Evaluation of animal-drawn weeders for smallholder maize production in Zimbabwe

E Mbanje
AGRITEX Institute of Agricultural Engineering, PO Box BW 330, Harare, Zimbabwe

S J Twomlow
ICRISAT, PO Box 776, Bulawayo, Zimbabwe

D H O'Neill
International Development Group, Silsoe Research Institute, Silsoe, Bedford, MK45 4HS, UK

ABSTRACT
To improve weed management at the smallholder level, a plethora of animal- drawn weeding equipment is now available, with very little technical information available to the farmer on their performance. This paper presents the draught characteristics of a range of weeders, commonly used in southern Africa. Comparable results were obtained in terms of weeding efficiency for the ox-drawn cultivators, but the light three-tined cultivator had a lower work rate and was more difficult to control on heavier soils than the traditional five-tined cultivators. In sandy soils the ducks foot tines performed better than the reversible tines in terms of yield responses, and vice versa for the heavier clay loam soils. The single animal-drawn tool-bar had the second lowest weeding efficiency and required the most supplemental hand weeding. In terms of yield responses and returns to hours spent weeding, post-emergent ridge weeding out-performed all of the other weeding treatments. The results help explain the reasons why many farmers are unwilling to use the traditional five-tine weeders currently available.

INTRODUCTION

The greatest demands for mechanical energy in smallholder crop production operations are for land preparation and weed control; either or both of these can limit the output of a smallholder farmer (O'Neill, 1998) and "bottlenecks" can occur. For land preparation, the limit is due primarily to the intensity of the energy demand, whilst for weed control, the duration of the demand is likely to be more restricting (Rogan & O'Neill, 1993). The maximum rate of working (i.e. power developed) that may be expected of a smallholder is around 40 watts over a 6-hour working day (Dibbits, 1993) and this energy (totaling around 0.9 MJ) would be applied rather inefficiently through the use of simple hand tools, typically hoes of various designs (e.g. see Anon, 1992). Although weeding is associated with a lower power demand than land preparation, the time available may be inadequate to complete the task. This contributes to the drudgery of long hours of weeding, a burden borne mainly by women who must accommodate weeding within their schedule of domestic tasks. Chatizwa (1997) has investigated the use of hoes for weeding in Zimbabwe. He concluded from studies using female subjects that, of four designs, the garden hoe was the best overall tool in terms of work rate and effort, but that the badza, the local traditional tool, had the highest weeding efficiency.

To improve weed management, a plethora of animal-drawn weeding equipment is now available, with limited technical information available to farmers and extension agents on their performance (e.g. Kwiligwa et al., 1994; Mbanje and O'Neill, 1997). This paper presents the performance results of a range of weeders and weeding practices that are common in southern Africa.

METHODS AND MATERIALS

Site location

Field experiments were conducted at Domboshawa Training Center (DTC) and the Institute of Agricultural Engineering (IAE), Mashonaland East, in northern Zimbabwe during the 2000/2001 cropping season. The climate at each site is characterized by a unimodal rainy season from October/November to March, when most of the rain falls as sporadic 'heavy' convectional storms, followed by a long dry season from April to May. The 15-year seasonal average rainfall is 478 mm (1982-1996) at DTC, with a range of 260 to 1150 mm, similar to rainfall patterns at IAE. Table 1 shows the rainfall for 2000/2001.

Table 1. Rainfall distribution (mm) at DTC and IAE for the 2000/2001 seasons

Site	October	November	December	January	February	March	Total
DTC	1	57	272	91	301	309	1031
IAE	56	64	271	71	332	320	1114

The soils at DTC are deep, coarse-grained granitic sands (82% sand, 13% silt, 5% clay) with a plant-available water capacity of less than 12% by volume, which means that crops grown on these soils are prone to drought, as any excess water quickly drains below the plant rooting zone (Vogel, 1994). These are typical soils that are found in most smallholder farming areas of Zimbabwe. Dependent on the availability and condition of draught animals and implements, they are normally cultivated annually to a depth of 80 to180 mm (Koza et al., 2000). The second site, IAE, has deep red clay loams (>60% clay), typical of much of Zimbabwe's commercial agricultural land (Elwell, 1986).

