| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a International Rice Research Institute (IRRI), MCPO, P.O. Box 3127, 1271 Makati City, Philippines
b IANR, Dep. of Agronomy, Univ. of Nebraska, Lincoln, NE 68583-0915 USA
adobermann2{at}unl.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionGeneral discussionREFERENCES |
|---|
Abbreviations: AEN, agronomic efficiency of applied fertilizer N • DAT, days after transplanting • DS, dry season (January–April) • EWS, early wet season (May–August) • IEN, internal efficiency of N • INS, indigenous N supply • LTCCE, long-term continuous cropping experiment • LWS, late wet season (September–December) • PEN, physiological efficiency of applied N • PI, panicle initiation • REN, recovery efficiency of applied N • WS, wet season
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionGeneral discussionREFERENCES |
|---|
Yield declines appear to be less common at sites with double cropping of rice and yield levels of 4 to 7 Mg ha-1 (Dawe et al., 2000), a cropping system found on about 14 million ha of land area in Asia. The LTCCE, however, serves the purpose of studying the sustainability of intensive rice cropping at its greatest level of crop intensification so that trends observed in this system may be early warning indicators of what could happen in the future in farmers' fields. The objective of the LTCCE is to achieve maximum and stable annual grain production on a sustainable basis in relation to N fertilizer use. Therefore, starting in 1991, significant changes in crop management were made to reverse the yield decline and to achieve yield levels that approach the yield potential of modern rice varieties. In this paper, we attempt to identify the causes of recent yield increases in the LTCCE after a period of decline and examine how these factors relate to the previous hypotheses about causes of yield decline. This analysis is made difficult by the simultaneous implementation of several changes. We will focus on the following questions:
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionGeneral discussionREFERENCES |
|---|
Since January 1963, the 1-ha field used for the present LTCCE has been under continuous rice cultivation. From 1963 to 1968, 13 rice crops were grown in the same field at an average cropping intensity of 2.5 crops per year (IRRI, 1967). We only report data for the period after 1968, because varieties grown from 1963 to 1968 had variable yield potential, and cropping practices varied. Variety `IR8' was grown in each crop cycle starting with the 1966 DS. Other varieties were changed regularly to include the highest-yielding, most disease- and insect-resistant germplasm available at each point in time, but all varieties or breeding lines grown from 1968 to 1991 were high-yielding, semidwarf indica varieties with a similar harvest index and yield potential.
From 1968 to 1991, the experiment was conducted in a split-split-plot randomized complete block design with N rates as main plots, six varieties or elite breeding lines as subplots, and broadcast sowing or transplanting as crop establishment methods (sub-subplots). Nitrogen main plots (25 by 25 m) were surrounded by permanent bunds to minimize soil, fertilizer, and floodwater movement. Three rice crops were grown each year (dry season, January–April; early wet season, May–August; late wet season, September–December). In each crop cycle, the soil remains water-saturated during the first 14 days after transplanting (DAT). Thereafter, a floodwater depth of 5 to 10 cm is maintained until the field is drained prior to harvest. Reflooding of the field for land preparation starts shortly after harvest, so that to our knowledge, during 1968 to 1991 the soil rarely dried.
Rates and application schemes of fertilizer nutrients in the LTCCE were initially selected to achieve maximum yield based on results of field experiments conducted in the 1960s (Cassman et al., 1995). Nitrogen rates were 0 (N0), 50 (N1), 100 (N2), and 150 (N3) kg N ha-1 in the DS and 0, 30, 60, and 90 kg N ha-1 in the wet season (WS), except in the 1968 DS and the WS crops grown from 1968 to 1974 (Table 1) . Typically, two-thirds of the total amount of fertilizer N was incorporated into the soil at transplanting and one-third applied topdressed at 5 to 7 d before the panicle initiation (PI) stage. The N topdressing before PI was applied on the same day to all varieties in the experiment based on the development stage of one standard variety. In most years, all treatments received 26 kg P ha-1 for each rice crop, while K applications of 50 kg K ha-1 crop-1 began in 1980.
|
Management of the LTCCE and Measurements during 1991 to 1995
Beginning with the 1991 EWS crop and continuing until 1995, various changes in the design of the experiment and crop management practices occurred (Table 1). Although some of these changes were caused by the need to renovate irrigation infrastructure at the site or unavoidable technical constraints, a number of changes were implemented to reverse the yield decline observed from 1968 to 1991 and to improve the quality of field management and measurements.
