- Email: [email protected]
Influence of flood depth and duration on growth of lowland rice weeds, Cote d’Ivoire
Crop Protection 20 (2001) 691}694
Short communication
In#uence of #ood depth and duration on growth of lowland rice weeds, Cote d'Ivoire R.J. Kent , D.E. Johnson * West Africa Rice Development Association (WARDA), 01 BP 2551, Bouake& , Ivory Coast Natural Resources Insititute, University of Greenwich, Chatham Maritime, Kent ME4 4TB, UK Received 9 February 2001; received in revised form 15 February 2001; accepted 15 February 2001
Abstract The growth of weeds from the soil seedbank and sown Echinochloa colona and E. crus-pavonis was studied in relation to di!erent water depths (0, 2 4 and 8 cm) and #ood durations (2, 4 and 7 days in 7). Deeper #ooding increased plant numbers in Ammannia prieriana, Sphenoclea zeylanica and Heteranthera callifolia, but decreased plant numbers in E. colona and E. crus-pavonis and had no e!ect on Fimbristylis littoralis. Increased #ood duration increased plant numbers in H. callifolia, had no e!ect on S. zeylanica or A. prieriana, but decreased plant numbers in S. xlicaulis, E. colona and E. crus-pavonis. Weed biomass at 28 days followed a similar pattern, with increased depth resulting in higher biomass of H. callifolia and A. prieriana, but lower biomass of S. xlicaulis, E. colona and E. crus-pavonis. Deeper but intermittent #ooding stimulated growth of A. prieriana, and F. littoralis. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Water; Flooding; Weed control
1. Introduction Weed competition is one of the most important causes of yield loss in rice, and is estimated to be 15}21% in the tropics, with the largest losses occurring in the rainfed ecologies (Oerke et al., 1994). A major bene"t of improved water control in lowland rice production is the increased ability to manage weeds by #ooding. In Bangladesh, maintaining transplanted rice "elds weed free resulted in a 60% increase in grain yield where soils were saturated, compared to a 3% increase where the "eld had been continuously #ooded to a depth of 5}10 cm (Ahmed et al., 1980). In India, weed numbers and biomass were reduced by about 35% in rice #ooded to a depth of 10 cm for the duration of the crop, compared to plots where the soil was only saturated (Misra et al., 1981). Also in India, across a range of land preparation methods 7 cm #ooding in the early vegetative phase of rice reduced weed numbers and weight considerably (Reddy and Reddy, 1999). Studies have demonstrated the importance of water control in limiting weed infestation in lowland rice "elds * Corresponding author. E-mail address: [email protected] (D.E. Johnson).
in Co( te d'Ivoire where poor water control was closely correlated with greater weed growth, and in consequence greater yield losses due to weeds (Becker and Johnson, 1999). The potential of #ooding as a component of integrated weed management strategies can be enhanced if the response of key lowland weed species to #ooding is better understood. This paper presents the results of simple controlled #ooding experiments designed to explore the e!ect of #ood depth and duration on important lowland rice weeds in West Africa.
