African Crop Science Journal African Crop Science Society ISSN: 1021-9730 EISSN: 2072-6589 Vol. 7, Num. 4, 1999, pp. 455-463

African Crop Science Journal, Vol. 7. No. 4,  pp. 455-463, 1999                                                           

The effect of 2,4-D, dicamba and hand weeding on witchweed density and maize yield

O. A. Chivinge, D. Musambasi and  I. K. Mariga
University of Zimbabwe, Crop Science Department, P. O. Box MP167, Mt. Pleasant, Harare, Zimbabwe

Code Number: CS99036

ABSTRACT

An on-farm experiment was conducted in Chinyika resettlement area during the 1994/95 and 1995/96 rainy seasons to determine the effects of 2,4-D, dicamba and hand weeding on Striga asiatica density and maize grain yield.  Three sites were established during the 1994/95 season while an extra site was added during  the 1995/96 season.  Dicamba   suppressed S. asiatica more than the other treatments at 8 and 10 weeks after crop emergence (wace) while 2,4-D was as effective as dicamba at 12 wace.  In a drier season like the 1994/95, handweeding  two times was more economic than applying the herbicides.  Dicamba gave a marginal rate of return of 91% over 2,4-D during the wetter 1995/96 season.  When the data were combined over sites, maize grain yields were lower in plots treated with dicamba and highest in plots hand weeded at 2 and 5 wace during the 1994/95 season.  However, maize grain yields were similar in all the weeding treatments during the 1995/96 season.

Key Words: Herbicides, marginal returns, Striga asiatica, weed management

RÉSUMÉ

Un essai en milieu réel a été conduit dans la zone de repeuplement de Chinyika pendant les saisons pluvieuses de 1994/95 et 1995/96 pour déterminer des effets de 2, 4D, dicamba et du sarclage à la main sur la densité de Striga asiatica et du rendement du maïs.  Les sites ont été établis durant la saison 1994/95 et des sites additionnels ont été établis pendant la saison 95/96.  Dicamba a supprimé le Striga asiatica plus que d’autres traitement à  8 et 10 semaines après émergence (SAE) alors que 2, D-D a été efficace comme dicamba à 12 SAE.  Pendant la saison sèche de 1994/95, le sarclage à la main a été plus économique plus que l’application des herbicides.  Dicamba a donné un revenu marginal de 91% plus que 2, 4-D pendant la saison pluvieuse de 1995/96.  Quand les données des sites ont été combinées, le rendement en grains du maïs a été bas dans les parcelles traitées au dicamba, plus élevé dans les parcelles sarclées à la main à 2 et 5 SAE en saison 94/95.  Cependant les  rendements du maïs ont été les mêmes dans les traitements au sarchage pendant la saison 95/96.

Mots Clés: Herbicides, revenu marginal, Striga asiatica, gestion des mauvaises herbes

Introduction

The major biotic constraint to increased cereal (maize, sorghum and millet) production in the smallholder (SH) sector of Zimbabwe is infestation by the parasitic witchweeds (Striga asiatica (L.) Kuntze  and Striga forbesii Benth) (Karima et al., 1998). Striga forbesii is more damaging than S. asiatica but is less abundant and  hence of less economic importance.  These parasitic weeds cause irreversible crop damage even before they emerge.  On germination, the striga seedlings attach to the crop roots and become major sinks for the crops’ photosynthates resulting in yield losses which vary from 20-100% (Abayo et al., 1998).  Damage to the crop manifests itself in the form of reduced internodes and lack of floral parts culminating in no ear or head formation in the infected cereals.

To be effective against S. asiatica, a herbicide should control the parasite at incipient attachment so that damage is prevented or minimised. The phenoxycarboxylic acids including 2,4-D (2,4 -dichlorophenoxy acetic acid) have high activity at the germination stage of most broad leaved weeds including S. asiatica (Ransom et al., 1995; Gressel et al., 1996). Wilson-Jones (1953; 1956) demonstrated the value of the herbicides MCPA (methyl-chorophenoxy acetic acid) and 2,4-D not only for selective killing of  emerged S. hermo-nthica but also as a treatment a few weeks after sowing to prevent or at least reduce, germination and attachment of the parasite.  2,4-D has occasionally been shown to prevent S. asiatica attack, thus reducing damage to the maize crop but success depends on rainfall availability to wash  the chemical into the soil at the critical time of the weed germination.  Some reports suggest a crop yield benefit  even when 2,4-D is used to destroy S. asiatica long after emergence (Tattersfield and  Cronin, 1958; Eplee, 1991).