Because statistical differences between yields of direct-seeded and transplanted treatments were small, the direct-seeded subplot treatment was eliminated in the 1991 WS, and all rice was transplanted thereafter. The current experimental design is therefore a split-split-plot randomized complete block with N rate as main plots and three varieties as subplots and sub-subplots. From 1991 to 1995, IR72 was grown as the standard reference variety together with two elite breeding lines (IR59682-132-1-1-2 and IR60819-50-3-3). These varieties had similar yield potential to those grown before 1991, but have greater yield stability due to improved resistance to major pests. Selected varieties represented the best available genotypes, but only varieties with a similar maturity were used. Mean days to maturity for the varieties used from 1970 to 1972 was 124 d (±4 SD), as compared with 115 d (±1 SD) in 1992–1993. More uniform maturity and elimination of broadcast-seeded subplots facilitated greater precision in water management and timing of N applications in relation to the stage of crop development (Cassman et al., 1995).
Nitrogen rates were increased in 1991–1992, the number of split applications was changed from two to three or four, and the timing of N applications was modified (Table 1). In general, a more even spread of N addition was implemented, including a late N dose at flowering in the DS crop. In 1993 and 1994, chlorophyll meter (SPAD 502, Minolta, Ramsey, NJ) readings of the uppermost fully expanded leaf (Y-leaf) were used to determine the timing of N topdressings based on thresholds established by Peng et al. (1996). In the 1994 and 1995 DS, ZnSO4 was applied as a blanket application to all plots to provide 10 kg Zn ha-1.
In years with the normal triple-crop system, soil does not have time to dry in the 15- to 20-d fallow period between harvest and transplanting (Fig. 1) . However, only two rice crops were grown in 1991, 1993, and 1994 because the deep tube well and irrigation conveyance system were reconstructed during an extended fallow period after the DS crop in 1991 and 1993. In 1994, not enough seed of two new lines was available to produce two WS crops. Therefore, in years with double cropping, only one WS crop was transplanted in July and harvested in October. Surface soil dried thoroughly during the extended fallow of 70 d between the 1993 DS and WS crops, and there was partial drying during this period in 1991 and 1994. After harvest of the single WS crop in those years, the soil remained flooded from October to December until puddling operation of the next DS crop (Cassman et al., 1996).
|
Beginning with the 1991 DS, a more detailed sampling procedure was implemented. Dates of occurrence of important growth stages, complete pesticide application profiles, and soil moisture status between crops were recorded. Plant N accumulation was measured for each crop. Grain yields were obtained from a central 5-m2 harvest area at harvestable maturity. A 12-hill plant sample was collected at physiological maturity, when 90 to 95% of all grains had lost their green color. This sample was used to determine nutrient concentrations in plant tissue. Grain and straw subsamples from the 12-hill sample were dried to constant weight at 70°C. Straw yields were estimated from the oven-dry grain yield of the 5-m2 harvest area and the grain-to-straw ratio of the 12-hill sample. All yields (Mg ha-1) are reported at a standard moisture content of 140 g H2O kg-1 fresh weight. Nitrogen concentrations in grain and straw were measured by micro-Kjeldahl digestion, distillation, and titration (Bremner and Mulvaney, 1982).
Crop Modeling and Statistical Analysis
The crop simulation models ORYZA1 (Kropff et al., 1993) and WOFOST 7.1 (van Diepen et al., 1989; Boogaard et al., 1998) were used to estimate potential yields of IR72 for all rice crops grown from 1968 to 1995 using the actual weather data collected daily. The term potential yield is defined as the maximum attainable yield of a variety, restricted only by the season-specific climatic conditions and assuming optimal N and water supply. Dry season potential yields predicted by WOFOST 7.1 tended to be 0.5 to 1.0 Mg ha-1 lower than those estimated by ORYZA1, whereas the early wet season yields were not significantly different between the two models. Both models predicted similar total aboveground dry matter yields, but differed in their treatment of dry matter partitioning. Values of potential yields cited in this paper are averages predicted by the two crop models. The attainable N-limited yield, simulated by WOFOST 7.1, was defined as the potentially attainable yield based on the actual measured N accumulation and climatic conditions with no other constraints to growth.
