SciELO - Scientific Electronic Library Online

vol.71 número2Interacción Genotipo × Ambiente en el Rendimiento de Semilla de Raps (Brassica napus L.) en ChileEficacia Comparativa de Distintas estrategias en el Manejo de Malezas en Trigo índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados



Links relacionados


Chilean journal of agricultural research

versión On-line ISSN 0718-5839

Chilean J. Agric. Res. vol.71 no.2 Chillán jun. 2011 

Chilean Journal of Agricultural Research 71(2) April-June


Organic Versus Conventional Methods of Fertilization and Weed Control in a Long Term Rotation of Cereals in Semiarid Spain

Comparación de Métodos Convencionales y Orgánicos de Fertilización y Control de Malezas en una Rotación de Larga Duración de Cereales en Secano, España.

Gabriel Pardo1, Joaquín Aibar2, Pilar Ciria3, Carlos Lacasta4, Juan Antonio Lezaún5, and Carlos Zaragoza6

1ETSIA. Universidad de Sevilla, Carretera de Utrera, km 1, 41013 Sevilla. España. Corresponding author (
2Escuela Politécnica Superior de Huesca, 22071 Huesca, España.
3Escuela Universitaria de Ingenierías Agrarias, Campus Universitario, 42004 Soria, España.
4Centro de Ciencias Medioambientales CSIC, Finca La Higueruela, 45530 Santa Olalla, Toledo, España.
5Instituto Técnico y de Gestión Agrícola SA ITGA, El Sario, Carretera El Sadar, 31006 Pamplona, España.
6Centro de Investigación y Tecnología Agroalimentaria, Apartado 727, 50080 Zaragoza, España.


Under semiarid conditions the response of cereal crops to chemical fertilizers and weed control practices with herbicides is often reduced. In fact, the economic profitability of agricultural production in many dryland regions is critically affected by high costs of inputs and low crop yields. As a solution, cropping systems like organic farming, obtaining similar yields and promoting environmental sustainability by reducing fertilizer and herbicides, could be an alternative to conventional systems. In this study, 23 trials were performed in five semiarid regions of Spain during 5 yr to compare different fertilizers and weed control methods on durum wheat (Triticum durum Desf.) and barley (Hordeum vulgare L.) yields. The following rotation pattern was developed on the experimental plots: fallow-barley-ground ploughed vetch (Vicia sativa L.)-durum wheat. Plots either received organic fertilizer, chemical fertilizer, or no fertilizer. In addition, three levels of weed control were applied in cereal plots: flex-tine harrow tillage, conventional herbicide, and no weeding. The results indicated that neither the fertilization nor the weed control have effect on the yield crop.

Key words: Sustainability, ecological agriculture, crop rotation, compost, flex-tine harrow.


En condiciones de clima semiárido, la agricultura de cereal tiene una reducida rentabilidad debido a los escasos rendimientos y elevados costes de abonos y herbicidas. Dado que las condiciones climáticas impiden incrementar estos rendimientos, la viabilidad económica y la sostenibilidad ambiental de la agricultura en estas regiones pasan por reducir el coste de los insumos o por conseguir precios del producto superiores mediante la obtención de un certificado ecológico para su comercialización. En este artículo se presentan los resultados de 23 ensayos de un experimento de 5 años localizado en cinco zonas semiáridas representativas de la agricultura española de secano. El objetivo del trabajo fue comparar los efectos de diferentes métodos de fertilización y control de malezas -ecológicos y convencionales- sobre la producción. Para ello se diseñó una rotación compatible con las técnicas de agricultura ecológica: barbecho- cebada (Hordeum vulgare L.)- veza (Vicia sativa L.) como abono verde-trigo duro (Triticum durum Desf.). En las parcelas destinadas al cereal se estudiaron tres niveles de fertilización: testigo sin fertilizar, fertilización orgánica y química. Además, se aplicaron tres niveles de control de malas hierbas: testigo sin desmalezar, grada de varillas flexibles, y aplicación convencional de herbicidas. Los resultados mostraron que ni el tipo de fertilizante ni el control de malezas tuvieron efectos sobre los rendimientos, por lo que para estas áreas la agricultura orgánica puede ser una solución viable.

Palabras clave: Sostenibilidad, agricultura ecológica, rotación de cultivos, compost, grada de varillas flexible.


