Zambezia (1990), XVII (ii).WEED SCIENCE TECHNOLOGICAL NEEDS FOR THECOMMUNAL AREAS OF ZIMBABWEO. A. CHIVINGEDepartment of Crop Science, University of ZimbabweIN ZIMBABWE THE agricultural sector is divided into three levels existingside by side: a relatively small number of large-scale commercial farmersnumbering about 6 000; a larger number of small-scale commercial farmersnumbering about 8 600; and a very large number of small-scale non-commercial farmers, also known as communal area farmers, numberingabout 4,3 million (Zimbabwe, 1985, 132). For the purposes of this articlethe small-scale commercial farming sector will not be considered as itscrop production cannot be clearly distinguished as being either commercialor communal in nature. There is a considerable difference between thecontributions of the large-scale commercial farmer and the communalarea farmer to crop production in the national economy. The communalarea farmers comprised about 77 per cent of the population in 1982(Zimbabwe, 1984), but they produce enough food crops to feed the ruralpopulation only if there is no drought; and their contribution to marketedproduce is only between 1 and 42 per cent depending on the crop (GrainMarketing Board, 1987). On the other hand, large-scale commercial farmers,whose arable lands average 2 200 hectares in size (including plantedpastures but excluding natural pastures) (Zimbabwe, 1985, 132), producebetween 58 and 98 per cent of marketed crops (ibid., 139-51).The differences in farm size and contribution to marketed producereflect the different technologies in weed management which in the large-scale commercial sector has led to a 30 per cent increase in crop yieldsince the early 1960s (Tattersfield, 1982). The methods of weed controlcurrently being used depend on the land's physical features, soil conditions,and economic, technical and human resources at the disposal of the farmer.Research specifically in weed technology improvement started in1965 with the formation of the Weed Research Team at Henderson ResearchStation (Henderson Research Station, 1967). However, most of the workdone there was oriented towards the needs of the large-scale commercialfarmers with only a few elements applicable to the communal area farmingsector. (One such example is the ox-drawn cultivator, originally meant forthe large-scale commercial farming sector before the use of tractors,which eventually reached the communal area farmer.) This orientation isshown by the type of research programmes described in the AnnualReports of the Weed Research Team issued from 1969 to 1983 which werepublished by the Department of Research and Specialist Services of the133134WEED SCIENCE TECHNOLOGICAL NEEDSMinistry of Agriculture. During this period there were great advances inweed technology such as the use of sophisticated tractor-mount weeders,herbicide spraying booms and aerial spraying equipment. Consequentlyweed technology for the large-scale commercial farmer is now very good,but the communal area farmer lags behind and so this article deals withthe needs of this sector and focuses upon the following: the status andeffectiveness of weed management strategies; the main problems in currentweed management strategies; and current and future research needs.HAND-WEEDING OR HOEINGThis is the most widely used method of weed control practised by thecommunal area farming sector. Weeds close to the crop plants are hand-pulled, while weeds further from the crop plants are removed by hoeswith iron blades attached to a wooden or iron handle. These hoes areoften used while sitting or squatting and slowly moving ahead, but oneswith longer wooden handles are used while stooping. The efficiency ofthese operations is low, requiring 20(M00 man hours per hectare (Gill,1982). The choice of a big or small hoe is governed by the wetness of thesoil, soil type, the type of weed, the growth stage of weeds and crops andthe type of crop. For example, groundnuts (Arachis hypoged) need a smallhoe while cotton (Gossypium hirsutum) and maize (Zea mays) need abigger hoe.This method is slow, labour intensive, cumbersome and inefficient.Chemical weed