86.1%, and 18.5%. A common QTL was observed in chromosome 1 for the three quantitative traits (low Na, high K, and low Na-K ratio) putatively associated with salinity tolerance. This segment in chromosome 1 contained the saltol gene which explained a range of 64.3-

Cross-ecosystems research


Table 3. QTLs identified by interval analysis for K and Na absorption and Na-K ratio in the shoot. IRRI, 1998. Index Marker interval K absorption P3/M9-8 - Saltol RG375 - P4/M3-2 P1/M1-3 - P2/M1-3 Na absorption P3/M9-8 - Saltol P1/M5-3 - P3/M9-1 P1/M10-6 - P1/m7-10 P1/M3-10 - P1/M3-8 Na-K ratio P3/M9-8 - Saltol G291 - P1/M7-8 P1/M1-3 - P2/M1-3 Chromosome no. LOD

195-209 206-200 3 - 93

1 4 12

17.2 5.3 3.4

195-209 39 -188 88 -67 27 -25

1 3 3 10

14.5 3.2 3.0 4.0

195-209 207-65 3 - 93

1 10 12

14.5 3.6 3.1

80.2% of the total phenotypic variation with LOD >14.5. A common QTL was also detected on chromosome 12 for traits Na and Na-K ratio, which respectively explained 21.2% and 18.5% of the phenotypic variation and LOD of 3.46 and 3.14. Fine mapping the salinity tolerance gene. Fine mapping involves the saturation of the chromosome region with molecular markers where the salinity tolerance gene was detected and the use of larger population size to approximate the true linkage value of the molecular markers and the gene. The 275 F8 RILs were phenotyped for salinity tolerance in a controlled environment in the IRRI phytotron. The phenotyping was done by two separate experiments. First was the visual reaction of the RILs to salinity using the modified standard evaluation system, and the other was the absorption of Na and K and Na-K absorption ratio in the shoot. The mechanism for salinity tolerance in rice is the ability to absorb less Na+ and take up more K+ to maintain a good Na-K balance in the shoot. The screening tests were maintained at 29/21 C day/night temperature, a minimum relative humidity of 70% during the day and natural daylight. Final phenotyping will be in a naturally saline field in Iloilo Province, Philippines. STS markers near the salinity tolerance gene were used for a parental survey. Of the 32 STS markers surveyed in chromosome 1, only 7 were polymorphic even after 20 restriction enzyme digestions. Genotyping the 275 F8 RIL populations with polymorphic STS markers will be continued and microsatellite markers will also be used to saturate the region containing the salinity tolerance gene.

Salt tolerance in Pokkali. The Indian land race Pokkali has been frequently used as a donor of salinity tolerance in rice breeding. The salt tolerance of Pokkali is related to its capacity to maintain a high K-Na ratio in the shoot. The factors modulating this ratio are not understood at the physiological, cellular, or molecular levels. The recent discovery of a wheat gene, hkt1, encoding a high-affinity K transporter, provided an entry point into the molecular study of salt tolerance in Pokkali. This approach was given additional impetus when a mutant form of wheat hkt1 was reported to confer salinity tolerance in yeast. We found that marker RZ405 on chromosome 6 of the rice genetic map had high sequence homology with wheat hkt1. We used PCR to amplify part of the hkt1 locus from both Pokkali and IR29, a salt-sensitive cultivar. When the two amplicons proved to be of the same size, we used more than 30 randomly selected restriction endonucleases in an unsuccessful attempt to find a polymorphism between amplicons. After the approach of randomly choosing restriction endonucleases failed, we isolated the entire hkt1 gene from an IR36 genomic library and used the sequence information to amplify the entire gene from Pokkali and IR29. We found a small number of single-base differences between the Pokkali and IR29 alleles, including one that was detectable through use of the restriction endonuclease NdeI. Digestion of the IR29 and Pokkali amplicons confirmed that NdeI was able to detect a polymorphism between the alleles. We then used this enzyme to examine the segregation of the IR29 and Pokkali alleles in an F8 mapping population derived from IR29/Pokkali. We examined 40 lines (from a population of 277 RILs) that were at least as saltsensitive as IR29 and 40 lines that were at least as salt-tolerant as Pokkali. All 80 RILs contained only the Pokkali allele, a total bias against the IR29 allele. No DNA marker at any other locus showed comparable bias. Other DNA markers on chromosome 6 showed quite balanced distributions of IR29 and Pokkali alleles among the 80 lines. Furthermore, the waxy locus on chromosome 6 showed unbiased segregation. We concluded that the Pokkali allele at the hkt1 locus makes little or no contribution to the salt tolerance of Pokkali. However, it competes so strong-


IRRI program report for 1998

ly against the IR29 allele at some stage in plant growth that no copy of the IR29 allele survives. All but one of the point mutations distinguishing the hkt1 alleles of IR29 and Pokkali appear to be trivial. The exception is the mutation that created or destroyed the NdeI site mentioned above. The mutation is actually a frame-shift mutation that is predicted to alter the last 60 of the ~510 amino acids of the transporter protein. The IR29 protein is predicted to contain 30 amino acids fewer than the Pokkali allele as a result of early translational termination. Such a mutation may not be lethal in itself and could survive in the homozygous state (as in IR29) but it may be sufficiently deleterious that it could be at a severe disadvantage relative to the Pokkali allele. However, we cannot eliminate the possibility that the loss of the IR29 allele of hkt1 is not due to the locus itself but to a closely linked locus. We examined the pedigree of IR29 and asked IRRIs Genetic Resources Center for stored seeds. PCR and NdeI digestion traced the abnormal allele back to Peta, one of the most commonly used cultivars in IRRIs breeding program. The absence of the Peta allele from most of the IR cultivars derived from Peta confirms that the allele is associated with a very unfavorable phenotype.

S.K. Katiyar,31 B.C. Huang,32 and J. Bennett The Asian rice gall midge Orseolia oryzae is the most important dipteran pest of rice. It is most commonly found in eastern, central, and southern India; in the wet zone of Sri Lanka; and in southern China. Breeding for resistance to gall midge has been hindered by the existence of many biotypes of the insect. IRRI and several collaborating NARS insti-

tutes in Asia have used molecular tools to 1) tag several of the resistance genes for use in DNA MAS and 2) provide insight into the biodiversity of the insect. Tagging is accomplished by bulked segregant analysis, usually of the F3 or F4 generation of a cross between a gall-midge-resistant cultivar and a sensitive cultivar. Wherever possible, the sensitive cultivar is an elite cultivar into which gall midge resistance must be introgressed. The F3 or F4 population is phenotyped for gall midge resistance in the greenhouse or field and at least 40 lines that are true-breeding for resistance or true-breeding for susceptibility are identified. DNA is extracted from the 80 lines and from the two parents for amplification by random amplified polymorphic DNA (RAPD) analysis. RAPD bands that are found in the resistant parent and the resistant lines but are absent from the susceptible parent and the susceptible lines are noted, as are bands with the opposite pattern of appearance. These bands are cloned and mapped, and either used as markers for the resistance genes or as pointers to the location of even closer markers. This approach was used successfully to tag the Gm2, gm3, Gm4, Gm5, and Gm6 genes. MAS kits were developed for Gm2 and Gm6 and were under development for gm3, Gm4, and Gm5 (Table 4). Assam land race ARC5984 contains the Gm5 gene that is effective in India and probably another gene (Gm?) that is effective in China. The mapping population used to map Gm5 will be used in 1999 to map the Gm? gene. Because gall midge resistance is occasionally overcome by changes in the gall midge population, the goal is to pyramid two or more resistance genes in key elite cultivars. In southern China, the goal is to move both Gm6 and Gm? into hybrid rice by per-

Table 4. Use of marker-aided selection (MAS) for rapid generation of near-isogenic lines for gall midge resistance in elite backgrounds. IRRI, 1998. Gene Gm1 Gm2 gm3 Gm4 Gm5 Gm6 Gm? Donor W1293 Phalguna RP2068 Abhaya ARC5984 Duokang #1 IR36 Status of MAS Mapping population Gene mapped Gene tagged Gene mapped Gene tagged Gene mapped Tagging in progress Targeted cultivars (recurrent parents) Kranti, Swarna in India Kranti, Swarna in India Kranti, Swarna in India Kranti, Swarna in India Restorers for hybrid rice in China Restorers for hybrid rice in China Sri Lankan cultivars

Cross-ecosystems research


forming MAS on key restorer or maintainer lines. In India, the goal is to pyramid Gm2 and Gm4, and later gm3 and Gm5, into elite inbred lines of the rainfed lowlands of eastern and central states and hybrid rice grown in southern states. Gm6 is effective against all four gall midge biotypes known in China but is ineffective against any of the biotypes of India, Nepal, and Sri Lanka. This result is consistent with the result of DNA fingerprinting of gall midge isolates in those countries. DNA fingerprinting, accomplished by AFLP analysis, revealed that the gall midge of China, Laos, and the northeast Indian state of Manipur are closely related among themselves but only distantly related to gall midge of the rest of the Indian subcontinent, which are closely related among themselves. There appears to be two species or subspecies of gall midge in Asia, instead of one as previously believed. This information is of crucial importance in planning where new resistance genes should be sought and where they should be deployed. Wide hybridization

ile and set only a few seeds upon selfing. Backcrossing subsequently produced more than 3,000 backcross (BC2/BC3) progenies. Those progenies will be evaluated at WARDA and IRRI for different target traits and at IRRI for the possible introgression of genes from O. glaberrima into O. sativa.

F. Abbasi, M. Ashgar, J. Talag, D. S. Brar, K. Fukui,34 and Z. Li A series of hybrids from crosses of cultivated rice and various wild species, representing BBCC, CC, CCDD, EE, FF, GG, and HHJJ genomes, have been produced. In addition, monosomic alien addition lines (MAALs) and introgression lines derived from various crosses involving cultivated and wild species are also available. We used genomic in situ hybridization (GISH) to characterize parental chromosomes in the wide-cross derivatives involving crosses of O. sativa with four wild speciesO. officinalis (CC), O. australiensis (EE), O. brachyantha (FF), and O. ridleyi (HHJJ). Total genomic DNA of the wild species was used as a probe for in situ hybridization with the somatic chromosome preparations of wide-cross derivatives. Characterization of O. sativa/O. brachyantha derivatives. Total genomic DNA of O. brachyantha was digested with EcoRI and labeled with biotin-14dATP. This DNA was used as a probe in hybridization with the somatic chromosome propagations of F1/BC1 and MAALs. The enzyme maceration method was used to obtain well-spread somatic chromosomes for in situ hybridization. In the F 1 hybrid, all the 12 chromosomes of O. brachyantha could be identified from sativa chromosomes. The hybridization signal appeared dark brown due to probe hybridization and unlabeled sativa chromosomes appeared light blue. In a similar experiment, MAALs (MAAL 6 and MAAL 12) derived from O. brachyantha were used for in situ hybridization. The extra chromosome of O. brachyantha showed dark brown hybridization signal, making it clearly distinguishable from the remaining 24 chromosomes of O. sativa. The technique characterized introgression of alien chromosome segments from brachyantha into the sativa genome.

D.S. Brar, R. McNally, J. Molinawe, G.S. Khush, and M. Jones33 An IRRI-West Africa Rice Development Association (WARDA) collaborative project, Interspecific Hybridization between African and Asian Rice Species, seeks to 1) transfer resistance to African rice gall midge (ARGM) and rice yellow mottle virus (RYMV) and tolerance for abiotic stresses and weed competitiveness from O. glaberrima to highyielding indica cultivars of O. sativa. Seeds of 18 accessions of O. glaberrima were received from WARDA. Some of those had been identified as resistant to ARGM and RYMV and as possessing weed competitiveness. Preliminary screening at IRRI showed a few accessions having increased tolerance for nematodes (Meloidogyne graminicola), Fe toxicity, Al toxicity, and P deficiency. Hybrids were produced with IR64 and BG90-2 and O. glaberrima accessions TOG 5674, 5675, 5681, 5697, 5860, 6216, 6472, 6508, 6589, 6629, 6631, 7235 and 7291. The hybrids were highly sterile and were backcrossed using BG90-2 and IR64 as recurrent parents. The BC1F1 were also highly ster94 IRRI program report for 1998

Characterization of O. sativa/O. australiensis derivatives. The genomic DNA of O. australiensis was used as a probe in in situ hybridization to characterize the parental chromosomes of the plant having 2n = 27 chromosomes derived from anther culture of the F1 hybrid (O. sativa/O. australiensis). The probe produced uniform labeling pattern over the entire length of all 14 O. australiensis chromosomes (bluish-green), whereas 13 chromosomes of sativa appeared light blue after Giemsa staining. The O. australiensis chromosomes showing the hybridization signal appeared bluish-green under blue light excitation allowing the identification of all O. australiensis chromosomes, whereas the O. sativa chromosomes appeared blue due to counterstaining with 4-6-diamidino-2-phenylindole (DAPI). The same cell showed green fluorescence of fluorescein isothiocyanate (FITC) only on the 14 chromosomes of O. australiensis. Characterization of O. sativa/O. officinalis derivatives. We also used fluorescent in situ hybridization (FISH) to identify parental chromosomes and to locate rDNA loci in the F1 hybrid of an elite breeding line from O. sativa/O. officinalis (accession 100896). Total genomic DNA of O. officinalis and 45S rDNA used as probe were labeled by random primer labeling technique with biotin-16dUTP and digoxigenin-11-dUTP. Posthybridization involved stringent washing of the slides with 2X SSC at 42 C and 0.1X SSC at 60 C. Following FISH, and using total genomic DNA of O. officinalis as probe, parental chromosomes of O. officinalis could be distinguished in the F1 hybrid from the sativa chromosomes. The bluish-green FITC signal appeared on O. officinalis chromo-

somes when the hybrid cell was counterstained by DAPI. This was depicted more clearly when the same cell was exposed to blue light and only O. officinalis chromosomes were seen as green. Hybridization with 45S rDNA probe showed reddish fluorescent signal of rhodamine on two chromosomes of O. officinalis and on one chromosome of O. sativa.

R.K. Aggarwal,35 D.S. Brar, J. Talag, S. Nandi, N. Huang, and G.S. Khush We used AFLP markers to determine phylogenetic relationships among Oryza species. Seventy-seven accessions of 23 Oryza species, five related genera, and three outgroup taxa were included in the analysis. AFLP data were analyzed to study species relationships using different clustering algorithms, and the resulting phenograms were tested for stability and robustness. AFLP analysis revealed a large number of distinct, scoreable fragments per primer pair. A total of 1,191 polymorphic markers were obtained using five AFLP primer combinations (Table 5). The level of polymorphism was much lower within species (~2% in O. minuta to 21% in O. officinalis) and also between species carrying similar genome(or genomes) (~20% for HHJJ genome to 35% for BB/BBCC genome species). Table 5 shows the distribution of AFLP markers for different genomes of Oryza, related genera, and outgroup taxa. In general, more markers were obtained for allotetraploid Oryza species (av 46) than for their diploid relatives (av 33).

Table 5. Distribution of AFLP markers detected for different primer pairs, across genomes of Oryza, related genera, and outgroup taxa. IRRI, 1998. Primer Total Average number of markers/ pair polyindividual over used morphic markers Diploid Tetraploid scored Oryza Oryza Related Outgroup species species genera taxa P1/M1 P1/M2 P1/M3 P1/M4 P2/M5 Av 234 260 218 255 224 238 28 39 32 39 28 33 166 40 52 48 52 39 46 231 33 43 39 47 34 39 196 37 49 39 44 40 42 209 Average number of markers/individual/ genome in Oryza AA BB CC EE FF GG BBCC CCDD HHJJ

22 36 33 37 29 31 157

22 37 36 37 22 31 154

31 38 34 36 26 33 165

32 37 32 40 33 35 174

25 39 18 40 26 30 148

36 45 40 44 34 40 199

39 50 53 52 38 46 232

40 51 48 47 42 46

41 56 43 57 37 47

Total 1191

228 234

Cross-ecosystems research


Neis genetic distances (D values) showed a linear increase from within species to between species and were most striking between different genomes. Within species, distances ranged from 0.024 (O. meyeriana) to 0.213 (O. officinalis) with the exception of two accessions of O. brachyantha, which showed a much higher value (0.301). The average D value between species carrying similar genome(s) was ~2.8 times (0.293) greater than within species (0.104), while between genomes of Oryza, it was 0.737. The results of AFLP analysis suggest common ancestry to the genus Oryza and indicate that q evolution in Oryza has followed a polyphyletic path wherein multiple lineages underwent independent divergence after separation early in the evolution from a common ancestor or pool of related taxa; q newly assigned genomes, GG for O. meyeriana and HHJJ for O. ridleyi complexes, are among the most diverged in the genus; q CCDD tetraploids are of relatively ancient origin among the officinalis complex; q O. malampuzhaensis, O. indandamanica, O. alta, and O. grandiglumis are diverged enough to deserve species status; and q O. brachyantha is the most diverged species in the genus. Rice transformation

for the Cry1Ab toxin from Bacillus thuringiensis (Bt) under control of the maize PEPC promoter. One advantage of the PEPC promoter is that it is active only in green tissues and thus its use will result in only limited accumulation of Bt toxin in the rice grain. The resistance of cry1Ab-transformed and control plants was tested at vegetative and flowering stages with neonate and 10-d-old larvae of two stem borer species, Chilo suppressalis and Scirpophaga incertulas. Vegetative-stage cry1Ab-transformed plants were highly resistant to neonate larvae of both species but the cry1Ab-transformed and control plants did not differ in resistance to 10-d-old larvae (Table 6). At flowering stage, the survival of neonate S. incertulas larvae, but not 10-d-old larvae, was significantly lower on the cry1Ab plants than on control plants. Survival of C. suppressalis larvae of both age classes did not differ significantly on cry1Ab-transformed and control plants. Our results suggest that Bt toxin genes under control of the PEPC promoter alone will not provide satisfactory control of rice stem borers.

O. Azzam, E.L. Coloquio, Z. Flores, A. Klti,37 J. Ftterer,37 and I. Potrykus Tungro is a complex disease associated with rice tungro spherical virus (RTSV) and rice tungro bacilliform virus (RTBV). A green leafhopper (Nephotettix virescens) transmits and spreads tungro throughout South and Southeast Asia. IRRIs strategy to breed for a durable resistance to tungro has encompassed incorporating vector and virus resistance genes by conventional breeding approaches. Because no natural resistance to RTBV

F. Alinia,36 B. Ghareyazie, L. Rubia, J. Bennett, and M. Cohen We transformed an Iranian aromatic variety, Tarom Molaii, with a construct containing a synthetic gene

Table 6. Percent larval survival of two stem borer species at 7 d after infestation on cry1Ab-transformed and control plants. IRRI, 1998. Vegetative stage Species Larval age cry1Aba C. suppressalis Neonate 10-d-old Difference Neonate 10-d-old Difference 1.25 + 1.25 62.50 + 11.09 61.25 ** 1.25 + 1.25 22.50 + 7.50 21.25 ** Control 61.25 + 13.44 75.00 + 9.57 13.75 25.00 + 5.40 35.00 + 5.00 10.00 Difference cry1Aba Control 13.75 + 4.30 77.50 + 4.79 63.75 ** 5.00 + 2.04 40.00 + 8.16 35.00 ** Difference 5.00 5.00 Flowering stage

60.00 ** 18.75 + 10.87 12.25 82.50 + 4.79 63.75 ** 23.75 ** 17.50 + 1.44 12.50 47.50 + 7.50 30.00 *

S. incertulas

12.50 * 7.50

Mean SE of percentage live larvae of 20 larvae plant-1. n = 4 plants.


IRRI program report for 1998

has yet been transferred to commercial rice varieties, work was initiated to induce RTBV resistance in rice using several antiviral transgenic strategies. The strategies included expression of RTBV open reading frame 1 (ORF1), putative viral coat protein, viral replicase, mutated fragments of the replicase, and production of different parts (sense and antisense) of ORF4. We used particle bombardment of the rice scutella and rice suspension cells to produce fertile trangenic TP309 and Kinuhikari plants (japonica rice varieties). Two sources of virus inocula were used to evaluate R1 progeny resistance to RTBV and RTSV. Nineteen transgenic lines with complete copies and 32 lines with rearranged copies of the transgenes were evaluated. Using either the greenhouse or Famy virus populations, none of the lines tested showed resistance to RTBV or RTSV. Almost all the inoculated plants exhibited severe symptoms (stunted growth and leaf discoloration) 20 d post-inoculation (dpi) and viral coat protein titers, as measured by the enzymelinked immunosorbent assay (ELISA), were comparable with those from the nontransgenic control plants. Titers varied among individual plants from different lines. Some had high-RTSV and -RTBV5 titers while others had low-RTSV but high-RTBV titers. Based on the ELISA results and visual scores, none of the tested lines recovered at 40 dpi. The average symptom severity was about 7 per line, indicating that most plants exhibited stunted growth and leaf discoloration. The 32 remaining lines, which mostly had antisense RNA constructs of ORF4, were tested using only the greenhouse virus population and results were similar to those obtained earlier. None of the lines showed resistance to RTBV 20 dpi and plants did not recover after 40 dpi. Although no RTBV-resistant lines were found, several new antiviral strategies were identified. These will enhance the low expression of transgene that was observed throughout these experiments and should induce a resistance response. Organization of the International Molecular Breeding Program (IMBP) Rice breeding programs or biotechnology laboratories, or both, of 30 institutions from 14 Asian countries plus Egypt in Africa have shown interest in participating in IMBP. Fifty-four scientists (12 ge-

neticists and 42 breeders) were nominated by the participating institutions to coordinate and execute the proposed activities. A total of 177 diverse rice lines were recommended as the materials for the breeding efforts. Of those, 75 elite rice lines are currently grown on more than 60% of the rice lands in Asia and will be used as the core gene pool lines. The rest of the accessions, possessing unique alleles for specific target traits, will be used as donor gene pool lines for introgression and identification of desirable QTLs. These lines (except those from India) were planted in an IRRI screenhouse for seed increase. Seed will be sent to all participating NARS in May 1999. Matching funds were provided by China and Thailand to support the research activities in the first phase of the project.

Exploiting biodiversity for sustainable pest management

Rice pests reduce yields and yield stability. However, some pest management practices, such as chemical insecticides, are harmful to the environment. Other practices, such as host plant resistance based on single genes, may contribute to additional yield instability. The effectiveness of rice pest management can be improved by increasing bio-diversity. Research in this project extends from molecular genetics to landscape ecology. DNA sequence variation in rice tungro viruses O. Azzam, P.Q. Cabauatan, M. Arboleda, I. Uyeda,19 M. Isogai,19 T. Omura,20 H. Hibino,20 H. Koganezawa,21 and U. Melcher22 Several biological variants of RTSV and RTBV have been identified at IRRI since 1993 and characterized based on their reaction to differential hosts.

The maintained variant for RSTV, RTSV-A, differs from RTSV-Vt6 in virulence on TKM6 rice variety. To understand the resistance-breaking mechanism in the RTSV-Vt6-TKM6 system, and to identify genes that can confer a more broad and durable resistance to the virus, RTSV-Vt6 was characterized

Cross-ecosystems research


molecularly. Extracted viral RNA was amplified by reverse transcriptase polymerase chain reaction (RT-PCR). The PCR products were then cloned and the complete nucleotide sequence was determined from multiple clones that covered the whole genome. The full-length sequence of 12,171 nucleotides was compared with the published sequence of RTSV-A. The overall identity between the two strains was respectively 90% and 95% at the nucleotide and amino acid levels. The many sequence differences were scattered over the entire genome and no potential hypervirulent region was identified. A full-length infectious cDNA of RTSVA or RTSV-Vt6 will be needed to identify proteins involved in the virulence of strain RTSV-Vt6.

Three biological variants for RTBV, G1, G2, and Ic, which originated from the same IRRI greenhouse virus isolate, were cloned and sequenced. These variants showed differential reactions on rice varieties FK135 and TN1. The genome sizes of the three variants differed but were respectively 95% and 99% identical at the nucleotide and amino acid levels. The analysis of restriction endonuclease maps for the RTBV genomes identified EcoRV, PstI, and EcoRI as potential enzyme markers to differentiate between the Ic-G1 pair and the other isolates. Alignment of these sequences with published RTBV sequences (Phi-1, Phi-2, and Phi-3), which originated from the same IRRI greenhouse isolate, and with a recently published Malaysian sequence (Serdang), showed that the cysteine-rich region of ORF3 is the most variable region (Fig. 5). This region contains a cysteine-histidine motif that is conserved among retroelements and is thought to bind RNA during packaging of the pregenomic RNA to separate it from cellular RNAs as the template for reverse transcription. The observed changes in amino acid sequences were conserved in the G1, Ic, and Serdang isolates as compared with the other isolates. This suggests that the cysteine-rich region may affect the virus infection cycle and the phenotypic differences among the variants.

5. Comparison of the predicted amino acid residues no. 842-948 in the cysteine-rich region of the seven RTBV isolates. Phi-1 (EMBL Acc. no. X57924), Phi-2 (EMBL Acc. no. M65026), Phi-3 (EMBL Acc. no. D10774), Serdang (EMBL Acc. no 076470). Amino acid substitutions are underlined and in bold letters. IRRI, 1998.

IR64 deletion mutants for functional genomics H. Leung, C. Wu, M. Baraoidan, and G. Khush Rice, with the smallest genome of all cereals, is the focus of an international sequencing effort and is likely to be completely sequenced within a decade. The DNA sequences represent an enormous pool of markers and genes for rice improvement through MAS or transformation. A full exploitation of this wealth of information will not be possible, however, until the biological functions encoded by the sequenced DNA are understood. We began a systematic production of mutants with the goal of assigning available sequenced DNA to biological functions revealed by the mutations. Focus is on the production of deletion lines, which will allow physical detection of mutations in the genome using available expressed sequence tags (ESTs) and other cloned DNA sequences. We reason that the nontransgenic deletion mutants will meet needs not met by transgenic insertion mutants


IRRI program report for 1998

and will accelerate trait discovery in rice through the use of advancing genomic technologies (Fig. 6). We concentrate on the production of mutants in IR64, which carries many valuable agronomic traits and tolerance for biotic and abiotic stresses. Thus, creating mutations in IR64 can facilitate detection of phenotypic changes in important agronomic traits. Diepoxybutane (DEB, 0.004-0.006%) and fast neutrons (FN, 33 Gy) were used to generate more than 30,000 M3 lines with potential deletions and

chromosomal rearrangements. A gamma ray (GR, 250 Gy)-treated population was also produced to compare the spectrum and molecular characteristics of the mutations produced by different mutagenic treatments. Chlorotic or albino mutants were observed respectively in 8 and 10% of M2 lines in the DEB and GR populations. More than 3,200 M2 lines were screened with four diverse blast fungus isolates and susceptible plants were detected in 0.4% of the M2 populations. Disease-lesion-mimic mutants were detected at 0.1% in the DEB population.