In the last 30 yr, the profitability of extensive dryland agriculture in Spain has decreased gradually. While the price of grain has decreased, the costs of energy and inputs like chemical fertilizers and herbicides required for farming have increased. As a consequence, the economic viability of the majority of farms under conventional agricultural system critically depends on the European Union Common Agriculture Policy (CAP) support in the form of direct payments.

These problems have given rise to new productive models such as organic farming and integrated agriculture. These production systems can be profitable while providing environmental benefits by reducing or avoiding the use of chemicals and preventing other agricultural problems like pesticide resistance and soil residues (Pardo et al., 2009a). Furthermore, the reduction in agrochemicals can improve the sustainability of the system in the long term. In this context, crop rotation and mechanical weed control are recommended to maintain crop productivity and reasonable weed densities.

In Spain, grain production levels on semiarid drylands cannot compete with central and northern Europe because of low rainfall (Angás et al., 2006). However, organic farming could easily be developed in these drylands since chemical treatments have little impact over the yield (Lacasta and Meco, 2006; García-Martín et al., 2007), and sometimes they can be even negative (Van Herwaarden et al., 1998). On the other hand, the effect of herbicide applications is limited because the soil conditions are not adequate under semiarid climate conditions. In other cases when rainfall is more abundant, herbicides controls weeds but the crop yields do not increase because of a low density of weeds (Lacasta and Meco, 2006). Therefore, farmers do not use herbicides if the humidity and agronomic conditions are not positive (Rodríguez-Pérez and Díaz-Salazar, 2002).

On the contrary, some authors have suggested that eliminating the use of chemicals will slowly impoverish the soil of nutrients (García-Martín et al., 2007) and will lead to a progressive proliferation of weeds (Nogueroles and Zaragoza, 1999) with the subsequent negative effects on crop production.

Even though organic production techniques may be easy to introduce, there is a need to develop alternatives for fertilization and weed control to correctly evaluate the impacts of their introduction under semiarid conditions in the long term. The evidence in Spanish and European literature are incomplete because the existent works are limited temporally (covering a small number of years) and spatially (very local studies).

The main objective of this paper was to study the effect of different types of fertilization and weed control -typical of conventional and organic production systems- over the cereal production in five semiarid dryland plots representative of Spanish agriculture, where the same crop types were maintained from 1996 to 2003.


Trial locations, rainfall and soil types
The experimental trials began in 1996-1997, although the present analysis only includes the results of the last 5 yr. The trials were performed in five different regions where the rainfed agriculture is especially relevant. The specific locations of field experiments are the following: Huesca (42º06' N, 0º23' W, 430 m.a.s.l.), Navarra (42°27' N, 1°39' W, 380 m.a.s.l.), Soria (41º34' N, 3º1' W, 860 m.a.s.l.), Toledo (40º26' W, 490 m.a.s.l.) and Zaragoza (42º17' N, 2º25' W, 454 m.a.s.l.). The annual average rainfall for all locations ranged between 391 and 560 mm, with a semiarid moisture index (Thornthwaite, 1948). Table 1 show the precipitation registered during the crop season (November to June) in trial locations. The possibility of repeating the experiment in five different locations is one of the most interesting and novel aspects of the study, because it allows to replicate the trial under different climatic conditions and soil types representing the characteristics of rainfed agriculture in Northern Spain. The relevant soil parameters of each trial are shown in Table 2.

Table 1. Precipitation registered during crop season in trials.

Table 2. Soil characteristics in each trial.

Crop rotation and plant material
At present, the agriculture in all studied regions is based in a conventional production system, following a rotation fallow-cereal (durum wheat, Triticum durum Desf., or barley, Hordeum vulgare L.). For the purposes of our study and according to organic system principles, we modify the usual rotation by introducing vetch one of the years and incorporating it on the soil by ploughing, in order to supply N to the soil. Furthermore, stubble and cereal straw were also incorporated in the soil after harvest to maintain or even increase organic matter in soil in a sustainable way.

In each location the trials were performed on two contiguous plots in order to have a winter crop every year, where a 4 yr rotation was established (Table 3). The data used in this study included the last 5 yr of the rotation (1998-2003, Table 3). This period was chosen because any possible differences in cereal output among treatments would be more obvious after the first 2 yr of the trial, which were necessary to reconvert from the previous conventional management.