control is 20-30 per cent more efficient in controllingweeds compared to hoeing or weeding by hand (Gill. 1982). In most casestimely weed control is rarely achieved. Delaying weeding until weeds havealready inflicted adverse effects on the growth and development of thecrop plant is a wasteful operation. Due to untimely weeding operationsand the low efficiency of this method of weed control there are yieldlosses and in some cases total loss of yield. Weeding frequency variesfrom one ecological region to another depending on rainfall distribution,soil type and fertility, the condition of land preparation at planting thecrop and the weed flora. Most crops need at least two weedings.Most communal area farmers use family labour, and when it is insuf-ficient they hire labour. However, hired labour may be unavailable whenneeded at a critical time, such as the November-January period when theplanting of late crops and weeding of early planted crops is taking place.For those farmers with heavy soils excessively wet conditions may notpermit efficient weeding to be done resulting in long periods of crop-weedcompetition and yield reduction (Table I)- Another problem with weedingand hoeing by hand is that there are morphological similarities of someweeds with certain crops. Finger millet (Eleusine coracana) and rapokoO. A. CHIVINGE 135Table IMEAN SEED COTTON YIELDS AT HENDERSON RESEARCH STATIONWeed-free full seasonWeed-free first 2 weeks onlyWeed-free first 6 weeks onlyWeed-free first 4 weeks onlyWeed-free first 2 weeks onlyWeedy full seasonWeedy first 8 weeks onlyWeedy first 6 weeks onlyWeedy first 4 weeks onlyWeedy first 2 weeks onlyLeast significant difference (5%)Yields1967/8Drierseason2 8522 7832 7332 444NilNil68015902 1423 074596(kg per ha)1968/9Wetterseason2 9913 1888889935Nil2 0432 5073 3872 909737Source: Abridged from Schwerzel and Thomas (1971).grass (£ indica) are very similar and difficult to distinguish particularly atthe early stages of growth before flowering; other examples are Shamvagrass (Rottboellia cochinchinensis*) in maize and stoCkrose (Hibisus spp.) incotton. Because of such similarities these weeds commonly escape removal.Another problem with this method of controlling weeds is that it cannotdeal effectively with parasitic perennial and annual weeds which reproducevegetatively. Striga spp. comprise the most common group of parasiticweeds causing considerable economic crop damage to maize and pearlmillet (Pennisetum typhoides). The haustoria of parasitic weeds penetratethe food conduction tissues of the roots of the crop plants, and it isimpossible to relieve the crop plants from such an intimate and damagingrelationship by weeding or hoeing.The most troublesome perennial weeds are Cyperus esculentus, C.rotundus, Cynodon dactylon and Imperata cylindrica. They are very persistentand difficult to control because they propagate by underground rhizomes,stolons and tubers. They readily regrow after hoeing from depths beyondthe reach of conventional hand tools. Other methods are needed for theireffective control.MECHANICAL WEED CONTROLMechanical weed control has been improved by the introduction of theplough, the spike-tooth harrow and the animal-drawn tyne cultivator. The136WEED SCIENCE TECHNOLOGICAL NEEDSplough is used primarily for land preparation, tilling land to depths greaterthan 15 cm from the soil surface. Spike-tooth harrows are used as secondarytillage equipment tilling land to not deeper than 15 cm. In addition to landpreparation these two implements also control weeds. The tyne cultivatoris more specifically used for controlling weeds. The main constraint to theuse of animal-drawn cultivators is their availability and, to a lesser extent,draft-power problems.The tyne cultivator is very efficient at removing weeds, particularlywhen they are below 15 cm tall, but its main disadvantage is that it israther heavy for the draft animals. This is particularly so early in thesummer when animals are in poor condition because of the lack of grazingpasture during the dry months. As a result some farmers opt for the lessefficient but lighter plough with its shear removed.Weed