DNA sequences from rice and other cereals

Rice mutant bank


Assign DNA to mutant phenotypes

Screen for agronomic traits: gain or loss of functions

Discover new traits

Identify new genes Evaluate allelic diversity and functions

Improved rice productivity with NARS partners: breeding, marker-aided selection, transformation
6. Deletion mutant bank in the framework of functional genomics. Initial focus is on using candidate genes in the disease defense pathways to characterize the mutants.

Cross-ecosystems research


The phenotypes observed in M2 indicate a rich source of mutations present in the IR64 collection. These mutant lines are being mass increased in the M3 generation for screening of a variety of traits. To develop a system for reverse genetics, candidate disease defense genes are being used to survey a panel of DNA from individual deletion lines selected for altered response to BB and blast. The empirical mutation rate (about 0.1%) observed in rice mutants induced by chemicals or ionizing irradiation suggests that each mutant probably harbors 10-20 mutated sites. The goal is to produce 40,000 independent deletion lines to give a 95% chance of detecting a mutation in most genes (except homozygous lethals). We expect to advance all the DEB mutant lines to M3 (about 19,000 lines) by early 1999. Additional disease-susceptible and disease-resistant mutants will be identified through the ongoing screening for gain and loss of disease resistance. Bacteria with disease antagonism and plant growth-promoting properties Z. Chen, R. Pamplona, E. Regalado, B. Cottyn, and T.W. Mew Bacteria were isolated from seeds and blast lesions collected from farmers fields in Iloilo and Rizal provinces, Philippines, and plated on various media. Morphologically distinct bacterial colonies were selected and studied for cell morphology, pathogenicity, antagonism to the pathogens Rhizoctonia solani and Fusarium moniliforme, plant growthpromoting ability, and genetic diversity. The 63 Bacillus spp. isolates from blast lesions showed a higher proportion of antagonism and stronger antagonism than the 32 Gram negative isolates from the seed washes. None of the isolates from blast lesions had plant growth-promoting properties. Eight isolates from the seed washes promoted seedling growth and four isolates promoted taproot growth. Some isolates from the seed washes reduced seedling growth. Five of the eight growth-promoting isolates from the seed washes also showed antagonism to pathogens, indicating that some Gram negative bacteria on seeds have multiple beneficial activities. Forty-one bacterial isolates, 18 positive and 23 negative for antagonism by the dual culture test, were selected for generating DNA markers diagnos-

tic for antagonistic isolates of Bacillus amyloliquefaciens. Based on their RAPD fingerprints, those isolates were found to cluster at 65-99% similarity with 8 isolates that were previously identified as B. amyloliquefaciens by the fatty acid methyl ester (FAME) method. In a preliminary test, 124 RAPD primers were screened for their ability to differentiate four isolates that included 2 antagonistic and 2 nonantagonistic B. amyloliquefaciens. Four primers (AM-18, AM-04, AA-01, AA-09) amplified fragments that were specific to the antagonistic isolates. Further testing with 37 isolates (16 antagonistic and 21 nonantagonistic) showed that AM-18 was the most effective in differentiating the isolates. AM-18 amplified an approximately 1,100 bp band with all antagonistic isolates except G327, G458, and G460, and was absent from the nonantagonistic isolates (Fig.7). These results indicate the usefulness of RAPD primers in identifying antagonistic strains of B. amyloliquefaciens. Heritability of Bt tolerance in the striped stem borer F. Alinia, F. Gould,23 and M.B. Cohen It is important to anticipate the development of resistance to new insecticides and useful to estimate the rate at which insect resistance can develop. Many groups are working to produce rice transformed with insecticidal toxin genes from Bacillus thuringiensis (Bt) for improved protection against stem borers. It has already been demonstrated that insect pests of several crops can evolve resistance to Bt toxins. We examined the heritability of tolerance for a Bt toxin in a population of the striped stem borer, Chilo suppressalis (Lepidoptera: Pyralidae). We estimated the narrow-sense heritability of tolerance for a Bt toxin, Cry1Ab, with a half-sib design in which each of 20 striped stem borer males was mated to two females. Two-thirds of the progeny of each cross were tested for survival on an artificial diet containing 0.03% Cry1Ab toxin by weight, and one-third for survival on a control diet. Mortality was scored 5 d after diet infestation. Heritability was calculated using standard formulas with the assumption that mortality was a threshold character with tolerance for Cry1Ab as the underlying continuous variable.


IRRI program report for 1998

3054 bp 2036 bp 1636 bp 1018 bp

7. RAPD analysis of DNA from antagonistic (+) and nonantagonistic (-) Bacillus amyloliquefaciens isolates with primer AM-18. IRRI, 1998.

Table 7. Mortality data used in calculation of heritability of Cry1Ab tolerance in a striped stem borer population. IRRI, 1998. Mortality (%) Male Female Cry1Ab diet 1 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 90 89.6 89.8 79.2 85.0 82.1 73.3 55.8 64.6 40.0 86.7 63.3 43.3 81.7 62.5 75.0 43.3 59.2 58.3 56.7 57.5 60.0 45.0 52.5 26.7 71.7 49.2 31.7 63.3 47.5 Control diet 0 3.3 1.7 3.3 0 1.7 0 0 0 6.7 6.7 6.7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean 1 2 Mean Male Female Cry1Ab diet 40.0 35.0 37.5 50.0 23.3 36.7 25.6 46.7 36.1 33.3 38.3 35.8 18.0 50.0 34.0 50.0 15.0 32.5 31.7 25.0 28.3 36.7 18.3 27.5 28.3 18.3 20.8 8.3 16.7 12.5 Control diet 0 0 0 6.7 0 3.3 0 0 0 0 0 0 13.3 26.7 20 6.7 3.3 5 3.3 0 1.7 0 0 0 0 0 0 0 6.7 3.3 Mortality (%)


1 kb marker 43343445244344142535236441640245936636538 78 84 447344345444267335462396+ 229+ 45+ 61+ 81+ 82+ 83+ 417+ 437+ 346+ 455+ 458+ 460+ 426+ 412+ 327+ 59 + 69 + B. subtilis 9-16










Cross-ecosystems research


Mean family mortality on the Cry1Ab-treated diet ranged from 89.8 to 12.5% (Table 7), and the effect of male parent on larval mortality was highly significant. There was no correlation between mortality on Cry1Ab and control diet (r=-0.02, P=0.90, n=20 males), suggesting that differential mortality of the families on Cry1Ab diet was due to their differential sensitivity to the toxin used. The narrowsense heritability for the test population was 0.52, meaning that 52% of the variation in mortality on the Cry1Ab diet was attributable to genetic differences. This is a relatively high value in comparison with those from other studies of insect heritability to insecticides, and indicates the importance of carefully designed resistance management strategies to delay the evolution of Bt resistance in striped stem borer populations. Invertebrate biodiversity in a Philippine farmers field K.G. Schoenly, H.D. Justo, Jr.,2 A.T. Barrion, M.K. Harris,24 and D.G. Bottrell25 Most invertebrate biodiversity studies in tropical rice systems have focused on the fauna of the plant canopy. Rarely have the canopy and the floodwater been studied collectively. This study compared temporal trends in community structure of invertebrates inhabiting the terrestrial (plant canopy) and aquatic (floodwater) portions of the rice ecosystem. Rank abundance curves and indices of community structure were applied to invertebrate time-series data collected from a farmers irrigated rice field in Calauan, Laguna Province, Philippines. Canopy and floodwater invertebrates were vacuumand strainer-sampled at roughly weekly intervals from seedling to harvest for a total of 8 sampling dates. The cumulative samples included 202 taxa and 9,570 individuals for the plant canopy and 180 taxa and 84,905 individuals for the floodwater. Rank abundance curves (Fig. 8) revealed q a 10,000-fold range in invertebrate abundance with many abundant taxa remaining abundant over all crop stages; q lower evenness (equitability) of invertebrate abundances in the floodwater than the plant canopy, due in large part to the numerical dominance (75%) of two crustaceans [Heterocypris luzonensis Neale, and Eucyclops serrulatus (Fischer)];

the preponderance of natural enemies as the largest guild in number of taxa in both the canopy and floodwater, followed by herbivores, detritivores, and tourists (nonpredatory taxa with no known interaction except as prey to ricefield predators); and herbivore- and detritivore-dominated faunas respectively typified early crop periods in the canopy and floodwater, followed by predatordominated faunas in both systems at mid- and late-crop stages. Rates of community turnover generally increased with crop age in both the canopy and floodwater faunas, with the former increasing faster than the latter.

Effect of water management on sheath blight spread S. Savary, L. Willocquet, N. Castilla, D. Flura,26 and L. Huber26 Water for agricultural use is becoming scarce worldwide, and growing rice with less water is a major challenge. A 1998 DS study at IRRI examined how water-saving measures in growing rice affect sheath blight epidemics. The spread of sheath blight in transplanted rice with two levels of water management (continuously flooded and flash flooded) was compared. The water level in the continuously flooded crop was maintained at about 5 cm above the soil surface. In the flash-flooded crop, water was introduced to the field regularly to maintain a saturated soil. Artificial disease sources were established (i.e., rice hills were inoculated with sheath blight) at weekly intervals in each water management level. Inoculations were done at 34, 41, and 48 d after transplanting in three different sites in each water treatment. The disease foci that developed were monitored at weekly intervals. Sheath blight severity on the leaves of the eight hills adjacent to an inoculated hill was assessed weekly from 1 to 3 wk after inoculation. Sheath blight severity was higher (Table 8) in the flash-flooded crop, i.e., when water saving was attempted. The effect of water management was significant (P<0.05) at each observation date, for all three generations. The effect of water management on sheath blight was attributed to the differences in amount of free water present on the canopies. Free water on the canopy was monitored using a four-point leaf wetness rating scale on three layers


IRRI program report for 1998

Mean and range of invertebrate abundances A. Canopy 10,000.000 1,000.000 100.000 10.000 1.000 0.100 0.010 0.001 0.000
D (9%) > H(30%) > E (59%) > T (2%) > 23 58 82 165 202




0 100,000.000 10,000.000 1000.000 100.000 10.000 1.000 0.100 0.010 0.001 0.000
D (11%) > H(26%) > E (62%) > T (1%) > 2 12











B. Floodwater 75% 90% 95%











Rank order of invertebrate abundance

8. Ranked-abundances and functional-group composition of canopy (A) and floodwater (B) invertebrates for each of eight crop stages. Numbers above each curve give the sampling date (d after transplanting, DT), number of taxa, and number of individuals. For clarity of presentation, curves were truncated at 10 individuals per taxon. Counts of canopy and floodwater taxa are numbers per 60 (initial) rice hills per sampling date. IRRI, 1998.

Table 8. Sheath blight severity at two levels of water managementa and three generations of foci. IRRI, 1998 DS. Generation 1b Observationa (wk) FF 1 2 3

Generation 2 FF 0.325 0.488 0.875 CF 0.091 0.133 0.259

Generation 3 Mean FF 0.756 1.134 1.152 CF 0.175 0.138 0.087 0.434 0.615 0.868 0.370 0.105 0.141

CF 0.029 0.043 0.077

0.221 0.224 0.578

Observations were made at weekly intervals from 1 to 3 wk after inoculation of source hill. bFF = flash-flooded, CF = continuously flooded.

Cross-ecosystems research


Weight of free water (mg cm-2) 5 Generation 1 Generation 2 Generation 3 Flash flooded Continuously flooded a

0 5 Generation 1 Generation 2 Generation 3 b

0 5 Generation 1 Generation 2 Generation 3 c

0 98






Julian days

9. Estimated weight of free water at 0630 h over time in a flash-flooded and a continuously flooded rice crop. a) upper layer; b) middle layer; c) lower layer of the canopy. Horizontal bars indicate the monitoring of three generations of sheath blight foci. IRRI,1998 DS.


IRRI program report for 1998

of the canopy throughout the experiment. The rating scale was 0 = dry, 1 = with a few scattered big droplets, 2 = a thin film of small droplets, and 3 = a dense film of small droplets. Visual observations were made daily at 0630h. Leaf wetness at that time of the day was considered to represent the amount of dew and guttation water formed on the canopy. The rating scale was translated in water weights and the weight of free water on all layers of the canopy was higher in the flash-flooded than in the continuously flooded crop (Fig. 9), particularly true in the beginning of the experiment. Differences in leaf wetness were stronger in the upper and middle layers of the canopy. This experiment suggested that attempts to save water could favor a disease such as sheath blight. A similar reasoning could apply to other diseases, such as blast or bacterial leaf blight.

Assessing opportunities for N2 fixation in rice

The project goal is to make rice amenable to symbiotic N2 fixation. In collaboration with scientists at the National Institute for Agrobiological Resources, Japan, and the French Institute of Agronomical Research/National Center for Scientific Research (INRA/CNRS), this activity examined q the occurrence and expression of the homologues of legume early nodulin (ENOD) genes in rice, and q the ability of rice to perceive rhizobial Nod factors. Homologues of legume ENOD genes in rice P.R. Reddy, J.K. Ladha, D.S. Brar, and H. Kouchi27 Investigations of legume symbioses have identified critical genetic components that are important for the accomplishment of symbiosis, but the presence of those components has not been assessed in rice. We initiated investigations on identification and characterization of homologues of ENOD genes in rice to determine the genetic predisposition of rice toward rhizobial infection. Eighty rice accessions from 23 Oryza species were examined for the presence of the homologues of ENOD genes. Southern blot analysis of genomic DNA was performed on nylon membranes using

cDNAs of legume ENOD genes as probes to visualize the homologues of ENOD genes in rice. The membranes were subsequently autoradiographed, and differences in the hybridization signal obtained were measured semi-quantitatively by scanning densitometry. For the estimation of hybridization differentials, the values obtained with a particular ENOD probe were normalized relative to the total hybridization signal intensity of the bands derived from the soybean genomic DNA probed with the same ENOD cDNA. Southern analyses revealed a widespread distribution of homologues of ENOD genes across all genomes of rice. The degree of cross-hybridization of the legume ENOD genes with sequences in the genomes of various species, however, suggests that the homologues of ENOD genes are conserved to various extents in different Oryza species and related genera. Hybridization differentials derived from semi-quantitation of the hybridization signals indicated that the homologues of ENOD2 are relatively better conserved in O. sativa, O. rufipogon, O. longistaminata, O. nivara, and O. perennis. ENOD5 is better conserved in O. punctata, O eichingeri, O. minuta, and O. malampuzhaensis; ENOD12 in O. australiensis and O. brachyantha; ENOD14, ENOD40, and ENOD55 in O. brachyantha; ENOD70 in O. australiensis, and ENOD93 in O. sativa, O. rufipogon, O. nivara, O. rhizomatis, O eichingeri, O. minuta, O. latifolia, and O. malampuzhaensis (Fig. 10). The fact that ENOD gene homologues exist widely in dicots and monocots is evidence that these homologues have arisen from a common ancestral plant. The ENOD genes in legume plants apparently acquired nodulespecific symbiotic functions as a result of gene duplication after the divergence of dicots and monocots. Characterization of the homologues of legume ENOD93 from rice P.M. Reddy, J.K. Ladha, and H. Kouchi27 The presence of homologues of ENOD genes in a wide variety of rice species denote that the biological functions of early nodulins may be diverse and not restricted to nodule organogenesis alone. Hence, we set out to characterize the rice homologues to ascertain how closely they are related to their leg-


Cross-ecosystems research



Os Os Os Or Or Or Or Or O1 O1 O1 Ob1 On On On Op1 Op2 Op2 Op2 Op2 Op4 Op4 Op4 Oe Oe Oe Oe Oe Om Om Ob Ob 23.1 9.4 6.6 4.4



EN OD 14

Ob Ob

EN OD 40

Ob Ob

2.3 2.0


10. Southern blots of DraI-digested DNA of Oryza species that gave strong hybridization signals with 32P-labeled legume ENOD cDNA probes. O. sativa (Os genome AA; Accessions IR31917, IR56, IR64); O. rufipogon (Or AA; 105908, 105909, 105910, 106412, 106423); O. longistaminata (Ol AA; 103886, 103890, 103902); O. barthii (Obr AA; 101937); O. nivara (On AA; 103407, 105721, 106185); O. perennis (Opr AA; 104823); O. punctata (Op2 BB; 103896, 104064, 105690, 105980); O. punctata (Op4 BBCC; 100884, 101409, 104975); O. eichingeri (Oe CC; 101424, 105181, 105182, 105408, 105413); O. malampuzhaensis (Om BBCC?; 105223, 105328); O. brachyantha (Ob FF; 101232, 94-10482). IRRI, 1998.



39 46 38


85 92 84


115 105 96

11. Comparison of deduced amino acid sequences of OsENOD93a and OsENOD93b of O. sativa showing homology to GmENOD93. Asterisks denote the amino acids in OsENOD93a and OsENOD93b that are identical to GmENOD93. IRRI, 1998.

ume counterparts. We initially characterized the homologues of infection-related legume ENOD93 gene in rice (O. sativa). A rice (O. sativa var. Nipponbare) cDNA library derived from poly(A)+ RNA of suspension-cultured cells, as well as its genomic library, were screened. We used digoxigenin-labeled rice-expressed se-

quence tags (ESTs) that showed similarity to soybean ENOD93, as probes to isolate cDNA and genomic clones of the homologues of GmENOD93 from rice. The library screening was done on nylon membrane replicas. Positive clones were plaque-purified, phage DNA was isolated, and cDNA and ge-


IRRI program report for 1998

EN OD 55

OsEN93Ha (0.8 Kb)

OsEN93Hb (0.6 Kb) 25S rRNA

12. Expression of OsENOD93a and OsENOD93b homologues in O. sativa roots, etiolated and green leaves, and calli treated with heptameric chitin oligomer (upper panel). rRNA bands stained with methylene blue were used as standards to confirm that about equal amounts of total RNA were added to each lane (lower panel). The numerals within parentheses indicate the length of the transcripts in kilobases. Lane 1, roots; lane 2, etiolated leaves; lane 3, green leaves; lane 4, suspension culture; lanes 5-7, suspension culture treated with chitin oligomer heptamer, 1 g ml-1) for 15, 60, and 120 min, respectively. IRRI, 1998.

cates that they have a special function in roots. It is yet to be determined, however, whether their expression increases, similar to that in soybean, when rice root tissues are infected by bacteria. OsENOD93a expression was extremely high in suspension-cultured cells, whereas that of OsENOD93b was similar to the level found in roots (Fig. 12). Abundant expression of OsENOD93a in suspension-cultured cells suggests that the induction of OsENOD93a has a crucial function in undifferentiated rice tissues. Because chitin oligomers are structurally similar to rhizobial Nod factors and can induce expression of early nodulin genes such as ENOD40 in Glycine soja roots, it was important to determine whether the application of chitin oligomers to suspensioncultured cells could alter the expression of OsENOD93a or OsENOD93b. Northern analysis demonstrated that, similar to that of legume ENOD93, the expression of both OsENOD93a and OsENOD93b was not enhanced by the application of chitin oligomer (heptamer, 1 g ml-1; Fig. 12). Ability of rice to perceive rhizobial nod factors P.M. Reddy, J.K. Ladha, S.K. Datta, K. Datta, M.C. Ramos, F. Maillet, R.J. Hernandez, L.B. Torrizo, and N.P. Oliva Proteins encoded by the activated nod genes of rhizobia aid the synthesis of extracellular lipochitooligosaccharide signal molecules, known as Nod factors, which play a pivotal role in determining the fate of the symbiotic interaction. Purified Nod factors were shown to elicit a number of responses on the roots of legumes, such as deformation of root hairs, cortical cell divisions, and the formation of nodule primordia, as well as nodules. In addition, Nod factors mediate transcriptional activation of early nodulin genes, such as ENOD2, ENOD5, ENOD12, ENOD40, and rip1, the products of which are involved in the early steps of the nodulation processes. These findings demonstrate that Nod factors play a crucial role in initiating the processes leading to N2-fixing nodule development in legumes. Determining whether rice can respond to Nod factors could lead to strategies that would make rice amenable to develop a N 2-fixing endosymbiotic association with rhizobia. Hence, we

nomic inserts were subcloned into pBluescript II SK+ and sequenced by the dideoxy-chain termination method. The DNA sequences were compared with sequences deposited in the GenBank database using the BLAST program, and additional computer analyses were performed using the GCG package. For northern analysis, total RNA was electrophoresed, blotted onto nylon membranes, and hybridized with 32P-labeled cDNA probes. The investigations revealed that rice possesses two different homologues of the GmENOD93. Analysis of the cDNA clones of rice homologues showed that OsENOD93a has an ORF with a coding sequence homology of 58.2% to GmENOD93, whereas the ORF of OsENOD93b has displayed a homology of 42.3% (Fig. 11). Total RNAs from various tissues of rice were subjected to RNA transfer blot analysis to examine the organ-specific expression of OsENOD93a and OsENOD93b. Those exhibited different organspecific expression patterns (Fig. 12). In intact rice tissues, OsENOD93b was most abundantly expressed in roots and at much lower levels in etiolated and green leaves, whereas the expression of OsENOD93a was low in roots and etiolated leaves, and not detected in green leaves. Predominant expression of the rice homologues in rice roots indi-

Cross-ecosystems research


initially tried to determine whether Nod factors are recognized by rice. MtENOD12 gene expression was shown to be an excellent molecular marker for Nod factor perception in alfalfa. To ascertain whether Nod factors are perceived by rice, we used the MtENOD12 promoter fused to the gusA (uidA) reporter gene and generated transgenic plants of three rice varieties (Taipei 309, Chinsurah Boro II, and IR58) carrying the MtENOD12-GUS fusion gene. Transgenic rice plants were first incubated in fresh Fahraeus medium, supplemented with or without ammonium nitrate (10 mM) for a period of 4-6 d prior to subjecting the roots to 6-24-h or 48-h treatment with a mixture of sulfated and nonsulfated NodNGR factors. NodNGR factors (10-6-10-9 M) were added directly to the medium containing either transgenic plants in tubes or excised root segments derived from the transgenic plants. Roots were stained for histochemical localization of GUS at various times during NodNGR factor incubation. No GUS expression was normally observed under N2-limiting conditions in the roots when not treated with NodNGR factors. On rare occasions, roots displayed faint GUS activity, which was strictly confined to parenchyma tissue surrounding vascular elements in stele. However, as much as 80% of the roots of the transgenic plants grown in N 2 -limiting conditions, upon treatment with NodNGR factors, showed conspicuous GUS expression. It was specifically in cortical parenchyma in the elongation zone behind the root tip and at the sites of lateral root emergence (Table 9, Fig. 13 a-e,

Ex Ep
En Pc

Yp Cs

En Ep Pc

Ep Ex C1 Ex Cs Yp

13. Histochemical localization of GUS activity in calli and roots of transgenic rice carrying the MtENOD12-GUS in response to NodNGR factors. Nod factor-elicited expression of GUS in the roots of transgenic (a)-(f) Taipei 309, (i)-(n) Chinsurah Boro II and (o)-(s) IR58, and (g)-(h) transgenic calli of Taipei 309. IRRI, 1998.

i-k, and o-s). In contrast, roots of transgenic rice grown in the presence of combined N, whether or not supplemented with NodNGR factors, showed

Table 9. Expression of the MtENOD12-GUS in roots of Taipei 309 in response to different concentrations of NodNGR factors. Root systems of three transgenic plants were analyzed. Excised roots were incubated for 24 h in the N2-free medium containing different concentrations of elicitors prior to visualizing GUS expression. IRRI, 1998. Concentration (M)a Elicitor 0 NodNGR factors NodNGR factors (sulphated) NodNGR factors (nonsulphated) N, N,N,Ntetraacetylchitotetraose

10-9 9/18 (50 %) [++] 7/17 (41 %) [+] 12/26 (46 %) [+] 0/21 (0 %)

10-8 8/15(53 %) [++] nd nd 0/16 (0 %)

10-7 12/17 (71 %) [+++] nd nd 0/14 (0 %)

10-6 15/18(83 %) [++++] 13/17 (77 %) [+++] 11/15 (73 %) [+++] 0/20 (0 %)

0/15 (0 %) 0/19 (0 %) 0/23(0 %) 0/15 (0 %)

Values depict number and percentage of roots expressing MtENOD12-GUS fusion in cortical parenchyma in response to the treatment with elicitors. Intensity of GUS expression: (+) low; (++) medium; (+++) high; (++++) very high. nd: not determined.


IRRI program report for 1998

no GUS activity at all. Taken together, these findings clearly demonstrated that the rice roots perceive NodNGR factors, and these rhizobial signal molecules are able to efficiently mediate the activation of MtENOD12-GUS fusion in the rice roots only when the plants were starved of N. Our study showed that the MtENOD12 promoter is activated in rice by rhizobial Nod factors. The findings imply that rice has a mechanism to perceive Nod factors and also possesses a signal transduction system to enable subsequent activation of the legume early nodulin promoter. As in legumes, the presence of excess combined N in the growth medium inhibited the expression of the symbiosisrelated gene MtENOD12 in rice in response to Nod factors. This implies that Nod factor action on MtENOD12 expression in rice is controlled by the N-status of the plant, suggesting that at least a part of the N 2 -mediated regulatory mechanism(or mechanisms) responsible for symbiotic responses in legumes is conserved in rice. Previous work found that NodNGR factors were unable to induce root hair deformation in rice (IRRI program report for 1996). In the current study, Nod factors failed to elicit MtENOD12-GUS expression in epidermal cells including root hairs. The apparent inability of epidermal cells to respond to Nod factors may be due to the absence of a putative cell reception, or receptors, that perceive them, or the repression of Nod signal transduction,or both, in these cells. In contrast to Nod factors, chitooligosaccharide alone was found to be inactive in promoting MtENOD12 expression in rice roots. This finding suggests that the Nod factor-mediated MtENOD12 expression in roots is not simply due to the action of chitooligosaccharides released after degradation of Nod factors by plant enzymes, and that the structural features of Nod factors are clearly responsible for eliciting the expression of MtENOD12 in rice. Taken together, our findings suggest that the genetic machinery regulating nodule development in legumes is at least partially conserved in rice. It is, therefore, essential that future studies be extended at the cellular and molecular levels to identify why rhizobia-induced symbiotic responses do not fully occur in rice. This could lead to genetically engineering rice to form a more intimate symbiotic association with rhizobia.