Table 3. Crop rotations during the experimental trials.


The plant material used for trials was durum wheat 'Antón', during 1998-1999, 1999-2000 and 2002-2003. Barley 'Graphic' was used in 2000-2001, and 'Hispanic' in 2001-2002 in all locations. The vetch (Vicia sativa L.) in all years was the cv. Senda. The cereal and the vetch were sown in November and the harvest was in June-July, except in Soria, where it was extended to August. The vetch was ploughed in the soil in May.

Experimental design
The experimental design for the cereal unit in each location and year was a split-plot with four replicates in Toledo and Zaragoza, and three replications in Huesca, Navarra and Soria. There were two factors (fertilization and weeding) and three levels for each factor in an experimental plot of 91 m2. The same type of fertilizer and weed control was used in the same physical space from the beginning of the experiment (year 1996) to evaluate the constant effect of both treatments through time. The adjacent plot remained as fallow or vetch, following the description in Table 3. The treatments of the different factors were the following:

Main factor (fertilization). F1: Control, without fertilization; F2: Organic fertilizer: 2500 kg ha-1 compost (sheep manure and cereal straw) applied at pre-planting (33.3% moisture, 57.4% organic matter, 27.8% C, 2.9% N, 1.4% P2O5, 5.1% K2O and 0.89% Na). Compost was provided by Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT) in Soria, with adequate characteristics for organic farming; F3: Chemical fertilizer: 100-60-60 N-P2O5-K2O kg ha-1 in 1998-1999 and 1999-2000 and 70-60-60 kg ha-1 in 2000-2001, 2001-2002 and 2002-2003. The fertilizer was applied at pre-planting with a rate of 32-60-60 N-P2O5-K2O kg ha-1 and the rest of N was applied as top dress at tillering.

Secondary factor (weed control): E1: Control, without weed control; E2: Mechanical weed control with harrow: one superficial pass along the sowing lines with a light flex-tine harrow (Hatzenbichler, St. Andrä, Austria), at tillering of the cereal. The regulation was set for the maximum weed control while avoiding appreciable crop damage according to previous work (Pardo et al., 2009a); and E3: Chemical weed control: herbicides chosen according to the existing weeds, aiming for selectivity and efficiency at an authorized dose and at the right moment (Table 4). The herbicides were selected according weed flora, crop (barley or wheat) and growth stages of both weeds and crop. Herbicides were appliedusing a manual sprayer especially designed for these trials. The predominant weed species were Papaver rhoeas L., Fumaria officinalis L., Anacyclus clavatus (Desf.), Veronica hederifolia L., Sinapis arvensis L., Convolvulus arvensis L., Galium aparine L., Polygonum aviculare L., Lolium rigidum Gaudin and Avena spp.

Table 4. Herbicide active ingredient and applied doses rates in trials.

Statistical analysis
The statistical analysis was performed according to the design established in each field using the SYSTAT 7.0 statistical package (SYSTAT, 1997). The data satisfy normality and variance homogeneity. When we detected a significant effect of the studied factor we separated the measurements using a Tukey test (p < 0.05).


In this section we discuss the results corresponding to 23 of 25 total existing trials (5 yr × 5 locations). Data from Huesca 2001-2002 were eliminated because of poor emergence and data from Soria 2002-2003 because of excessive irregularity. In both cases technical problems were caused by machinery breaks. With respect to the rest of trials, 14 correspond to durum wheat crop (1999, 2000, and 2003) and nine correspond to barley crop (2001 and 2002). To facilitate discussion of results, separately data obtained for fertilization and weed control treatments are presented.

In the majority of the cases, the results indicate that chemical or organic fertilizers were not effective to significantly increase crop yields (Table 5). In relation with unfertilized plot, chemical treatment only increased yields in five occasions, and in one case the yield significantly decreased. The effect of the organic fertilizer was lower also in relation to the unfertilized plot: yield only increased in two cases and the effect was never negative.

Table 5. Yield according to location and fertilization treatment.