control by tillage equipment has several disadvantages (Tattersfieldand Cronin, 1958; Chivinge, 1984), notably about 5 per cent crop damageeach time any implement passes through the land, especially when theanimals are not well trained; the failure to remove intra-row weeds; the nearimpossibility of adequately controlling some weeds bigger than 15 cm; andthe inability to employ mechanical implements when crops are about 60 cmtall for fear of crop damage. On the other hand, mechanical weed control iscomparatively faster and less labour intensive than hand-hoeing.CROP ROTATIONBy and large very little crop rotation is practised in the communal areasbecause of the big maize or cereal-legume rotation ratio. Most farmers donot grow many legumes any more or if they grow them the proportion isextremely small. The only legumes commonly grown are groundnuts andsome cowpea (Vigna angulata). Allied to the limited crop rotation hasbeen the use of more or less the same system of weed control. This hasresulted in ecological shifts in the weed species in response to thoseapproaches which relied heavily on a single method of control. Thislimited crop rotation has led to some weeds becoming more widespreadand persistent; examples are rapoko grass, upright starbur (Acantho-spermum hispidum D.C.), pigweed (Amaranthus hybridus L.), witchweed(Striga spp.) and the wandering jew (Commelina benghalensis L.).CULTURAL WEED CONTROLThis involves the use of clean certified crop seeds, optimum plant popu-lations and plant arrangement, crop cultivars adapted to the ecologicalregion, optimum planting dates, maintenance of sufficient soil fertility,availability of adequate moisture, and use of competitive crops. Thecommunal area farmer has problems in putting some of these methodsO. A. CHIVINGE 137into practice. Optimum planting dates are difficult to attain because rainfallmay be delayed. There might be enough moisture to encourage weed butnot crop germination. The result is that some weeds germinate before thecrop and yet the farmer might not be in a position to remove these weedsbefore planting the crop. Adequate moisture may also be difficult tomaintain because of mid-season droughts. As there are no irrigation facili-ties the crop may suffer while the weeds, which are mostly C4 plants,*survive relatively well under those conditions. Optimum planting popu-lations are rarely achieved. In fact most farmers use about half of therecommended plant populations (Agronomy Institute, 1985).Most communal farmers do not use the optimum amount of fertilizerfor their soil, largely because of lack of soil analysis and shortage of cash.Some farmers do not fertilize their crops at all. Farmers prefer to fertilizecotton and maize, but sunflower (Helianthus annus), sorghum (Sorghumbicolof), finger millet and pearl millet rarely get fertilized. Despite this lackof fertilization the weeds always grow very vigorously, depriving the cropof the necessary growth resources.The only common cultural practice is the use of clean certified seed andrecommended cultivars. Maize and cotton are the two main crops grownwith clean seed by farmers. Certified seed for other crops such as groundnutsare available but are not widely used.Because of these constraints most crops are not vigorous enough tocompete with the weeds for limited nutrients. Cultural methods of weedcontrol are not being utilized to maximum effect and yields are depressed.CHEMICAL WEED CONTROLHerbicide usage in Zimbabwe started in the early 1950s (Tattersfield andCronin, 1958). By 1955 yield increase in maize due to chemical weedcontrol, particularly with 2,4-D (2,4-dichlorophen-oxyacetic acid), hadbeen noted in large-scale commercial farms; but communal area farmershave not yet caught up with herbicide usage (Table II). The advantagesand disadvantages of using herbicides by communal area farmers havebeen categorized as follows (Sharman, 1970; Parker, 1972; Hammerton,1974; Chivinge, 1984):AdvantagesŁ Elimination of early crop-weed competition leading to higher yields.Ł Reduction of time spent on weeding, thus giving time for otherfamily duties.