Implementing ecoregional approaches to improve natural resource management in Asia

IRRI, in 1995, accepted the task of leading the ecoregional initiative for the humid and subhumid tropics and subtropics. The focus of the ecoregional approach is conservation and management of natural resources to develop sustainable food production systems, with attention to socioeconomic factors in biophysically defined ecoregions. In 1998, the ecoregional project became part of IRRIs research program. There have been three major sets of research activities: 1. Development of an operational research and development model for natural resource management (NRM) in a partnership mode involving NARS, the international agricultural research centers (IARCs), advanced research institutions (ARIs), and nongovernment organizations (NGOs). The first pilot ecoregion was established in the Red River Basin in 1998. 2. Development of systems approaches and methodologies q for biotic stress characterization; q assessment of the status of rice demand and supply, and the relationship between demand and use of natural resources; and q exploring land use options and supporting a network of study sites in different ecoregions in four countries. 3. Developing participatory approaches to NRM from field or farm to the regional level and fostering closer linkages between NRM research and development efforts in the uplands of the Red River Basin. Characterization of tropical lowland rice production situations and injury profiles S. Savary, L. Willocquet, F.A. Elazegui, P.S. Teng, and N. Castilla

A protocol for characterizing patterns of rice cropping practices and injuries due to pathogens, insects, and weeds was used for six tropical Asian sites cov-

Cross-ecosystems research


Table 10. Production situationsa (PR1-PR5) for lowland rice characterized across tropical Asia. IRRI, 1998. PR PR1 PR2 PR3 PR4 PR5 PR6

MF High Low Low High High High

FP Long Long Medium Short Medium Short

WCP Herb NW Hand Hand Herb NW

IU Medium Medium Low High Medium Medium

HU Medium Low Low High Medium Low

FU Low Low Low High High High

DS Low Low High Low High Medium

WE Low Low Low Low High High


PC Rice Rice W/B W/B Rice Rice

Y (t ha-1) 4.8 4.6 3.5 6.7 3.8 3.9

Mineral fertilizer input (MF), fallow period duration preceding the rice crop (FP), weed control practices (WCP), insecticide use (IU), herbicide use (HU), fungicide use (FU), drought stress (DS), water excess (WE), crop establishment method (PM), crop preceding the rice crop (PC), and mean yield estimated in each cluster of production situations (Y). PRs are described by their predominant characteristics (medians) only, e.g., DSR implies that this is the most frequent crop establishment method in a cluster. Attributes are described with reference to the other clusters of production situations, e.g., high means that the attribute has a higher median than others in the cluster. For WCP, herb = herbicides, hand = hand-weeding, NW = no specific direct weed control measures; for PC, W/B = wheat or barley.

Table 11. Injury profilesa (IN1-IN5) for lowland rice characterized across Asia. IRRI, 1998. INs BLB SR ShR ShB BS LB NB PH RWM LF DH WH WA WB Y (t ha-1) 4.6 3.9 3.5 4.3 3.3


None Present None None None

H M None None None

L M H H None

H L M H None


None M H None L

None L H L None





M M H L None



a BLB = bacterial leaf blight, SR = stem rot, ShR = sheath rot, ShB = sheath blight, BS = brown spot, LB = leaf blast, NB = neck blast, PH = planthopper, RWM = rice whorl maggot, LF = leaffolders, DH = deadheart, WH =whitehead, WA = weed infestation above the rice crop canopy, WB = weed infestation below the rice crop canopy, Y = mean yield estimated in each cluster of production situations. INs are described with reference to the other clusters of injury profiles, e.g., high means that the considered attribute has a higher mean in the considered injury profile than in the others. L = low, M = medium, H = high.

ering a wide range of lowland rice environments. Data used were collected from several hundred individual rice fields. Production situations and injury profiles. Multivariate, nonparametric analysis of the survey data provides a sound, reasonable basis for characterizing the biotic stresses and their effects on rice yields. A limited number of patterns of cropping practices, as represented by the six production situations, PR1 to PR6, emerge in common across the surveyed sites. Table 10 summarizes the characteristics of the production situations and the corresponding rice yields. The overall mean of farmers yields was 4.1 t ha-1, which agrees with commonly reported yields for lowland rice in tropical Asia. Five injury profiles (IN1 to IN5) were additionally identified (Table 11). Figure 14 shows the distribution of production situations at the six surveyed sites. Figure 15 shows that three of the five injury profiles are particularly frequent: IN1 (stem rot and sheath blight); IN2 (bacterial leaf blight,







14. Ecoregional display of production situations derived from cluster analysis on an array of 15 categorized attributes of the patterns of cropping practices. The analyses involve several seasons in six sites: 2 cropping seasons in FAIZ, 2 cropping seasons in CLUZ, 3 cropping seasons in MD, 1 cropping season in HGZ, 2 cropping seasons in LAG, and 2 cropping seasons in ILO.


IRRI program report for 1998







15. Ecoregional display of injury profiles derived from cluster analysis on an array of 10 quantitative measurements of injury levels.

Most of the progress in closing the gap between actual (4.1 t ha-1) and attainable yield should be expected from better management, especially water supply. Injuries due to pests should be seen as secondary factors in their yield-reducing effects, compared with alleviation of yield-limiting factors. This approach in characterization enables broad domains to be identified for pest management strategies that are specific to production situations across the region. Linking characterization results with experimental yield loss measurements. A series of experiments were conducted where injuries due to pests (pathogens, insects, and weeds) were manipulated at levels of production representing the lowland production situations characterized in the sur-

planthoppers, and leaffolders); and IN3 (sheath rot, brown spot, leaf blast, and neck blast). Current importance of rice pest injuries by production situations. Certain injury profiles are distinctly associated with particular production situations at a regional scale (Fig. 16). Production situations PR1, PR2, and PR4, which occur at sites CLUZ and HG, represent situations where comparatively high attainable yields may be achieved in favorable irrigated environments. These production situations are associated with injury profile IN1, where sheath and stem diseases predominate (Fig. 16a). PR3 reflects a poorly endowed, rainfed lowland situation, e.g. at sites FAIZ and ILO, and is associated with injury profiles IN3 and IN4, which are characterized by many injuries. PR5 and PR6, which occur at sites MD, LAG, and ILO, are favorable irrigated environments but with water management problems. The corresponding attainable yields are average. These production situations are associated with injury profile IN2. Injury profiles and yield levels are strongly linked, but in a complex pattern, suggesting that no one specific injury profile is related per se to the variation of yield. Therefore, the yield-reducing effect of a given injury profile should be addressed at the specific scale of a given production situation rather than at the regional level (Fig. 16c). The corresponding production situations and injury profiles concur with weather patterns in their description of variation in yield levels.

Axis 2 1 0.5
IN4 a PR1


0 0.1 0.2 1 0.5 0 0.1 0.2 1

c b



PR6 Y1 Y5 PR3 Y2 PR5 Y3 Y4 PR1 PR2




Y4 IN3 Y3 IN1


0.5 0 0.1 0.2 --1 --0.5


Y2 IN2

Y5 IN4

0 Axis 1


16. Display of linkages between injury profiles IN and production situations PR (a); between production situations and increasing levels of actual, categorized, yield levels Y1 to Y5 (b); and between actual yield levels and injury profiles across the region (c), using correspondence analyses.

Cross-ecosystems research


Yield losses (%) 100 Mean survey Median survey IPM threshold 10

0.1 WB WA WH DH WM BS RTD PB Injuries

17. Estimated current yield losses due to selected rice pests in tropical Asia. Average losses (%) due to injuries have been computed under three scenarios: injuries set at their mean values across the region; injuries set at their median values across the region; injuries set at published IPM threshold values. Note that the vertical axis is on a logarithmic scale. The 'combined' bars refer to losses due to injuries combined into regionwide injury profiles based on means, medians, or thresholds.


BLB SHB Combined

vey. A principal component regression model adequately described the variation in actual yield (R2 = 0.978; F = 2269; n = 445 plots). The yield-reducing effect of some injuries was found to vary with changing, e.g. improving, production situations. For instance, yield reductions due to sheath blight, weed infestation, and tungro disease tend respectively to increase, remain stable, and decrease with increasing attainable yields. Principal component regression allows for estimation of yield losses attributable to individual injuries (Fig. 17). The conclusions were: q Weeds (WA, WB) appear to be the most important constraint across the region. q Yield losses computed for all injuries combined do not add up to the losses derived individually from each injury, indicating a less-than-additive effect of most injuries on yield losses. q Losses due to brown spot are high when means are considered but not so when medians are used. That indicates the importance of this disease in some production situations. q The use of published integrated pest management (IPM) thresholds generally underestimates losses caused by most pests. IPM thresholds blatantly exaggerate the

importance of pests when considered in a combined situation. Our results indicate that sheath blight, brown spot, and leaf blast cause important losses (between 1 and 10%) at the regional scale. Among the insect injuries, only whiteheads caused by stem borers appear to be of relevance (2.3% yield losses). However, these injuries do not match in importance with those caused by weeds. Both types of injuries due to weeds, WA and WB, cause about 20% yield losses when considered individually. When all mean injuries are combined into a single mean injury profile, at a regional attainable yield of 5.5 t ha1, a mean yield loss of 37.2 % is estimated. Two other types of injuries need mention. The first pertains to injuries that are controlled by current technologies used by farmers, particularly host plant resistance (e.g., to blast, bacterial leaf blight). The second pertains to injuries caused by harmful agents that are able to spread over large distances and cause considerable damage locally (e.g., tungro). Our results indicate that, regionwide, losses caused by tungro are low. The importance of individual pests in the different injury profiles associated with the six production situations are summarized in Table 12.


IRRI program report for 1998

Table 12. Importance of injuries by production situations. IRRI, 1998. Production situation Key components of the corresponding injury profilesa SR ShB PH WH WA WB ShR ShB BS LB NB DH WH WA WB BLB ShB LB PH WH WA WB Importance of injuries

A multiple-pest, production situation-specific model to simulate yield losses of rice in tropical Asia L. Willocquet, S. Savary, L. Fernandez, and P.S. Teng A simple model was developed to simulate rice yield losses due to insects (deadhearts, whiteheads), weeds, and disease (sheath blight) in a range of production situations, making use of information generated by empirical and mechanistic approaches over the last decade.




Medium Very high Low High Medium Medium High Medium Very high Low Low High Medium Medium High Low High Medium Low Medium Very high Very high

a SR = stem rot, ShB = sheath blight, PH = planthoppers, WH = whiteheads, WA = weed infestation above the rice crop canopy, WB = weed infestation below the canopy, ShR=sheath rot, BS = brown spot, LB = leaf blast, NB = neck blast, DH = deadhearts, BLB = bacterial leaf blight.

The model, which is based on daily time-steps, and considers 1 m2 area of rice crop, consists of two linked components. The first component simulates the dynamics of rice biomass, i.e., its daily partitioning into leaves, stems, roots, and panicles. The second component simulates the dynamics of tillering, tiller maturation, panicle formation, and tiller death. State variables and simulated processes allow accounting for the effects of production situations (i.e., yield-limiting factors) and injuries (i.e., yieldreducing factors) on rice crop growth and yield (Fig. 18).

Crop biomass component

Tiller number component

Growth Dev. stage

Pool of biomass

Number of vegetative tillers

Dry weight of leaves

Dry weight of stems

Dry weight of grains

Dry weight of roots

Number of reproductive tillers

Coupling point for damage due to pests Coupling point for the effect of production situation

18. Schematic representation of the rice crop growth model. IRRI, 1998.

Cross-ecosystems research


Coupling functions representing damage mechanisms due to stem borers, weeds, and sheath blight were developed and parameterized using published and experimental data. The damage mechanisms considered for sheath blight injury were decrease in leaf area index (LAI) and leaf senescence caused by leaf and sheath lesions. Deadheart damage mechanism was accounted for as a daily removal of vegetative tillers, numerically linked with a corresponding loss of dry matter of leaves and sheaths. The overall, field-level effect of whitehead damage was simulated as the nonfilling of the injured panicles. The overall effect of weed infestation is simulated by a reduction of rice crop growth, using a functional relationship between rice growth-rate reduction factor and weed biomass derived from published data.

Field experiments were designed whereby, for a given production situation, rice plots free of pests and those injured by pests (alone or in combination) are established. Crop growth and yield, environmental factors, and pest injuries were monitored. Data from control plots were used to calibrate parameters for attainable yield simulation, while data from injured plots were used to test the simulation of yield losses due to pests. Simulation of attainable and actual yield. The parameters required to simulate attainable growth and attainable yield were calibrated in three production situations (PS) in irrigated conditions commonly found in the Philippines and Vietnam: PS1 and PS2 transplanted rice (IR72) with N fertilizer rates of 30 kg ha-1 and 110 kg ha-1, and PS3directseeded rice with N applied at 60 kg ha-1. The grain yields were respectively 556, 642, and 583 g m-2 for PS1, PS2, and PS3. Figure 19 shows plots of the simulated against observed grain yield for the 18 (PS injury) combinations. Simulated yield estimates falling within 10% of the observed yield were considered acceptable, given that an experimental error of at least 10% is usually associated with estimation of rice yields. In most cases, the damage mechanisms due to sheath blight, deadhearts, whiteheads, weeds, and combined injuries were well accounted for by the model, except for four casesPS1: WD and

COMBI; PS2: WH and COMBI. The poor performance of the model with respect to damage due to weeds in PS1 and to whiteheads in PS2 was reflected in the two corresponding COMBI treatments as well. Sensitivity analyses. The model calibrated for PS2 was run at four injury levelsno injury, M, 2M, and 4M, where M corresponds to the mean injury level across a population of 450 farmers fields surveyed in tropical Asia. One injury was considered at a time. Results from sensitivity analyses for varying levels of damage are shown in Figure 20. Increasing levels of sheath blight injury were associated with decreased LAI, and subsequently decreased dry weight of panicle and yield (Fig. 20a). Sheath blight simulation resulted in yield losses ranging from 5.7 to 24.6%; maximum sheath blight severity had losses ranging from 10 to 40%. The simulated yield loss at mean sheath blight level is similar to that estimated from empirical experiments. Increasing levels of deadheart injury were also associated with decreased LAI, and subsequently decreased panicle dry weight and yield (Fig. 20b), but the effect on LAI and panicle weight was small, e.g., a simulated yield loss of <1% at the highest level of injury considered. This is in agreement with previous reports, which conclude that deadhearts do not cause severe damage unless a high incidence is reached, and also holds true for irrigated, direct-seeded rice with medium fertilization. Whitehead injury did not at all

Y sim 600 500


400 300 200 100 100

PS2, WH PS2, COMBI Y sim 1:1 Y obs - (10%*Yobs) Y obs + (10%*Yobs)




400 Y obs




19. Simulated versus observed yield in a 1997 experiment at IRRI. Each dot represents one injury treatment within one production situation, including the (noninjured) control treatments. IRRI, 1998.


IRRI program report for 1998

LAI 7 SHBf = 0 6 Sheath blight 5 4 3 2 1 0 7 DH = 0 6 Deadhearts 5 4 3 2 1 0 7 6 Whiteheads 5 4 3 2 1 0 7 6 Weeds 5 4 3 2 1 0 YLm = 22.20% YL2m = 38.53% YL4m =59.10% WDf = 0 WDf = 300 g m-2 WDf =150 g m-2 YLm = 3.20% YL2m = 6.40% YL4m = 12.80% WH = 0 WH = 3.2% WH = 6.4% WH = 12.8% YLm = 0.24% YL2m = 0.48% YL4m =0.99% DH = 2.1 % DH = 4.2 % DH = 8.4 % YLm = 5.7% YL2m = 13.4% YL4m =24.6% SHBf = 10% SHBf = 20% SHBf = 40%

-2 PANW (g m ) 700 600 500 400 300

200 100 0 700 600 500 400 300

200 100 0 700 600 500 400 300

200 100 0 WDf = 600 g m-2 700 600 500 400 300

200 100 0

14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 86 90 94 98

Days after crop establishment 20. Simulated LAI and panicle dry weight in a rice crop damaged by injuries at different levels. YLm: simulated relative yield loss when injury level is set to M; YL2m: simulated relative yield loss when injury level is set to 2xM; YL4m: simulated relative yield loss when injury level is set to 4xM; SHBf: maximum sheath blight severity; DH: percent of deadhearts; WH: percent of whiteheads; WDf: final weed biomass. IRRI, 1998.

Cross-ecosystems research


affect simulated LAI, and caused proportional-toinjury simulated yield losses (Fig. 20c). Increased weed infestation levels were associated with lower LAI and panicle dry weight (Fig. 20d). The model developed is sufficiently flexible to account for diverse production situations and injury mechanisms. In most cases, simulated growth and yield were close to observed values, and fell within the acceptance range. Field experiments in the Philippines, India, and China will provide data sets for further testing the model under other production situations. The model can also be adapted for other injuries, such as those due to brown planthopper, bacterial leaf blight, or defoliators. The Systems Research Network (SysNet) for ecoregional land use planning R. Roetter, A. Laborte, C.T. Hoanh, P. Cabrera, C. Lopez, B. Nunez, and P.S. Teng The Systems Research Network (SysNet) project started in late 1996 with the objectives of developing q scientific-technical methodology for exploring land use options using crop models and expert systems, and q operational methodology for supporting a network of sites representing various ecoregions in Asia SysNet consists of four main NARS partnersIndia (Indian Agricultural Research Institute), Malaysia (Malaysian Agricultural Research and Development Institute), Philippines (University of the Philippines Los Baos, Mariano Marcos State University, and PhilRice), and Vietnam (Cuu Long Delta Rice Research Institute)and IRRI. Collaborating scientists are from the Wageningen University and Research Centre, The Netherlands. SysNet case studies were set up at the subnational level in India, Malaysia, Philippines, and Vietnam. A common framework and ecoregional approach was adopted that bring together models, expert systems, and multiple-goal analysis to provide land-use options for local problems identified by stakeholders and related to competition for limited resources. The core of this approach is multiplegoal analysis, which allows one to weigh competing goals of stakeholders in a region.


The sequence of nine steps for methodology development followed by SysNet is illustrated in Figure 21. In 1998, the project completed the first iteration of testing, generalizing, and refining the Land Use Planning and Analysis System (LUPAS) for each case study (steps 6-8). LUPAS, designed by SysNet for the purpose of exploratory land-use analysis, consists of three databases and four major components plus geographic information systems (GIS) as a supportive component (Fig. 22). The core of LUPAS is an optimization model based on the linear programming technique for exploring land-use options. The other three components relate to assessment of resource availability, estimation of crop yield and livestock production, and input-output relations for the various agricultural production activities. The starting points for the analysis are policy views and development plans that are translated into alternative sets of objectives (scenarios) for a given region. Conversion of these objectives into mathematical equations (objective functions) constitutes the first step of developing a multiple-goal linearprogramming model (MGLP) for quantifying tradeoffs among conflicting goals and identifying optimum land-use allocation for a given set of objectives and constraints. The generic crop growth simulation model (WOFOST) for annual crops was improved, calibrated, and evaluated for rice cultivars OM997, IR72, and IR64, among others, in support of yield estimation. SysNet case studies. Each of the SysNet study sites has its own set of biophysical and socioeconomic characteristics, NRM issues, and regional development goals, as summarized in Table 13. Applying the LUPAS methodology. The LUPAS methodology was applied to each of the case studies. Table 14 shows preliminary results based on a limited set of four objectives for Ilocos Norte Province, Philippines. Two scenarios were considered (no water-sharing and water-sharing) assuming existence of an efficient irrigation network connecting all the municipalities of the province. It was further assumed that alternative production technology with higher resource-use efficiency than the current


IRRI program report for 1998

New case study with: -new approach and/or -new region


Learning from NARS and stakeholders


Outline LUPAS structure


Collecting and processing data


Training NARS on LUPAS components Integrating LUPAS components by NARS and stakeholders Testing the LUPAS with NARS and stakeholders Generalizing and tailoring the LUPAS to case studies Refining the LUPAS for each case study


Iterative process



Validating and continuing present case 9)studies and starting new case studies

21. Steps in SysNet methodology development. IRRI, 1998.

Data on biophysical resources

Data on socioeconomic resources



Land use options and achievements

Data on policy views and development plans

22. Components of the Land Use Planning and Analysis System (LUPAS). IRRI, 1998.

Cross-ecosystems research


Table 13. Natural resource management (NRM) issues and regional development goals of the SysNet sites. IRRI, 1998. SysNet site Haryana State, India NRM issues Lowering of water table in the northeast. Waterlogging, flooding, and salinization in the central part. Rural-urban migration due to low farm incomes. Competition for agricultural land from urban and industrial expansion. Reduced farm labor. Federal policy for the region to remain as the countrys rice bowl. High diversification of dry season cropping. High use of agricultural chemicals and water in dry season. Upland cultivation on steep slopes, causing soil erosion. Leaching groundwater pollution by nitrate and biocide residues. Conversion of agricultural land for urban use. Intensification of rice cropping. Crop diversification. Farmers and policymakers development objectives are at variance. Regional development goals Intensify cereal production. Solve water problems. Manage salinization Increase farmers income.

Kedah-Perlis region, Malaysia

Intensify and increase rice production. Increase nonfood production. Increase labor use efficiency. Reduce use of agrochemicals by improving resource use efficiency. Increase farmers income. Intensify rice production Increase cash crop production. Increase employment in agriculture. Increase input use efficiency. Increase farmers income.

Ilocos Norte Province, Philippines

Can Tho Province, Vietnam

Intensify and increase rice production. Diversify agriculture. Increase farmers income.

Table 14. Example of MGLP outputs, Ilocos Norte Province, Philippines: goal achievements without and with water sharing. Maximize rice production (A) Maximize other production (B) Maximize employment (C) Maximize income (D)



Without water sharing 1 Rice production (t) 2 Other production (t) 3 Employment (mandays) = 4 Total income (000 P) With water sharing 1 Rice production (t) 2 Other production (t) 3 Employment (mandays) = 4 Total income (000 P)

620,088 50,876 10,135,591 5,798,165

147,414 1,868,771 6,907,862 13,010,818

335,376 597,226 11,953,708 9,268,957

138,797 1,412,644 6,300,537 31,557,859

887,230 0 12,551,854 7,299,089

156,059 2,448,860 8,294,347 16,206,893

382,201 938,753 15,553,056 13,375,791

218,593 2,009,068 9,346,436 39,083,256

farmers practices will be available in the future. Model results are presented for the zero round, i.e., a model run that optimizes individual goals without imposing restrictions set by minimum requirements for satisfying other goals. Results with both scenarios indicate that farmers income will be lowest

if the province opts for maximizing rice production. Incomes will be highest by allocating more land for cultivating the more profitable crops such as garlic, tomato, and onion. Effective water-sharing would increase individual goal achievements by 20-30%.


IRRI program report for 1998

Progress of unreported projects

Ricea way of life for the next generation of rice farmers M. Bell
q q

Through a multisectoral international think tank, developed an R&D approach for increasing the impact of engineering in agricultural and rural development. Facilitated a postproduction chain stakeholders meeting in the Philippines. Identified needs, concerns, and opportunities within the postproduction industry in the Philippines and a greater understanding of the wider series of issues for all participants. This meeting led to the formation of the Philippine Rice Postproduction Consortium (consisting of the NFA, BPRE, CEAT/UPLB, PhilRice, and IRRI). Organized multisectoral stakeholders meetings and formalized partnership modes for improved dissemination channels to farmers through NARS and NGOs in Thailand and Bangladesh. Identified key agronomic limitations in northeast Thailand and engineering limitations in Bangladesh. Trained Thai NGO collaborators in rice production principles. Have observed large buy-in to the concept of multisectoral partnerships. Defined components of yield gap project in Bangladesh. Designed questionnaires for surveys and conducted data collection training. Began surveys for the rapid rural appraisal for general characterization of eight villages in Bangladesh. Will use results to identify farm households for the intensive survey and to estimate technical efficiency differentiation across farms. Established field demonstrations involving NGO, government, and IRRI-Rainfed Lowland Rice Research Consortium for promotion to farmers (short-maturity rice-chickpea rotation in Bangladesh, and leveling, improved nutrition, and weed management in northeast Thailand). Participated in successful NGONARS field day, which highlighted the successful outcome of selected technologies in Thailand. The number of farmers interested in

the technologies for 1999 has increased tenfold. Information technologydeveloped knowledge packages on technology change options. Developed, field tested, and identified improvements for the draft TropRice decision support system. Developed and placed on the IRRI intranet a Web site about IRRI engineering technology. Published proceedings of engineering think tank, promoted the recommended approach through several journals and newsletters, and gave five presentations in various international venues. Facilitated the availability for demonstration of the improved SG800. Completed and made available the blueprints for the stripper gatherer SG800 Mk III. Developed databases of drying systems, transplanters, hand tractors, power tillers, and combine harvesters available in the market, including their features, capacity, cost, and manufacturers. Began evaluating impact of introduced postharvest machinery (rice flour mill, rice dehuller for glutinous rice processing, and portable rice micromill) on family welfare and womens groups. Completed the IRRI component of the Impact Assessment and Evaluation Group (IAEG) project on CGIAR and NARS Postproduction Innovations in the Philippines and Vietnam. Produced a video titled Fertile Feedback: Postharvest Innovations through Partnerships to demonstrate that the impact of public-sector R&D can be accelerated by research partnerships with manufacturers and users. Began case study in Thailand to look at the effectiveness of the multisectoral approach to knowledge transfer. Identified relevant data sources. Conducted basic experiments on the effects of variety, drying air temperature, and relative humidity on fissuring, milling quality, and drying performance. Developed low-cost laboratory equipment especially for studying rice fissuring for drying and readsorption fissuring studies and training.

Cross-ecosystems research


Tested the stripper combine SC1500 during the 1998 DS. Shattering loss was recorded at about 1.6%, at a forward speed of about 4 kph. Tested and optimized the biomass-sensor pendulum-meter in drum-seeded IR72 rice for in-field, site-specific use at mid-tillering, booting, and milky grain stages. Parameterized and validated two tillering models and compared them with a newly developed tiller model based on relative crop growth rate.