For durum wheat, data by location showed that only in the case of Navarra-2003 there were significant differences for the chemical fertilizer. The reason of having not important differences between fertilized and unfertilized plots for the alternative treatments considered could be found in the amount and distribution of rainfall. Certainly, yields obtained in these areas critically depend on the rainfall (Angás et al., 2006; Moret et al., 2007) and the effect of applied inputs is low (García-Martín et al., 2007). Furthermore, the incorporation of cereal straw and the rotation with vetch may also explain the results. In fact, the rotation with buried vetch in the previous season appeared to have provided enough nutrients for the scarce yields obtained under these semiarid conditions. Some authors like Ernst (2006), Odhiambo and Bomke (2001) and Talgre et al. (2009) suggest that burying a leguminous plant can provide between 10 and 200 kg N ha-1 for the following crop. In Zaragoza location, for example, N present in vetch aerial biomass varied from 22 to 50 kg N ha-1 (Pardo et al., 2009b), without considering N in root biomass. Thus, the effect of the additional 100 kg N might be masked or exceed the quantity of N already in the soil if we consider the amount of N given by vetch. According to Campbell et al. (1991), Thorup-Kristensen et al. (2003), and Tonitto et al. (2006), when the potential crop yields are low, the inclusion of legumes in rotations can even compensate the total elimination of other fertilizers without abating yields. Some other studies located in semiarid areas of Europe (Betrán and Pérez-Bergés, 1997) and Australia (Van Herwaarden et al., 1998) indicate that not positive response was observed to chemical fertilization (without previously ploughing a leguminous plant), and even the wheat yield decreases at higher doses of applied N.

Organic fertilizer significantly increased wheat yield (although with little differences in absolute values) in Navarra and Toledo (1999), with respect to chemical fertilizer (in Navarra also with respect to control plot). Thus, this treatment can be useful in some situations although it is difficult to determine a priori whether it is better than the chemical fertilizer. In any case, increase in yield could be not enough to compensate the costs of treatment application (Pardo et al., 2009a).

Some authors underline that in dry years production is higher with compost than with chemical fertilizer, but the opposite occurs in wet years (Maiorana et al., 1992). On the contrary, in our case this expectation was not accomplished, because the best yield with organic fertilizer was obtained in Navarra (1999), under not especially dry conditions (Table 2). Other authors like Gonzalez et al. (1992) also obtained better yields under wet conditions, but with much higher fertilizer doses (20 t ha-1).

With respect to barley trials (2001 and 2002), there were more situations (four in total) of positive response to chemical treatment than in the case of wheat. However, when production was especially low (Zaragoza 2001 and 2002, Huesca and Toledo 2001) yields were only equal or lower compared with compost or no fertilizer, showing that the amount of 100 kg N ha-1 could be excessive in very dry years (as it is explained above for the case of wheat). Yields were higher using chemical fertilizer when they were around 3000 kg ha-1 (Navarra 2001 and 2002, Toledo and Soria 2002). This result agrees with Angás et al. (2006), located in another semiarid area of Spain (Southern Lérida), where the yield improved by increasing the N dose, although the experiment was focus on a barley monoculture without vetch incorporation, thus one can expect that soil reserves are lost faster.

The compost had only a little effect in specific cases (Soria and Toledo 2001, Navarra 2002). This result is consistent with García-Martín et al. (2007), who did not find a positive response with the compost treatment in other semiarid areas of Southern Spain. However other analyses found an effect of the compost use in the long term (Eghball and Power, 1999), but that was not our case.

The results for barley presented here do not agree completely with the durum wheat results but the following points should be considered: first, barley was grown after a fallow period and 3 yr after vetch was incorporated, while the wheat was grown immediately after vetch. Thus, in theory barley had a lower amount of nutrients than wheat and then had a better response to the chemical fertilizer. Second, the biological cycle of barley is shorter than that of wheat so the former requires that nutrients should be available earlier. Finally, although it seems not very probable, unfertilized plots or just fertilized with compost could start to be deficient in nutritional elements after 7 yr, as suggested by García-Martín et al. (2007).

Summarizing previous discussion, and in spite of some particular cases, the global data of trials indicated that fertilization treatments studied do not appear to be justified for the conditions of the experiments. Under the semiarid climate of the studied sites the rotation and the nutrients remaining after the harvest were enough to maintain the relatively attainable low production levels through time (< 2500 kg). Focusing on the profitability of fertilizer treatments, only in one especially rainy year in Toledo location (2002) of 23 cases, the increase in crop yield with chemical fertilizer is justified (Pardo et al. 2009a). The use of compost is not profitable in any case.