An alternative to a soil-acting herbicide which has to reach the parasite directly from the soil, is one applied to the host which is then translocated to the  roots and exuded to damage the parasite at the point of attachment (Ransom et al., 1995; Gressel et al., 1996).  Dicamba (3,6-dichloro-methoxybenzoic acid) has sometimes been used successfully in this way to control S. asiatica on both maize and sorghum in the USA (Eplee and  Norris, 1987) but results have not been consistent.  For instance, Odhiambo and Ransom (1993) reported that use of dicamba to control striga is only effective for a few weeks.  Dicamba is readily absorbed by leaves, stems and roots and translocated throughout the plant, accumulating in areas of high metabolic activity.  Since S. asiatica is an active sink for water, carbohydrates and nutrients from a host plant (Olivier, 1996), it is feasible that systematic herbicides applied to host foliage would be translocated from shoots to roots and subsequently to the S. asiatica (Award et al., 1991).  Timing of application is therefore a critical factor not only in achieving optimum control of the witchweed but also to ensure crop safety.

Mechanical control of  S. asiatica in traditional systems is generally confined to hoeing and ox-drawn cultivation, but this only removes S. asiatica during the course of weeding.  This operation is critical to the containment of  S. asiatica seed production in many systems.  Where it is not practical because of  labour bottle-necks, it is common to find that S.asiatica populations build-up rapidly to an intolerable level.  Weeding out of  S.asiatica plants before they flower is effective in reducing the soil seed bank (Eplee, 1991). But, it is labour intensive and for the smallhold farmer the large number of weeds to be cleared out at frequent internals makes the task difficult, and often beyond the farmers’ resources.

The objective of our study was to determine the effects of 2, 4-D, dicamba and hand hoeing on S. asiatica management under smallhold agriculture, such as in the Chinyika resettlement area of Zimbabwe.

Materials and methods

Three sites were used in the 1994/95 season, two in Govakova and one in Chinyudze, areas which lie between latitudes 18º 02” and 18º 17” East and longitudes 32o 9” and 32o 24” South.  During the 1995/96 rainy season, four sites were used of which two were in Chinyudze and the other two in Govakova.  The trial had four treatments replicated four times in a randomised complete block design (RCBD).  The treatments were hand weeding at 2 and 5 weeks after crop emergence (wace), hand weedings at 2, 5 and 8 wace, and hand weeding at 2 wace followed by application of dicamba as full-cover spray at 3 wace, and hand weeding at  2 wace followed by application of 2,4-D full-cover spray at 8 wace. 

The land was ploughed using an ox-drawn plough and the furrows opened using a hand hoe. Planting was done at the on-set of rains in the months of November and December depending on location.  The maize seeds were planted 20 mm deep at a spacing of 0.3 m along the  row by 1.0 m between the rows, giving a plant population of 33,333 plants ha^-1.  Basal fertiliser was applied at planting (300 kg ha^-1 Compound D (8%N, 14%P₂O₅,  7%K₂O)) while 250 kg ha^-1  ammonium nitrate (34.5% N) was applied 6 WACE.

Each plot had a gross plot size  of 32.4 m² while the net plot size was 14.4 m².  To minimise exo-drift the plots within the same block were separated by 1 m wide alleys.  The blocks were 1.5 m apart and each block measured 27 m long by 6 m wide.  Hand weeding at 2 wace was done using a hand hoe to remove all weed species.  Dicamba (480 g a.i. ha^-1) was applied at a rate of 500 ml and 2,4-D at 750 g a.i. ha^-1).  The knapsack sprayer was calibrated to deliver 250 1 ha^-1 of the spray mixture.

At 8 wace, a 1 m square quadrant was cast in the net plot at random and the area it covered marked with metal pegs.  The number of emerged S.asiatica plants within the marked area were then recorded at 8, 10 and 12 wace.  Data were subjected to logarithmic transformation before analysis of variance.

Maize grain was harvested at crop maturity (5 months after planting) and the field moisture content of the grain determined using a grain master moisture metre.  The grain yield was   adjusted to 12.5% moisture content.  Partial budgets and marginal analysis were done to determine the profitability of the assessed technologies in both seasons (CIMMYT, 1988).