The total N accumulation in grain and straw from plots that did not receive N fertilizer (N0) was considered to provide an estimate of the indigenous N supply (INS), which includes supply of N from soil pools as well as other indigenous sources such as biological N2 fixation, rainfall, and irrigation. Nitrogen use efficiencies were estimated using the differences between N-fertilized treatments and the unfertilized control, as defined by Cassman et al. (1998). Terms used are AEN, agronomic efficiency of applied N (kg grain yield increase per kg N applied); REN, apparent recovery efficiency of applied N (kg N taken up per kg N applied); PEN, physiological efficiency of applied N (kg grain yield increase per kg fertilizer N taken up); and IEN, internal N efficiency (kg grain per kg N taken up).
Long-term yield trends were plotted for the highest-yielding variety in each year because no single variety was grown throughout the whole experimental period. The averages of all three varieties were not used because before 1991, one variety (IR8) became increasingly affected by tungro. Note, however, that potential yield and N uptake data shown for the period from 1991 to 1995 are for cv. IR72 only because the crop models used were not calibrated for other varieties grown. Linear regression of yields (in logarithmic form) against a time trend variable was performed for the period 1968 to 1991:
 | (1) |
Quadratic N response models of the form
 | (2) |
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionGeneral discussionREFERENCES |
|---|
|
|
|
Grain Yield and Nitrogen Uptake during 1991 to 1995
The yield decline observed until the 1991 DS was reversed in both DS and WS crops. Beginning with the 1991 EWS crop grown after the first dry fallow period, grain yields in all treatments that received N fertilizer increased to levels similar to those measured during the initial years of the experiment (Fig. 3 and 4) . A maximum grain yield of 9.4 Mg ha-1 was measured in the 1993 DS (variety IR59682-132-1-1-2), as compared with 9.7 Mg ha-1 in the 1970 DS (IR661-1-140-3-2) or 9.9 Mg ha-1 in the 1971 DS (IR24). Note that the high yield in 1993 was achieved with a 115-d variety, whereas those grown in 1970–1971 required 127 d from sowing to maturity. Dry season yields in both the N2 and the N3 treatments increased similarly (Fig. 3). Early WS yields increased to levels of about 6 Mg ha-1. Three crops grown per year after 1991 had a total average annual grain production of 19 Mg ha-1 yr-1 in the N2 and N3 treatments, which was close to the annual production from 1968 to 1972 (Table 3) . Growing two rice crops per year restored the annual grain production only up to levels achieved with triple cropping about 15 yr ago (14 Mg ha-1 yr-1). The yield reversal on an annual basis only occurred in treatments with N applied (Table 3).
|
|
|
In contrast to the large effect of a dry fallow period on the INS of the subsequent WS crop, there was no detectable residual effect on the INS in the DS crop grown in the following year. Dry season grain yields in N0 were generally similar to those in the DS of years with the normal triple-crop system before 1991 (Fig. 1 and 4, Table 2). Average N accumulation in N0 plots was 42 kg N ha-1 in DS crops succeeding a triple-crop year (1991 and 1993 DS) and 43 kg N ha-1 in DS crops succeeding a double-crop year (1992, 1994, and 1995 DS). In years with double cropping, prolonged flooded fallow periods occurred after the harvest of the WS rice crops (October–December 1991, 1993, and 1994), so that little soil aeration occurred before the DS crop was planted.
Using the analytical approach of Witt et al. (1999), the relationship between grain yield and N accumulation indicates that N deficiency limited growth of plants in the N0 and N1 treatments, whereas plant N supply was not a major growth limitation in the N2 and N3 treatments after 1991 (Fig. 5) . Dry season data points of N0 and N1 treatments fell close to the line of maximum dilution of N in the plant (YND, high IEN), confirming that N was the major growth-limiting factor in these treatments. Most of the DS data points for the N2 and N3 treatments were close to the simulated optimal nutrition situation. In the WS, N was less diluted in the plant in N0 and N1 treatments than in the DS (lower IEN), but most of the data points from the N2 and N3 treatments also were close to the line of optimal nutrition. The fact that many of the highest yields fell close to the line of optimal IEN is further evidence that there was no significant yield loss from pests or abiotic stresses. However, both in DS and WS, some data points fell close to the line of maximum accumulation of N in the plant (YNA). These are situations in which unusual events decreased grain yield. They include lodging damage due to a typhoon (N3 in the 1994 DS) or selected cases of increased disease incidence and severity at the highest N level (N3 in the 1992 and 1993 WS).