2. Materials and methods Flooding experiments were conducted at WARDA's research station in M'be, Co( te d'Ivoire. Plastic basins (50;40;18 cm) were "lled with a soil taken from irrigated lowland rice "elds that had been dry ploughed and harrowed. Basins were "lled to a depth of 20 cm with dry soil, water was added and the soil in each basin &puddled' by hand. Basins were drained and half of the area shallow sown (5 mm) with one line each of Echinochloa colona (L.) Link (0.5 g) and Echinochloa crus-pavonis (Kunth) Schultes (1 g). The sown area was hand weeded to leave only the sown species while the other half of the
0261-2194/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 1 - 2 1 9 4 ( 0 1 ) 0 0 0 3 4 - 5
692
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
R.J. Kent, D.E. Johnson / Crop Protection 20 (2001) 691}694
Table 1 Emergence (transformed counts) and plant biomass after 28 days, means and signi"cance levels from anova combined over three experiments Flood depth (cm) 0
2
Flood duration (days out of 7) 4
8
S.E.$
2
4
7
S.E.$
(a) Main ewects on pldant number (log#1) C. diwormis 1.070 1.222 H. callifolia 0.453 0.842 S. zeylanica 0.127 0.294 S. xlicaulis 0.433 0.273 E. colona 1.353 0.855 E. crus-pavonis 1.900 1.800
1.217 0.942 0.414 0.243 0.767 1.719
1.257 0.961 0.375 0.273 0.720 1.703
0.0443 0.0449 0.0417 0.0374 0.0571 0.0454
1.192 0.764 0.285 0.459 1.073 1.867
1.179 0.751 0.279 0.336 0.929 1.815
1.203 0.883 0.344 0.122 0.769 1.660
0.0383 0.0389 0.0361 0.0324 0.0494 0.0393
(b) Main ewects on plant biomass C. diwormis 0.306 H. callifolia 0.017 S. zeylanica 0.006 S. xlicaulis 0.019 E. colona 0.850 E. crus-pavonis 1.436
0.447 0.303 0.022 0.003 0.469 0.989
0.311 0.244 0.031 0.000 0.133 0.864
0.0490 0.0427 0.0442 0.0042 0.0880 0.0967
0.400 0.088 0.008 0.006 0.492 1.504
0.298 0.138 0.006 0.004 0.433 1.188
0.531 0.308 0.115 0.006 0.423 0.613
0.0424 0.0370 0.0383 0.0036 0.0762 0.0848
0.575 0.147 0.114 0.000 0.344 1.117
basin was left to natural weed growth from the soil seed bank. Drainage holes in the side of the basins at "xed heights and the use of rubber plugs allowed water levels to be maintained at the desired depth for each treatment. The experiment was laid out as a randomised complete block design with four replicates; each basin represented a sample unit. Treatments were water depth (four levels: 0, 2, 4 and 8 cm) and #ood duration (three levels: continuous, 4 days out of 7, 2 days out of 7). At 28 days, plants were harvested, oven dried and weighed. The experiment was repeated three times in the period from May to October 1999. After each experiment the soil was thoroughly mixed and puddled before re-sowing. Data were analysed by analysis of variance with data combined over experiments. Plant counts were log transformed before the analysis.
3. Results 3.1. Emergence Natural weed growth was dominated by seven species: the broadleafs Ammannia prieureana Guill. and Perr., Sphenochlea zeylanica Gaertner, Spilanthes xlicaulis and Heteranthera callifolia Rchb. Ex Kunth; the sedges Cyperus diwormis and Fimbristylis littoralis Gaudich.; and the grass Echinochloa crus-pavonis. In view of the clear di!erences between species in their responses to #ooding, each is considered separately (see Table 1). Ammannia prieureana (Fig. 1a): At 28 days, deeper #ooding increased plant number (P(0.01), but the main e!ects of #ood duration were not signi"cant. There were signi"cant interaction e!ects (P(0.01) between #ooding depth and duration (Table 2), plant number was least
where the soil was temporarily saturated (2 days out of 7) and greatest with deep (8 cm) but temporary #ooding. Plant biomass was increased by deeper #ooding (P(0.01) but #ood duration had no signi"cant e!ect. Plant biomass was highest in treatments with #ooding (4}8 cm) and periodic drying. Sphenochlea zeylanica (Fig. 1b): Increased #ood depth increased plant number (P(0.01) but #ood duration had no signi"cant e!ects. Weed biomass was extremely low, but there were no signi"cant treatment e!ects at 28 days. Heteranthera callifolia (Fig. 1c): Plant numbers were increased by #ood depth (P(0.01) and duration (P(0.05), re#ecting a clear preference for #ooded conditions. Biomass after 28 days was substantially higher in the continuously #ooded treatments (4 and 8 cm). Spilanthes xlicaulis (Fig. 1d): Both #ood depth and duration (P(0.01) signi"cantly a!ected emergence with numbers reduced by deeper #ooding or increased #ood duration. Plant numbers in the no-#ood treatment where the soil was allowed to dry for 5 days out of 7, were signi"cantly higher than all but one of the remaining treatments. Weed biomass was extremely low, but there were no signi"cant treatment e!ects at 28 days. Cyperus diwormis: Plant number was least in the no-#ood treatment which was allowed to dry 5 days out of 7. There appeared to be little or no e!ect of the other treatments on plant number. Treatment e!ects on plant biomass after 28 days were more apparent, both #ood depth and duration having signi"cant e!ects (P(0.01). The greatest biomass was recorded in the shallow (2 cm) continuous #ooding treatment. Fimbristylis littoralis: Flood treatments had no signi"cant e!ect on plant number, but #ood depth (P(0.01) and duration (P(0.01) had signi"cant e!ects on "nal
57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
R.J. Kent, D.E. Johnson / Crop Protection 20 (2001) 691}694
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
693
Fig. 1. Emergence of species from seed bank (a}d); and Echinochloa species (e}f) under di!erent #ooding regimes.