Socioeconomic studies for technology impact, gender, and policy analysis M. Hossain
q q

Updated world rice statistics database to 1997. Began integration of biophysical and socioeconomic variables at subnational levels for Bangladesh and Myanmar. Prepared a manuscript, A Rice Village Saga: The Three Decades of Green Revolution in the Philippines (by Y. Hayami and M. Kikuchi) for copublication by Macmillan Press and IRRI. Analyzed the Central Luzon and Laguna longterm household survey data and prepared a research monograph for publication in 1999, which highlights the evolution of rice production technologies and their impact on the rural household economy. With these new technologies, the Philippines strengthened its comparative advantage in rice cultivation from 1966 to 1979. But this advantage began eroding in 1986 because of the decline in the world price of rice, stagnating yield, and higher prices of land and labor. By 1994, the country had lost its comparative advantage. Assessed changes in the rural household economy for 1987 and 1995 for Thailand. Surveyed eight villages in Bihar, India, to understand the importance of rice in the household economy and to analyze constraints to the adoption of new technologies. Processed in electronic files household survey data for four villages in Myanmar and six villages in eastern Uttar Pradesh. Completed a study for Vietnam on the impact of modern rice varieties on rice productivity,

socioeconomic equity, and poverty, which showed that the adoption of modern rice varieties was almost complete in villages with access to physical infrastructure and irrigation facilities. Their adoption contributed to an almost 20% reduction in the unit cost of rice production. Their adoption has also helped to alleviate poverty and better distribute rural incomes because of the more equal distribution of landholding and absence of crop-sharing tenancy. Generated information on the impact of breeding strategies on the diversity of rice cultivars through household surveys in Madhya Pradesh and Uttar Pradesh (India), Bangladesh, and the Philippines. Began a survey for updating IRRIs database on the release of rice varieties and their genealogy. Published a book, Impact of Rice Research (edited by Pingali and Hossain), emanating from the international conference of the same name held in Bangkok, 3-5 Jun 1996.

Program outlook
Research activities under the program for 1999 will focus on q linking gene function to molecular map and genomic sequences; q transferring resistance against brown planthopper to elite lines; q developing tolerance for stem borer and gall midge in collaboration with NARS partners; q transferring disease resistance and tolerance for adverse soil conditions from wild species to rice; q implementing MAS to transfer multipledisease resistance genes to elite cultivars; q developing a model for sheath blight management and for deploying biocontrol agents for diseases; q developing strategies for field testing of transgenic plants, using the endogenous resistance gene Xa21 as a case study; q analyzing factors that contribute to yield gaps; q using information technology tools for knowledge packaging and for demonstrating technology options for farmers;


IRRI program report for 1998

generating land-use options for selected sites to demonstrate sustainable NRM at regional levels; assessing economic values of IRRIs germplasm-related research, analyzing household-level data to assess the impact of breeding strategies on the biodiversity of rice cultivars; and evaluating impact of improved rice technologies on income distribution and poverty in the Philippines.

In addition to these focused research activities, the Cross Ecosystems Research Program has initiated a new Functional Genomics project to be fully implemented in 2000. The overall goal of this project is to develop genetic resources and research capacity to apply information and reagents provided

by structural genomics to understand complex biological functions. A genomewide experimental approach will help IRRI find new genes and dissect metabolic pathways important for increasing rice productivity. By creating genetic resources for trait discovery, IRRI can be in a strong position to use the vast genomics databases and to ensure accessibility of these resources to the rice-growing world. In 1999, we have defined the scope and activities of the project and have made progress in preparing the genetic resources (mutants, isogenic lines, mapping populations) essential for functional genomics. We have also explored partnerships with both private and public sectors. We expect the Functional Genomics project to become the engine for trait discovery and synthesis that can be used by ecosystem-based programs for solving specific production constraints.

Cross-ecosystems research


Research programs Rice genetic resources: conservation, safe delivery, and use

CONSERVATION OF RICE AND BIOFERTILIZER GENETIC RESOURCES 124 Germplasm and information exchange 124 Genebank management 126 Germplasm characterization 126 Data management 127 Training 127 Conservation of biofertilizer germplasm 128 Biosystematic studies of wild rices 129 Dynamic systems of genetic conservation 129 DELIVERY OF GENETIC RESOURCES: THE INTERNATIONAL NETWORK FOR GENETIC EVALUATION OF RICE (INGER) 130 1998 INGER nurseries 130 Processing and distribution of test materials for the Upland Rice Research Consortium and Breeding Network 130 Additional seed distribution 131 Seed increase for INGER nurseries 131 Preparation of the 1999 INGER nurseries 131 Utilization of the 1997 INGER entries 131 THE INTERNATIONAL RICE INFORMATION SYSTEM 131 The Genealogy Management System 131 The Data Management System 132 ICIS applications 133 Linkages 133 SEED HEALTH TESTING SERVICES 133 PROGRAM OUTLOOK 135

Rice genetic resources: conservation, safe delivery, and use 123

Rice genetic resources: conservation, safe delivery, and use

The IRRI Medium-term Plan for 1998-2003 consolidated genetic resources activities into four projects in this new program. It contains researchsupport and research components of germplasm conservation and evaluation, as well as seed health activities. Studies on the biosystematic relationships of the genus Oryza are included. A new project for development of the International Rice Information System (IRIS), which is part of a broader database development agenda between IRRI and the International Maize and Wheat Improvement Center (CIMMYT), is also included.

common 5-10 yr ago. Factors that contributed to genetic erosion are 1) rapid change of the rice ecosystem from deepwater rice to irrigated rice (boro), 2) extensive human disturbance of wild rice habitats, 3) intensive culture of fish, 4) intensive cattle grazing and harvesting of grasses for cattle feed, and 5) overexploitation of agricultural land.

Conservation of rice and biofertilizer genetic resources

Germplasm and information exchange M.T. Jackson, B.R. Lu, G.C. Loresto, and S. Appa Rao More than 8,150 samples of Oryza sativa (7,409) and various wild species (745) were received from 16 countries for long-term conservation in the International Rice Genebank. Much of that germplasm was collected by partner national agricultural research systems (NARS).

Extension staff members trained by IRRI collected more than 50 samples of cultivated rice in west central Bhutan. Some 255 samples of cultivated rice collected in eastern Bhutan were multiplied at the Renewable Natural Resources Research Centre, Bajothang.

Extension workers collected 438 samples of wild species in seven provinces and 298 samples of cultivated rice in three provinces between Dec 1997 and Feb 1998. One extension worker in Stung Treng took the initiative of collecting 31 samples of traditional varieties in an area formerly held by the Khmer Rouge. IRRI received 620 samples of wild species and 492 samples of cultivated traditional varieties for long-term storage.

B.R. Lu assisted genebank staff members of the Bangladesh Rice Research Institute to collect wild species in the southern and western districts. Ten samples were collected. Only one wild species, O. rufipogon, one weedy type, and two related genera, Hygroryza aristata and Leersia species, were found. Farmers reported that wild types were more

New areas along the Atlantic coast, from the Colorado River to Manzanillo, were identified for collecting O. latifolia. Plants and insects associated with wild Oryza and information of water and soils were also collected. The O. latifolia samples were


IRRI program report for 1998

characterized and a morphological analysis made. A detailed study of a population of O. glumaepatula from Rio Medio Queso (near the frontier with Nicaragua) is ongoing.

Almost 300 samples collected from Aceh in Sumatra and in central Kalimantan were received at IRRI for long-term conservation. B.R. Lu helped Indonesian colleagues collect wild rices in central Sulawesi. The group found a population of O. meyeriana completely shaded under trees in a mountainous area, the first reported collection of this species in central Sulawesi. Local farmers refer to it as padi kakaju (forest chicken rice). Thirty-two upland and lowland rices were also collected.

the time of the visit. No one was allowed to gather seeds or panicle samples from the field because the farmers had not performed their rice harvest rituals. However, some farmers provided one or two panicles per variety. Twenty-one distinct varieties were collected. Farmers grow from five to seven varieties in one field. Seeds and vegetative stocks of O. officinalis were also collected in Tijirak village, about 25 km from Kuching. Villagers called this species padi pipit, meaning bird rice. Extension workers and staff members from the Malaysian Agriculture and Research Development Institute (MARDI), Seberang Perai, collected 260 samples of cultivated rice in the interior of Sibu, Sarawak. During the year, a total of 315 samples of cultivated rice and 4 samples of O. officinalis were collected.

S.A. Rao collaborated with extension personnel and Ministry of Agriculture and Forestry officials to collect more than 2,150 samples of cultivated rice and eight wild species from the southern regions and northern provinces. More than 3,870 samples of cultivated rice and 29 samples of wild species were received at IRRI. The Lao rice collection is the second largest component in the International Rice Genebank.

An Instituto Nacional de Investigao Agronmica (INIA) team collected 72 accessions of cultivated rices at 23 sampling sites. The samples were highly variable, with outstanding traits such as flood tolerance, low water requirement, aromatic, lodging resistance, tolerance for attacks of birds, long grain, early maturity, and high yield. Eleven accessions of O. longistaminata were also collected. One accession was an apparent hybrid between O. longistaminata and O. sativa. Herbarium specimens of the wild species were made.

Collecting activities during the second quarter of 1998 were part of training for field collection. One of the highlights of field training was the collection of natural hybrids between O. longistaminata and O. sativa. Research staff members also collected 30 samples from the middle-west area. Duplicates of one O. longistaminata and 151 samples of cultivated rice were received at IRRI for long-term conservation.

Extension workers from Shan, Chin, Kachin, and Rakhine states collected 693 samples. About 70% of those were sent to IRRI. More than 210 samples are currently being multiplied at the Central Agricultural Research Institute (CARI), Yezin. Forty samples of wild species were collected in Dec 1998.

Collection at an upland area in Sri Aman in Jan 1998 was part of a short course on field collection of rice germplasm for research and extension staff members of the Department of Agriculture of Sabah and Sarawak. The rice crop was ready for harvest at

The curator of the National Plant Genetic Resources Center (NPGRC) sent five samples of O. longistaminata to IRRI for seed multiplication and conservation. Those samples were collected from Oshana, Omusati, and Oshikoto north-central Na-

Rice genetic resources: conservation, safe delivery, and use 125

mibia. After seed multiplication, IRRI will return a set of samples to NPGRC and the Southern African Development Community (SADC) Plant Genetic Resources Center in Zambia.

A total of 232 samples of cultivated rice were collected during 1998. IRRI received 723 samples collected in 1995 and 1996 for long-term conservation.

B. R. Lu assisted national genebank staff members with collection of wild rices in the central and western regions. The team collected 73 samples of wild species and 2 samples of cultivated rices. About 50 ha of pure O. rufipogon were found near Ajigara Lake, Kapilbastu District. Populations of O. rufipogon and O. nivara were abundant in the lower Terai region near the Indian border, and O. granulata thrives in the evergreen forest at an elevation of 400-500 m.

IRRI received nine samples of wild species collected in 1997. Genebank management M.T. Jackson, F. de Guzman, R. Reao, and S. Almazan Germplasm multiplication and rejuvenation activities encompassed 2,080 new introductions of O. sativa, and 3,480 O. sativa accessions already stored in the genebank collection. Seed from almost 85% of the newly received samples was harvested for long-term storage. Samples (439) of wild species were successfully multiplied in the genebank screenhouses. Most of the samples were received during the year, including one new species, O. neocaledonica. Protocols were modified for managing accessions of the allozamous African wild rice species, O. longistaminata. We added 4,758 accessions to the Base Collection. The seed viability of about 15,000 accessions in the Active Collection was monitored (6 yr after the previous germination test). In addition, initial viability was determined for about 5,700 samples that had been prepared for long-term storage. Under the terms of the Memorandum of Understanding (No. 58-5402-8M-F063) with the United States Department of Agriculture-Agricultural Research Service, which was renewed in Sep 1998, an additional 12,108 samples were sent for black-box duplicate storage at the National Seed Storage Laboratory, Fort Collins, Colorado. Germplasm characterization M.T. Jackson, R. Reao, and S. Almazan Vegetative and reproductive characters were scored during 1998 wet season (WS) for about 1,700 accessions of O. sativa and 190 accessions of O. glaberrima. Postharvest characterization was completed on 1,000 accessions that had been grown during 1997 WS. Wild species accessions (418) were also characterized.

A collecting team from the Philippine Rice Research Institute (PhilRice) invited B.R. Lu to join them in collecting wild rice species in the mountainous regions of Palawan. The team collected five samples of O. meyeriana, known locally as peacock rice because farmers use it as bait for trapping peacocks. Farmers also use the roots and underground parts of O. meyeriana to cure diarrhea and stomach ache. The Philippine team collected 264 samples of cultivated rice in Surigao, Abra, and Quirino provinces. About 190 samples of cultivated rice were sent to IRRI for long-term storage.

Collaborators sent IRRI 200 samples of cultivated local varieties collected in 1997. Collecting activities for 1998 concentrated in the upland areas of the north and northeast. About 240 samples of cultivated and 20 wild rices were collected.

Collection work in March was part of a short training course on field collection. About 16 upland and 2 lowland local cultivars were collected in remote villages of the mountainous central region. A vegetative stock of O. rufipogon was collected at the edge of a farmers field in an irrigated area. Farmers called this species co lung, meaning cattle feed.


IRRI program report for 1998

Data management M.T. Jackson, A.P. Alcantara, and E. B. Guevarra The International Rice Genebank Collection Information System (IRGCIS) was fully operational. Two powerful servers (with data backup capabilities) were acquired to separately house the IRGCIS database and application. In terms of data maintenance, 27 batches comprising 9,206 samples were added to the system, and morphoagronomic data on 7,176 accessions were encoded. The Systemwide Information Network for Genetic Resources (SINGER) database provides a subset of IRGCIS data, and is accessible through the Internet (http://www.cgiar.org/singer). Data on more than 82,650 accessions were refreshed in SINGER in February. Training G.C. Loresto, S. Appa Rao, B.R. Lu, and M.T. Jackson Through support from the Swiss Agency for Development and Cooperation (SDC), GRC offered incountry training courses as well as on-the-job training at IRRI in field collection techniques, data management, genetic resources conservation and management, and isozyme analysis.

trainee was a lecturer in the Royal University of Agriculture, three trainees were from the Technical Office of the Department of Agronomy, and seven were from CIAP. The data management staff at IRRI developed the Cambodia Rice Information System (CamRIS) for managing the Cambodian rice collections. The IRRI staff also discussed the use, modification, and enhancement of CamRIS, and the management of existing data of the Cambodian rice collection.

A short course on data management and documentation for 17 genetic resources staff members was held at the National Bureau of Plant Genetic Resources (NBPGR), New Delhi, 26 Oct-7 Nov. Each participant used their own germplasm data to design and develop an information system in Microsoft Access 7. Most of them handled multiple crop data. GRC data management staff members visited the genebank at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) to assist scientists in refining their germplasm databases.

A 1-wk course on germplasm collection and conservation was given at Tingtibi, Zhemgang, to 17 extension and research workers from the east-central regionShemgang, Trongsa, and Gaylegphug districts. Because access to those areas is difficult and requires several weeks of trekking, the trained extension and research personnel assigned to the areas will facilitate the collection of rice germplasm. It is anticipated that extension agents assigned in the remote areas near the border with India will be able to find and collect any wild species that may occur in Bhutan.

A training course on field collection for 23 agricultural extension officials was held 21-26 Sep at the Northern Regional Agricultural Station, Houay Khot, Luang Prabang. After the training, extension personnel and the Lao collection team collected early-maturing varieties in the northern provinces.

A 2-wk short course on data management was conducted at the Cambodia-IRRI Australia Project (CIAP) office in Phnom Penh 26 Feb6 Mar. One

An in-country training on field collection 20-28 Apr at Mandriambero had 12 participants. Three were extension workers of the Ministry of Agriculture while the others were technicians from the National Center for Applied Research on Rural Development (FOFIFA) research stations where collection was planned for 1998-99. A 2-d on-the-job training on field characterization was held at the Mahitsy Research Station. A technician from FOFIFA who had on-the-job training at IRRI served as resource person and supervised the traineeseight technicians who participated in the 1997 short course in Mahajanga, and

Rice genetic resources: conservation, safe delivery, and use 127

one plant breeder from Marovoay Research Station. Those trainees joined the 1998 field collection trainees to share their experiences with the group.


A short course on field collection of rice germplasm was held 16-23 Feb at the Agriculture Institute, Sarawak. Participants were from Sabah (5), Sarawak (14), and Penang (1). Field practicals were organized in villages in Kuching, Sebangkoi, Betong, and Sri Aman. The Iban tribes grow mostly upland varieties in those areas. Samples collected during the field practicals were divided into two portions. One set with the corresponding passport data was brought to the MARDI genebank at Seberang Perai, and the other set was left in the Agriculture Research Center as part of their working collection. MARDI will multiply the samples and send duplicates to IRRI. One of the highlights of the course was the planning session wherein the participants were grouped according to their proximity to each other and assigned a priority area for collection. That session established priorities for 1998-99 collecting activities in Sabah and Sarawak.

Three research staff members from the China National Rice Research Institute (CNRRI) and the Chinese Academy of Agricultural Sciences (CAAS), and one from PhilRice attended a 3-wk on-the-job training on genetic resources conservation and management 10 Aug-4 Sep. The Chinese researchers also studied the morphological variation of three species using herbarium specimens. Three Indonesian research staff members from the Research Institute for Food Crop Biotechnology, Bogor, Indonesia, and one from PhilRice were trained in data management and documentation 7-23 Sep. The trainees used data from their institutes to develop prototype rice information systems. A Vietnamese researcher from the Cuu Long Delta Rice Research Institute underwent a 2-mo onthe-job training on isozyme analysis. A researcher from the Vietnam Agricultural Science Institute (VASI) received on-the-job training on isozyme analysis. Conservation of biofertilizer germplasm J.K. Ladha and T. S. Ventura The biofertilizer germplasm collection at IRRI has 1,392 accessions of Azolla, N 2-fixing bacteria, Rhizobium sp., cyanobacteria, and aquatic legume species. The collection is held by the Soil and Water Sciences Division. The Azolla collection is maintained as shoot-tip and liquid cultures, the cyanobacteria on agar slants, and the bacterial strains by lyophilization, deep freeze, or as agar slants. The aquatic legumes are conserved as seeds. Researchers in 18 countries requested a total of 230 samples during 1998. Management of the biofertilizer collection was enhanced by work on development of the Biofertilizer Information System (BIOFIS), which is scheduled for completion in 1999. Data standards were brought into line with other centers holding N2-fixing microorganisms, to permit the development of a CGIAR systemwide database, under the auspices of the Systemwide Genetic Resources Program (SGRP) and coordinated by the International Center for Agricultural Research in the Dry Areas (ICARDA).

Training at the Central Agricultural Development and Training Center, 14-19 Sep had 23 extension workers from Kayah, Kayin, and Tanintharyi states and 4 research staff members from the CARI seed bank as participants. Extension workers collect in the remote villages, and the CARI researchers coordinate and gather the collected samples along with the corresponding information from the different townships.

A 9-d course was conducted 16-24 Mar for 15 research and extension service staff members at the Agricultural Research Center for Coastal Southern Central Vietnam. Field practicals were conducted in remote villages of the Ba Na and Ure tribes. Those minority groups grow traditional varieties in the upland and mountainous areas. The participants collected 16 upland varieties and two lowland varieties with their corresponding passport data.


IRRI program report for 1998

Biosystematic studies of wild rices B.R. Lu, M.T. Jackson, A. Juliano, and M.E. Naredo Chromosome counts of all wild rice species in O. officinalis, O. ridleyi, and O. meyeriana complexes and accessions in O. brachyantha were completed. About 33% of wild rice accessions in the International Rice Genebank collection have chromosome counts. The diversity of 34 accessions of the Australian wild rice O. meridionalis was studied using morphology and isozymes. The accessions from Queensland generally showed a pattern different than the accessions from Northern Territory and Western Australia. The genetic diversity of 52 accessions of the African wild rice O. barthii from 14 countries was estimated through morphological, isozyme and random amplified polymorphic DNA (RAPD) studies. Cluster analyses of morphological data and RAPD markers revealed one accession from Niger as clearly different from the others. The 52 accessions were polymorphic for 14 (70%) of the 20 isozyme loci examined. Cluster analysis of these data showed a tight grouping of 48 accessions, and four accessions that formed a distinct cluster at a similarity level of 0.87. O. barthii was hybridized with other AA genome Oryza species, and more than 29,100 spikelets pollinated. Hybrids between O. barthii/O. barthii and O. barthii/O. glaberrima showed fertility higher than 70%. Among the other interspecific hybrids, only those with O. longistaminata showed higher than 5% fertility. Hybrids with O. meridionalis showed the lowest fertility (<0.2%). These results demonstrate the strong reproductive barriers between O. barthii and the other AA genome wild rices. Meiotic analysis of O. barthii and its intra- and interspecific hybrids was undertaken at metaphase I in the pollen mother cells. All samples showed regular meiosis and high chromosome pairing, which was not significantly lower than in the respective parental species. That indicated that the genome of O. barthii has not become significantly differentiated from the other wild rice species studied. However, strong reproductive barriers have developed between these species.

Meiotic pairing of O. ridleyi and O. longiglumis, and their intra- and interspecific hybrids was analyzed at metaphase I in pollen mother cells. All had regular meioses with full pairing, indicating high chromosome homology among the two species. Dynamic systems of genetic conservation J.L. Pham, S.R. Morin, M. Calibo, S. Quilloy, and M. Bellon Seed collections and socioeconomic surveys were completed in India in partnership with the Indira Gandhi Agricultural University and the NBPGR (New Delhi). The analysis of the passport data from the seed collection on the Bastar Plateau (Madhya Pradesh, India) demonstrated the high diversity of rice varieties in the region. Among the 106 variety names found in the region, 12 account for 50% of the total number of samples collected, and 36 for 80% of the samples. Socioeconomic surveys showed that farmers in the Bastar Plateau and Chhatisgarh areas choose varieties based on the interplay between soil quality and fertility, slope, erratic rainfall patterns, and basic household needs. The outcome is a complex, well-understood system of matching specific varieties with soil and plot types (and with other varieties) that reduces risk and enhances household welfare. A database of the results from the socioeconomic survey in the rainfed lowland system was developed in the Philippines in partnership with PhilRice. The analysis of microsatellite polymorphism of 178 accessions from the Cagayan Valley confirmed earlier isozyme data, and the original contribution of the Wagwag varieties to the overall diversity of the region. It also showed that traditional varieties are an important component of the on-farm diversity in this area as they contain a number of alleles, which are not present in the modern varieties cultivated in the region. Thus, 12 microsatellite alleles were specific to the 67 samples of traditional varieties studied, while only 2 alleles were specific to the 111 samples of modern varieties. The analysis of the variety naming process in the Cagayan Valley was completed. It showed that farmers categorize varieties in a limited number of

Rice genetic resources: conservation, safe delivery, and use 129

groups, which helps them manage varietal diversity. Duration is one of the main criteria used by farmers. Variety names change, and those changes also indicate genetic changes. IRRI and PhilRice organized the distribution of rice seeds to the Cagayan farmers who had contributed to the surveys. About 2 t of seeds representing 20 modern and 8 traditional varieties were distributed. Due to El Nio in 1997, typhoon Loleng in 1998, and low seed storage capacity, a number of farmers totally lost their own seed, making this operation highly appreciated. Farmers were asked which varieties they were currently growing, as well as for the previous three seasons, to evaluate their changes since a 1997 survey. A clear trend toward modern varieties was observed. In Vietnam, in partnership with the Hu University of Agriculture and Forestry, seed collection and a socioeconomic survey were completed in the upland and irrigated ecosystems. In the four irrigated villages, 69 rice samples, representing 17 variety names, were collected. Seventy samples, representing 17 variety names, were collected in five upland villages. Duplicate samples were sent to VASI, Hanoi. A field trial in Hu was conducted to study samples collected at the end of the winter crop season 1995-96. A field trial at IRRI allowed study of the samples from the 1996 summer crop season. Analysis of the microsatellite polymorphism was completed at IRRI for 77 accessions from the rainfed lowland ecosystem. Vietnamese scientists completed the second phase of a biotic survey in the rainfed-lowland ecosystem. Analysis of the results of the first phase demonstrated the difference in the biotic environment between the coastal and inland sites. Encoding of the data from the socioeconomic surveys in the rainfed lowland ecosystem was completed.

tem-oriented trials were planned with an augmented block design to assess adaptability to irrigated, upland, rainfed lowland, upland, and deepwater environments. Nurseries for multilocation evaluation of hybrid rice and fine-grained aromatic rice were also organized. Stress-oriented nurseries were composed to screen for soil salinity-alkalinity tolerance and resistance to rice blast and tungro.

Two hundred and fifty-eight nursery sets were dispatched to 28 countries. More than 88% of the sets went to countries in Asia, and the remaining 12% to countries in Sub-Saharan Africa, Latin America, and Europe. Within Asia, South and Southeast Asia received 78% of the INGER nurseries. More than 700 INGER entries were sent to national programs of Indonesia, Myanmar, Philippines, Thailand, Bangladesh, and India for evaluation and use.

The ecosystem-oriented nurseries were tested at key test sites and the stress-oriented nurseries were evaluated at hot spots for rice blast and tungro. The most widely tested nurseries were the irrigated observational nursery (IIRON), evaluated at 48 test sites in 26 countries, and blast (IRBN) screening conducted at 42 sites in 24 countries. Except for the rainfed lowland observational nursery (IDRON) and tungro sets, the nurseries were generally assessed at more than 20 experimental sites worldwide. Processing and distribution of test materials for the Upland Rice Research Consortium and Breeding Network B. Courtois, S.W. Ahn, and M. Laza In collaboration with the IRRI Plant Breeding, Genetics, and Biochemistry Division (PBGB) and network cooperators, 119 upland rice breeding lines from China, Indonesia, Myanmar, Bangladesh, Brazil, and IRRI were tested for viability, cleaned, and treated with chemicals to meet quality and quarantine requirements. Eight test sets were dispatched to Indonesia, Thailand, Vietnam, Philippines, China, Myanmar, and India for evaluation.

