| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departamento de Protección Agrícola y Forestal, Universidad Nacional Agraria, Managua, Nicaragua, APT 453
Corresponding author (freddy{at}ibw.com.ni)
Received for publication May 2, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Abbreviations: CT, conventional tillage • DAS, days after sowing • LD50, lethal dose for 50% of the animals • MT, minimum tillage • NT, no tillage
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Intensive soil tillage for crop production can deplete soil organic matter and nutrients and require inputs like fertilizer, pesticides, and water to maintain productivity and quality (Thurston, 1994). Crop production using conservation tillage systems has gained widespread acceptance in Nicaraguan agriculture during the past two decades (Tapia and Camacho, 1988). Benefits commonly reported from NT plus mulching include reductions in soil erosion and fuel and labor requirements as well as improved soil water retention (Lal, 1989). In addition, NT systems are reported to influence the occurrence and proliferation of weeds, insect pests, and diseases. However, one of the main problems with NT is its dependence on the use of herbicides (Lal et al., 1990).
There are differences in weed control strategies between reduced tillage systems and CT systems. In CT systems, soil-applied herbicides are widely used while in reduced tillage systems, the use of postemergence herbicides are more common. When tillage is reduced, pre-emergence and postemergence herbicides are the principal methods of weed control (Buhler and Oplinger, 1990; Powell and Renner, 1999). Integration of mechanical weeding and reduced use of herbicides has been approached by other authors (Buhler et al., 1995; Teasdale, 1993).
Excessive tillage can be detrimental to common bean production. According to Powell and Renner (1999), this crop is more susceptible than soybean [Glycine max (L.) Merr.] to tillage and traffic compaction. Research showing the impact of reduced tillage on common bean performance has been contradicting. Sandoval-Avila et al. (1994) evaluated seven white bean cultivars in different tillage systems, including NT, and found that mean seed yield was not affected by tillage systems. Skarphol and Corey (1987) found bean yields with NT to be comparable to or greater than those obtained with CT. Xu and Pierce (1998) stated that bean yield was not affected by tillage in the first year of a 3-yr experiment but was reduced in subsequent years compared with moldboard plowing. Liebman et al. (1995) tested reduced tillage combined with mulching systems, which resulted in greater weed infestation and lower crop yield. In contrast, Mascianica et al. (1986) studied the effects of tillage and wheat (Triticum aestivum L.) residues on snap bean growth and morphology and concluded that NT systems yielded pod levels that were equal to or exceeded those of CT.
Swanton et al. (1993) stated that shifts toward grass, perennial weeds, and volunteer crop plants had been observed under conservation tillage. Derksen et al. (1995) found differences in composition of weed communities among tillage systems before herbicide application. However, Hooker et al. (1997) did not find an influence of tillage on the relative proportion of annual broadleaf weed species and stated that weeds were effectively managed with reduced herbicide inputs in conservation tillage systems. Powell and Renner (1999), working with reduced tillage systems, found differences in weed populations between tillage systems. Weed control in common bean was improved in NT systems compared with CT systems. Buhler (1998) stated that changing the tillage system will change the distribution and density of weed seeds in agricultural soils.
During the last 10 yr, paraquat (1,1'-dimethyl-4,4'-bipyridinium ion) has been widely used as a postemergence herbicide because it is effective and easy to apply. Paraquat is applied postemergence directly to weeds using a knapsack sprayer with a bell-shaped nozzle shield, which permits selective application to the weeds and not to the bean crop. The application is made when bean has reached a height of approximately 30 cm. The use of paraquat is controversial because it is extremely toxic to humans and is considered to be one of the 12 most dangerous pesticides in the world (Akobundu, 1987). This demonstrates the need for effective weed control measures that combine the use of agronomic and mechanical practices but also consider the use of herbicides with low toxicity that are suitable in common bean production.