Ł C4 plants produce the first products of photosynthesis with two 4-carbon compoundswhich makes them more efficient under conditions of high temperature and strong light.138WEED SCIENCE TECHNOLOGICAL NEEDSŁ Can be used under wet conditions.Ł Reduced tillage system.Ł Flexibility in the crops grown.DisadvantagesŁ They may be costly.Ł The availability of spraying equipment might be a problem.Ł Sufficient technical knowledge is needed.Ł Herbicides with a wide safety margin may be limited.Table IIHERBICIDE SALES IN ZIMBABWE, 1979-1986Season1979/801980/11981/21982/31983/41984/51985/6AverageLarge-scalecommercialfarms98,2398,4297,4898,1698,0098,0397,8098,02Peasantfarmers0,740,951,081,411,101,151,351,11Sales (_%)Others1,030,631,440,430,900,820,850,87Total100100100100100100100100Source: From information supplied by agro-chemical companies.INTEGRATED WEED CONTROLThis involves combining two or more methods to control weeds, such asremoval of inter-row weeds by a tyne cultivator followed by hand or hoeremoval of intra-row weeds. However, the impact of this practice on theimprovement of crop yields is doubtful as intra-row weeds may not beremoved soon enough to prevent competition with crop plants.CURRENT RESEARCHResearch aimed specifically at the needs of the communal area farmerstarted in 1982. The programmes included: weed surveys in communalareas; research into weeding systems which would fit into communal areafarming systems; research into herbicide carry-over problems in croprotations; surveys of soil types found in communal areas; and screeningherbicides specifically for use in sandy soils which are typical of mostO. A. CHIVINGE 139communal areas. Up to now the main problem with the current research isthat of applicability, as problems such as low rainfall and socio-economicconstraints which are found in the communal areas are not found onHenderson Research Station where most of the research is concentrated.RESEARCH NEEDS FOR THE COMMUNAL AREA FARMERAbout 75 per cent of the rural population spends more time battling withweeds than with any other operation from mid-December to mid-February.These farmers have not adopted chemical weed control to any significantextent. Over a period of seven years herbicide usage by the communalfarmer has averaged 1,11 per cent compared to 98,02 per cent for the large-scale commercial farming sectors (Table 10- The Table on its own does notexplain why there is such a disparity. However, Chivinge (1984) indicatedthat the scale of operation and the lack of cash, technical know-how,equipment and herbicides were some of the reasons why herbicidetechnology is not implemented in communal areas. These explanations arenot exhaustive and other constraints more important than these might beresponsible for the poor adoption of chemical weed control in these areas.Table III compares the cost of herbicides and hand weeding in cottonand maize. This Table shows that it is economically viable to use herbicidesbut certain considerations involving a farmer's choice have not beentaken into account; these include: yield levels to be attained; the elementof risk of crop failure; application equipment problems; whether the labourforce is technically competent; and the socio-economic conditions of thefarmer concerned. Only after considering all these factors can the use ofherbicides be considered economic or not. It has been reported that theuse of herbicides has resulted in yield increases of up to 55 per cent inmaize and 75 per cent in cotton under Zimbabwean communal farmingconditions (Anon., 1984) and yet the use of herbicides is still very low.In spite of the apparent 'economic benefit', the communal area farmerdoes not seem eager to adopt herbicide technology. This implies thatthere is either something wrong with the technology or the way it is beingintroduced. It is not currently known whether it is the appropriatetechnology for these farmers. Herbicide technology has been introducedto the communal area farming sector in the same manner as in the large-scale commercial farming sector and this is perhaps partly why it seemsto be failing. Blackie (1982) has pointed out that Zimbabwean farmers willadopt any technology