Delivery of genetic resources: The International Network for Genetic Evaluation of Rice (INGER)
1998 INGER nurseries S.W. Ahn, C. Toledo, V. Lopez, and R. Reao More than 800 diverse entries from various NARS and international agricultural research centers were organized into 10 INGER nurseries. Five ecosys-


IRRI program report for 1998

Additional seed distribution Rice scientists worldwide requested 608 seed samples, which were processed and distributed to 23 countries. The best entries in various ecosystems, sources of resistance to pests and diseases, or differential varieties for insect pests and diseases were most frequently requested. Seed increase for INGER nurseries Seven hundred and twenty-one entries were multiplied during the 1997-98 dry season (DS). Those included 243 incoming nursery nominations to 10 INGER nurseries. Those will be included as test materials in the year 2000 INGER nurseries. In addition, 27 blast monogenic lines were successfully multiplied in the screenhouse. Preparation of the 1999 INGER nurseries Five ecosystem-based and four stress-oriented INGER nurseries were tested for viability and seed health and processed for dispatch in 1999. Utilization of the 1997 INGER entries The number of 1997 INGER entries used by participating NARS for further yield evaluation and hybridization is given in Table 1. A total of 219 entries from ecosystem-based nurseries and 108 from stress-oriented nurseries were used as parents in the varietal improvement programs of 14 countries. Breeding lines most frequently used in hybridization are listed in Table 2. Some 412 entries were further evaluated in advanced yield tests of NARS. Those included in follow-up yield trials in more than three countries are given in Table 3. Yield performance of IR62141114-3-2-2-2, IR65610-105-2-5-2-2-2, and Qing Liu Ai No. 1 was further assessed in advanced yield trials in six countries.

Table 1. Utilization of 1997 INGER global nursery entries by participating NARS. Entries used (no.) Region and country Yield testing East Asia China Korea Southeast Asia Cambodia Malaysia Myanmar Philippines Thailand Vietnam South Asia Bangladesh Bhutan India Nepal Pakistan Sri Lanka West Asia and North Africa Iran Latin America Surinam Hybridization

16 80 10 133 7 55 57 120 16 138 24 44 2 23 -

42 9 1 4 160 10 62 51 26 59 29 2 14 14

breeding, evaluation, and utilization data have been identified as major constraints to developing knowledge-intensive improvement programs in many crops. The IRIS is a database system for the management and integration of global information on rice genetic resources and crop improvement. IRIS is part of a multicenter initiative known as the International Crop Information Systems (ICIS). IRIS provides all rice researchers access to the latest international information linked unambiguously to specific germplasm. IRIS allows scientists to benefit from, and participate in, the development and deployment of new, knowledge-intensive crop improvement systems that link information to the seeds being exchanged. The Genealogy Management System The core of the IRIS structure is a common genealogical data model called the Genealogy Management System (GMS). The central idea of the model is unique identification of germplasm and management of the homonyms and synonyms that arise naturally in the process of germplasm development and utilization.

The International Rice Information System

C.G. McLaren Ambiguity in germplasm identification, difficulty in tracing pedigree information, lack of integration between genetic resources, and characterization,

Rice genetic resources: conservation, safe delivery, and use 131

Table 2. INGER entries frequently used in 1997 as parents in NARS varietal improvement programs. Designation Ecosystem-oriented nurseries IR67017-13-3-3 IR67039-115-3-1 IR67423-42-2-3-3 IR65509-22-1-2-1R IR66233-234-2-1-2 IR65617-52-2-3-3-2-3 IR65912-31-2-4-2-3-1 RP2095-5-8-31 Nanjing 63056 Pusa Basmati 1 B6144 IR62141-114-3-2-2-2 IR65610-105-2-5-2-2-2 IR67423-53-2-3-3-2 IR62127-55-1-2-2-3 Stress-oriented nurseries K479-2-3 K39-96-1-1-1-2 Baldo IR-BB7 Hexi 10 Yungen 9 K459-34-1-1-3


Frequency of use

Countries where useda

IRRI IRRI IRRI IRRI IRRI IRRI IRRI India China India Indonesia IRRI IRRI IRRI IRRI India India Italy IRRI China China India

5 5 5 4 4 4 4 3 3 3 3 3 3 3 3 4 3 3 3 3 3 3


BGD = Bangladesh, CHN = China, IND = India, IRN = Iran, KOR = Korea, MYS = Malaysia, MYA = Myanmar, PAK = Pakistan, PHL = Philippines, THA = Thailand, SUR = Surinam, VNM = Vietnam.

Table 3. INGER entries frequently tested in NARS advanced yield trials, 1997. Designation IR62141-114-3-2-2-2 Qing Liu Ai No. 1 B3632F-TB1 IR65610-105-2-5-2-2-2 B6149F-MR-7 Nanjing 63056 BG1639 IR63872-3-1-3-3-1 IR65617-52-2-3-3-2-3 IR67423-42-2-3-3 132 Origin IRRI China Indonesia IRRI Indonesia China Sri Lanka IRRI IRRI IRRI Vietnam Frequency of use 10 8 7 7 6 6 6 6 6 6 6 NARSa BGD, IND, MYS, PAK, THA, VNM BGD, IND, MYS, PAK, THA, VNM BGD, CAM, NPL, PHL, THA BGD, CAM, IND, MYS, PAK, VNM BGD, CAM, NPL, PHL, THA BGD, IND, PAK BGD, IND, PAK, VNM BGD, IND, PAK, THA, VNM BGD, CAM, PAK, THA, VNM BGD, CAM, IND, IRN, MYA BGD, IND, MYA, PAK, THA

a BGD = Bangladesh, CAM = Cambodia, CHN = China, IND = India, IRN = Iran, KOR = Korea, MYS = Malaysia, MYA = Myanmar, NPL = Nepal, PAK = Pakistan, PHL = Philippines, THA = Thailand, SUR = Surinam, VNM = Vietnam.

The GMS functions are: q Assign and maintain unique germplasm identification. q Retain and manage information on genealogy. q Manage nomenclature and chronology of germplasm development. The genealogical core links to various applications required by individual users. Some of the applications will add genealogical data to the GMS,

while others will be analytical, displaying genealogies or calculating coefficients of parentage, for example. The Data Management System The Data Management System (DMS) is the component of IRIS that manages environmental data, germplasm characterization, and evaluation data for


IRRI program report for 1998

genetic resources and crop improvement projects. DMS links these data to genotype information in the GMS and to location information in the location manager as well as providing links to other specialized data sources. DMS also allows integration of data from different studies, thus permitting a broad range of queries across trials or types of factors and variates. The functions of the DMS are: q Store and manage documented and structured data from genetic resources, variety evaluation, and crop improvement studies. q Link data to specialized data sources such as GMS, location, and climate databases. q Facilitate inquiries, searches, and data extraction across studies according to structured criteria for data selection. ICIS applications ICIS applications are programs that access the database either for the purpose of installing or updating the database or for extracting and analyzing information in the database. They can be written by any ICIS user. As long as they only access the database through the supplied access functions, they will be portable between different local installations and even between different crop databases. Current applications include data management tools for installing ICIS databases, updating local databases to the central database, and launching user programs. Other applications allow users to browse the database and search for records, display pedigrees, and calculate coefficients of parentage. The External Pedigree Input Tool allows the capture of germplasm and genetic resources information. The Set Generation Module allows breeders and

germplasm evaluation specialists to compose lists of crosses, selections or entries for development and evaluation while automatically updating the database and having immediate access to all information known about the germplasm. Linkages The ICIS structure allows databases such as IRIS to link germplasm evaluation data to climate databases and geographical information systems. IRIS also links to SINGER and the USDA Germplasm Resources Information Network. Work is under way to speed up these links via the Internet and to connect to Rice Genes and other international genome databases. IRIS now permits scattered information generated on rice germplasm to be integrated, linked to sources of seed, and put to work in plant improvement. Its distributed design allows all partners to have access to the latest information and participate fully in the development of knowledge-intensive crop improvement programs.

Seed health testing services

T.W. Mew, S.D. Merca, C.C. Huelma, P.G. Gonzales, and J.O. Guevarra The Seed Health Unit, under supervision of the Philippine Plant Quarantine Service, processed 59 incoming rice seed shipments with 16,065 seed lots (421.3 kg) during 1998 (Table 4). Their consignment is seen in Table 5. Among the quarantine objects, weed seed contamination was only 0.8%, mainly due to Echinochloa spp. Insect-affected seed lots accounted for 4.7%, mainly due to Sitophilus oryzae and S. granarius. Blotter test of 556 incom-

Table 4. Origin of incoming seed shipments to IRRI, 1998. Region East Asia Europe Latin America Oceania South Asia Southeast Asia Sub-Saharan Africa West Asia and North Africa Total Countries (no.) 3 2 3 1 4 8 4 3 28 Shipments (no.) 13 4 3 1 4 22 6 6 59 Seed lots (no.) 3,997 29 161 6 258 10,982 186 446 16,065 Weight (kg) 45.1 .6 2.5 1.7 5.9 344.8 17.4 3.3 421.3

Rice genetic resources: conservation, safe delivery, and use 133

Table 5. Distribution of incoming rice seed shipments in 1998. Division APPAa EPPb GRCc INGERd IRGe PBGBf Total Shipments (no.) 4 3 8 27 17 59 Seed lots (no.) 364 235 76 8,419 6,971 16,065 Weight (kg) 38.9 1.6 8.3 210.2 162.3 421.3

a Agronomy, Plant Physiology, and Agroecology. bEntomology and Plant Pathology. cGenetic Resources Center. dInternational Network for Genetic Evaluation of Rice. eInternational Rice Genebank. fPlant Breeding, Genetics, and Biochemistry

ing seed lots revealed that the brown spot fungi (Drechslera oryzae) affected 85% of the seed lots, followed by sheath rot (Sarocladium oryzae) 51%, bakanae (Fusarium moniliforme) 43%, leaf scald (Gerlachia oryzae) 20%, kernel smut (Tilletia barclayana) 9%, and blast (Pyricularia oryzae) 8%. White tip (Aphelenchoides besseyi) affected 3% of the incoming seed samples. All the diseases, insects, and weeds found were collected and disposed of by autoclaving. All seed lots were subjected to postentry seed treatments of hot water (HWT) at 5257 oC for 15 min followed by fungicide slurry seed treatment of benomyl and mancozeb both at 0.1% formulated product by seed weight. A total of 376 phytosanitary certificates were issued in 1998 for outgoing shipments covering 67,597 seed lots (1,963.4 kg) (Table 6). The bulk of seed shipments to North America went to the National Seed Storage Laboratory, Fort Collins, Colo-

rado, for duplicate storage of the International Rice Genebank Collection. Sources of outgoing seeds were INGER 27,996 seed lots; PBGB 23,265 seed lots; IRG 15,886 seed lots; EPPD 400 seed lots; APPA 49 seed lots; and SWSD 1 seed lot. All outgoing seed lots were cleaned of quarantine objects. Routine seed health test of 1,475 outgoing seed lots revealed that S. oryzae affected 42% of the seed lots, followed by F. moniliforme 39%, G. oryzae 24%, T. barclayana 4%, and P. oryzae 2%. A. besseyi were found in 2% of the outgoing seed lots. Pre-export fumigation and HWT followed by benomyl and mancozeb seed treatment, both at 0.1% by seed weight, were administered on all outgoing seeds except for countries that do not accept fungicidal treatments. Post-entry crop health inspection was done on 4,747 entries in DS and 2,106 entry in WS. The major diseases observed were sheath rot affecting 87% of the entries followed by sheath blight affecting 20% in DS. In WS, sheath rot affected 28% and sheath blight 2%. Fewer diseases were observed in WS, which could have been due to the El Nio (dry) weather phenomenon. None of the diseases observed during the initial seed multiplication were introduced by incoming entries. A total of 9,718 entries were inspected for preexport crop health data on materials of PBGB. In DS, the diseases were sheath rot affecting 2% of entries and narrow brown leaf spot affecting 3%. In WS, narrow brown leaf spot affected 10% of the entries followed by bacterial leaf streak 5%, leaf scald 4%, sheath rot 3%, and sheath blight 2%. Rice blast was only observed on 1% of the entries.

Table 6. Distributions of rice seed shipments with phytosanitary certification by Seed Health Unit, IRRI, 1998. Region East Asia Europe Latin America North America Oceania South Asia Southeast Asia Sub-Saharan Africa West Asia and North Africa Total Countries (no.) 6 11 8 2 3 6 9 7 5 57 Shipments (no.) 95 42 16 25 14 85 74 9 16 376 Seed lots (no.) 13,638 533 4,038 14,695 2,384 15,164 12,385 1,250 3,510 67,597 Weight (kg) 499.6 24.9 105.6 18.4 22.1 595.6 560.2 41.5 95.5 1,963.4


IRRI program report for 1998

Researchers from China (2), Myanmar (1), and Bangladesh (1) were trained 1 Jul20 Sep in rice seed health for crop management.

Program outlook
Germplasm conservation and the safe delivery of germplasm to researchers worldwide are important and strategic components of IRRIs research agenda. Programmatic focus brings together the activities of the International Rice Genebank, INGER, and the Seed Health Unit and ensures that IRRI can correctly address those issues. Our research on biosystematics contributes to the more efficient management of the wild species, and

facilitates utilization of this germplasm in rice improvement. Policies for the on-farm conservation of rice varieties will emerge from GRCs research in this strategic area, which has been given prominence in both the Convention on Biological Diversity and the FAO Global Plan of Action for the Conservation and Sustainable Utilization of Plant Genetic Resources for Food and Agriculture. Access to information on conserved germplasm and its use in breeding will be enhanced as we continue to develop IRIS and ensure that the necessary strategic links to IRGCIS, SINGER, and the developing INGER information system are put in place.

Rice genetic resources: conservation, safe delivery, and use 135

Accelerating the impact of rice research

IRRIs ability to make positive and lasting contributions to poverty alleviation, food security, and sustainable management of natural resources depends on q the quality and relevance of our research, and q the effective evaluation, adaptation, and delivery of research products to users. New high-yielding rice varieties and knowledgebased rice technologies can increase the productivity and efficiency of rice farming. The program, Accelerating the Impact of Rice Research (IM), draws available research results from the research programs, synthesizes them into usable technologies, and evaluates and adapts them to specific rice-growing conditions in collaboration with national agricultural research systems (NARS). The IM program has four projects: q Strengthening partnership with NARS q Delivery of knowledge-intensive technologies (KIT): Crop and Resource Management Network (CREMNET) q Collecting, exchanging, and distributing knowledge and information about rice q Human capital development of NARS rice professionals

Strengthening partnership with NARS

East and Southeast Asia

Varietal improvement M. Sarom, O. Makara, P. Phon Hel, H. Yadana, and E. Javier

Farmers yields and acceptance of Cambodia-IRRIAustralia Project (CIAP)-released, improved local, long-duration rice varieties (CAR4, CAR5, and CAR6) were measured during 1995-98. The varieties performed well in wide-ranging environments and cultural practices. All exceeded the farmers best variety in yield (Table 1) and acceptability. CAR4 was the most preferred variety, with an average acceptance rate of 42%. The three varieties contribute significantly to higher production across 40% of the rice-growing area of Cambodia.

Farming systems agronomy C. Phaloeun, T. Vuthy, S. Sakkhunthea, and H.J. Nesbitt

A technology package involving variety, land leveling, soil management, fertilizer, and integrated pest management (IPM) was tested during 1995-98 in rainfed lowlands in southeastern Cambodia. The technology increased rice yields by as much as 140% and farmers income by 69%. However, farmers had to have off-farm income or ability to borrow money for investing in inputs to realize the full potential of the technology package.


IRRI program report for 1998

Table 1. Mean yield (t ha1) of CAR4, CAR5, CAR6, and a local variety in on-farm adaptive trials in Cambodia, 1995-97.a Variety CAR4 CAR5 CAR6 Local variety

1995 2.8 2.7 2.8 2.3 1.1 (23) 1.1 (17) 1.2 (23) 0.9

1996 2.6 2.5 2.5 2.3 1.0 (13) 0.9 (9) 0.9 (9) 0.9 3.1 3.0 3.1 2.7

1997 1.4 (15) 1.4 (11) 1.5 (15) 1.2 2.83 2.73 2.80 2.43

Av 1.2 (17) 1.1 (13) 1.2 (17) 1.0

Number in parenthesis indicates the percentage yield advantage over the local variety.

Table 2. Yield response of IR66 to cutting off leaves and stems at different crop stages. Mean yield (t ha-1) Treatment Cutting off leaves (IR66) Cutting off 50% stems (IR66) 2.8 3.1 2.5 3.3 Complete stem cutting (IR66) 0 1.5 0.5 3.0 Complete stem cutting (CAR11) 1.2 2.8 0.4 3.8

Cut off all seedling leaves Cut off 50% of leaves at tillering Cut off 50% of leaves at booting Control

3.1 3.2 2.6 3.3

Crop protection and farming systems (socioeconomics) P.G. Cox, G.C. Jahn, Mak Solieng, Chhorn Nel, Tuy Samram, K. Bunnarith, and P. Chanthy
Farmer participatory research for rat management. CIAP researchers worked with Catholic Relief Services during 1998 to identify weaknesses in farmers rat management practices and make improvements in them. Farmers in nine villages already used traps and baits, which were complemented by rat hunts. The work helped farmers understand that destroying rats in their burrows in offseason improved control of rat populations. Pest damage simulation. Four simulation trials in a glasshouse determined the types and levels of pest damage affecting rice yields at different crop stages. The results (Table 2) suggest that rice has considerable capacity to compensate for or recover from drastic damage at vegetative phase. Thus, it may not be necessary to manage pests that restrict

their damage to leaves before panicle initiation (PI). Likewise, pests that cut tillers, such as rats, are likely to reduce yields after PI. Cutting off 50% of tillers at stem elongation did not reduce yields of IR66, which has the ability to grow new tillers, suggesting that IR66 could tolerate fairly heavy gall midge damage before PI.

Integrated nutrient management N. Heer, P. White, and M. Sana

The recently introduced Cambodian Agronomic Soil Classification System (CASC) provides a simple means for agronomists and extension officers in Cambodia to identify and manage soils. A survey among 22 agronomists and extension officers in 6 provinces determined that CASC was used by all to q identify the soil where their research trials were located, and q help farmers identify their soils and determine appropriate fertilizer rates.

Accelerating the impact of rice research 141

Agricultural engineering J. Rickman, S. Bunna, and P. Sinath

Land leveling is used in Cambodia to improve water use efficiency and crop management in rainfed rice crops. Land leveling increased yields by more than 30%. Good land preparation and water management reduced weeding time from 21 to 5 d ha-1. Leveling also increased the opportunity for direct seeding with 1 d ha-1 used, compared with 30 d ha-1 for transplanting. Water use efficiency was also improved by use of water from higher elevation fields to establish and improve crops in lower fields. Yields in 1998 trials were compared in large fields (0.250.50 ha) with different degrees of levelness but with identical crop management and inputs. A strong correlation was found between yield and land levelness (Fig.1). Animals, 2-wheel, and 4-wheel tractors have been successfully used to level fields with harrows and leveling boards. These techniques required total water coverage in the field. Four-wheel tractors were effective in leveling both wet and dry fields. The cost of land leveling with a tractor ranged from $3 to $5 per centimeter of soil moved per hectare. Contractors are presently charging double that amount. If appropriate plowing techniques are adopted, re-leveling should not be necessary for at least 8-10 yr.

-1 Grain yield (t ha ) 4 3 2 1 0 0 1 2 3 4 5 Land levelness (SD in cm) 6 7

y=-278.63x + 3541.6 2 R =0.66

1. Crop yield as affected by land levelness. CIAP, 1998.

lines (IR69075-1-1-3-2-11, IR69726-116-1-3, and IR71031-4-5-5-1) were highly resistant to tungro while IR64 was almost completely damaged by the disease.

J. Schiller, B. Linquist, K. Fahrney, and NARS staff The Swiss Agency for Development and Cooperation (SDC)-funded Lao-IRRI project started in Aug 1990 and is currently in its third phase. In Nov 1998, the SDC indicated in-principle support for a Phase 4 that will extend the project to Dec 2002. Varietal improvement. By 1998, eight improved glutinous varieties had been released, five of which had IRRI genotypes in their parentage. Plant nutrition. Four varieties and N rates from 0 to 120 kg ha-1 were compared in field trials in 1998 dry season (DS). Combined analyses across sites indicated significant variety and N effects, and variety site interactions. The yield response to N was typically linear between 0 and 120 kg ha-1 for all varieties. Similar observations were made in rainfed lowland trials in wet season (WS). The response of TDK1 to N in WS rainfed lowland trials was evaluated at 20 sites between 1993 and 1998. Average yield without N was 0.3 t ha-1, lower in the southern and central Lao regions than in the north. In the south and central regions, the average yield response of TDK1 was linear between 0 and 90 kg N ha-1 (Fig. 2). In the north, yields did not increase with N rates above 60 kg N ha-1. Lower solar radiation may contribute to the lower yield in the north.

S. Tang and NARS staff The International Rice Molecular Breeding Program was organized and will be initiated in 1999. Collaboration on hybrid rice will continue. Seeds of 261 Chinese rice varieties were sent to IRRI for research purposes.

M. Syam and NARS staff A rice crop intensification program involving the growing of a third crop of rice annually was launched in 1998. About 120,000 ha of irrigated lowland areas in five provinces of Java and Bali were covered. About 90% of the area had been harvested by Dec 1998 with an average yield of 5.3 t ha-1. Field observations in Bali were that three IRRI


IRRI program report for 1998

-1 Grain yield (t ha ) 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 0 20 40 60
North South/Central y=0,0153x + 2,404 R2=0.9967 y=0,0002x + 0.0263x+2.693 R2=0.9968



N rate (kg ha1)

tilizer gave higher yields than the farmers practice of applying half the recommended rate (28 kg N ha-1) during planting. Wet-season trials on upland rice in Kalaw, Shan State, showed that locally produced liquid BioSuper fertilizer partially substituted for N requirement. A small credit scheme and community tree planting were successfully organized to support farmers. Fifty-eight Myanmar Agricultural Services researchers and extension workers participated in three in-country courses: Sustainability of the Community-based Natural Resource Management (25), Field Collection and Preservation of Rice Germplasm (23), and Introduction to Basic Geographic Information Systems (10). South Asia

2. Grain yield response of TDK1 in central and southern Lao PDR compared to northern Laos in rainfed lowland sites, 199398 WS.

Farming systems research. On-farm trials at rainfed lowland sites highlighted the potential for improving yields by use of simple technology packages based on use of improved varieties, fertilizer management, and agronomic practices. Average yield of adopters of the full package was 56% higher (3.4 t ha-1) than partial adopters (2.3 t ha-1), and 108% higher than nonadopters (1.6 t ha-1) in Champassak Province. Average net return of adopters of the full package was 43% greater ($510) than partial adopters ($356) and 96% higher than nonadopters ($259).

S. Bhuiyan and NARS staff During the year, scientists from IRRI spent more than 170 d working with scientists from the Bangladesh Rice Research Institute (BRRI) and other organizations on several collaborative projects. Those included yield gap research and the megaproject Poverty Elimination Through Rice Research Assistance (PETRRA), which will start in 1999.

R.K. Singh and NARS staff Initial results of research on farmers' participatory breeding indicate that farmers and breeders selection criteria based on agronomic traits did not differ much. One of the early impacts of the project was manifested in terms of changing the mind-set of breeders. The interaction with farmers, as well as social scientists involved in the project, helped breeders to better appreciate the multiplicity of farmers goals and the complexity of the environment. East, Central, and Southern Africa (ECSA) M. Gaudreau, V. Balasubramanian, and NARS staff The inaugural Steering Committee Meeting of ECSA Rice Research Network (ECSARRN) was held at Entebbe, Uganda. An ECSARRN project

T. Morinaka, K. Marooka, and NARS staff IRRI Hotline Vol.8 and News about Rice and People were published in Japanese and copies of each issue distributed to news media, policymakers, rice researchers, and persons concerned in international cooperation. Selected articles of Hotline and Science on Line on the IRRI home page were also translated into Japanese and distributed.

A. Garcia and NARS staff Split application of urea improved efficiency of applied N. Results of 1997-98 DS experiments on irrigated rice were that gypsum (sulfur) application, combined with 4-6 split applications of urea N fer-

Accelerating the impact of rice research 143

proposal was developed and submitted to the Association for Strengthening Agricultural Research in Eastern and Central Africa (ASARECA) for review and funding by potential donors.

sources. Simple tools are available to monitor crop N status and to apply fertilizers at the right time to meet crop demand.

M. Gaudreau and NARS staff Collaborative research programs. The United States Agency for International Development (USAID)-funded Madagascar-IRRI Environment and Agriculture Research Project (US$1.5 M for 3 yr) started in Jan 1998. A program to collect, characterize, and test the pathogenicity of a fungus that attacks striga was established and trials on striga control in upland rice were set up. An experiment to determine the fungi most effective against black beetle (Heteronychus) was established. Latin America and the Caribbean V. Balasubramanian and NARS staff The annual and Technical Advisory Committee meetings of the Caribbean Rice Industry Development Network (CRIDNet) and the annual meeting of the Caribbean Rice Association were held in Haiti in Feb 1998. IRRI provided limited research support to CRIDNet through rice germplasm and information exchange and provision of one chlorophyll meter each to Guyana and Cuba.