Table 2 Interaction e!ects of #ood depth and duration for Ammannia prieureana and Fimbristylis littoralis on plant numbers and biomass, means and signi"cance levels from anova combined over three experiments Plant no.s (log X#1)
Biomass (g)
Flood depth
Flood depth
Species
Duration
0 cm
2 cm
4 cm
8 cm
0 cm
2 cm
4 cm
8 cm
A. prieureana
7/7 days 4/7 days 2/7 days
1.328 0.860 0.888
1.461 1.190 1.456
1.142 1.600 1.593
0.133 0.092 0.058
0.233 0.317 0.408
0.400 0.508 0.425
7/7 days 4/7 days 2/7 days
1.524 1.516 1.335
1.556 1.440 1.424
1.180 1.439 1.561
1.207 1.695 1.654 * ns 0.2067 1.322 1.452 1.653 ns ns 0.0663
0.883 1.450 0.958
0.750 0.458 0.625
0.383 0.400 0.850
0.342 0.425 0.367 * ns 0.0718 0.308 0.425 1.083 * * 0.165
Flood depth Duration S.E. $ F. littoralis
Flood depth Duration S.E. $
57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
694
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
R.J. Kent, D.E. Johnson / Crop Protection 20 (2001) 691}694
dry weights. There was a signi"cant interaction between depth and #ood duration (P(0.01) (Table 2); deeper #ooding and draining appears to stimulate growth, while growth was greatest where soil was saturated 4 days out of 7. 3.2. Sown treatments Echinochloa colona (Fig. 1e): Increased depth of #ooding reduced plant number and biomass (P(0.01). Flood duration reduced plant number (P(0.01) but had no signi"cant e!ect on biomass. E. crus-pavonis (Fig. 1f): The experiment shows a clear impact of #ood depth and duration on plant biomass (P(0.01) and emergence (P(0.01) both were lowest where #ooding was deeper of maintained longer. Constant #ooding resulted in the greatest depression of plant growth. Maintaining a constant depth of 2 cm reduced plant growth by over 60% compared to the case where the soil was #ooded for only 2 days out of 7. At this #ood duration, water depth had relatively little e!ect on growth.
4. Discussion The experiment demonstrates substantial di!erences in the response of common lowland rice weeds to #ooding. Flooding soils can be expected to reduce the emergence of some important weeds but encourages the growth of others. The response of S. xlicaulis and E. colona indicates a preference for dry conditions, while H. callifolia was favoured by increased duration and depth of #ooding. For some species, such as F. littoralis, #ooding may have a limited impact on seedling emergence, but it could be employed to check subsequent growth. Pons (1982) reported that submergence to 10}12 cm depth greatly reduced growth and emergence of F. littoralis, while Monochoria vaginalis was not a!ected by #ooding. In our study the presence of standing water encouraged growth in A. prieureana and H. callifolia. The greater problems posed by grassy weeds such as E. crus-pavonis, both in their high competitive ability and tendency to achieve high levels of infestation, mean that the disadvantages of #ooding, by encouraging aquatic weeds, may be outweighed by the potential to control these species. In
undeveloped lowland areas stream diversion or the building of shallow bunds to retain rainfall and surface run-o! may achieve this. It must be acknowledged that many rice farmers in West Africa have limited control over water even in irrigated "elds, which restricts their capacity to use #ooding as a weed control mechanism. Nonetheless, the results of this study show that with only shallow #ooding the growth of one of the most serious weeds can be greatly reduced. Constant #ooding to a depth of 2 cm or more could check the build-up-of E. crus-pavonis. This "nding is similar to that of Smith and Fox (1973), who report that the emergence of E. crus-galli, a close relative of E. crus-pavonis, was reduced by 90% by #ooding to 1.3 cm. A continuous #ood after herbicide application or hand weeding could largely prevent subsequent growth of this weed and reduce the need for further interventions.