which is suitable to their farming system. Freeman(1983) and Vega (1983) have reported that small-scale farmers (communalarea farmers) adopt new technologies when: they are aware of the tech-nologies; the technologies are useful to them; they are reasonably certainof the efficacy of the technologies; the technologies are economically140WEED SCIENCE TECHNOLOGICAL NEEDSTable IIIA COMPARISON BETWEEN THE COST OF HERBICIDE USAGE AND HAND-WEEDING IN MAIZE AND COTTON USING THE 1986 COST OF HERBICIDESPack size(litres')MaizeGardomil 500FW+ one hand-weedingGesagram 500FW+ one hand-weedingTwo hand-weedingsCottonCotogard 500FW+ Dual 750EC+ one hand-weedingThree hand-weedings5555Cost perpack (Z$)67,6091,35143,95163,20Dosage(litres per ha)3331Cost per ha(Z$)40,5615,0055,5654,8137,5092,31125,0086,3732,6437,50156,51212,00Source: Weeding figures are based on a recent survey carried out in Murewa by theauthor as follows: (1) An average of two weedings for maize and three for cotton. (2)Herbicide-treated lands have less weed pressure and so are charged Z$37,50 per ha.Untreated lands with heavy weed pressure are charged Z$62,50 per ha.feasible; they are sure of securing technical assistance, the necessarycredits, and a sufficient supply of herbicides.Work done in other developing countries on herbicide technology forthe communal area farmer may give some idea as to why the introductionof herbicides seems to fail in Zimbabwe. In Central America it has beenshown that herbicides can only supplement and not replace labour(Hammerton, 1974). Work done in Ghana, Nigeria and Zambia (Carson,1979; Akobundu, 1980; Parker and Vernon, 1982) has shown that herbicidesare cheaper than labour yet up to now the use of herbicides by the small-scale farmer in these countries is very limited. The main reason is that thesmall-scale farmer needs a systems or holistic approach in the introductionof any technology. Consideration should be given to the farmer's human^nd financial resources, his yield goal, the climatic and edaphic constraints,the impact that the new technology will have on the whole society and theO. A. CHIVINGE 141feasibility of adopting the technology. The farmer's decision as to whetheror not to adopt herbicide technology will depend on these factors. Asthese factors differ from one group of people to another, there is a need todevise weed science technologies suitable for each group.One of the biggest weaknesses in weed science technology for thecommunal area farmer in Zimbabwe is that the weed spectrum and itsimpact on crop production is not known. There has been no weed surveyin the peasant farming sector until recently yet at least four have beencarried out in large-scale commercial farms (Soane and Waister, 1963;Thomas, 1970; Budd, 1976; Chivinge, 1983). It is not known how competitiveweeds are under the farmer's level of management, climatic and edaphicconditions. Nobody has explained why farmers are using the methods ofweed control in current practice and whether these need modification orchanging. Table IV shows maize and groundnut yields in Chibi andMangwende Communal Lands. The two districts are in different ecologicalregions which is reflected in the yields. Their cultural practices in cropproduction are also different. The question is: should one recommend thesame weed science technology to the two rural areas or not? It is mostlikely that we would need a different package for each area and this can bedone only after studying the weed management component in addition tothe whole farming system, that is, by adopting a systems approach.Table JVMAIZE AND GROUNDNUT YIELDS IN TWO DIFFERENTECOLOGICAL REGIONSCropMaizeGroundnutsYieldsChibi (Region N)With cattle1,900,16No cattle0,950,15(tonnes per ha)MangwendeWith cattle3,240,74(Region Ha)No cattle2,100,54Source: Modified from Shumba (1984).CONCLUSIONThe communal areas of Zimbabwe have seen very little change in weedscience technology. Most of the research done in the past does not seemto have had any useful impact on improving weeding efficiency or increasingcrop yields in these areas. Consequently there is a need to undertakeresearch which has direct