Delivery of knowledge-intensive technologies: Crop and Resource Management Network (CREMNET)

V. Balasubramanian, A.C. Morales, and NARS staff Rice production technologies are becoming more location-specific, complex, and knowledge-intensive with movement toward high yields (8-10 t ha-1) in farmers fields. CREMNET is designed to facilitate the identification, free exchange, participatory evaluation, and promotion of promising technologies in rice farming. Techniques for real-time N management in rice The nutritional status of rice plants reflects the availability and uptake of nutrients from different

The chlorophyll meter, also called the SPAD meter, is a simple, portable diagnostic tool that monitors crop N status in the field. When properly calibrated to locally important rice varieties and crop-growing conditions, the SPAD meter serves as an efficient tool for developing need-based, variable-rate N application on rice crops. A SPAD threshold value of 35 works well for semidwarf indica varieties in DS transplanted rice (TPR) systems in the Philippines. The value is reduced to 32 for WS TPR when solar radiation is low. However, the SPAD threshold values for TPR have to be kept at 35 for kharif (wet) season and 3738 for rabi (dry) season in India to obtain high yields. This is probably due to higher solar radiation during both seasons in India. For wet-seeded rice (WSR) in the Philippines, a SPAD threshold of 29-30 is optimum for broadcast WSR with a planting density of 800 productive tillers m-2, and 32 for row WSR with 650 productive tillers m-2. Thus, the critical SPAD value is inversely related to plant density, which varies with crop establishment methods. We observed three outcomes with the use of the SPAD meter: q An increase in grain yield but with a higher fertilizer N use. N use efficiency (NUE) was the same for both the farmers practice and the SPAD method, indicating efficient fertilizer use by farmers. However, grain yield can be further increased with additional N application, e.g., Nueva Ecija, Philippines, 1996 DS (Table 3). q A saving in N fertilizer use without reducing the grain yield. Here the NUE is higher for the SPAD method than for the farmers practice, indicating the need for improving farmers N management practice, e.g., Nueva Ecija, Philippines, 1996 WS; Sukamandi, Indonesia, 1996 DS (Table 3). q An increase in grain yield and a reduction in N fertilizer use. In this case, farmers NUE is low. Improvement in farmers fertilizer use practice is needed to save N fertilizer and to


IRRI program report for 1998

Table 3. Comparison of chlorophyll meter (SPAD) method with farmers practice for N management in rice in four countries, 1996. N used (kg ha-1) Increase or decrease in N use (%) + 19 28 33 52 Grain yield (t ha-1) 3.7 6.0 6.7 3.2 4.2 4.2 4.8 6.6 6.5 5.3 6.5 7.1 Increase or decrease in yield (%) + 11 + 0.1 0.8 + 10

Site and season


Nueva Ecija, Philippines 1996 DS Nueva Ecija, Philippines 1996 WS Sukamandi, Indonesia 1996 DS New Cauvery Delta, TN, India 1996 DS Cai Lay, Vietnam 1996 WS WSR

Control Farmers practice SPAD Control Farmers' practice SPAD Control Local recommendation SPAD Control Local recommendation SPAD

0 126 150 0 101 73 0 90 60 0 125 60

Control Local recommendation SPAD

0 120 70


2.8 4.0 4.0


increase rice yields, e.g., New Cauvery Delta, India, 1996 DS; Cai Lay, Vietnam, 1996 WS (Table 3). The cost of a SPAD meter is US$1,400, which individual farmers cannot afford. However, field researchers, extension soil specialists, crop consultants, and farmer cooperatives can purchase SPAD meters to monitor crop N status and advise farmers on N fertilization. CREMNET has distributed 38 chlorophyll meters to various collaborators in Asia and the Caribbean for evaluation and adaptation to local conditions. Use of the SPAD meter technique by NARS is increasing in Bangladesh, India, Indonesia, Myanmar, Philippines, and Vietnam.

Farmers generally use leaf color as a visual indicator of the rice crops need for N fertilizer. A Japanese prototype color chart was used by IRRI and PhilRice to develop an inexpensive leaf color chart (LCC) to determine the N fertilizer needs of rice crops. The chart contains six gradients of green color from yellowish green (No. 1) to dark green

(No. 6). It is calibrated with the SPAD meter and used effectively for guiding N application in rice fields. A simple instruction sheet in the local language helps the farmer determine the correct time of N application to rice crops. Use of the LCC promotes timely and efficient use of N fertilizer and can help minimize the pollution of surface and groundwater. The chart can be successfully adapted and used in irrigated and favorable rainfed rice ecosystems. CREMNET will extend this technology to ricegrowing countries of Asia, Africa, Latin America, and the Caribbean. Farmers in the Philippines and Vietnam were eager to use the simple and handy LCC to manage N in their rice crops. The Philippine Department of Agriculture distributed 15,000 color charts nationwide to agricultural extension agents and farmer-cooperators. About 5,000 charts were distributed to farmers in the Mekong Delta area of Vietnam. The Vietnamese farmers claim that adopting the color chart saved them 20-25% of N fertilizer, avoided lodging, and increased grain yield in WSR (Table 4).

Accelerating the impact of rice research 145

Table 4. Comparison of leaf color chart (LCC) method with farmers practice for grain yield and N use efficiency in the Philippines and Vietnam, 199698. N used (kg ha-1) Increase or decrease in N use (%) 52 Grain yield (t ha-1) 3.2 4.2 4.2 Increase or decrease in yield (%) + 0.4

Site and season


Nueva Ecija, Philippines 1996 WS; 17 farmers Nueva Ecija, Philippines 1997 WS; 12 farmers Cai Lay, Vietnam 1996 summer; WSR; 14 farmers Cai Lay, Vietnam 1997-98 DS; WSR; 14 farmers

Control Farmers' practice LCC

0 101 48

Control Farmers' practice LCC

0 97 87


3.5 4.5 4.3


Farmers' practice LCC

107 64


4.9 4.9

+ 0.8

Farmers' practice LCC (3 critical values)

88 64


7.0 7.1

+ 1.4

Collecting, exchanging, and distributing knowledge and information about rice

I. Wallace, G. Hettel, B. Hardy, S. Inciong, M. Movillon, and staff IRRI is a major disseminator of rice research information. Activities include q Creating, producing, and disseminating information materials that cover rice research and related issues, that create public awareness, and that are accurate, interesting, and useful. q Improving the publishing and disseminating of IRRI research results and promoting global exchanges of rice research information among scientists. q Making rice research information accessible electronically. q Maintaining the IRRI Library and Documentation Service as the worlds major repository of rice literature and facilitating access to the collection by rice scientists worldwide. q Serving as a convener, clearinghouse, and forum for dialogue among IRRI partners and IRRI in setting program strategies and

priorities, planning rice research activities, sharing results, and promoting discussion on institutional and policy issues. Maintaining IRRI Riceworld as the worlds leading museum devoted to rice culture.

Public awareness and general publications


The public awareness program is aimed at raising global awareness of the importance of rice research and IRRI research products. Activities focus on donors, policymakers, news media, NARS, advanced research institutions, nongovernment organizations (NGO), farmers organizations, and the general public. Many journalists visited IRRI in 1998. At least 180 articles on rice and about IRRI were featured in leading publications in Sweden, Japan, Germany, Thailand, France, Hong Kong, Saudi Arabia, Republic of Korea, Malaysia, Italy, Lao PDR, Vietnam, United Kingdom, India, Bangladesh, USA, Myanmar, Australia, Canada, Indonesia, and the Philippines. These articles resulted from the story ideas provided by IRRI.


IRRI program report for 1998

Six radio interviews were organized for IRRI staff to describe the Institutes major programs and key research issues on two Philippine radio stations, BBC World Service Radio, Dutch National Radio, Australian Broadcasting Corporation, and the Science and Technology Report of the United States. National and foreign TV film groups visited IRRI to produce special segments or documentaries about rice and the Institute. Articles about IRRI were featured on numerous worldwide web sites. IRRI participated in four exhibitionsthe International Cooperation Days in Japan (Oct), the University of the Philippines Los Baos (UPLB) Loyalty Day in the Philippines (Oct), the anniversary of the International Fund for Agricultural Development (IFAD) in Rome (Feb), and the Frankfurt Book Fair (Oct). The Institute gained special recognition from the Far Eastern Economic Review: winning a Silver Award for Asian Innovations.

Five issues of the newsletter Hotline were produced in English and made available on the web. A corporate report, IRRI 1997-1998: Biodiversity: Maintaining the Balance, was published and distributed. New editions of Facts about Cooperation (FAC) for 25 donors and rice-producing countries were published. A Japanese version of the FAC booklet for Japan was also published. A revised edition of Facts about IRRI was produced. A Chinese version of the IRRI slide show, Filling the Worlds Rice Bowl, was produced. Scientific publishing Nineteen titles were produced and distributed10 IRRI books, 7 IRRI discussion papers, a new publications catalog, and the Program Report for 1997. A bibliography is provided in the section on publications and seminars near the end of this program report. Three of the IRRI books were dual imprints with the Pacific Basin Study Center, the Thailand Development Research Institute, and Kluwer Academic Publishers. A significant set of papers on nutrient use efficiency in rice cropping systems appeared in a special issue of Field Crops Research (Elsevier). A complete revamping of the International Rice Research Notes (IRRN) started in 1998 with the ap-

pointment of an editorial board of scientists. The new IRRN will debut in April 1999 with a new look and features. IRRN has also been included in Cornell Universitys Essential Electronic Agricultural Library, which is a stand-alone compact disk library available only to scholars and students in developing countries. IRRN is also available on the Internet at http://www.cgiar.org/irri/irrn.htm. IRRIs efforts on the worldwide web resulted in significant accomplishments. More than 50,000 user sessions (more than 200,000 successful hits) were recorded on the IRRI homepage (http:// www.cgiar.org/irri). It had electronic versions of the IRRN, the 1997 Program Report, Bt Gene Information Bulletin, Rolling MTP 1999-2001, Sandiwa (new monthly IRRI newsletter), and more than 200 abstracts from 10 recent IRRI conferences and workshops. The Riceworld site (http://www. riceworld.org) was redesigned and relaunched during International Centers' Week. The Riceweb site (http://www.riceweb.org) was recognized as a highquality educational site by USA Today and Dow Jones, among others. Successful credit card sales of IRRI books (through the German book vendor TRIOPS) began on the Internet via the IRRI homepage. Orders for more than 100 books were received from around the world. Library and documentation service IRRI Library continued providing scientific information about rice to scientists globally. An External Program and Management Review panel noted that The Library is unparalleled in the world.

The Rice Bibliography grew by 5,480 references, of which 66% were in English, 14% each in Chinese and Japanese, and the remaining 6% in other languages. The online portion of the Rice Bibliography covers literature from 1970 onward, more than 172,000 references. The online Library Catalog, containing mostly book entries, grew to more than 33,000 bibliographic records, including 92 new dissertations on rice, from 10 different countries. The Library now has 1,370 active serial records and a growing collection of maps from many countries.

Accelerating the impact of rice research 147


The Library received requests for information by email, from countries as far away as Uruguay and Mozambique. Most of these external correspondents had used the Librarys Web site and were following up with specific queries. Efforts were made to strengthen information ties with NARS partners, most notably during extended visits to IRRI Library by librarians from Indonesia, Japan, and Pakistan. The IRRI librarian provided library training in Bhutan and visited colleagues in Indonesia, Philippines, and Republic of Korea.

the Consultative Group on International Agricultural Research (CGIAR). The Library provided web clients with a frequently revised list of forthcoming conferences of interest, often with electronic links to the organizers. Ties were also established with an on-line patent delivery service. IRRI visitors, conferences, and workshops IRRI welcomed about 71,000 visitors from 35 countries during 1998. Many came to observe IRRIs research activities and to explore the Riceworld Museum and Learning Center. Visitors included 2 heads of state, 13 state ministers, 10 ambassadors and members of the diplomatic corps, 7 donor representatives, and 16 representatives of international organizations. IRRI hosted or cosponsored 31 international and regional conferences, workshops, symposia, meetings and reviews, in which 1,249 nationals from 33 countries participated (Table 5).

IRRI Library is fast becoming a library without walls as outside electronic resources form an important part of the collection. Clients now have access to a vast array of information that the IRRI Library does not own in the traditional sense, and that is not located within the library building. Additional links to electronic journals, newspapers, bookstores, libraries, and reference sources were added during the year. Connections were expanded to the web sites of partner institutes and organizations such as CAB International, the Food and Agriculture Organization, and research centers within and outside

Human capital development of NARS rice professionals

R. Raab and staff IRRI, as a research and training institution, develops human resources of NARS to improve their capability for rice research. More than 12,000 degree

Table 5. International and regional conferences, workshops, symposia, and meetings hosted or cosponsored by IRRI in 1998. Date 12-16 Jan Title Workshop on Quantification of Yield Losses due to Rice Pests and Analysis of Survey Data in Plant Protection II Sysnet Workshop on Exploratory Land Use Planning Methodology for Ilocos Norte Province, Philippines Planning and Implementation Workshop for Crop Loss Assessment and Rice Straw Management Research Sysnet Workshop on Multiple Goal Land Use Planning for Can Tho Province, Vietnam Think Tank: Increasing the Impact of Engineering in Agricultural and Rural Development Venue IRRI Participants (no.) 29 Countries (no.) 7

14-21 Jan



19-22 Jan



16-26 Feb



26-28 Feb




IRRI program report for 1998

Table 5 continued. Date 24-25 Mar Title Annual Workshop of the IRRI-India Participatory Plant Breeding Project Land Use Systems Analysis Methodology for Haryana State, India Management of Rodent Pests in Southeast Asia Prioritization of Rice Research in Asia Land Use Systems Analysis Methodology for Kedah-Perlis Region Inaugural Meeting of the Project Development and Use of Hybrid Rice in Asia SysNet International Workshop on Exchange of Methodologies in Land Use Planning Review and Planning Workshop for Themes I and II of IPM Network Scaling Methodologies in Ecoregional Approaches for Natural Resource Management Workshop on Increasing Water Productivity and Efficiency of Rice-based Cropping Systems First Technical Committee Meeting of the Project Development and Use of Hybrid Rice in Asia IRRI-ACIAR Workshop on Strategic Research on Gender Issues in Ricebased Household Economy Workshop on USAID-sponsored IRRI-US University Collaboration Natural Resource Management in the Chao Phraya Basin: an Ecoregional Approach Irrigated Rice Research Consortium/ IPMNet/INMNet Joint Technical and Steering Committee Meetings Upland Rice Research Steering and Technical Committee Meetings Third IPM Network Steering Committee Meeting Steering Committee Meeting of the Project Development and Use of Hybrid Rice in Asia Nutrient Research in Rainfed Lowland Rice Venue India Participants (no.) 26 Countries (no.) 1

23-30 Mar



6-8 Apr



20-22 Apr 03-09 May

IRRI Malaysia

24 34

9 1

11-13 May



15-19 Jun



17-20 Jun




22-24 Jun




29-31 Jul




10-13 Aug



24-28 Aug




2-3 Sep



16 Sep



21-25 Sep

People's Republic of China Thailand



21-25 Sep



22-24 Sep



6-8 Oct



12-15 Oct




Accelerating the impact of rice research 149

Table 5 continued. Date 15 Oct Title Rainfed Lowland Rice Research Consortium 9th Steering Committee Meeting Ecoregional Approach for Planning and Management of Land and Other Natural Resources in the Ayerwaddy Delta Technical Workshop on Ecoregional Approaches for Natural Resource Management in the Red Basin, Vietnam Workshop on Rice Tungro Management Center-Commissioned External Review (CCER) of the Rainfed Lowland Rice Program Research Methods for Studying Weed Succession Venue Thailand Participants (no.) 14 Countries (no.) 7

20 Oct



9-10 Nov



9-11 Nov 16-20 Nov


45 55

8 4

23-27 Nov



30 Nov-2 Dec Sysnet Technical Review Workshop 1-3 Dec Workshop on Genetic Improvement for Rice in Water-limited Environments

Thailand IRRI

25 47

3 12

Table 6. Participants in degree and postdegree training at IRRI, 1998. PhD scholars Alam, Muhammed Murshedul Biswas, Jatish Chandra Chharom, Chin Oberthur, Thomas Muhsin, Muhammad Alinia-Gerdroudbari, Faramarz Fallah, Allahyar Nouanthasing, Lasay Ramanantsoanirina, Alain Marie Justin Razafinjara, Aime Lala Eow, Boon Tiak Htet, Kyu Htut, U Tin Thet, Khin Maung Winn, Tun Khadka, Yajna Gajadhar Abbasi, Fida Mohammad Arif, Muhammad Asghar, Muhammad Faiz, Faiz Ahmad Hussain, Fayyaz Ijaz, Muhammad Rillon Jr., Guillermo S. Tado, Caesar Joventino M. Trillana, Nemesio U. Dirie, Ahmed Mohamoud Lersupavithnapa, Boontium Cuong, Ngo Luc Ngo, Ngoc Hung Nguyen, Van Hong

Bangladesh Bangladesh Cambodia Germany Indonesia Iran Iran Lao PDR Madagascar Madagascar Malaysia Myanmar Myanmar Myanmar Myanmar Nepal Pakistan Pakistan Pakistan Pakistan Pakistan Pakistan Philippines Philippines Philippines Somalia Thailand Vietnam Vietnam Vietnam

Tran, Chi Thien Truong Van, Tuyen Cui, Kehui Fu, Binying Jianli, Wu Li, Luping Lu, Wenjing Tu, Jumin Yahai, Lu Yueqiu, He Zhong, Daibin Zhong, Xiaoyan Zhong, Xuhua Ziqi, Wang Baisakh, Niranjan Dey, Moul Kaur, Jatinder Mathan, Natarajan Mitra, Sudip Nath, Palash Deb Esfahany, Masoud Naoyoshi, Kawano Okada, Kanako Ojha, Gana Pati Bernardo, Eleuterio Noel Nghia, La Tuan Deka, Nivedita Lourdusamy, Gabriel Stephen Abdullah, Buang Bhandari, Hum Nath

Vietnam Vietnam China China China China China China China China China China China China India India India India India India Iran Japan Japan Nepal Philippines Vietnam India India Indonesia Nepal


IRRI program report for 1998

Table 6 continued. Trolove, Stephen Neil Borines, Lucia Cabuslay, Gloria De Los Reyes, Jeannelyn Lumbo, Susanita Linwattana, Grisana Sripongpankul, Krishnapong Graw, Stephen Le, Cam Loan Le, Thi Phuong Tran, Thi Ut MS scholars Sattari, Majid Inthavong, Soulaphone Rasabandit, Sengpaseuth Andriantsimialona, Dodelys Rakotomalala, R. Mbolarinosy Adhikari, Chiranjibi Upadhyay, Bhawana Abao, Jr., Elias B. Ong, Marilyn A. Phengchanh Somphet Le Van, Lang Ngo, Dang Phong Nguyen, Thi Phong Lan Thach, Thi Ngoc Anh Truong, Thi Ngoc Chi Shimizu, Akifumi Varghese, Pulickal Bhattarai, Kiran Alcantara, Jovencio Madamba, Reina Suzette Maghirang, Reycel DM Ulat, Victor Jun New Zealand Philippines Philippines Philippines Philippines Thailand Thailand United States Vietnam Vietnam Vietnam Nondegree on-the-job trainees Dema, Kezang Golinowski, Shawn Paul Richards, Barbara Schnupf, Mirjam Gao, Lizhi Hu, Fengyi Ma, Zhenrong Wang, Ying Xie, Xiaobo Kurniawan, Hakim Minantyorini Setyowati, Mamik Adachi, Shimpei Jeong, Eung-Gi Lakmaitry, Khamsone Abro, Abdul Haque Askari, Ejaz Bhand, Amir Ali Oad, Gulshan Lal Shah, Zahoor Hussain Newingham, Ma. Cristina V. Barnard, Katherine Patricia Anh, Ta Hoang Do Tuan, Khiem Mai, Duong Thi Hong Minh, Vo Quang Nguyen, Thach Can Nguyen, Van Tao Tran, Danh Suu Interns Manio, Denise Chung, Carrie-Lee Golinowski, Shawn Paul Gray-Donald, James

Iran Lao PDR Lao PDR Madagascar Madagascar Nepal Nepal Philippines Philippines Lao PDR Vietnam Vietnam Vietnam Vietnam Vietnam Japan India Nepal Philippines Philippines Philippines Philippines

Bhutan Canada Canada Canada China China China China China Indonesia Indonesia Indonesia Japan Korea Lao PDR Pakistan Pakistan Pakistan Pakistan Pakistan Philippines United Kingdom Vietnam Vietnam Vietnam Vietnam Vietnam Vietnam Vietnam

Australia Canada Canada Canada

scholarships, postdegree on-the-job training, and short-term group training fellowships have been provided to NARS scientists since 1961. Degree and postdegree training Degree and postdegree training provides rice scientists with opportunities to pursue a PhD or MS degree or acquire relevant skills through an on-the-job training or an internship with IRRI scientists. In 1998, IRRI extended degree and postdegree training opportunities to 126 scientists, mainly from the Asian region (Table 6). Fifty-six scholars and fellows completed their programs in 1998 (Table 7). Development of short-term courses Short-term group courses provide NARS scientists with an opportunity to update their knowledge and

skills on specific areas of rice science. Eleven group courses in 1998 attracted 144 NARS scientists from 18 countries. Genetic Evaluation and Utilization (GEU) training at Ubon, Thailand, focused on the rainfed ecosystem. An IRRI plant breeder was the course coordinator. Development of new training methodology IRRI intensified efforts to improve its capability for distance education during the year. Implementation of an International Development Research Centrefunded project, Application of Distance Learning Technologies to Human Capital Development in NARS, was initiated. IRRI, Simon Fraser University, and the Commonwealth of Learning collaborate in this project. IRRI staff members conducted an assessment of connectivity capabilities and re-

Accelerating the impact of rice research 151

Table 7. Scholars and trainees who completed training at IRRI during 1998.a Type I Country PhD MS PhD MS ND Type II Type III Total

Africa Madagascar Somalia Subtotal Asia Bangladesh Bhutan Cambodia China Indonesia Iran Japan Korea Lao PDR Malaysia Nepal Pakistan Philippines Thailand Vietnam Subtotal Europe Germany United Kingdom Subtotal North America Canada Subtotal Total

0 0 0

0 0 0

0 1 1

1 0 1

0 0 0

1 1 2

0 0 0 3 0 0 0 0 0 0 0 0 1 0 0 4

0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 2

1 0 1 0 0 1 0 0 1 1 1 2 4 0 0 12

0 0 0 0 0 0 0 0 1 0 1 0 2 1 2 7

0 1 0 5 3 0 0 1 1 0 0 5 1 0 7 24

1 1 1 8 3 1 1 1 3 1 2 7 9 1 9 49

0 0 0

0 0 0

1 0 1

0 0 0

0 1 1

1 1 2

0 0 4

0 0 2

0 0 14

0 0 8

3 3 28

3 3 56

a Type 1 = MS and PhD scholars, thesis research at IRRI; Type II = MS and PhD scholars, coursework and thesis at IRRI; Type III = on-the-job or nondegree training.

quirements of a recipient site in Hyderabad, India. Preparations continued for the first IRRI online course on Experimental Design and Data Analysis (EDDA) in collaboration with Simon Fraser University. The whole course will be ready for pre-testing by summer 1999.

q q

Integrated pest management (IPM)(1st ed.) Rice tungro disease identification and management (one for extension officers, one for farmers) Problem-based technology generation for rainfed lowland environments-Indonesia offering

Collaborative in-country training IRRI develops training materials for use by both trainers and trainees in group courses. Training materials were produced for courses on q GEU for rainfed lowland rice ecosystems q Instructional video production (supplementary readings) Collaborative in-country courses were initiated in 1989 to help develop NARS institutions indigenous training capability and to complement IRRI training activities at headquarters. Fifteen collaborative in-country courses were offered in 1998, training 257 NARS scientists (Table 8).


IRRI program report for 1998

Table 8. Participants in collaborative group training courses. IRRI, 1998. Course and date Participants (no.) /country Course and date Participants (no.) /country 16 Indonesia

Regional Integrated Pest Management 17 Aug-9 Oct

Host: 3 2 2 1 1 Host: 8 1 3 1 2 2 26

Philippines Cambodia Lao PDR Philippines Colombia Madagascar Thailand Cambodia Ghana Lao PDR Surinam Tanzania Thailand

EDDA/IRRISTAT in Chlorophyll Meter and Leaf Color Chart Technology 28 Sep-3 Oct IRRISTAT Training 14-16 Oct IRRISTAT Training 19-21 Oct Introduction to GIS Training Course 19-30 Oct Experimental Design and Data Analysis with IRRISTAT 22-30 Oct Presentation Skills Course for Lao Trainers and Scientists 26 Oct-6 Nov

6 Thailand 10 Lao PDR 9 Myanmar 8 Bangladesh

Rice Production Research (Course 6) 6 Oct-27 Nov


17 Lao PDR

Consortium or Network Problem-based Technology Genera16 Indonesia tion for Rainfed Lowland Environments 9-22 Apr
Subtotal 16 10 Iran

Training on Data Management and 17 India Documentation for Genetic Resources 26 Oct-7 Nov Refresher Course on Identification and Management of Rice Tungro Disease for DA Technicians 27 Oct Rice Tungro Disease Identification 28 Oct 29 Oct Subtotal Total 17 Philippines

National Transformation and Molecular Analysis of Transgenic Plants 17-27 Aug

Field Collection and Conservation of Rice Germplasm, 14-19 Sep IRRISTAT Training 16 Sep

27 Myanmar 5 Cambodia

38 Philippines 35 Philippines 215 257

Regional in-country courses are conducted collaboratively with a NARS institution for an international group of trainees. Two regional courses were conducted for 26 trainees from 9 countries in Asia and AfricaIPM by the National Crop Protection Center of the University of the Philippines Los Baos and Rice Production Research course by the Pathum Thani Rice Research Institute in Thailand. National in-country courses are adapted to the specific needs of the requesting country. Twelve national courses were conducted in 9 countries, upgrading the skills of 142 rice scientists. Moreover, a farmers version of one of the courses (Rice Tungro Disease Identification and Management) trained 73 farmers. Network and consortium courses are requested and funded by research networks and consortia. The course on Problem-based Technology Generation for Rainfed Lowland Environments conducted in

Bogor and Pati, Indonesia, for the Rainfed Lowland Rice Research Consortium. IRRI also assists national systems in formulating their own training programs. Training needs assessments were made and training plans were developed for Cambodia and Papua New Guinea.

Program outlook
The Strengthening Partnership with NARS project will continue to support country and regional collaborative projects for strengthening NARS capabilities and collaborative research. The IRRI liaison offices in different countries will help develop new projects and identify funding sources; monitor progress of all collaborative research activities; facilitate in-country training, workshops and conferences, and visits of scientists; and assist in the exchange of information and knowledge among rice

Accelerating the impact of rice research 153

scientists. IRRI-NARS research dialogues and planning meetings will be held in India, Indonesia, Nepal, Philippines, and Sri Lanka in 1999. CREMNET will continue to evaluate the chlorophyll (SPAD) meter, LCC, controlled-release urea, and urea briquette deep placement in four countries. Farmers will be surveyed to gain increased understanding of their knowledge and rationale in fertilizer use on rice in Bangladesh and India. Information and knowledge exchange activities will continue and ways and means of efficiently disseminating information about rice and IRRI will be developed. Web activities will assume even greater importance as IRRI sites are redesigned to make them more attractive and easier to search, while more content is added from IRRI programs, centers, and divisions. Web-based sale of IRRI publications will likely increase as customers become aware of a credit card option as part of a shopping basket concept. IRRI will publish around 20 titles in 1999 and International Rice Research Notes (IRRN) will expand its content under a new editor-in-chief and editorial board. The Riceworld Museum will unveil as many as nine new exhibits, including displays on the CGIAR and IRRI-NARS collaboration, as well as a revamped look at some IRRI milestones of the past four decades. A complete overhaul of library acquisitions is planned with new staff and computer software expected early in 1999. In addition, the Library will extend its activities into Lao PDR, sign an information memorandum of understanding with partners in Korea, and add new online services such as FirstSearch, which provides access to more than 60 databases and document delivery suppliers.