One goal in Nicaraguan agriculture is to develop ecologically based cropping systems that emphasize the inherent strengths of the systems. To achieve this goal, the effect of tillage intensity and weed control strategies on common bean yields and weed growth was examined during a 3-yr period (1994–1996). The hypothesis was that different tillage intensities combined with weed control strategies could affect bean plant growth and yield and would also affect weed density and biomass. To test this hypothesis, field experiments were designed with combinations of treatments (tillage intensity and weed control strategies).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Surveys of the site showed that weeds associated with common bean were predominantly dicots, mainly belonging to the Asteraceae family. The most abundant weed species were showstar [Melampodium divaricatum (L.E. Rich) DC], hairy beggarticks (Bidens pilosa L.), little button [Melanthera aspera (Jacq) L.C.], spiny amaranth (Amaranthus spinosus L.), Mexican pricklepoppy (Argemone mexicana L.), Mexican fireplant (Euphorbia heterophylla L.), Mexican clover (Richardia scabra L.), and red tasselflower [Emilia sonchifolia (L.) DC ex Wight]. Important monocots were knotroot foxtail [Setaria geniculata (Lam.) Beauv], goosegrass [Eleusine indica (L.) Gaertner], large crabgrass [Digitaria sanguinalis (L.) Scop], thistle (Cenchrus pilosus H.B.K.), johnsongrass [Sorghum halepense (L.) Pers.], and bermudagrass [Cynodon dactylon (L.) Pers.].
Cumulative precipitation during the 3 yr is shown in Figure 1. Rainfall was low in 1994 and evenly distributed throughout the cropping season. In 1996 it was higher than in the preceding years. A large proportion fell in October when common bean is easily damaged by heavy rain. Rainfall in 1995 was midway between that of 1994 and 1996 but above average for La Com-pañia. The low rainfall in November 1995 facilitated bean harvest.
|
The experiment was a randomized split-plot design in four blocks with soil tillage as the main plot factor and weed control practice as the subplot factor. Levels of tillage were NT, MT, and CT. Weed control practices were mulching, mechanical, and chemical. The same treatment was repeated in the same plot for 3 yr. The subplots (weed control) were 19.2 m2, each with eight rows that were 6 m long and 0.4 m apart. The four central rows constituted the sampling plot, leaving a border of 0.5 m at the ends. The area sampled in each plot was 8 m2.
The experiments were sown in the second half of the rainy season each year. Maize (Zea mays L.) was sown as a preceding crop at the beginning of the rainy season (May–June). Residues from maize were used for mulching when the treatment was required. Soil tillage in maize production was in accordance with the cultivation carried out in the second season for common bean production. Weeds in maize were controlled manually, and fertilizers were applied according to the standard recommendations.
Soil preparation was carried out according to the tillage systems. In the NT treatment, weeds were cut using a machete, and weed residuals were left in the field covering the soil. In MT, weeds were cut and weed residuals were left in the field. Thereafter, furrows for sowing were made 40 cm apart with a tractor-pulled implement. In CT, soil preparation consisted of five land operations. First the land was disk-plowed. Thereafter, it was disk-harrowed two times. At the second harrowing, a beam was attached behind the disc harrow to level the soil. Finally, furrows were made for sowing. Glyphosate [N-(phosphonomethyl) glycine] was applied before sowing, using a calibrated manual sprayer, at a rate of 1.5 L ha-1 commercial product. Bean was sown into a weed-free seedbed in all tillage systems.
In Nicaragua there are basically three methods of soil preparation: NT, MT, and CT. The major part of the research within this field has dealt with NT and CT (Tapia and Camacho, 1988). No tillage, as practiced at a small-scale Nicaraguan farm, consists of cutting the weeds with a machete at approximately 5 cm from the soil surface. Weed residues are left in the field as cover to protect the soil. Soil is not disturbed or broken. Sowing is done by placing seeds in a hole made with a pointed stick. Minimum tillage is a drastically reduced method of seedbed preparation that is intermediate between NT and CT. When using MT, farmers clear the land from vegetation using machete, and weed residuals are burned or expelled out from the field. Shallow furrows for sowing are made 40 cm apart with an ox-pulled plow, an ard. This practice leaves strips of undisturbed land between the bean rows. Seeds are sown by hand and covered to ensure good emergence. Conventional tillage is the use of disk ploughing, followed by harrowing to prepare a seedbed; thereafter, seeds are sown manually in furrows.
The bean variety used was ‘DOR-364’, which has a Type II growth habit (bushed indeterminate with creeping growth). This variety flowers 35 DAS and is harvested about 84 DAS. The color of the pod is cream when mature, and the grain is dark red and kidney shaped. This variety has been reported to be well adapted to the Pacific area (CNIGB, 1995).