Weed control
In the first 4 yr of experiments neither the application of herbicide nor the mechanical controls were effective to improve the yields (Table 6). Moreover, the treated plots did not achieve higher significant yields than the control plots in any case. The reason of this effect can be found in the scarce presence of weeds rendering weed control treatment unnecessary (or even negative) in the majority of the cases.

Table 6. Yield depending on weed control treatment in each location.

In general, mechanical weed control did not result effective, and harrowing even reduced production in two cases (Navarra 1999 and Zaragoza 2001) since it did not reduce weed density but could caused crop damages (there were not measured but they were observed in both trials). This result agrees with other authors who also registered a decrease in production with several mechanical treatments due to crop damages and/or ineffective weed control (Rasmussen and Rasmussen, 2000; Barberi et al., 2000; García-Martín et al., 2007).

In the first 4 yr the weed treatments were effective in some cases (data not shown) but yields were not increased by herbicide application or harrowing (Huesca 2001, Navarra 2001). In these cases the presence of high weed density did not diminish the final crop yields.

Only in the last year (2003) for durum wheat, the lower weed density in the chemical or mechanical weeded plots achieved a higher yield in some sites (Toledo, Navarra and Zaragoza). An increased spring rainfall made the herbicides more efficient (data not shown), increasing yield in three locations. In Zaragoza, the mechanical treatment also increased yield, but less than the herbicide. This exceptional case -over 23 trials- underlines the difficulty of increasing output using this tool.

The previous results show the efficacy of the rotation cereal-fallow and cereal-ploughing vetch as green manure had a very positive effect impeding the presence of weeds associated to cereal. Thus initially weed densities were maintained at relatively low levels in the majority of the cases. Therefore even in the case that weed controls (chemical or mechanical) could eliminate weeds without damaging crops (which is almost impossible due to phytotoxicity or uprooting plants respectively), yields will rarely increase because of these low densities (see control plot results). On the other hand, mechanical control was not effective when weeds were too much developed and dry conditions reduced the effect of herbicide. As a result, both chemical and mechanical controls are clearly limited in productive terms and the ratio control-damage was not positive enough. Therefore, the weed control methods studied are not justified under our trial conditions.

In summary, in view of little reduction in yields resulting from not fertilizing and not weeding, the organic agricultural system can be an adequate alternative under the semiarid climate conditions of studied areas. In any case, the inputs reduction will conduct to economic savings. Moreover, if the crop yields can be sold at higher prices in ecological markets, farmers could improve the economic profitability of their exploitations (Pardo et al., 2009a).


The chemical or organic fertilization methods increase yields rarely. Only when the precipitation regime is high the treatments are effective, but this is not frequent in semiarid regions. Furthermore, the weed control treatments (mechanical or chemical) are little effective to get better yields under trial conditions.

As a consequence, the proposed rotation cereal-vetch-cereal-fallow,which includes burying the stubble and cereal straw, is enough to achieve the adequate fertility levels and to maintain a limited density of weeds obtaining the usual crop productions in the region. With this management, the farmer may accomplish the European rules for organic products and, at the same time, increase the sustainability of the agricultural system by reducing the external inputs.


Angás, P., J. Lampurlanés, and C. Cantero-Martínez. 2006. Tillage and N fertilization effects on N dynamics and barley yield under semiarid Mediterranean conditions. Soil and Tillage Research 87:59-71.         [ Links ]

Barberi, P., N. Silvestri, A. Peruzzi, and M. Raffaelli. 2000. Finger-harrowing of durum wheat under different tillage systems. Biological Agriculture and Horticulture 17:285-303.         [ Links ]

Betrán, J.A., y M. Pérez-Bergés. 1997. Respuesta del cereal en secano al abonado N P K. Información Técnica 30. 16 p. Dirección General de Tecnología Agraria, Gobierno de Aragón, Zaragoza, España.         [ Links ]

Campbell, C.A., M. Schnitzer, G.P. Lafond, R.P. Zetner, and J.E. Knipfel. 1991. Thirty year crop rotation and management practices effects on soil amino nitrogen. Soil Science of America Journal 55:739-745.         [ Links ]

Eghball, B., and J.F. Power. 1999. Composted and non-composted manure application to conventional and no-tillage systems: corn yield and nitrogen uptake. Agronomy Journal 91:819-825.         [ Links ]