Acknowledgements The authors are grateful to Mariko Mariame and Kouadio Benoit for their assistance with the "eld work. This paper is an output from a project partly funded by the UK Department for International Development (DFID).
References Ahmed, N.U., Hoque, Z.M., Khan, A.H., Khan, S.A.A., 1980. E!ect of weed management and water regime on the yield of farmers' boro rice. Int. Rice Res. Newslett. 5 (4), 23. Becker, M., Johnson, D.E., 1999. Rice yield productivity in irrigated systems of the forest zone in Cote d'Ivoire. Field Crops Res 60, 201}208. Misra, A., Tosh, G.C., Nanda, K.C., 1981. E!ects of herbicides and water management regimes on weeds and grain yields of transplanted rice in India. Int. Rice Res. Newslett. 6 (5), 20}21. Oerke, E.C., Dehne, H.W., SchoK nbeck, F., Weber, A., 1994. Crop Production and Crop Protection*Estimated Losses in Major Food and Cash Crops. Elsevier, Amsterdam, p. 808. Pons, T.L., 1982. Factors a!ecting weed seed germination and seedling growth in lowland rice in Indonesia. Weed Res. 22, 155}161. Reddy, B.S., Reddy, S.R., 1999. E!ect of soil and water management on weed dynamics in lowland rice. Indian J. Weed Sci. 31 (3/4), 179}182. Smith, R.J., Fox, W.T., 1973. Soil water and the growth of rice and weeds. Weed Sci. 21 (1), 59}63.
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
Short communication
In#uence of #ood depth and duration on growth of lowland rice weeds, Cote d'Ivoire R.J. Kent , D.E. Johnson * West Africa Rice Development Association (WARDA), 01 BP 2551, Bouake& , Ivory Coast Natural Resources Insititute, University of Greenwich, Chatham Maritime, Kent ME4 4TB, UK Received 9 February 2001; received in revised form 15 February 2001; accepted 15 February 2001
Abstract The growth of weeds from the soil seedbank and sown Echinochloa colona and E. crus-pavonis was studied in relation to di!erent water depths (0, 2 4 and 8 cm) and #ood durations (2, 4 and 7 days in 7). Deeper #ooding increased plant numbers in Ammannia prieriana, Sphenoclea zeylanica and Heteranthera callifolia, but decreased plant numbers in E. colona and E. crus-pavonis and had no e!ect on Fimbristylis littoralis. Increased #ood duration increased plant numbers in H. callifolia, had no e!ect on S. zeylanica or A. prieriana, but decreased plant numbers in S. xlicaulis, E. colona and E. crus-pavonis. Weed biomass at 28 days followed a similar pattern, with increased depth resulting in higher biomass of H. callifolia and A. prieriana, but lower biomass of S. xlicaulis, E. colona and E. crus-pavonis. Deeper but intermittent #ooding stimulated growth of A. prieriana, and F. littoralis. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Water; Flooding; Weed control
1. Introduction Weed competition is one of the most important causes of yield loss in rice, and is estimated to be 15}21% in the tropics, with the largest losses occurring in the rainfed ecologies (Oerke et al., 1994). A major bene"t of improved water control in lowland rice production is the increased ability to manage weeds by #ooding. In Bangladesh, maintaining transplanted rice "elds weed free resulted in a 60% increase in grain yield where soils were saturated, compared to a 3% increase where the "eld had been continuously #ooded to a depth of 5}10 cm (Ahmed et al., 1980). In India, weed numbers and biomass were reduced by about 35% in rice #ooded to a depth of 10 cm for the duration of the crop, compared to plots where the soil was only saturated (Misra et al., 1981). Also in India, across a range of land preparation methods 7 cm #ooding in the early vegetative phase of rice reduced weed numbers and weight considerably (Reddy and Reddy, 1999). Studies have demonstrated the importance of water control in limiting weed infestation in lowland rice "elds * Corresponding author. E-mail address: [email protected] (D.E. Johnson).