|
|
That the timing of N topdressings was more congruent with crop demand in the 1992 to 1995 DS compared with the late phase of the yield decline period is consistent with measurement of greater plant N accumulation, more efficient AEN and REN, optimal IEN in N2 and N3 treatments (Fig. 5), and the increased slope of the N response functions (Fig. 6). We presume that the greater precision of N management reduced N losses from the applied fertilizer from volatilization and denitrification (Fillery et al., 1986; Buresh and De Datta, 1990). Previous work at IRRI has shown that reducing the proportion of total N fertilizer applied at planting and increasing the number of N topdressings can result in significant improvements in AEN, yield, and protein content of rice (Cassman et al., 1994; Peng et al., 1996; Perez et al., 1996). Other factors that may have contributed to greater fertilizer-N efficiency after 1991 include better control of floodwater depth in combination with greater precision in the timing of N application (particularly at PI and flowering stages) and prophylactic treatment for sheath blight and stemborer.
Contribution of Different Factors to the Yield Reversal
Because changes in management practices after 1991 were coincident with changes in climatic factors that influenced crop yields, the separate contributions of climate and management factors to the yield reversal were estimated for IR72, which was the only variety grown throughout the 1989 to 1995 period. Comparison of potential and actual grain yields of IR72 during this period allows partitioning of the effects of climate, increased N rate, greater N use efficiency, and dry fallow periods using the following approach for the DS:
 |
 |
 |
 |
We performed similar calculations for the EWS crop, one for fallow years (1991, 1993, and 1994) and one for nonfallow years (1992 and 1995), and compared both with the EWS crops during the 1989 to 1991 period. The YRT calculated as above was 2.66 Mg ha-1 in the fallow years and 1.75 Mg ha-1 in the nonfallow years. The net effect of climate and different transplanting dates was a YRC of 0.39 Mg ha-1 in fallow years and 0.89 Mg ha-1 in the nonfallow years. Applied N increased from 90 kg N ha-1 before the yield reversal to 105 kg N ha-1 in N3 in the fallow years and to 115 kg ha-1 in N3 in the nonfallow years. After the yield reversal, however, yields in the N2 treatment were higher than those in the N3 treatment in both fallow and nonfallow years (YN2b > YN3b, Table 2). Hence, there was no contribution of YRN to the yield reversal in the EWS (YRN = 0) so that YRR was 2.27 Mg ha-1 in the fallow years and 0.85 Mg ha-1 in the nonfallow years. We attribute the difference between these two numbers (1.42 Mg ha-1) to the effect of fallow on YRT, whereas the YRR in nonfallow years (0.85 Mg ha-1) resulted solely from improved crop management.
Summarizing, in the DS, 54% of YRT was due to climate (increase in solar radiation), 16% to the increase in N rate, and 30% to other improvements in crop management. In years with a fallow period before the WS, 15% of YRT in the WS was due to climate, 53% due to the fallow, and 32% due to other improvements in crop management. In years with the normal triple-crop system, 51% of YRT in the EWS was due to climate and 49% due to improvements in crop management.
It is possible that the relative contribution of climate to YRT was overestimated by using the 1989 to 1991 baseline period, because solar radiation was unusually low in these years (Fig. 3). The YRC as defined above was greater than if we had used mean daily solar radiation values for the entire late phase of the yield decline, from 1980 to 1991, for comparison with the 1992 to 1995 period. This was not possible, however, because IR72 was not grown before 1989, and the ORYZA simulation model requires genoptype-specific coefficients to estimate potential yields.