IRRI will maintain about 85 scientists in its degree and postdegree training programs. About 30 scientists should complete their studies by the end of 1999. Fourteen regular and three regional courses for a maximum of 200 scientists are planned. New courses will be designed and developed during 1999: 1) Modern Rice Farming and 2) Transgenic Rice: Production and Deployment with Special Reference to Sheath Blight and Rice Stem Borer Resistance. Collaborative in-country courses for 1999 will include Problem-based Technology Generation for Rice Environments in collaboration with CIAP for rice scientists in Cambodia; Engineering for Rice Agriculture in Bangladesh; Scientific Writing and Presentation and Community-based Natural Resource Management (CBNRM) in Bhutan; Strategic Planning for Effective GO-NGO Collaboration in support of an ongoing project in northeast Thailand; and CBNRM in Myanmar in collaboration with the Southeast Asian Regional Center for Graduate Study and Research in Agriculture (SEARCA). IRRI will improve its training strategy through an in-depth review during 1999 and will seek direction for its distance education initiatives through a meeting of distance and open learning experts. Staff skills on training in the distance mode will be improved through training and teamwork with information and communication technology professionals. Collaboration with universities and other training institutions will be explored and used to retool IRRI staff to enhance quality of training for NARS scientists.


IRRI program report for 1998

Affiliations of collaborating researchers


Institute for Rice, Indonesia. Rice Research Institute, Philippines. 3Zhejiang Agricultural University, China. 4Soil and Water Management Research Institute, India. 5Pathum Thani Rice Research Center, Thailand. 6National Institute for Soils and Fertilizers, Vietnam. 7Cuu Long Delta Rice Research Institute, Vietnam. 8Vietnam Agricultural Science Institute, Vietnam. 9Malaysian Agricultural Research and Development Institute, Malaysia. 10Department of Agriculture, Thailand. 11Narendra Deva University of Agriculture and Technology, India. 12Orissa University of Agriculture and Technology, India. 13Ubon Rice Research Center, Thailand. 14Texas Tech University, USA. 15Nagoya University, Japan. 16University of Adelaide, Australia. 17Cornell University, USA. 18International Food Policy Research Institute, USA. 19Hokkaido University, Japan. 20National Agriculture Research Center, Japan. 21Shikoku National Agricultural Experiment Station, Japan. 22Oklahoma State University, USA. 23North Carolina State University, USA. 24Texas A&M University, USA. 25University of Maryland, USA. 26Institut national de la recherche agronomique, France. 27National Institute of Agrobiological Resources, Japan. 28Agricultural Genetics Institute, Vietnam. 29Harvard Medical School, USA. 30Massachusetts General Hospital, USA. 31Indira Gandhi Agricultural University, India. 32Guangdong Academy of Agricultural Sciences, China. 33West Africa Rice Development Association, Cot d'Ivoire. 34Hokuriku National Agricultural Experimental Station, Japan. 35Centre for Cellular Molecular Biology, India. 36Rice Research Institute, Iran. 37Institute of Plant Sciences, The Swiss Federal Institute of Technology, Switzerland. 38University of Tokyo, Japan. 39University of Hanover, Germany. 40Thai Nguyen University, Vietnam. 41Katholieke Universiteit Leuven, Belgium.

Acceleratingof collaborating researchers 155 Affiliations the impact of rice research

Publications and seminars

Institute publications
Books Allelopathy in rice. 1998. 154 p. IRRI 1997-1998. Biodiversitymaintaining the balance. 1998. 58 p. Pest management of rice farmers in Asia. 1998. 245 p. Rainfed lowland rice: advances in nutrient management research. 1998. 304 p. Sustainability of rice in the global food system. 1998. 404 p. Program report for 1997. 1998. 176 p. Impact of rice research. 1998. 428 p. Advances in hybrid rice technology. 1998. 443 p. Periodicals/serials International rice research notes, vol. 23, nos. 1-3. IRRI discussion paper series., nos. 23-36. Agricultural Engineering Bell MA, Douthwaite B, de Padua D, Rickman JF. 1998. Rising to the challenge: increasing the impact of engineering in Asia. In: Proceedings of the International Agricultural Engineering Conference. Bangkok (Thailand): Asian Institute of Technology. p 1012-1018. Bell MA, Rickman JF, Castro EC Jr., Aclan LB, McNamara J. 1998. Precision land leveling for rice production in Asia. In: Proceedings of the International Agricultural Engineering Conference. Bangkok (Thailand): Asian Institute of Technology. p 257-264.

Borlagdan PC, Douthwaite B, Alihamsyah T, Bell MA. 1998. Rainfed lowland rice crop establishment in Central Java: practices, problems, and opportunities. In: Hsiu-Ying Lu, Jih-Min Sung, Ching Huei Kao, editors. Proceedings of the 3rd Asian Crop Science Conference. Taichung (Taiwan): The Chinese Society of Agronomy. Agronomy, Plant Physiology, and Agroecology Auld BA, Watson AK, Mabbayad MO, Ciotola M, Hetherington SD. 1998. Biocontrol of weeds with fungi in developing nations: the role of bioherbicides. Abstracts. Biological control of weeds workshop. 7th International Congress of Plant Pathology, Edinburgh, UK. Banoc DM, Yamauchi A, Kamoshita A, Wade LJ. 1998. Genotypic variation in developmental plasticity of rice seminal root system under the complex anaerobic-aerobic soil water transitions. Jpn. J. Crop Sci. 67: 422-423. Bronson KF, Cassman KG, Wassmann R, Olk DC, van Noordwijk M, Garrity D. 1998. See Soil and Water Sciences. Cassman KG, Peng S, Olk DC, Ladha JK, Reichardt W, Dobermann A, Singh U. 1998. Opportunities for increased nitrogen-use efficiency from improved resource management in irrigated rice systems. Field Crops Res. 56:7-39. Castella JC, Husson O, Le Quoc Doanh, Ha Dinh Tuan. 1998. Implementing the ecoregional approach in the Red River Basin uplands

Publications and seminars


(Vietnam): the Mountain Agricultural Systems (SAM) Project. In: Proceedings of the Ecoregional Planning Workshop for the Red River Basin, Vietnam, 6-8 Oct 1997, Ministry of Agriculture and Rural Development, Hanoi, Vietnam. Centeno HGS, Sheehy JE. 1997. Simulating potential rice yield at different environments. A case study. In: Manton M, Phelon A, Virji H, editors. Proceedings of the Workshop on Climate Variability, Agricultural Productivity, and Food Security in the Asian Monsoon Region, Bogor, Indonesia. Report No. 2. p 29-34. Centeno HGS, Hammer GL, Sheehy JE, Stone R. 1998. The impacts of ENSO on rice production in the Asia-Pacific region. In: Proceedings of the 23rd Conference of Forest and Agriculture Meteorology, New Mexico, USA. Centeno HGS, Hammer GL, Sheehy JE, Stone R. 1998. The impacts of ENSO on rice production in Central Luzon. In: Proceedings of the Annual Scientific Meeting, Federation of Crop Science Societies of the Philippines, Cebu City, Philippines. De Luna LZ, Watson AK, Paulitz TC. 1998. Infection structures of Curvularia tuberculata and C. oryzae on Cyperaceae weeds. Phytopathology 88:S135. Dobermann A, Cassman KG, Mamaril CP, Sheehy JE. 1998. see Soil and Water Sciences. Edmeades GO, Bolanos J, Banziger M, Ribaut J-M, White JW, Reynolds MP, Lafitte HR. 1998. Improving crop yields under water deficits in the tropics. In: Chopra VL, Singh RB, Varma A, editors. Crop productivity and sustainabilityshaping the future. Proceedings of the 2nd International Crop Science Congress. New Delhi: Oxford and IBH. p 437-451. Fukai S, Kwon YW, Sinclair TR, Wade LJ. 1998. Suggestions for potential authors from editors. Plant Prod. Sci. 1: 300. George T. 1998. Nutrient decision-aids for the transition to high value production systems in erosion-free Asian uplands. In: Proceedings of the. 16th World Congress of Soil Science, 20-26 Aug 1998. Montpellier (France): CIRAD.[in CD-ROM]

Henderson S, von Caemmerer S, Farquhar GD, Wade LJ, Hammer GL. 1998. Correlation of carbon isotope discrimination and transpiration efficiency in genotypes of C4 Sorghum bicolor in glasshouse and field. Aust. J. Plant Physiol. 25: 111-123. Ito O. 1998. Strategy for rice production technologies at IRRI. In: Sustainable agricultural development compatible with environmental conservation in Asia. Japan: Japan International Research Center for Agricultural Sciences. p191-202. Kamoshita A, Fukai S, Muchow RC, Cooper M. 1998. Genotypic variation for grain yield and grain nitrogen concentration among sorghum hybrids under different levels of nitrogen fertiliser and water supply. Aust. J. Agric. Res. 49:737-747. Kamoshita A, Fukai S, Muchow RC, Cooper M. 1998. Physiological processes determining genotypic variation for grain yield and grain nitrogen concentration among sorghum hybrids under different nitrogen fertiliser conditions. I. Hybrids with similar phenology. Aust. J. Agric. Res. 49:1267-1276. Kamoshita A, Fukai S, Muchow RC, Cooper M. 1998. Physiological processes determining genotypic variation for grain yield and grain nitrogen concentration among sorghum hybrids under different nitrogen fertiliser conditions. II. Hybrids with contrasting phenology. Aust. J. Agric. Res. 49:12771286. Katayama K, Ito O, Rao TP. 1998. Seedling characteristics and retention of current photosynthates in leaves in relation to initial growth in pigeon pea (Cajanus cajan L. Millsp.) and cowpea (Vigna sinensis Endl.). Soil Sci. Plant Nutr. 44(3):477-480. Khush GS, Peng S, Virmani SS. 1998. see Plant Breeding, Genetics, and Biochemistry. Kiniry JR, Landivar JA, Witt M, Gerik TJ, Calvero J, Wade LJ. 1998. Radiation-use efficiency response to vapour-pressure deficit for maize and sorghum. Field Crops Res. 56: 265-270. Ladha JK, Kirk GJD, Bennett J, Peng S, Reddy CK, Reddy PM, Singh U. 1998. see Soil and Water Sciences.


IRRI program report for 1998

Lafitte HR. 1998. Research opportunities to improve nutrient-use efficiency in rice cropping systems. Field Crops Res. 56: 223-236. Latore J, Gould P, Mortimer AM. 1998 The spatial dynamics and critical patch size of annual plant populations. J. Theor. Biol. 190:277-285. Masangkay RF, Mabbayad MO, Paulitz TC, Watson. AK. 1998. Host range of Alternaria alternata f. sp. sphenocleae causing leaf blight of Sphenoclea zeylanica. Can. J. Bot. (in press) Mortimer AM. 1998. The need for studies on weed ecology to improve weed management. In: Labrada R, editor. Expert consultation on weed ecology and management. Rome: Food and Agriculture Organization. p 15-22. Mortimer AM. 1998. Weed management and farmer decision makingthe role of ecological studies. In : Ali Abdul Hamid, Lum Keng Yeang, Tosiah Sadi, editors. Integrating science and people in rice pest management. Kuala Lumpur: Malaysian Agricultural Research and Development Institute. p 117-126. Olk DC, Cassman KG, Mahieu N, Randall EW. 1998. Conserved chemical properties of young humic acid fractions in tropical lowland soil under intensive irrigated rice cropping. Eur. J. Soil Sci. 49:337-349. Olofsdotter M. 1998. Allelopathy for weed control in organic farming. In: El Bassam N,. Behl RK, Prochnow B, editors. Sustainable agriculture for food, energy and industrystrategies towards achievement. London: James & James Science Publisher. p 453-465. Peng S, Cassman KG. 1998. Upper thresholds of nitrogen uptake rates and associated nitrogen fertilizer efficiencies in irrigated rice. Agron. J. 90:178-185. Peng S, Laza RC, Khush GS, Sanico AL, Visperas RM, Garcia FV. 1998. Transpiration efficiencies of indica and improved tropical japonica rice grown under irrigated conditions. Euphytica 103:103-108. Ram S, Chauhan RPS, Singh BB, Singh VP. 1998. Response of wet season rice (Oryza sativa) to nitrogen and phosphorus in partially reclaimed sodic soil. Indian J. Agric. Sci. 67(11): 514-517.

Rao TP, Ito O. 1998. Differences in root system morphology and root respiration in relation to nitrogen uptake among six crop species. JARQ 32(2): 97-103. Raybould AF, Mortimer AM, Bullock JM, Gray AJ. 1998. Stress tolerance and weediness. Wareham, Dorset (UK): Institute of Terrestrial Ecology. 162 p. Sanetra CM, Ito O, Virmani S, Vlek PLG. 1998. Remobilization of nitrogen from senescing leaves of pigeon pea (Cajanus cajan (L.) Millsp.): genotypic differences across maturity groups? J. Exp. Bot. 49(322): 853-862. Sheehy JE, Dionora MJA, Mitchell PL, Peng S, Cassman KG, Lemaire G, Williams RL. 1998. Critical nitrogen concentrations: implications for high-yielding rice (Oryza sativa L.) cultivars in the tropics. Field Crops Res. 59: 31-41. Sheehy JE, Mitchell PL, Beerling DJ, Tsukaguchi T, Woodward FI. 1998. Temperature of rice spikelets. Thermal damage and the concept of a thermal burden. Agronomie 18: 449-460. Singh G, Singh VP, Singh OP, Singh RK. 1998. Production potential of various cropping systems in flood-prone areas of Uttar Pradesh. Indian J. Agron. 42 (1): 9-12. Singh U, Ladha JK, Castillo EG, Punzalan G, Tirol-Padre A, Dequeza M. 1998. Genotypic variation in nitrogen use efficiency in medium- and long-duration rice. Field Crops Res. 58: 35-53. Singh VP, Sovyanhadi Y. 1998. Kinetics of phosphate fixation in acid sulfate, iron-toxic and neutral soils. Oryza 35(2): 95-105. Srivastava PC, Gangwar MS, Singh VP. 1998. Adsorption-desorption of Zn in Mollisols and their relationship with the uptake of fertilizer applied zinc by rice. Commun. Soil Sci. Plant Anal. 30: 3-4. Turkelboom F, Trbuil G. 1998. A multiscale approach for on-farm erosion research: application to northern Thailand highlands. In: Penning de Vries FWT, Fahmuddin A, Kerr J, editors. Soil erosion at multiple scales: principles and methods for assessing causes and impacts. Wallingford (UK): CAB International and International Board for Soil Research and Management. p 51-71.

Publications and seminars


Wade LJ. 1998. Research balance, impact delivery and future challenges in crop-soil management systems. Jpn. J. Crop Sci. 67: 310-311. Wade LJ, George T, Ladha JK, Singh U, Bhuiyan SI, Pandey S. 1998. Opportunities to manipulate nutrient by water interactions in rainfed lowland rice systems. Field Crops Res. 56: 93-112. Wade LJ, Quintana L, Amarante ST, Naklang K, Harnpichitvitaya D, Singh AP, Sengar SS, Parihar SS, Singh G, Wihardjaka A, Mazid MA. 1998. Nutrient-water interactions in diverse soils of the rainfed lowland rice ecosystem in Asia. In: CD-ROM Proceedings of the World Congress of Soil Science, 19-26 Aug 1998. Symposium 14, Paper 7. Montpellier (France). Ying J, Peng S, Yang G, Zhou N, Visperas R.M., Cassman KG. 1998. Comparison of high-yield rice in a tropical and subtropical environment. II. Nitrogen accumulation and utilization efficiency. Field Crops Res. 57(1): 85-93. Ying J, Peng S, He Q, Yang H, Yang C, Visperas RM, Cassman KG. 1998. Comparison of high-yield rice in a tropical and subtropical environment. I. Determinants of grain and dry matter yields. Field Crops Res. 57(1): 71-84. Zhang C, Peng S, Bennett J. 1998. Glutamine synthetase of roots and leaves in response to nitrogen application at different growth stages in field-grown rice. J. Plant Nutr. 21(4): 625633. Ziska LH, Moya TB, Wassmann R, Namuco OS, Lantin RS, Aduna JB, Abao E Jr, Bronson KF, Neue HU, Olszyk D. 1998. see Soil and Water Sciences.

Entomology and Plant Pathology Alam SN, Cohen MB. 1998. Detection and analysis of QTLs for resistance to the brown planthopper, Nilaparvata lugens, in a doubledhaploid rice population. Theor. Appl. Genet. 97: 1370-1379. Alam SN, Cohen MB. 1998. Durability of brown planthopper, Nilaparvata lugens, resistance in rice variety IR64 in greenhouse selection studies. Entomol. Exp. Appl. 89:71-78.

Chen XM, Line RF, Leung H. 1998. Genome scanning for resistance-gene analogs in rice, barley, and wheat by high-resolution electrophoresis. Theor. Appl. Genet. 97: 345355. George MLC, Nelson RJ, Zeigler RS, Leung H. 1998. Rapid population analysis of Magnaporthe grisea by using rep-PCR and endogenous repetitive DNA sequences. Phytopathology 88: 223-229. Heong KL. 1998. IPM in developing countries: progress and constraints in rice IPM. In: Zalucki MP, Drew RAI, White GG, editors. Pest managementfuture challenges. Proceedings of the Sixth Australasian Applied Entomological Research Conference, Brisbane, Australia, 29 Sep-2 Oct 1998. Australia: University of Queensland. p 68-77. Heong KL, Schoenly KG. 1998. Impact of insecticides on herbivore-natural enemy communities in tropical rice ecosystems. In: Haskell PT, McEwen P, editors. Ecotoxicology: pesticides and beneficial organisms. London: Chapman and Hall. p 381-403. Heong KL, Escalada MM, Huan,NH, Mai V. 1998. Use of communication media in changing rice farmers pest management in South Vietnam. Crop Prot. 17(5): 413-425. Luo Y, Teng PS, Fabellar NG, TeBeest DO. 1998a. The effects of global temperature change on rice leaf blast epidemics: a simulation study in three agroecological zones. Agric. Ecosyst. Environ. 68: 187-196. Luo Y, Teng PS, Fabellar NG, TeBeest DO. 1998b. Risk analysis of yield losses caused by rice leaf blast associated with temperature changes above and below for five Asian countries. Agric. Ecosyst. Environ. 68: 197-205. Mew TW, Merca SD, Gonzales P, Guevarra J, Huelma C. 1998. Epidemiology of seed-borne fungal diseases. In: Proceedings of the International Workshop on Management of Seed-borne Diseases for Food Production in the 21st Century, 25-27 Mar 1998, Nagoya Congress Center, Japan.


IRRI program report for 1998

Michel VV, Mew TW. 1998. Effect of a soil amendment on the survival of Ralstonia solanacearum in different soils. Phytopathology 88: 300-305. Ona I, Vera Cruz CM, Nelson RJ, Leach JE, Mew TW. 1998. Epidemic development of bacterial blight on rice carrying resistance genes Xa-4, Xa-7, and Xa-10. Plant Dis. 82(12): 13371340. Satapathy MK, Teng PS, Anjaneyulu A. 1998. Rice tungro disease and its management. In: Upadhay K, Mukerji KG, Rajak RL, editors. IPM systems in agriculture. Vol. 3. Cereals. New Delhi, India: Aditya Books Pvt. Ltd. p 105-130. Savary S, Elazegui FA, Teng PS. 1998. Assessing the representativeness of data on yield losses due to rice diseases in tropical Asia. Plant Dis. 82 (6): 705-709. Savary S, Willocquet L, Castilla N. 1998. Effects of some environmental factors on rice sheath blight epidemiology. In: Proceedings of the International Conference of Plant Pathology, Edinburgh, 9-16 Aug 1998. Vol 1. Savary S, Elazegui FA, Willocquet L, Teng PS. 1998. Changing production situations in rice and implications for plant pathology. In: Proceedings of the International Conference of Plant Pathology, Edinburgh, 9-16 Aug 1998, Vol 1. Schoenly KG, Justo HD Jr, Barrion AT, Harris MK, Bottrell DG. 1998. Analysis of invertebrate biodiversity in a Philippine farmers irrigated rice field. Environ. Entomol. 27(5): 1125-1136. Shi Z, Christian D, Leung, H. 1998. Interactions between spore morphogenetic mutations affect cell types, sporulation, and pathogenesis in Magnaporthe grisea. Mol. Plant Microbe Interact. 11: 199-207. Sigsgaard L. 1998. Natural control of Helicoverpa armigera in sorghum-pigeon pea intercropping. In: Proceedings of the Sixth Australasian Applied Entomological Research Conference, University of Queensland, Brisbane, Australia, 29 Sep-2 Oct 1998. 340 p. Teng PS, Batchelor WD, Pinnschmidt HO, Wilkerson G. 1998. Simulation of pest effects

on crops using pest-crop models: the potential for decision support. In: Tsuji G, Hoogenboom G, Thornton PK, editors. Understanding options for agricultural production. Dordrecht, The Netherlands: Kluwer Academic Publishing. p 225-270. Theunis W, Aguda RM, Cruz WT, Decock C, Peferoen M, Lambert B, Bottrell DG, Gould FL, Litsinger JA, Cohen MB. 1998. Bacillus thuringiensis isolates from the Philippines: habitat distribution, -endotoxin diversity, and toxicity to rice stem borers (Lepidoptera: Pyralidae). Bull. Entomol. Res. 88: 335-342. Tiongco ER, Chancellor TCB, Villareal S, Magbanua M, Teng PS. 1998. Studies on the effectiveness of roguing as a means of controlling rice tungro virus disease. J. Plant Prot. Trop. 11: 45-52. Tu J, Ona I, Zhang Q, Mew TW, Khush GS, Datta SK. 1998. see Plant Breeding, Genetics, and Biochemistry. Way MJ, Islam Z, Heong KL, Joshi RC. 1998. Ants in tropical irrigated rice: distribution and abundance, especially of Solenopsis geminata (Hymenoptera: Formicidae). Bull. Entomol. Res. 88: 467-476. Xie GL, Mew TW. 1998. A leaf inoculation method for detection of Xanthomonas oryzae pv. oryzicola from rice seed. Plant Dis. 82: 1007-1011. Yu X, Hu C, Heong KL. 1998. Parasitization and preference characteristics of egg parasitoids from various habitats to homopterans. Acta Entomol. Sin. 41: 41-47. Zeigler RS. 1998. Recombination in Magnaporthe grisea. Annu. Rev. Phytopathol. 36: 249-275. Genetic Resources Center Bellon MR, Pham JL, Sebastian LS, Francisco SR, Loresto GC, Erasga DS, Sanchez P, Calibo MA, Abrigo G, Quilloy SE. 1998. Farmers perceptions of varietal diversity: implications for on-farm conservation of rice. In: Smale M, editor. Farmers, gene banks and crop breeding. Dordrecht, The Netherlands: Kluwer Academic Publishing. p 95-108.

Publications and seminars


Juliano AB, Naredo MEB, Jackson MT. 1998. Taxonomic status of Oryza glumaepatula Steud. I. Comparative morphological studies of New World diploids and Asian AA genome species. Genet. Res. Crop Evol. 45: 197-203. Lu BR. 1998. Diversity of the rice gene pool and its sustainable utilization. In: Zhang Aolur, Wu Sugong, editors. Floristic characteristics and diversity of East Asian plants. Beijing: China Higher Education Press and Berlin: Springer-Verlag. p 454-460. Lu BR, Naredo MEB, Juliano AB, Jackson MT. 1998. Taxonomic status of Oryza glumaepatula Steud. III. Assessment of genomic affinity among AA genome species from the New World, Asia and Australia. Genet. Res. Crop Evol. 45: 215-223. Lu BR, Naredo MEB, Juliano AB. 1998. A biosystematic study of the genus Oryza (Poaceae). Abstract of monocots II. Sydney, Australia. p 74. Lu BR. 1998. Diversity of rice genetic resources and its utilization and conservation. Chin. Biodiversity 6 (1): 63-72. Mew TW, Merca SD, Gonzales P. Guevarra J, Huelma C. 1998. see Entomology and Plant Pathology. Morin SR, Pham JL, Sebastian LS, Abrigo G, Erasga DS, Bellon MR, Calibo MA, Sanchez P. 1998. The role of indigenous technical knowledge in on-farm conservation of rice genetic resources in Cagayan Valley, Philippines. In: People, earth and culture. Los Baos, Laguna, Philippines: Philippine Council for Agriculture, Forestry and Natural Resources Research and Development. PCARRD-NCCA Book Series No. 165. p 137-150. Naredo ME, Juliano AB, Lu BR, Jackson MT. 1998. Taxonomic status of Oryza glumaepatula Steud. II. Hybridization between New World diploids and AA genome species from Asia and Australia. Genet. Res. Crop Evol. 45: 205-214. Pham JL, Sebastian LS, Sanchez P, Calibo MA, Quilloy SE, Bellon MR, Francisco SR, Erasga DS, Abrigo G, Loresto GC. 1998. Genetic data and analysis in the Philippines on-farm. In: Jarvis DI, Hodgkin T, editors.