in Co( te d'Ivoire where poor water control was closely correlated with greater weed growth, and in consequence greater yield losses due to weeds (Becker and Johnson, 1999). The potential of #ooding as a component of integrated weed management strategies can be enhanced if the response of key lowland weed species to #ooding is better understood. This paper presents the results of simple controlled #ooding experiments designed to explore the e!ect of #ood depth and duration on important lowland rice weeds in West Africa.
2. Materials and methods Flooding experiments were conducted at WARDA's research station in M'be, Co( te d'Ivoire. Plastic basins (50;40;18 cm) were "lled with a soil taken from irrigated lowland rice "elds that had been dry ploughed and harrowed. Basins were "lled to a depth of 20 cm with dry soil, water was added and the soil in each basin &puddled' by hand. Basins were drained and half of the area shallow sown (5 mm) with one line each of Echinochloa colona (L.) Link (0.5 g) and Echinochloa crus-pavonis (Kunth) Schultes (1 g). The sown area was hand weeded to leave only the sown species while the other half of the
0261-2194/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 1 - 2 1 9 4 ( 0 1 ) 0 0 0 3 4 - 5
692
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
R.J. Kent, D.E. Johnson / Crop Protection 20 (2001) 691}694
Table 1 Emergence (transformed counts) and plant biomass after 28 days, means and signi"cance levels from anova combined over three experiments Flood depth (cm) 0
2
Flood duration (days out of 7) 4
8
S.E.$
2
4
7
S.E.$
(a) Main ewects on pldant number (log#1) C. diwormis 1.070 1.222 H. callifolia 0.453 0.842 S. zeylanica 0.127 0.294 S. xlicaulis 0.433 0.273 E. colona 1.353 0.855 E. crus-pavonis 1.900 1.800
1.217 0.942 0.414 0.243 0.767 1.719
1.257 0.961 0.375 0.273 0.720 1.703
0.0443 0.0449 0.0417 0.0374 0.0571 0.0454
1.192 0.764 0.285 0.459 1.073 1.867
1.179 0.751 0.279 0.336 0.929 1.815
1.203 0.883 0.344 0.122 0.769 1.660
0.0383 0.0389 0.0361 0.0324 0.0494 0.0393
(b) Main ewects on plant biomass C. diwormis 0.306 H. callifolia 0.017 S. zeylanica 0.006 S. xlicaulis 0.019 E. colona 0.850 E. crus-pavonis 1.436
0.447 0.303 0.022 0.003 0.469 0.989
0.311 0.244 0.031 0.000 0.133 0.864
0.0490 0.0427 0.0442 0.0042 0.0880 0.0967
0.400 0.088 0.008 0.006 0.492 1.504
0.298 0.138 0.006 0.004 0.433 1.188
0.531 0.308 0.115 0.006 0.423 0.613
0.0424 0.0370 0.0383 0.0036 0.0762 0.0848
0.575 0.147 0.114 0.000 0.344 1.117
basin was left to natural weed growth from the soil seed bank. Drainage holes in the side of the basins at "xed heights and the use of rubber plugs allowed water levels to be maintained at the desired depth for each treatment. The experiment was laid out as a randomised complete block design with four replicates; each basin represented a sample unit. Treatments were water depth (four levels: 0, 2, 4 and 8 cm) and #ood duration (three levels: continuous, 4 days out of 7, 2 days out of 7). At 28 days, plants were harvested, oven dried and weighed. The experiment was repeated three times in the period from May to October 1999. After each experiment the soil was thoroughly mixed and puddled before re-sowing. Data were analysed by analysis of variance with data combined over experiments. Plant counts were log transformed before the analysis.