In all cropping seasons, improvements in crop management (YRR) contributed 30 to 49% to the yield reversal. The two major changes in crop management occurring after the 1991 DS were (i) increased number and better timing of N topdressings and (ii) prophylactic control of diseases and stemborer (Table 1). The specific contribution to YRT from the changes in pest management practices cannot be estimated because detailed observations and measurements of incidence and severity of insect and disease damage were not made before 1991 and no unsprayed control plots were included after 1991. However, our previous analysis suggested that pests were not a major factor involved in the yield decline until 1991 (Cassman et al., 1995). The yield trends shown in Figure 4 are based on the highest-yielding variety in each season so that a possible breakdown in yielding ability or resistance to pests of one variety does not affect this time series. Diseases such as sheath blight may cause significant yield losses in high-N treatments, but rarely at the yield levels observed in the LTCCE during the late 1980s (Cu et al., 1996). During 1991 to 1995, yield losses due to pests and other factors were generally small. Therefore, also considering the high N use efficiencies (Table 2, Fig. 5), we argue that YRR mainly represents the yield increase achieved by improved congruence between N supply and crop N demand; i.e., better splitting and timing of N applications.
|
General discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionGeneral discussionREFERENCES |
|---|
A decline in the INS due to structural changes and conservation of soil organic matter (cause 4 above) was proposed as a major soil process contributing to the gradual yield decline. The DS yield without applied N decreased by about 2.0 Mg ha-1 between 1968 and 1972 (Fig. 4). A yield decrease in N0 treatments of this magnitude represents a reduction in plant N accumulation from indigenous sources of about 25 kg N ha-1 crop-1. The cause of this decrease and whether it has equally occurred in the N1 to N3 treatments remains to be determined, but it is not associated with a decline in soil organic matter content (Cassman et al., 1995). Follow-up studies on characterization of two soil organic matter fractions, representing about 20% of the total soil organic matter content, confirmed the increasingly less humified nature of soil organic matter under increasing intensity of irrigated rice cropping. Accumulation of phenolic compounds in the youngest humus fractions appears to occur in soils that remain flooded and anoxic throughout most of the year as in the LTCCE (Olk et al., 1996, 1998, 1999). Increased phenol content of humic acid fractions may reduce the rate of N mineralization per unit of organically bound N in the soil (Olk et al., 1996), but a relationship between soil N mineralization and organic matter composition has not been established yet.
Our recent observations and measurements in the LTCCE also provide insight about the role of other putative causes of yield decline. While genotype and climate affected seasonal fluctuations and short-term trends in yield observed in the LTCCE, these factors do not explain the long-term yield decline from 1970 to 1991 (Fig. 4). Likewise, our observations during the period of yield increase (1991–1995) do not support the hypothesis that B toxicity or Zn deficiency had much effect on the yield decline or its reversal. The irrigation water source has not been changed during the course of the LTCCE and is not among those known for occasional high B concentrations at the experimental site (Cayton, 1985). Using the same irrigation water supply, it was possible to achieve increased yields after 1991. Seedling beds were treated with a ZnO suspension before transplanting from 1981 to 1989. Despite this treatment, which is effective for alleviating Zn deficiency at the IRRI farm, there was no reversal of the yield decline during this period. Zinc sulfate was first applied in 1994, but yields approaching potential levels had already been realized in 1992 and 1993. Therefore, Zn deficiency does not appear to be a contributing factor to the yield decline.
In addition to the climatic effects, improved crop N supply was the major factor responsible for increased rice yields after 1991, providing new evidence that the yield decline observed until 1991 was caused by a gradual decline in N uptake and N use efficiency. Three factors may have contributed to the improved N nutrition associated with the yield reversal: (i) increased N supply from fertilizer to the root system (increased N rate and timing of N topdressings more congruent with crop demand), (ii) increased N supply to the root system from indigenous sources (increased N mineralization), or (iii) increased root N uptake capacity (improved root health and alleviation of other abiotic or biotic stresses affecting root N uptake). Although we did not measure root N uptake rates or other parameters of root health (iii), the changes in crop management probably had little impact on root development or function in the DS or in the WS in years when triple cropping was practiced.
Greater N use efficiency was achieved by reducing the amount of N applied at planting, by using a greater number of N topdressings (including a late N application at heading in the DS), and through changes in other aspects of crop management, such as prophylactic pest control. We caution, however, that the N response functions shown in Fig. 6 may not show the true change in N use efficiency over time, because they are based on fixed-location N treatment plots. It is common practice to estimate N response functions across N levels in a long-term experiment, but for studying changes over time, this assumes no residual effect of the respective N treatments on the indigenous N supply; i.e., any change that occurs in the N0 plots (intercept, YN0) is due to a process that occurs equally in the N1 to N3 treatments. The lack of an appropriate control N0 treatment to allow estimation of the INS in treatments with different fertilization history is a major limitation of current experimental designs for long-term studies that seek to quantify changes in response to fertilizer inputs over time and identify the causes for such changes.