Results

Density of emerged S. asiatica plants.  Dicamba suppressed S.asiatica emergence more than the other treatments at  8, 10 and 12 wace at Govakova during the 1994/95 season (Table 1).  The number of  S. asiatica plants in plots treated with dicamba were 6-12% lower than those  in plots treated with 2,4-D at 8 wace.  At the second site (Sibanga), hand weeding at 2, 5 and 8 wace had 7-30% lower S. asiatica numbers compared to the other three treatments at 12 wace (Table 1).  At Chinyudze the number of  S. asiatica plants that emerged were similar in all the treatments during the 1994/95 season (Table 2).

Table 1. The number of S.asiatica plants that emerged during the 1994/95 season at Chinyudze and Sibanga - 8, 10 and 12 weeks after maize emergence

Figures in parentheses are the actual means.  NS=not significant, WACE = weeks after crop (maize) emergence

Table 2.  The number of  Striga asiatica plants that emerged during the 1994/95 season at Govakova

Figures in parentheses are the actual means.  WACE = weeks after crop (maize) emergence

A highly significant (P<0.01) site by treatment interaction was obtained when the data were combined over sites at 8 and 12 wace during the 1994/95 season.  There was no site by treatment interaction at 10 wace during the same season (Table 3).  There was a 94% reduction in the number of S. asiatica plants that emerged in plots treated with dicamba compared to the rest of the treatments.

Table 3. Number of  Striga asiatica plants that emerged at 10 weeks after crop emergence (WACE) during the 1994/95 season

Pooled data across sites.  Figures in parentheses are the actual means. WACE = weeks after crop (maize) emergence

When the data collected from the 1995/96 season were combined over sites there was a highly significant (P<0.01) site by treatment interaction at 8, 10 and 12 wace. Application of dicamba resulted in fewer emerged S.asiatica plants at 8 and 10 wace at Chinyudze, Govakova and Sibanga (Tables 4 and 5).  However, at Dengedza, applying 2,4-D at 8 wace and hand weeding at 2, 5 and 8 wace were as effective as applying dicamba when the number of emerged S.asiatica plants were recorded at 12 wace (Table 4).  In the same season, 2,4 -D significantly reduced the number of live S.asiatica plants at 12 wace by between 85-92% at all the sites except at Dengedza (Tables 4 and 5).

TABLE 4. The number of Striga asiatica plants that emerged during the 1995/96 season at Chinyudze and Dengedza

Figures in parentheses are the actual means, and WACE = weeks  after crop (maize) emergence

Table 5. The number of Striga asiatica plants that emerged during the 1995/96 season at Govakova and Sibanga

Figures in parentheses are the actual means, and WACE = weeks  after crop (maize) emergence 

Maize grain yield.  Maize grain yield was similar at all sites in both season.  During the 1994/95 season maize grain yield was very low, averaging 719 kg ha^-1 while in the 1995/96 season the average was 4997 kg ha^-1.  However, data on maize grain yields combined over sites in the 1994/95 season showed that the highest yield occurred in plots hand weeded at 2 and 5 wace  (Table 6). The lowest yield of 355 kg ha^-1 was obtained in plots treated with dicamba (Table 6).  During the 1995/96 season combined analysis over sites resulted in similar grain yields.

Table 6. Maize grain yields combined over sites during the 1994/95 and 1995/96 seasons

WACE = Weeks after crop (maize)  emergence

Economic analysis. There was a 10% adjustment for yield.  Field price of maize during the 1996 marketing year was Z$0.90 kg^-1, and labour was charged at Z$167.35 ha^-1. Using dicamba to control S.asiatica during the drier 1994/95 season resulted in a net loss of Z$196.6 ha^-1 while applying 2,4-D at 8 wace and hand weeding gave net benefits of Z$128.92 ha^-1 and Z$170.20, respectively (Table 7).

Table 7. Partial budget for the 1994/95 season                   

Key :        1  =  hand weeding at 2 and 5 weeks after crop (maize) emergence (wace)
2  =  hand weeding at 2,5 and 8 wace
3  =  hand weeding at 2 wace then 2,4-D at 8 wace
4  =  hand weeding at 2 wace then dicamba at 3 wace

Hand weeding at 2 and 5 wace had a marginal rate of return of 101. 5 % due to applying 2,4-D at 8 wace (Table 8). Applying dicamba resulted in the highest net benefit of Z$5316.65 ha^-1 compared with Z$4737.03 ha^-1 obtained when two hand weedings were done (Table 9). 