The experiment involved four commercially available types of ox-drawn cultivators, fitted with different arrangements of reversible and ducks foot tines (Figures 1 and 2). These were (i) the standard BS41 five tined cultivator (Figure 3), (ii) the BS221 cultivator with hilling blades (Figure 4), (iii) the Zimplow lightweight cultivator with three tines, and (iv) a single

 

animal-drawn tool bar with a 0.3 m wide ducks foot sweep. Weeding with these implements was compared with the traditional farmers' practices of overall hand hoe weeding, weeding with the plough-share (mouldboard removed) and an improved practice of post emergent ridge weeding using the plough with the mouldboard attached (Riches et al., 1997). Full details of the different weeding treatments, including tine arrangements, are given in Table 2.

Table 2. Details of the different weeding treatments, including tine arrangements

Treatments	Description
BS2212R2H1D	Standard BS221 with hilling blades, 2 reversible tines and a single ducks foot tine
BS2213D2H	Standard BS221 with hilling blades and three ducks foot tines
BS415R	Standard BS41 with 5 reversible tines
BS414R1D	Standard BS41 with 4 reversible tines and single ducks foot tine
BS415D	Standard BS41 with 5 ducks foot tines
ZLW2R1D	Light weight 3 legged tine with 2 reversible tines and single ducks foot tine
ZLW3D	Light weight 3 legged tine with 3 ducks foot tines
CTB	Contil single animal drawn tool bar with sweep tine attached
SHARE	Standard VS8 plough with mouldboard removed, share only
MB	Standard VS8 plough with mouldboard attached
Hand	Traditional overall hand weeding with hand hoe (badza or jembe)

The trial was laid out in a fully randomized design; each plot was 10 m wide by 25 m long and replicated three times at each site. The site was uniformly spring ploughed using a standard ox-drawn plough. Maize hybrid SC513 test crop was planted in 0.9 m rows at both sites during the third week of December 2000, and thinned to an in-row spacing of 0.3 m after crop emergence. The crop was top-dressed with 56 kg/ha N (150 kg ammonium nitrate) in a split application at six and ten weeks after planting. No basal fertilizer was applied, in accordance with smallholder farmer majority practice. Weeding operations were planned to be undertaken at two and six weeks after emergence, with supplemental hand weeding within the crop row undertaken after each animal-drawn weeding operation.

During each weeding operation, the depths and widths of soil disturbance for each weeder were recorded (four measurements per plot), as were draught forces (eight measurements per run for a total of eight runs per plot). Effective working time per run, turning time and total time spent on each plot were measured, from which work rates and draught power characteristics were calculated. Crop records included maize plant population at harvest, grain yield at 12% moisture content, and the number of barren plants to check if there was any serious crop damage during the weeding operations. Estimates of weed biomass cover were made on a whole plot basis prior to and after the first weeding, and at harvest. Data were analyzed by conventional ANOVA for the design and comparisons of means by t-test at the 0.05 level of probability.

RESULTS AND DISCUSSION

The above average rainfalls (Table 1) experienced at both sites in February and March meant that it was possible to carry out only a single weeding operation during the second week of January 2001. The most common weeds at both sites were Cynodon dactylon, Richardia scabra, Galinsoga parviflora, Commelina benghalensis and Eleusine indica. C. dactylon was particularly prevalent on the sandy loam and caused numerous stoppages at weeding, as the rhizomes collected around the implements and reduced their performance.

Table 3 summarizes the key draught characteristics of the different animal-drawn weeding treatments and the labour inputs required for both the animal-drawn and supplemental hand weeding operations. Although the lowest draught forces were measured for the CTB, significantly lower ( P<0.001 ) than any of the other animal drawn implements, its work rate was very similar to the commercially available three and five tined cultivators. However, its weeding efficiency was the second worst after the BS2213D2H (Figure 5), necessitating the greatest amount of supplemental hand weeding (Table 3), and resulted in significantly lower ( P<0.001 ) yields than either of the plough-based systems (Table 4). Despite the significant differences in draught characteristics, all of the implements tested were well within the draught power capabilities of communal oxen (Koza et al . 2000).