With regard to the INS, soil drying during fallow periods increased N mineralization (INS) and contributed to the restoration of the grain yields in the WS of years when only two crops were grown. Enhanced soil aeration or soil drying is known to increase N mineralization (van Gestel et al., 1993). In our case, it is likely that this has caused a more complete decomposition of soil organic matter in years with double cropping and reduced the proportion of N stored in organic pools that contribute little to the available N supply for uptake by rice. Recent studies in an experiment located adjacent to the LTCCE on the same soil type have demonstrated enhanced C and N mineralization due to dry soil tillage after harvest coupled with a dry fallow period of two months. In this tillage system, there was an increase in the INS and rice N accumulation, particularly during the vegetative growth stage (Witt et al., 1998). However, the dry fallow effect was short-term in the LTCCE; it had no residual effect on subsequent DS rice crops, and the INS returned to levels that were considerably lower than in the early years of the experiment. In crop cycles that did not follow a dry fallow, N accumulation in N0 remained relatively constant at 50 to 60 kg N ha-1 (Fig. 1). This amount of N is comparable to the estimated N contributions of 56 kg N ha-1 crop-1 from biological N2 fixation in the soil-floodwater system, irrigation water, and atmospheric inputs (Cassman et al., 1995), so the soil N balance may have reached an equilibrium.
It remains to be seen whether rice yields in the LTCCE can be maintained at levels close to the climate- and genotype-specific yield potential, because the underlying mechanisms causing long-term changes in rice yield and in the response to N need further clarification. If there was a decline in INS or any other gradual change in soil processes that affect morphological or physiological characteristics of the root system related to N uptake, we did not mitigate the cause of this process over a longer time period, but only the outcome, by adjusting the timing and amount of applied N and through short-term increases in INS due to soil drying. Moreover, although reducing cropping intensity from three to two rice crops per year coupled with a dry fallow period improved the INS in the LTCCE, which indicates the potential to reduce N fertilizer requirements, on an annual basis, total rice yields were about 4 Mg ha-1 yr-1 lower in double-cropping years than in triple-cropping years (see Treatment N2 in Table 3). At issue, then, is how often soils in a triple-cropped rice system should be drained or fallowed to enhance soil organic matter turnover, reoxidize reduced compounds, and optimize both yield and N fertilizer use efficiency without sacrificing a major reduction in total rice production over the long term.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 28, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionGeneral discussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Yang, S. Peng, R. C. Laza, R. M. Visperas, and M. L. Dionisio-Sese Yield Gap Analysis between Dry and Wet Season Rice Crop Grown under High-Yielding Management Conditions Agron. J., August 11, 2008; 100(5): 1390 - 1395. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. F. Pampolino, E. V. Laureles, H. C. Gines, and R. J. Buresh Soil Carbon and Nitrogen Changes in Long-Term Continuous Lowland Rice Cropping Soil Sci. Soc. Am. J., May 1, 2008; 72(3): 798 - 807. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. deB. Richter Jr., M. Hofmockel, M. A. Callaham Jr., D. S. Powlson, and P. Smith Long-Term Soil Experiments: Keys to Managing Earth's Rapidly Changing Ecosystems Soil Sci. Soc. Am. J., March 12, 2007; 71(2): 266 - 279. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. R. Angeles, S. E. Johnson, and R. J. Buresh Soil Solution Sampling for Organic Acids in Rice Paddy Soils Soil Sci. Soc. Am. J., December 2, 2005; 70(1): 48 - 56. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. WATANABE, J. S. HANAN, P. M. ROOM, T. HASEGAWA, H. NAKAGAWA, and W. TAKAHASHI Rice Morphogenesis and Plant Architecture: Measurement, Specification and the Reconstruction of Structural Development by 3D Architectural Modelling Ann. Bot., June 1, 2005; 95(7): 1131 - 1143. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Dobermann, C. Witt, S. Abdulrachman, H. C. Gines, R. Nagarajan, T. T. Son, P. S. Tan, G. H. Wang, N. V. Chien, V. T. K. Thoa, et al. Soil Fertility and Indigenous Nutrient Supply in Irrigated Rice Domains of Asia Agron. J., July 1, 2003; 95(4): 913 - 923. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