relevance and applicability to these farmers. If142WEED SCIENCE TECHNOLOGICAL NEEDSZimbabwe is to make noticeable improvements in weed science technologyfor this group of farmers, a systems or holistic approach is needed.Investing in research, particularly with herbicides, without first of allunderstanding the constraints of the communal area farmer is a waste oftime, money, effort and human resources.AcknowledgementsThe author is grateful for the assistance given by Mr. C. Shumba forcritically evaluating this article and Miss K. Wilsnagh for typing themanuscript.ReferencesAGRONOMY INSTITUTE 1985 Annual Report: Summer 1984-5 (Harare, Min-istry of Agriculture, Dept. of Research and Specialist Services), 14-41.AKOBUNDU, 1.0. 1980 'Economics of weed control in African tropics andsub-tropics', in Proceedings of the 1980 British Crop Protection Conferenceon Weeds (Nottingham, UK, Boots Pure Drug Company), 911-20.ANON. 1984 Kahwa Pakuru Newsletter (Harare, Ciba-Geigy Zimbabwe).BLACKIE, J. M. 1982 'A time to listen: A perspective on agricultural policyin Zimbabwe', Zimbabwe Agricultural Journal, LXXIX, 151-6.BUDD, G. D. 1976 'The second survey of weeds of arable lands inRhodesia',Rhodesia AgriculturalJournal, LXXIII, 159-60.CARSON, A. G. 1979 'Weed competition and control in maize (Zea maysL.)', Ghana Journal of Agricultural Research, IX, 161-7.CHIVINGE. O. A. 1983 'A weed survey of arable lands in the commercialsector of Zimbabwe', Zimbabwe Agricultural Journal, LXXX, 139-41.CHIVINGE, 0. A. 1984 'The role of herbicide technology on the smallscale farms of Zimbabwe', Zimbabwe Agricultural Journal, LXXXI, 97-102.FREEMAN, D. M. 1983 'Technology transfer: Some strategic socialorganizational considerations with application to weed science', inCommunication of Weed Science Technologies in Developing Countries(St Louis, Weed Science Society of America), 75-84.GILL, H. S. 1982 'The role of hand and mechanical weeding in weedmanagement in the advancing countries', in Improving WeedManagement: Proceedings of the FAO/IWSS Expert Consultation onImproving Weed Management in Developing Countries (Rome, Foodand Agriculture Organization), 17-22.GRAIN MARKETING BOARD 1987 Reports and Accounts for the Year Ended31st March 1987 (Harare, The Board), 4-28.HAMMERTON, J. L. 1974 'Problems of Herbicide Use in Peasant Farming'(Las Vegas, Paper presented at the Weed Science Society of Americaannual meeting).O. A. CHIVINGE 143HENDERSON RESEARCH STATION 1967 Annual Report 1966/67 (Salisbury,Department of Research and Specialist Services).PARKER, C. 1972 'The role of weed science in developing countries',Weed Science, XX, 404-12.PARKER, J. M and VERNON, R. 1982 'Maize herbicides for small-scalefarmers in Zambia', Tropical Pest Management, XXVIII, 259-65.SOANE, C. M. and WAISTER, P. D. 1963 'A note on weeds of irrigated landsin the Sabi Valley', Rhodesia Journal of Agricultural Research, I, 47-52.SCHWERZEL, P. J. and THOMAS, P. E. L. 1971 'Weed competition in cotton',PANS: Pest Articles and News Summaries, XVII, 30-4.SHARMAN, C. 1970 'Herbicide development in peasant farming areas', inProceedings of the Tenth British Weed Control Conference (Nottingham,Boots Pure Drug Company), 685-8.SHUMBA, E. M. 1984 'Animals and the cropping sustems in communalareas', Zimbabwe Science News, XVIII, 99-102.TATTERSFIELD, J. R. 1982 'The role of research in increasing food croppotential in Zimbabwe', Zimbabwe Science News, XVI, 6-10, 24.TATTERSFIELD, J. R. and CRONIN, C. H. 1958 'Weed Control in Maize inSouthern Rhodesia with Special Reference to Use of Herbicides'(Victoria Falls, paper presented at the African Weed Control Confer-ence), 56-65.THOMAS P. E. L. 1970 'A survey of the weeds of arable lands in Rhodesia',Rhodesia Agricultural Journal, LXVII, 34-5.VEGA, M. R. 1983 'Constraints to weed science technology transfer indeveloping countries', in Communication of Weed Science Technologiesin Developing Countries (St Louis, Weed Science Society of America),93-104.ZIMBABWE 1984 1982 Population Census: A Preliminary Assessment(Harare, Central Statistical Office), 8.ZIMBABWE 1985 Statistical Yearbook 1985 (Harare, Central StatisticalOffice).