Table 8.  Dominance and marginal analysis for the four treatments during the 1994/95 season

The dominated treatments are not included in the calculation of the marginal rate of return; WACE = Weeks after crop (maize) emergence

Table 9. Partial budget for the 1995/96 season

Key :       1    =  hand weeding at 2 and 5 weeks after crop emergence (wace)
2   =  hand weeding at 2,5 and 8 wace
3   =  hand weeding at 2 wace then 2,4-D at 8 wace
4   =  hand weeding at 2 wace then dicamba at 3 wace

There was a 10% adjustment for yield in 1997.  Field price of maize during the 1997 marketing year was Z$1.20 kg^-1 and labour was charged at Z$185.95 ha^-1. Dominance analysis showed that hand weeding at 2 and 5 wace, and hand weeding at 2, 5 and 8 wace had higher costs  but lower net benefits than the treatments above them (Table 10) and were therefore excluded from marginal analysis.  Dicamba had a marginal rate of return of 90% over 2,4-D during the 1995/96 season.

Table 10. Dominance and marginal analysis for the four treatments during the 1995/96 season

Discussion

Density of emerged S.asiatica plants.  The fewer S.asiatica plants observed at 8 and 10 wace were attributed to the systemic nature of dicamba.  Systemic herbicides control S. asiatica plants at the point of incipient attachment and help to reduce the weed seed bank since fewer plants emerge.  Since S.asiatica is an active sink for water, carbohydrates and nutrients from the maize plants there is a high possibility that dicamba will be translocated from the maize’s vascular system to that of the parasite.  Basipetal translocation of  dicamba from the leaves of the maize plant to its roots and acropetal translocation in the attached S. asiatica plants has been demonstrated in the USA (Eplee, 1991).  Eplee (1991) also reported that although a large amount of the foliar applied dicamba remained in maize and sorghum shoots, sufficient concentrations were translocated to the S.  asiatica seedlings on roots of the two crops to cause death of the attached seedlings.  A rapid and extensive metabolism of the absorbed 14C-dicamba by both maize and sorghum was reported by Eplee (1991).  Therefore, the absorption of the herbicide through the soil cannot be ruled out since its effectiveness was sufficiently long.  This explains why dicamba was effective in suppressing the parasitic weed over a number of weeks.

Dicamba provides early protection to the crop by killing the S. asiatica seedlings before full  development (Odhiambo and Ransom, 1993). The use of this herbicide on maize grown in fields infested with S. asiatica helped to give the farmer an immediate benefit from the investment.  The long term effect of using dicamba is a reduction in the S. asiatica seed bank.  The adoption of this technology should not be a problem since marginal analysis showed that for every $1.00 invested in dicamba the farmer recovered the dollar plus an extra $0.91.  Therefore, the perception of most smallhold farmers that herbicides are expensive is misleading, there is certainly great potential of getting higher returns when farmers invest in dicamba.

Although 2,4-D resulted in fewer emerged S. asiatica plants at 12 wace, this is too late to protect the crop from S. asiatica damage and prevent the weed from flowering and setting seed.  Due to the limited mobility within the host plant, the critical time for applying 2,4-D should be at post-emergence stage to prevent S. asiatica from seedling.  Since S.asiatica germination and emergence is staggered, a single application of 2,4-D will not be effective because a fraction of the parasite will have flowered or will emerge well after the spray and escape its adverse effects.  However, it was also reported that 2,4-D can provide pre-emergent control of the weed and has some systemic characteristics (Wilson-Jones, 1953; 1956).  This would improve its efficacy and help to protect the crop.  The findings of this study showed that 2,4-D was effective at 12 wace suggesting that it may have some systematic characteristics, and may work in the same way as dicamba if applied earlier than 8 wace.

Higher costs associated with the hand weeding treatments are very prohibitive to the smallhold  farmers who are resource-poor.  This, coupled with the scarcity of labour in this sector due to rural-urban migration of young people for formal employment and education, makes hand weeding very unattractive.  In addition S. asiatica emerges in late January and early February when most farmers are through with the hand weeding operations in maize and are attending to other crops like groundnut, bambara nut, sunflower and millet, hence the weed is left to flower and set  seeds.  Furthermore, the parasitic weed emerges along the maize rows making its removal by using the hand hoe very difficult.