Sowing was done manually, with 40 cm between rows. In the NT treatment three seeds were sown together every 8 to 10 cm at a depth of 2 to 3 cm. In MT and CT, the seeds were evenly distributed in the furrows. In all three tillage systems, the plants were thinned after emergence to give a stand of 30 plants m-2 with equal space between plants. At sowing, N, P, and K fertilizers were applied in the rows at rates of 15, 39, and 13 kg ha-1 N, P2O5, and K2O, respectively.
Mulching consisted of an application of maize straw dry matter at a rate of 4500 kg ha-1. The application was carried out at sowing, and 160 dry plants of maize were placed parallel to the rows in each subplot. Maize plants were obtained from the previous crop. The mechanical control consisted of hand-weeding with a machete at 21 DAS. Efforts were made not to disturb the crop. Chemical control was done 21 DAS. A mixture of the herbicide fomesafen {5-[2-chloro-4-(trifluoromethyl)phenoxy]-N-(methylsulfonyl)-2-nitrobenzamide} and fluazifop-butyl {(±)-2-[[[5-(trifluorometyl)-2-pyridinyl]oxy]phenoxy]propanoic acid}, butyl ester, was applied at a rate of 1.5 L ha-1 (0.75 L ha-1 commercial product for each herbicide).
At harvest time, 10 plants were sampled from each subplot, and number of pods per plant and number of seeds per pod were recorded. In addition, hundred-seed weight was recorded. The number of bean plants was estimated at harvest time by counting plants in the subplot. Grain yield from each subplot was measured and adjusted to 14% moisture content.
A sample of 1 m2 was taken from each subplot at 42 (postflowering) DAS to determine weed abundance (individuals by species), number of individuals, and dry weight per group of plants (dicots and monocots). Plants from a subsample of 0.25 m2 were collected to record fresh weight of dicots and monocots. Four samples of 100 g of fresh material from each type of plant were dried at 60°C for 72 h to estimate dry weight.
The statistical analysis for all variables was done through ANOVA, and mean comparisons were made using Fisher's unprotected LSD at P 
0.05 (SAS, 1990). All data were tested for normality to determine if transformation was necessary. Data from weed density, weed dry weight, and net benefit were not distributed normally, and were therefore square-root transformed (X + 0.5). Year, tillage, and weed control effects and all interactions were tested for significance using the appropriate error term based on significant main effects and interactions. The interaction between tillage and weed control was observed in two variables (common bean yield and net income) during each year of the study. The main effects of tillage and weed control on bean canopy height and yield components will be discussed separately.
Cost and benefits of each treatment were compared using total budgeting, which includes costs and benefits for all treatments and costs that varied between treatments (i.e., cost of soil preparation and weed control practices). Inputs (seeds, fertilizers, and herbicides) were priced at the time of the experiments each year. Costs of each type of cultivation were determined using rental prices in the area. Labor was valued according to regulations of the Nicaragua Ministry of Work. Bean price was obtained at the market each year at the time of harvest (i.e., Dec.).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
|
|
|
|
Chemical treatment reduced total weed dry weight over the years (Table 4). Weed dry weight averaged across years under mechanical and chemical weed control was 60 and 72% less, respectively, compared with mulching. Only during 1995 did weed control methods lead to similar weed biomass. This contradicts the results of Teasdale et al. (1991) who argue that crop residues can influence weed populations in NT cropping systems because of the proximity of the residue to the site of seed germination on the soil surface. Residues of sorghum [Sorghum bicolor (L.) Moench] and maize have been shown to inhibit weed emergence and growth in cropping systems (Salmerón, 1996; Navarro, 1997).
Weed control in common bean under all three tillage systems was more successful when chemical weed control was used. In the 3-yr study, it was not possible to find a clear tendency showing how tillage systems affect weed dynamics (i.e., long-term effects). The interaction between tillage and years shows a tendency for weed biomass to be influenced by rainfall rather than tillage systems. Further information is required from more crop cycles to discover whether there is any tendency for tillage systems to influence weed dynamics.