Table 3. Draught performance characteristics, effective field capacities and labour inputs for the 11 different weeding systems (treatments) at DTC and IAE

	Draught, kN		Work rate oxen h/ha		Hand weeding h/ha
Treatments/sites	DTC	IAE	DTC	IAE	DTC	IAE
BS2212R2H1D	0.74	0.90	5.5	3.3	86.0	63.5
BS2213D2H	0.80	0.84	5.2	4.6	93.9	68.1
BS415R	0.85	1.03	5.4	3.3	92.2	57.5
BS414R1D	1.00	0.88	5.9	3.3	86.4	62.2
BS415D	0.77	0.89	5.2	3.3	78.0	72.1
ZLW2R1D	0.56	0.47	5.5	4.2	87.7	56.2
ZLW3D	0.65	0.56	6.1	4.2	90.8	72.8
CTB	0.14	0.34	5.9	3.8	98.3	76.1
SHARE	0.80	0.62	10.7	7.5	83.8	49.6
MB	0.91	0.59	11.7	7.9	81.1	47.6
Hand	n.d	n.d	n.d	n.d	175.5	158.1
SED	0.090	0.027	0.336	0.757	7.510	18.970
n.d. - no data collected from overall hand-weeded plots

 

Table 4. Maize yields, labour requirements # and returns on labour for 11 different weeding treatments at DTC and IAE

	Grain yield kg/ha		Barren plants/ha		Total weeding time h/ha		Returns on labour kg/h
Treatments/sites	DTC	IAE	DTC	IAE	DTC	IAE	DTC	IAE
BS2212R2H1D	484.0	1146.4	6208	450	96.9	70.2	5.0	16.3
BS2213D2H	582.4	1375.2	6750	475	104.4	77.3	5.6	17.8
BS415R	377.0	1548.2	5521	400	103.1	64.2	3.7	24.1
BS414R1D	836.8	1217.4	3292	300	98.2	68.8	8.5	17.7
BS415D	770.7	1049.4	6146	125	88.4	78.8	8.7	13.3
ZLW2R1D	815.7	1501.8	4063	1000	98.7	64.6	8.3	23.3
ZLW3D	1519.8	990.2	-250*	775	103.0	81.1	14.8	12.2
CTB	636.1	974.7	-1771	-450	110.1	83.6	5.8	11.7
SHARE	1566.2	1502.4	-1750	-525	105.3	64.6	14.9	23.3
MB	1618.6	1598.7	104	2375	104.4	63.5	15.5	25.2
Hand	1034.9	1183.8	3646	475	175.5	158.1	5.9	7.5
SED	36.3	311.1	4343	1469	6.96	8.17	1.30	1.78
# Total weeding time = oxen work rate (Table 3) x 2people + hand weeding (Table 3) * Negative values occur when plants have more than one cob, hence no. of cobs > no. of plants

Table 4 presents criteria that are important to farmers, although the importance of the returns on labour may be influenced by the availability of family labour (cf. hired). Yields are variable, particularly at DTC, but may not be representative of a typical year because of the heavy rainfall improving water availability in the sandy soils. The BS415R and ZLW3D give contrasting performances in terms of yield, with the former being associated with a (relatively) low yield at DTC and a high yield at IAE, whereas the latter has the opposite association. However, both these implements gave relatively high weeding efficiencies at both sites (Fig 5), which imply an interaction between the tine arrangements used and the soil types. It would appear that in sandy soils, when the reversible tines are replaced by the ducks foot tines on the traditional five-tined cultivator (BS415D) and the light-weight three-tined cultivator (ZLW3D), yield responses are significantly ( P<0.001 ) greater, although the number of barren plants may increase (Table 4). The opposite occurred on the heavier clay loam soil. This was different from the responses of the BS221 cultivator, which, as purchased (i.e. BS2212R2H1D), had significantly ( P<0.001 ) higher weeding efficiencies on both soils with reversible tines (Fig 5), than when they were replaced with ducks foot tines (BS2213D2H), but the latter set up gave the higher yields and better returns to overall weeding labour (Table 4). Of all of the weeding treatments, the CTB consistently performed the worst (Table 4), a reflection of its poor weeding efficiencies (Fig 5). Without discriminating between sites, the plough-based practices gave the best overall crop response. Because of the higher yields, the best returns to weeding labour are also associated with the plough-based practices.