Strengthening the scientific basis of in situ conservation of agricultural biodiversity onfarm. Options for data collecting and analysis. Proceedings of a workshop to develop tools and procedures for in situ conservation onfarm, 25-29 Aug 1997. Rome, Italy: International Plant Genetic Resources Institute. p 34. Zhang SZ, Lu BR, Hong DY. 1998. In situ hybridization and its application in studies on Oryza. Acta Phytol. Sin. 36(1): 87-96. Zhang ZY, Wen J, Lu BR. 1998. The structural features of leaf epidermis in Oryza and their systematic significance. Acta Phytol. Sin. 36(1): 8-18. Zhu J, Gale MD, Quarrie S, Jackson, MT, Bryan GJ. 1998. AFLP markers for the study of rice biodiversity. Theor. Appl. Genet. 96: 602611. Information Center Hettel GP, Hettel AA. 1998. Preserving Bhutans biological heritage. Tashi Delek 3(1): 28-35. Wallace I. 1998. Delivering information to rice scientists around the world. Inf. Dev. 14(4): 198-202. Plant Breeding, Genetics, and Biochemistry Alam MF, Datta K, Abrigo E, Vasquez A, Senadhira D, Datta SK. 1998. Production of transgenic deepwater indica rice plants expressing a synthetic Bacillus thuringiensis cryIA(b) gene with enhanced resistance to yellow stem borer. Plant Sci. 135: 25-30. Bennett J. 1998. Balancing needs for productivity and sustainability: genetic engineering of rice at IRRI. In: Managing biotechnology in a time of transition. The Hague: International Service for National Agricultural Research. 15 p. Bentota AP, Senadhira D, Lawrence MJ. 1998. Quantitative genetics of rice. III. The potential of a pair of new plant type crosses. Field Crops Res. 55: 267-273. Brar DS, McNally RE, Mendoza R, Senadhira D, Jones M, Khush GS. 1998. Gene transfer for tolerance to biotic and abiotic stresses from wild species into rice. In: Proceedings of the International Workshop on Breeding and


IRRI program report for 1998

Biotechnology for Environmental Stress in Rice. Sapporo, Japan. p 120-125. Datta K, Tu J, Oliva N, Muthukrishnan S, Datta SK. 1998. Transgenic rice for enhanced resistance to Rhizoctonia solani causing sheath blight disease. In: Larkin PJ, editor. Proceedings of the 4th Asia-Pacific Conference on Agricultural Biotechnology. Darwin, Australia. p 38-40. Datta K, Vasquez A, Tu J, Torrizo L, Alam MF, Oliva N, Abrigo E, Khush GS, Datta SK. 1998. Constitutive and tissue-specific differential expression of the cryIA(b) gene in transgenic rice plants conferring resistance to rice insect pest. Theor. Appl. Genet. 97: 20-30. Delos Reyes BG, Khush GS, Brar DS. 1998. Chromosomal location of eight isozyme loci in rice using primary trisomics and monosomic alien addition lines. J. Hered. 89:164-168. Fahim M, Dhanapala MP, Senadhira D, Lawrence MJ. 1998. Quantitative genetics of rice. II. A comparison of the efficiency of four breeding methods Field Crops Res. 55: 257-266. Hiroshi K. 1998. Development and utilization of hybrid rice in Malaysia [in Japanese, with English summary]. JIRCAS Res. Highlights 5: 37-38. Jain SM, Brar DS, Ahloowalia BS, editors. 1998. Somaclonal variation and induced mutations in crop improvement. Dordercht, The Netherlands: Kluwer Academic Publishers. 615 p. Julfiquar AW, Virmani SS. 1997. Inheritance of wide compatibility trait in rice (Oryza sativa L.). Bangladesh J. Life Sci. 9(2): 1-7. Khush GS. 1998. Strategies for increasing crop productivity. In: Chopra VL, Singh RB Anupam Verma, editors. Crop productivity and sustainability shaping the future. Proceedings of the 2nd International Crop Science Congress. New Delhi: Oxford and IBH Publishing Co. Pvt. Ltd.. p 19-43. Khush GS. 1998. Innovative approaches for improving the yield and grain quality of rice. In: Lu HY, Sung JM, Kao CH, editors. Asian Crop Science. Proceedings of the 3rd Asian Crop Science Conference. Taichung, Taiwan: Chinese Society of Agronomy. p 436-450.

Khush GS, Baenziger PS. 1998. Crop improvement: emerging trends in rice and wheat. In: Chopra VL, Singh RB Anupam Verma, editors. Crop productivity and sustainability shaping the future. Proceedings of the 2nd International Crop Science Congress. New Delhi: Oxford and IBH Publishing Co. Pvt. Ltd. p 113-125. Khush GS, Brar DS. 1998. The application of biotechnology to rice. In: Eves CL, Bedford BM, editors. Agricultural biotechnology in international development. Wallingford, UK: CABI Publishing. p 97-121. Khush GS, Sarkarung S. 1998. New plant type and breeding strategies for rainfed lowland rice. In: Rainfed rice for sustainable food security. Cuttack, Orissa, India: Central Rice Research Institute. p 44-52. Khush GS, Peng SB, Virmani SS. 1998. Improving yield potential by modifying plant type and exploiting heterosis. In: Waterlow JC, Armstrong DG, Fowden L, Riley R, editors. Feeding a world population of more than a billion people: a challenge to science. New York, USA: Oxford University Press. p 150-170. Ladha JK, Kirk G, Bennett J, Peng S, Reddy CK, Singh U. 1998. see Soil and Water Sciences. Lawrence MJ, Senadhira D. 1998. Quantitative genetics of rice. IV. A breeding strategy. Field Crops Res. 55:275-281. Li ZK, Pinson SRM, Paterson AH, Stansel JW. 1998. Genetic dissection of the source-sink relationship affecting fecundity and yield in rice. Mol. Breed. 4: 419-426. Ling DH, Tao LZ, Ma ZR, Zhang SP, Datta SK. 1998. Engineered male sterile transgenic plants of rice (Oryza sativa L.) with ps1barnase gene transformation by particle bombardment. Acta Genet. Sin. 25(5): 433-442. Peng SB, Laza RC, Khush GS, Sanico AL, Visperas RM, Garcia FV. 1998. see Agronomy, Plant Physiology, and Agroecology.

Publications and seminars


Perera ALT, Fahim M, Sriyoheswara S, Dhanapala MP, Senadhira D, Lawrence MJ. 1998. Quantitative genetics of rice. I. Evidence of unexploited genetic variation for yield and other quantitative characters in modern indica cultivars. Field Crops Res. 55: 245-256. Reddy PM, Ladha JK, Ramos MC, Hernandez R, Torrizo L, Datta SK, Datta K. 1998. see Soil and Water Sciences. Sanchez AC, Khush GS. 1998. Inheritance and linkage relationships of twenty-one genes in rice, Oryza sativa L. SABRAO J. 30: 51-60. Singh S, Singh ON, Singh RK, Sarkarung S. 1998. A shuttle breeding approach to rice improvement for rainfed lowland ecosystem in eastern India. In: Sustainable agriculture for food, energy and industry. London: James and James Science Publishers Ltd. p 150-155. Subudhi PK, Nandi S, Casal C, Virmani SS, Huang N. 1998. Classification of rice germplasm: III. High-resolution fingerprinting cytoplasmic genetic male-sterile (CMS) lines with AFLP. Theor. Appl. Genet. 96: 941-949. Subudhi PK, Virmani SS, Huang N. 1998. A TGMS-linked nuclear DNA marker as originated from the mitochondrial genome in rice (Oryza sativa L.). Heredity 80: 285-292. Subudhi PK, Borkakati RP, Virmani SS, Huang N. 1997. Molecular mapping of thermosensitive genetic male sterility gene in rice using bulked segregant analysis. Genome 40: 188-194. Tu J, Datta K, Alam MF, Fan Y, Khush GS, Datta SK. 1998. Expression and function of a hybrid Bt toxin gene in transgenic rice conferring resistance to insect pests. Plant Biotechnol. 15: 183-191. Tu J, Ona I, Zhang Q, Mew TW, Khush GS, Datta SK. 1998. Transgenic rice variety IR72 with Xa21 is resistant to bacterial blight. Theor. Appl. Genet. 7: 31-36. Virmani SS, Kumar I. 1997. Hybrid rice in Asia Pacific Region. Asian Seed 97 1(14): 1-37. Xu J, Constantino SV, Magpantay G, Bennett J, Sarkarung S, Huang N. 1998. Classification of rice germplasm. II. Discrimination of indica from japonica via analysis of amplicon length polymorphisms. Plant Cell Rep. 17: 640-645.

Yang D, Sanchez A, Khush GS, Zhu Y, Huang N. 1998. Construction of a BAC contig containing xa5 locus in rice. Theor. Appl. Genet. 97:1120-1124. Zhang CF, Peng SB, Bennett J. 1998. see Agronomy, Plant Physiology, and Agroecology. Social Sciences Dawe D. 1998. Re-energizing the Green Revolution in rice. Am. J. Agric. Econ. 80(5): 948-953. Hossain M, Diaz CP. 1997. Reaching the poor with effective microcredit: evaluation of a Grameen bank replication in the Philippines. Philipp. Sociol. Rev. 45(1-4): 89-119. Kam SP, Minh VQ, Tuong TP, Hoanh CT, Liew SC, Chen P. 1998. Remote sensing and GIS approaches to studying changes in rice cropping systems in the Mekong River Delta, Vietnam. In: Proceedings of the Fourth Seminar on GIS and Developing Countries, GISDECO98, Pretoria, South Africa. Kam SP, Hoanh CT. 1998. Implementing GIS in a decision support system for analyzing the balance between rice supply and demand. In: Proceedings of the Fourth Seminar on GIS and Developing Countries GISDECO98, Pretoria, South Africa. Liew SC, Kam SP, Tuong TP, Chen P, Minh VQ, Lim H. 1998. Application of multi-temporal ERS-2 synthetic aperture radar in delineating rice cropping systems in the Mekong River Delta, Vietnam. IEEE TGARSS IGARSS97 Special issue. Luo Y, Teng PS, Fabellar NG, TeBeest DO. 1998a,b. see Entomology and Plant Pathology. Palis FG. 1998. Changing farmers perceptions and practices: the case of insect pest control in Central Luzon, Philippines. Crop Prot. 17(7): 599-607. Price LL, Palis FG. 1998. Transformations in entitlement: land ownership and farming culture in Vietnam. Cult. Agric. 20(1): 12-20. Pandey S, Minh D. 1998. A socio-economic analysis of rice production systems in the upland of northern Vietnam. Agric. Ecosyst. Environ. 70: 249-258.


IRRI program report for 1998

Pandey S. 1997. Rainfed lowland rice research: challenges and priorities for the 21st century. In: Breeding strategies for rainfed lowland rice in drought-prone environments. ACIAR Proceedings No. 77. Australia: Australian Centre for International Agricultural Research. p.1-12. Pandey S, Lapar L. 1998. A microeconomic analysis of adoption of contour hedgerows in the Philippine uplands. In: Penning de Vries FWT, Angus F, Kerr J, editors. Soil erosion at multiple scales: principles and methods for assessing causes and impacts. Wallingford,UK: CAB International and International Board for Soil Research and Management. p 83-98. Pandey S, Lapar L, Waibel H. 1998. Factors determining the adoption of soil conservation practices: some implications for sustainability. In: Proceedings of the International Conference on Sustainable Agriculture for Food, Energy and Industry. London: James & James. p 270-275. Paris T. 1998. Technology and policy needs of poor women in Asian rice farming. Gender Technol. Dev. 2(2): 188-218. Pingali PL, Hossain M, Pandey S, Price LL. 1998. Economics of nutrient management in Asian rice systems. Field Crops Res. 56:157-176. Wade LJ, George T, Ladha JK, Singh U, Bhuiyan SI, Pandey S. 1998. see Agronomy, Plant Physiology, and Agroecology. Widawsky D, Rozelle S, Songqing J, Huang J. 1998. Pesticide productivity, host-plant resistance and productivity in China. Agric. Econ. 19: 203-217. Soil and Water Sciences Badaruddin M, Meisner CA, Razzaue MA, Timsina J, Bronson K, Razu RA, Hobbs P. 1998. Trends in intensive rice-wheat systems of South Asia. In: Agronomy abstracts. Madison, WI: American Society of Agronomy. p 54. Bijay-Singh, Bronson KF, Yadvinder-Singh, Hussain F. 1998. Chlorophyll meter-based nitrogen management for rice in Asia. In: Agronomy abstracts. Madison, WI: American Society of Agronomy. p 307.

Bronson KF, Cassman KG, Wassmann R, Olk DC, van Noordwijk M, Garrity D. 1998. Soil carbon dynamics in different cropping systems in principal ecoregions of Asia. In: Lal R, Kimble JM, Follett RF, Stewart BA, editors. Management of carbon sequestration in soil. Boca Raton, Florida: CRC Press. p 35-57. Bronson KF, Fillery IRP. 1998. Fate of nitrogen15-labelled urea applied to wheat on a waterlogged texture-contrast soil. Nutr. Cyc. Agroecosyst. 51(2): 175-183. Cassman KG, Peng S, Olk DC, Ladha JK, Reichardt W, Dobermann A, Singh U. 1998. see Agronomy, Plant Physiology, and Agroecology. Clement A, Ladha JK, Chalifour FP. 1998. Nitrogen dynamics of various green manure species and the relationship to lowland rice production. Agron. J. 90: 149-154. Damgaard LR, Revsbech NP, Reichardt W. 1998. Use of an oxygen-insensitive microscale biosensor for methane to measure methane concentration profiles in a rice paddy. Appl. Environ. Microbiol. 64: 864-870. Dobermann A, Adviento MAA, Pampolino MF, Nagarajan R, Stalin P, Skogley EO. 1998. Opportunities for in situ soil testing in irrigated rice. In: Proceedings of the 16th World Congress of Soil Science. Montpellier: International Society of Soil Science and Centre de cooperation internationale en recherche agronomique pour le developpement. Dobermann A, Cassman KG, Mamaril CP, Sheehy JE. 1998. Management of phosphorus, potassium and sulfur in intensive, irrigated lowland rice. Field Crops Res. 56: 113-138. George T, Buresh RJ, Ladha JK, Punzalan G. 1998. Recycling in situ legume-fixed and soil N in tropical lowland rice. Agron. J. 90: 429-437. Kirk GJD, George T, Courtois B, Senadhira D. 1998. Opportunities to improve phosphorus efficiency and soil fertility in rainfed lowland and upland rice ecosystems. Field Crops Res. 56: 73-92.

Publications and seminars


Kirk GJD, Santos EE, Findenegg GR. 1998. Phosphate solubilization by organic acid excretion. In: Proceedings of the 16th Congress of the International Society of Soil Science. Montpellier: International Society of Soil Science. Kronzucker HJ, Kirk GJD, Siddiqi MY, Glass ADM. 1998. Effects of hypoxia on 13NH + 4 fluxes in rice roots: kinetics and compartmental analysis. Plant Physiol. 116: 581-587. Ladha JK, Kirk GJD, Bennett J, Peng S, Reddy CK, Reddy PM, Singh U. 1998. Opportunities for increased nitrogen use efficiency from improved rice germplasm in lowland rice ecosystems. Field Crops Res. 56: 41-71. Ladha JK, Padre AT, Punzalan GC, Castillo EG, Singh U, Reddy CK. 1998. Non-destructive estimation of shoot nitrogen in different rice genotypes. Agron. J. 90: 33-40. Minh LQ, Tuong TP, Van Mensvoort MEF, Bouma J. 1998. Soil and water table management effects on aluminum dynamics of an acid sulphate soil. Agric. Ecosyst. Environ. 68(3): 255-262. Razafinjara AL, Kirk GJD. 1998. Interactions between silicon and ammonium-versus-nitrate nutrition in rice. In: Proceedings of the 16th Congress of the International Society of Soil Science. Montpellier: International Society of Soil Science. Reddy PM, Ladha JK, Ramos MC, Hernandez R, Torrizo L, Datta SK, Datta K. 1998. Nod factors activate MtENOD12 gene expression in rice. Plant J. 14: 693-702. Reddy PM, Kouchi H, Ladha JK. 1998. Isolation, analysis and expression of homologues of the soybean early nodulin gene GmENOD93 (GMN93) from rice. Biochem. Biophys. Acta 91214: 1-7. Reichardt W. 1998. Sustainability based on microbe-mediated nutrient-supplying capacities in tropical rice systems, In: Bassam NL, Behl RK, Prochnow B, editors. Sustainable agriculture for food, energy, and industry. Strategies towards achievement. Vol. 1. London, UK: James & James Science Publ. Ltd. p 466-469.

Sethunathan N, Neue HU, Parashar DC, Wassmann R, Rao VR, Adhya TK, Ramakrishnan B. 1998. Greenhouse effect and mitigation options. In: Mohanty SK et al, editors. Rainfed rice for sustainable food security. Cuttack, India. p 409-424. Shrestha RK, Ladha JK. 1998. Nitrate in groundwater and integration of nitrogen-catch crop in intensive rice-based system. Soil Sci. Soc. Am. J. 62: 1610-1619. Singh U, Ladha JK, Castillo EG, Punzalan G, Tirol-Padre A, Dequeza M. 1998. see Agronomy, Plant Physiology, and Agroecology. Tuong TP, Bhuiyan SI, Guerra LC, Barker R. 1998. Technology and management practices for increasing water productivity in rice-based systems: growing more rice with less water. In: Water in 2000efficient operation and management of irrigation systems. Proceedings of the 1998 World Day Symposium, Seoul, Korea, 20 Mar 1998. Seoul: Korean National Committee on Irrigation and Drainage and Rural Development Corporation. p 37-88. Tuong TP, Minh LQ, Ni DV, Van Mensvoort MEF. 1998. Reducing acid pollution from reclaimed acid sulphate soils: experiences from the Mekong Delta, Vietnam. In: Pereira LS, Gowing J, editors. Water and the environment, innovation issues in irrigation and drainage. Selected papers from the 1st Interregional Conference on Environment-Water: Innovative Issues in Irrigation and Drainage, Sep 1998, Lisbon, Portugal. Lisbon: Portuguese National Committee of the International Commission on Irrigation and Drainage. p 75-83. Wade L, Bhuiyan SI, Dobermann A, George T, Kirk GJD, Kondo M, Ladha J, Pandey S, Singh U, Tuong TP, Zeigler RS. 1998. see Agronomy, Plant Physiology, and Agroecology. Wassmann R, Neue HU, Bueno C, Lantin RS, Alberto MCR, Buendia LV, Bronson K, Papen H, Rennenberg H. 1998. Inherent properties of rice soils determining methane production potentials. Plant Soil 203: 227237.


IRRI program report for 1998

Ziska LH, Moya TB, Wassmann R, Namuco OS, Lantin RS, Aduna JB, Abao E Jr, Bronson KF, Neue HU, Olszyk D. 1998. Long-term growth at elevated carbon dioxide stimulates methane emission in tropical paddy rice. Global Change Biol. 4: 657-665.

Rice research seminars

Host plant resistance to the brown planthopper: where do we go from here? Dr. M. B. Cohen. Advances in development and use of hybrid rice technology for shifting the yield frontier. Dr. S. S. Virmani. Communicating pest management to millionsa case study in the use of printed materials and radio in Vietnam. Dr. K. L. Heong. Ok, I know its fun to surf buthow can I use the internet for real work?: exploiting the potential of the internet for communication, information dissemination, and training. Dr. R. Raab and Mr. I. Moore. Genetic studies on blast resistance in rice: IRRIJapanese government collaborative project. Mr. T. Imbe. Molecular mapping and marker-aided backcrossing of two bacterial blight resistance genes to the new plant type of rice. Dr. A. Sanchez. Biodiversity, discovery, molecular screening, and patenting. Dr. L. Lange, senior principal scientist, Novo Nordisk A/S, Bagsvaerd, Denmark and member, IRRI Board of Trustees. Functions of the organic matter phase in wetland rice soils: a promising target for soil quality assessments? Dr. W. Reichardt. Novel rapid methods for genomic sequencing. Dr. W. Szybalski, professor of oncology, McArdle Laboratory for Cancer Research, University of Wisconsin Medical School, Madison, Wisconsin, USA, and founder/ honorary editor (and former editor-in-chief) of the journal Gene. The database for extrapolating methane emissions from rice fields. Dr. R. Wassmann. Managing the worlds largest collection of rice genetic resources. Dr. M. Jackson.

Why we need the International Network for Genetic Evaluation of Rice (INGER). Dr. S.-W. Ahn. Genetic conservation: the role of rice farmers. Dr. J.-L. Pham. Rice variety classification among rice farmers in the Cagayan Valley, Philippines: its utility for on-farm conservation. Dr. S. Morin. Understanding the biosystematic relationships of wild rices. Dr. B. Lu. Disease resistance, rice mutants, and functional genomics. Dr. H. Leung. The ecology of plant virus disease. Dr. M. Thresh, chairman, Virus Epidemiology Committee, International Society of Plant Pathologists, UK. Plant breeding perspectives: balanced integration of approaches. Dr. A. Ashri, Jacob and Rachel Liss professor of agronomy, Faculty of Agriculture, The Hebrew University of Jerusalem, Israel. Operationalizing the ecoregional approach in the Uplands of the Red River Basin. Dr. J-C. Castella. Juggling science and management: seven years in the IRRI circus. Dr. R.S. Zeigler.

Division seminars
Agronomy, Plant Physiology, and Agroecology Weeds and crop production systemsa New Zealand perspective. Dr. K. Harrington, senior lecturer, Massey University, New Zealand. Methods in allelochemical discovery. Dr. A. Rimando, chemist, U.S. Department of Agriculture Modeling integrated weed management in directand water-seeded rice agroecosystems. Dr. B.P. Caton. A multiscale approach to on-farm erosion research: application to north Thailand highlands. Dr. G. Trbuil. Allelopathy in rice. Dr. M. Olofsdotter. Nitrogen availability in lowland rice soils. I. New research tools. II. Update on the rice-upland crop rotation field trial. Dr. D. Olk and Ms. M. Samson.

Publications and seminars


Plant hormones and their application in agricultural production. Prof. H. Saka, University Farm, Tokyo University, Japan. Ethylene-induced shoot elongation: a mechanism to avoid flooding stress. Prof. R. Voesenek, University of Nijmegen, The Netherlands. Photosynthate partitioning in rice and other studies of agro-environmental relevance. Prof. N.N. Barthakur, McGill University, Canada. Water stress and productivity: expansive growth of roots vs leaves, radiation interception in relation to canopy photosynthesis, and reproduction and harvest index. Prof. T.C. Hsiao, University of California at Davis, USA. Plant breeding strategies for rainfed lowland rice in northeast Thailand and Laos. Dr. S. Fukai, University of Queensland, Australia. Source-sink relations and grain filling of rice varieties. Dr. Jianchang Yang. Entomology and Plant Pathology Bt rice for control of yellow stem borer: initial impressions. Dr. J.S. Bentur. Evaluation of a cryIAb-transformed Iranian rice variety against lepidopterous pests and crylAb resistance development in striped stem borer. Dr. F. Alinia. Prey-mediated and direct effects of Bt proteins on immature Chrysoperla carnea. Dr. A. Hilbeck, Swiss Federal Research Station for Agroecology and Agriculture. Engineering resistance against bacterial blight of rice. Dr. Wen Yuan Song, University of California at Davis, USA. Plant Breeding, Genetics, and Biochemistry Comprehensive applications of DNA markers to rice improvement and genome research future research focus of GML. Dr. Zhikang Li. Prospects and challenges of commercializing hybrid rice technology through private sector in India. Mr. I. Kumar, general manager, Hybrid Rice International, Hyderabad, India.

Mapping genes controlling drought resistance in rice: progress, problems, and prospects. Dr. A. Price, lecturer, University of Aberdeen, United Kingdom. Doubled haploid production in wheat. Dr. N. Darvey, senior lecturer, Plant Breeding Institute, University of Sydney at Cobbity, Camden, Australia. The CGIAR Micronutrient Project: progress in breeding nutritionally superior staple. Dr. R. Graham, visiting scientist and professor, Plant Science University of Adelaide, and scientific coordinator, CGIAR Micronutrient Project. Detection of single-copy genes by fluorescent in situ hybridization (fish) and measurement of a copy number by extended DNA fiber fish. Dr. K. Fukui, Rice Genetic Engineering Laboratory, Hokuriku National Agricultural Experiment Station, Joetsu, Japan. New rice for Africa. Dr. M. Jones, plant breeder, West Africa Rice Development Association, Bouak, West Africa. Induced mutations in breedingmodification and improvement of industrial crops. Dr. A. Ashri, Jacob and Rachel Liss professor of agronomy, The Hebrew University of Jerusalem, Rehovot, Israel. Projects on genetic transformation at the Center for Applied Plant Molecular Biology at the University of Hamburg. Prof. H. Lorz, Center for Applied Plant Molecular Biology, University of Hamburg, Germany. Approaching hidden cells: egg cells, zygotes, and early embryogenesis of wheat and barley. Prof. H. Lorz, Center for Applied Plant Molecular Biology, University of Hamburg, Germany. Plant biotechnology and molecular breeding in Germanyproblems and prospects. Prof. H. Lorz, Center for Applied Plant Molecular Biology, University of Hamburg, Germany. Social Sciences Hybrid rice in Indias food economy of the 21st century. Dr. J. Janaiah, Directorate of Rice Research, Hyderabad, India.


IRRI program report for 1998

Peoples strategies of survival due to environmental changes, changes in resource use, and increase in population in two villages in the uplands of north Vietnam. Mr. C. Alther. Effect of technology transfer program participation on small farmers in Chile. Dr. C. Edmonds. Should the Philippines import or produce rice domestically? (A study of comparative advantage in rice production). Dr. J. Estudillo and Dr. M. Hossain. Farmers field school: IPM knowledge and farmers efficiency: a case study in Iloilo, Philippines. Dr. A. Rola, director, Institute of Strategic Planning and Policy Studies, College of Public Affairs, University of the Philippines Los Baos. A socioeconomic study of land use and crop production systems in Bac Yen District, uplands of northern Vietnam. Mr. T. Blohm, collaborative research fellow, University of Hanover.

Soil and Water Sciences On-farm evaluation of crop productivity and soil nitrogen-supplying capacity in a rice-wheat cropping system in Nepal. Dr. C. Adhikari. Agriculture and fertilizer scenario in India. Dr. V. Kumar. The role of rice roots in methane production and emission. Mr. Lu. Using crop simulation models to determine fertilizer applications in precision agriculture. Dr. R. Matthews. Physical and chemical characterization of soil organic matter. S. Suhi. How to use ASL. Ms. B. Mandac. Iodine, rice, and health. Dr. R. Graham. Modeling of N mineralization and immobilization in a flooded peat soil. Dr. P. Bloom. Spatio-temporal variability of acid sulfate soils in the plane of reeds: impact of soil properties, crop husbandry, and water management on the growth and yield of rice in relation to macro-topography. Dr. O. Husson, CIRAD Vietnam.

Publications and seminars