Ernst, O. 2006. Effect of the winter legume as cover crop on the yield and N response of a no till corn. Agrociencia 10:25-35.         [ Links ]

García-Martín, A., R.J. López-Bellido, and J.M. Coleto. 2007. Fertilisation and weed control effects on yield and weeds in durum wheat grown under rain-fed conditions in a Mediterranean climate. Weed Research 47:140-148.         [ Links ]

González, J.L., I.C. Benitez, I.M. Pérez, and M. Medina. 1992. Pig slurry compost as wheat fertilizers. Bioresource Technology 40:125-130.         [ Links ]

Lacasta, C., y R. Meco. 2006. Agricultura ecológica: Una alternativa económica y energética para los agrosistemas de cereal en el secano español. Tierras de Castilla León 128:86-99.         [ Links ]

Maiorana, M., G. Convertini, V. Di Vari, and V. Rizzo. 1992. Yield and quality wheat (Triticum durum Desf.) under continuous cropping after nine years of straw incorporation. European Journal Agronomy 1:11-19.         [ Links ]

Moret, D., J.L. Arrúe, M.V. López, and R. Gracia. 2007. Winter barley performance under different cropping and tillage systems in semiarid Aragón (NE Spain). European Journal of Agronomy 26:54-63.         [ Links ]

Nogueroles, C., y C. Zaragoza. 1999. Buenas prácticas agrícolas para el control de malas hierbas en agricultura ecológica. p. 185-205. In Fernández-Quintanilla, C., V.M. Garrido, y C. Zaragoza (eds.) Control integrado de las malas hierbas. Phytoma, Valencia, España.         [ Links ]

Odhiambo, J.J.O., and A.A. Bomke. 2001. Grass and Legume cover crop effects on dry matter and nitrogen accumulation. Agronomy Journal 93:299-307.         [ Links ]

Pardo, G., J. Aibar, J. Cavero, and C. Zaragoza. 2009a. Economic evaluation of cereal cropping systems under semiarid conditions: minimum input, organic and conventional. Scientia Agricola 66:615-621.         [ Links ]

Pardo, G., J. Cavero, J. Aibar, and C. Zaragoza. 2009b. Nutrient evolution in soil and cereal yield under different fertilization type in dryland. Nutrient Cycling in Agroecosystems 84:267-279.         [ Links ]

Rasmussen, K., and J. Rasmussen. 2000. Barley seed vigour and mechanical weed control. Weed Research 40:219-230.         [ Links ]

Rodríguez-Pérez, M., y J. Díaz-Salazar. 2002. Incidencias malherbológicas en la provincia de Ciudad Real. 5 p. In Proceedings of XXI Grupo de Trabajo Malas Hierbas y Herbicidas, Ciudad Real. Marzo 2002. Junta de Comunidades de Castilla-La Mancha, Ciudad Real, España.         [ Links ]

SYSTAT. 1997. SYSTAT for Windows, Version 7.0. SPSS Inc., Chicago, Illinois, USA.         [ Links ]

Talgre, L., E. Lauringson, H. Roostalu, and A. Astover. 2009. The effects of green manures on yields and yield quality of spring wheat. Agronomy Research 7:125-132.         [ Links ]

Thornthwaite, C.W. 1948. An approach toward a rational classification of climate. Geographical Review 38:55-94.         [ Links ]

Thorup-Kristensen, K., J. Magid, and L.S. Jensen. 2003. Catch crops and green manures as biological tools in nitrogen management in temperate zones. Advances in Agronomy 79:227-302.         [ Links ]

Tonitto, C., M.B. David, and L.E. Drinkwater. 2006. Replacing bare fallows with cover crops in fertilizer-intensive cropping systems: A meta-analysis of crop yield and N dynamics. Agriculture, Ecosystems & Environment 112:58-72.         [ Links ]

Van Herwaarden, A.F., G.D. Farguhar, J.F. Angus, R.A. Richards, and G.N. Howe. 1998. "Haying-off" the negative grain yield response of dryland wheat to nitrogen fertiliser. Biomass, grain yield, and water use. Australian Journal of Agricultural Research 49:1067-1081.
        [ Links ]

Received:   1 September 2010
Accepted: 21 January 2011