Powell and Renner (1999), working with reduced tillage systems, found differences between the systems concerning weed populations. They showed that weed control in common bean was improved in NT systems compared with CT systems. This contradicts the results presented here where CT worked better at reducing weed populations, independent of the weed control used. With CT the land was cultivated after clearing the vegetation, thus delaying the first weed generation. A possible reason for this contradictory result is that weed seeds in CT were distributed to different soil layers during cultivation, which ensures possible germination of only a part of whole seed potential in the soil and reinfection during the whole growth period. With NT and MT, seeds remain in the upper layers of the soil, ready to germinate once it starts to rain. In this system, weeds are abundant at the beginning of the crop cycle and there is considerable reduction in latter stages. Teasdale et al. (1991) showed that weed density increased after 1 yr of NT and after 2 yr of CT in a 4-yr experiment with application of treatments to the same plot.
The comparatively small effect of mulching on weeds in this experiment could be explained by a low residue biomass. The amount of mulch used was chosen in accordance with the expected amount of residue that can be produced by maize per unit area in a crop sequence strategy. Soil cover was approximately 50%, which was not enough to prevent weed seedling establishment. Weed density reduction is largely a function of soil cover by residues. According to Teasdale et al. (1991), there is not a reduction in weed density until soil coverage by residues reaches 42%, and 97% is required to reduce weed density by 75%.
Herbicides used in the experiments effectively control the main weed species that compete with common bean under Nicaraguan conditions. Fluazifop-butyl is a grass-specific herbicide that has been used post-emergence in several dicot crops with outstanding results (Romero, 1989). For male and female rats, the lethal dose of fluazifop-butyl for 50% of the animals (LD50) is 4.8 and 4.4 g kg-1, respectively (Ashton and Monaco, 1982). The other herbicide used, fomesafen, is recommended for controlling broadleaves in some legumes (ICI, 1986). It has a broad spectrum of control and must be applied during the third trifoliate leaf stage to avoid damage to the crop. The LD50 for fomesafen in rats is 8.7 and 7.0 g kg-1, respectively, for male and female rates (Ashton and Monaco, 1982). Thus, LD50 values of both herbicides are much greater than for the commonly used paraquat (120 mg kg-1). In the present experiment, chemical treatment performed well in controlling weeds during the 3-yr study in all three tillage systems, and particularly in CT.
Yield Components
Pods per plant, plants per unit area, and weight of hundred grains differed significantly between tillage and weed control systems. Minimum tillage resulted in more pods per plant than NT and CT and in heavier seeds than CT, which explains the larger yield obtained with MT. Bean plants per unit area were fewer with NT than with MT and CT (Table 5). Seeds per pod depend primarily on genetic characteristics (Campton, 1985) and did not differ between tillage systems.
|
Seed Yield
Seed yield of common bean did not differ between years but was affected by tillage system and weed control strategy. In addition, interactions of years x weed control, tillage x weed control, and a three-way interaction (year x tillage x weed control) were found.
Differences between tillage systems were most pronounced when chemical weed control was used. Bean sown in MT had a larger yield than in NT and CT when chemical weed control was used. On the other hand, mechanical weed control performed better in MT and NT than in CT (Table 6). Mechanical weed control, at the appearance of the third trifoliate, was comparable to results obtained with chemical weed control. Considering the cost of herbicides and problems that could arise as a consequence of their application, the use of mechanical weeding is preferable. Colquhoun et al. (1999) tested five cultivation implements for weed control in bean and concluded that use of either a brush hoe and/or shovel cultivator, preceded by flex-tine cultivation, allowed for weed control and a snap bean yield comparable to that obtained with broadcast herbicides.
|
The effect of tillage systems on common bean yield performance has varied in the literature. Alvarez-Solis et al. (1990) reported larger yields in NT than in CT plots. In contrast, Mulling et al. (1980) found larger yields of common bean in CT than in NT in a series of experiments. They showed that bean yield was greater in NT only during the first year of the study. Skarphol and Corey (1987), in accordance with the present work, stated that yields from NT were comparable to or greater than those obtained with CT in a series of experiments dealing with tillage and cover crop combinations. In addition, Sandoval-Avila et al. (1994) found no significant effect of tillage or tillage x year interaction on bean yield.