CONCLUSIONS

The results for two different soil types, albeit for a wetter than average season, help clarify some of the reasons why many farmers are unwilling to use the traditional five-tine weeders currently available, even after they have purchased them (Chatizwa and Ellis-Jones, 1997). In terms of weeding efficiencies, the light three-tined cultivator was very similar to the traditional five-tined cultivator but it had a lower work rate and was more difficult to control on the heavier soils, particularly when the reversible tines had been replaced with the ducks foot type. For heavier soils, farmers should use the reversible tines that are sold with most cultivators, whilst, on lighter sandy soils, farmers would be wise to use the ducks foot tines. However, for farmers who do not own a cultivator and draught animals are not a constraint, post emergent-ridge weeding with the mouldboard plough should be actively encouraged, as it not only gives good weed control, but also enhances crop yields. These may be further enhanced by tying the subsequent ridges to conserve moisture (Riches et al. , 1997).

 

ACKNOWLEDGEMENTS

The work reported here is part of a research project (R7352) funded by the UK Department for International Development (DFID) and the Government of Zimbabwe (GoZ). However, neither DFID nor GoZ can accept responsibility for any of the information or views expressed.

REFERENCES

Anon. 1992. Tools for Agriculture. Intermediate Technology Publications Ltd: London.

Chatizwa I. 1997. Mechanical weed control. Proceedings of the BCPC Conference-Weeds 1997, 1: 203-208.

Chatizwa I., Ellis-Jones J. 1997. Zimbabwe smallholder farmers: an assessment of the use and maintenance of tillage implements In: Improving the productivity of draught animals in sub-Saharan Africa. Proceedings of a technical workshop, Harare, 25-27 February 1997, eds. J Ellis-Jones, A Pearson, D O'Neill & L Ndlovu, pp. 131-134. Report IDG/97/7, Silsoe Research Institute: Silsoe, UK.

Dibbits H J. 1993. Human and draught animal power in crop production, past experiences and outstanding problems. In: Human and draught animal power in crop production", Proceedings of the Silsoe Research Institute/CEC/FAO Workshop, 18-22 January 1993, Harare, eds. D H O'Neill & G Hendriksen, pp. 11-25. FAO: Rome.

Elwell H A. 1986. Determinations of erodibility of a subtropical clay soil: a laboratory rainfall simulation experiment. Journal of Soil Science 37: 345-350.

Koza T., Ellis-Jones J., O'Neill D and Twomlow S. 2000. Enhancing the use of draught animal power by smallholder farmers in Zimbabwe. In: Optimizing DAP for cropping, pp. 5-13. Report IDG/00/22, Silsoe Research Institute: Silsoe, UK.

Kwiligwa, E. M., Shetto, R. M., Rees, D. J. and Ley, G. L. 1994. Weed management systems based on animal-drawn cultivators for maize production in the southern highlands of Tanzania. Soil and Tillage Research, 29: 383-395.

Mbanje E O'Neill D (1997) Database for animal drawn tillage implements. In: Improving the productivity of draught animals in sub-Saharan Africa. Proceedings of a technical workshop, Harare, 25-27 February, 1997, eds. J Ellis-Jones, A Pearson, D O'Neill & L Ndlovu, pp. 77-119. Report IDG/97/7, Silsoe Research Institute: Silsoe, UK.

O'Neill D H. 1998. Ergonomics issues in agricultural development. In: Global ergonomics, eds. P A Scott, R S Bridger & J Charteris, pp. 63-67. Elsevier Science Ltd: Oxford.

Riches C ., Twomlow S J and Dhliwayo H H. 1997. Low-input weed management and conservation tillage in semi-arid Zimbabwe. Experimental Agriculture 33: 173-187.

Rogan A and O'Neill D. 1993. Ergonomics aspects of crop production in tropical developing countries: a literature review. Applied Ergonomics 24 (6): 271-386.

Vogel H. 1994. Conservation tillage in Zimbabwe - Evaluation of several techniques for the development of sustainable crop production systems in smallholder farming. African Studies Series, A 11. Geographica Bernensia: Bern.