Maize grain yield.  The lower maize grain yields obtained during the 1994/95 season was attributed to the low and poorly distributed rainfall.  This had an adverse effect especially at silking, tasselling and grain filling when moisture requirements by the crop is very critical for the development of maize cobs.  The little moisture which was available was drawn up from the maize crop by the S. asiatica plants which have a high transpiration rate and low stomatal resistance to water movement.  As a result, the maize plants were severely wilted and little or no grain was formed.  There was adequate rainfall during the 1995/96 season and, consequently, pollen-silk synchronisation and grain filling was good resulting in higher grain yield than during the 1994/95 season.

Conclusions

Dicamba was  very effective and economic in suppressing S. asiatica in a good season.  There is therefore need to encourage the farmers to adopt  this technology.  The effectiveness of 2,4-D may be improved by timing its application since it displayed some systemic characteristics.  It was also found out that hand weeding twice or thrice was not economic for controlling S. asiatica. To minimise environmental contamination and to make agriculture sustainable, there is need to test whether reduced rates of the herbicides are as effective as the rates used in this experiment.

Acknowledgement

We are grateful to the Rockefeller Foundation Forum Programme for funding this study.

References

Abayo, G.O., English, T., Kanampiu, F.K., Ransom, J.K. and Gressel, J.  1998.  Control of parasitic witchweeds (Striga spp.) on corn (Zea mays) resistant to acetolactate synthase inhibitors.  Weed Science 46:459-466.

Award, A.E., Worsham, A.D., Corbin, F.T. and Eplee, R.E. 1991.  Absorption translocation and metabolism of foliarly applied 14c dicamba in sorghum (Sorghum bicolor) and corn (Zea mays) parasitized with Striga asiatica.  In: Proceedings of the Fifth International Symposium of Parasitic Weeds.  Ransom, J.K.,  Musselman, L.J., Worsham, A.D.  and  Parker, C. (Eds.), pp. 537-538.  Nairobi, Kenya.

CIMMYT, 1988. From Agronomic Data to Farmer Recommendations: An Economic Training Manual. pp. 1-79.  Completely revised edition. CIMMYT, Mexico, D.F.

Eplee, R.E. 1991.  Striga control methods.  In: Proceedings of the Eastern and Southern African Regional Maize Workshop, Nairobi and Kitale, Kenya, 18-22 September 1989. pp. 314-320.

Eplee, R.E. and Norris, R.S. 1987.  Control of  Striga  asiatica (L.)  Kuntze with systemic herbicides. In: Parasitic Flowering Plants.  Proceedings of the Fourth International Symposium on Parasitic Flowering Plants.  Weber, H. Chr. and Forstreutr, W. (Eds.), pp. 183-186.  Morburg (Germany, F.R.) : Phillips-Universite.

Gressel, J., Ransom, J.K. and Hassan, E.A. 1996. Biotech-derived herbicide-resistant crops for the third world.  Annual New York Academy of Science 792:140-153.

Karima, S.M.R., Islam, N. and Udin, M.N. 1998.  The Striga problem and its control measures: a review.  International Journal of Tropical Agriculture 13:13-23.

Odhiambo, G.D. and Ransom, J.K. 1993. Effect of dicamba on the control of Striga hermonthica in maize in Western Kenya.  African Crop Science Journal 1:105-110.

Olivier, A. 1996. The relationship between Striga hermonithica and its host: a synthesis.  Canadian Journal of Botany 74:1119-1137.

Ransom, J.K., Odhiambo, G.D. and Gressel, J. 1995. Seed dressing maize with imidazolinone herbicides to control Striga hermonithica (Del.) Benth.  Weed Science Society of America.  Abstracts 35:15.

Tattersfield, J.R. and Cronin, C.H.  1958.  Weed control in maize in Southern Rhodesia, with special reference to the use of herbicides.  In: Proceedings of African Weed Control Conference, Victoria Falls Hotel, Southern Rhodesia. pp. 56-61.

Wilson-Jones, W.K. 1953.  Further experiments on witchweed control.  I.  The effects of hormone weed killer applications on irrigated durra.  Empire Journal of Experimental Agriculture 21:331-334.

Wilson-Jones,W.K.  1956.  Weed control experiments in the Sudan. PANS 2: 36-38.

 ©1999, African Crop Science Society