Bean yield was markedly reduced when mulching was used as a weed control strategy in NT and MT. The amount of maize dry matter used (4500 kg ha-1) was probably not enough to provide effective weed control. With CT, the three weed control practices worked similarly.
Common bean yield was similar in NT and MT when mechanical and chemical weed control were used even though bean stands in NT (averaged across years and weed control) were reduced by 14.7 and 9.5% compared with CT and MT, respectively. Similar results were presented by Powell and Renner (1999) who found no differences in bean yield between tillage systems although there was a bean stand reduction of 10% in NT compared with six tillage systems that included plow, chisel, and zone till. Vyn et al. (1998) noticed a delay in soybean emergence, and a thinner population appeared to be associated with NT treatments. This was also observed in the present experiment as common bean stand at harvest was reduced in NT systems.
In 1994 and 1995, there were small variations in yield between weed control strategies. In 1996 chemical weed control gave the highest yield, which indicates that this method is more efficient than mulching or mechanical control in a situation with high weed pressure.
Mulching seems to be a good way of controlling weeds in common bean production when weed pressure is low. In the present experiments, levels of 450 to 650 weeds m-2 reduced common bean yield while levels observed in 1995 (50 weeds m-2) permitted a good yield, especially when CT was used.
Bean are prone to attack by several plant pathogens such as fungi web blight (Tanathephorus cucumeris) and common bacterial blight [Xanthomonas campestris pv. Phaseoly (Smith) Dye]. These diseases have long been a problem under Nicaraguan conditions, especially when CT is used (Tapia and Camacho, 1988). The occurrence of these diseases are favored by practicing monoculture and by excessive soil cultivation to prepare seedbeds and control weeds, which are more likely to occur in CT. The use of maize in sequence as a preceding crop for bean and the reduction of traffic operations for weed control reduce the possibility of bean infection by these diseases. In two of the experiments, the presence of leaf blight (Entyloma petuniae Speg. Basidiomycetes) was observed, especially in CT. However, the disease affected only the older leaves and had no effect on bean yield.
Economic Analysis
The average net benefit differed between years, tillage systems, and weed control strategies. A two-way interaction was found between year x tillage and tillage x weed control. In addition, a significant three-way interaction (year x tillage x weed control) was revealed.
Net benefit varied over the 3-yr study with the smallest in 1994 and the largest in 1995. In 1994 the net economic benefit of mulching as a weed control strategy was similar to mechanical and chemical weed control, but in 1996, mulching gave the smallest benefit. In 1996, the net benefit of chemical weed control was larger than mechanical control and mulching.
Net benefit varied over the three tillage systems. Smallest net benefits were obtained in CT using mechanical and chemical weed control and in NT when mulching was used to control the weeds (Table 7). In general, common bean production was profitable with MT using either mechanical or chemical weed control. However, there are large fluctuations between years. Differences in net income are based on market prices of common bean at the time of harvesting. In 1994 the market price of common bean was $0.3 kg-1, contrasting with 1995 ($1.0 kg-1) and 1996 ($0.6 kg-1).
|
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
The CT system was significantly less profitable than both MT and NT, which suggests that a system that reduces tillage could be an economic advantage for the farmer. Two major factors caused a lower net income in CT: somewhat smaller yield and five soil preparation operations (ploughing, harrowing twice, leveling, and furrowing). This is at least partly in accordance with results presented by other researchers, who argue that a reduction in tillage intensity does not mean an increase in yield (Sandoval-Avila et al., 1994). Instead, the advantages of these systems are to be found in other aspects such as lower production costs, conservation of resources, and less damage from insect pests and plant pathogens (Tapia and Camacho, 1988).
In situations where Nicaraguan farmers should restrict the use of paraquat, the mechanical control option (carried out at the appearance of the third trifoliate leaf) would be sufficient to ensure profitable production of common bean. Differences in net benefit between mechanical and chemical weed control were very small and far from consistent throughout all combinations of years and tillage systems. Thus, these results show that farmers should be advised to adopt an integrated approach to weed control with less intensive use of herbicides.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. E. Blackshaw, L. J. Molnar, G. W. Clayton, K. N. Harker, and T. Entz Dry Bean Production in Zero and Conventional Tillage Agron. J., January 1, 2007; 99(1): 122 - 126. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