3. Results 3.1. Emergence Natural weed growth was dominated by seven species: the broadleafs Ammannia prieureana Guill. and Perr., Sphenochlea zeylanica Gaertner, Spilanthes xlicaulis and Heteranthera callifolia Rchb. Ex Kunth; the sedges Cyperus diwormis and Fimbristylis littoralis Gaudich.; and the grass Echinochloa crus-pavonis. In view of the clear di!erences between species in their responses to #ooding, each is considered separately (see Table 1). Ammannia prieureana (Fig. 1a): At 28 days, deeper #ooding increased plant number (P(0.01), but the main e!ects of #ood duration were not signi"cant. There were signi"cant interaction e!ects (P(0.01) between #ooding depth and duration (Table 2), plant number was least
where the soil was temporarily saturated (2 days out of 7) and greatest with deep (8 cm) but temporary #ooding. Plant biomass was increased by deeper #ooding (P(0.01) but #ood duration had no signi"cant e!ect. Plant biomass was highest in treatments with #ooding (4}8 cm) and periodic drying. Sphenochlea zeylanica (Fig. 1b): Increased #ood depth increased plant number (P(0.01) but #ood duration had no signi"cant e!ects. Weed biomass was extremely low, but there were no signi"cant treatment e!ects at 28 days. Heteranthera callifolia (Fig. 1c): Plant numbers were increased by #ood depth (P(0.01) and duration (P(0.05), re#ecting a clear preference for #ooded conditions. Biomass after 28 days was substantially higher in the continuously #ooded treatments (4 and 8 cm). Spilanthes xlicaulis (Fig. 1d): Both #ood depth and duration (P(0.01) signi"cantly a!ected emergence with numbers reduced by deeper #ooding or increased #ood duration. Plant numbers in the no-#ood treatment where the soil was allowed to dry for 5 days out of 7, were signi"cantly higher than all but one of the remaining treatments. Weed biomass was extremely low, but there were no signi"cant treatment e!ects at 28 days. Cyperus diwormis: Plant number was least in the no-#ood treatment which was allowed to dry 5 days out of 7. There appeared to be little or no e!ect of the other treatments on plant number. Treatment e!ects on plant biomass after 28 days were more apparent, both #ood depth and duration having signi"cant e!ects (P(0.01). The greatest biomass was recorded in the shallow (2 cm) continuous #ooding treatment. Fimbristylis littoralis: Flood treatments had no signi"cant e!ect on plant number, but #ood depth (P(0.01) and duration (P(0.01) had signi"cant e!ects on "nal
57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
R.J. Kent, D.E. Johnson / Crop Protection 20 (2001) 691}694
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
693
Fig. 1. Emergence of species from seed bank (a}d); and Echinochloa species (e}f) under di!erent #ooding regimes.
Table 2 Interaction e!ects of #ood depth and duration for Ammannia prieureana and Fimbristylis littoralis on plant numbers and biomass, means and signi"cance levels from anova combined over three experiments Plant no.s (log X#1)
Biomass (g)
Flood depth
Flood depth
Species
Duration
0 cm
2 cm
4 cm
8 cm
0 cm
2 cm
4 cm
8 cm
A. prieureana
7/7 days 4/7 days 2/7 days
1.328 0.860 0.888
1.461 1.190 1.456
1.142 1.600 1.593
0.133 0.092 0.058
0.233 0.317 0.408
0.400 0.508 0.425
7/7 days 4/7 days 2/7 days
1.524 1.516 1.335
1.556 1.440 1.424
1.180 1.439 1.561
1.207 1.695 1.654 * ns 0.2067 1.322 1.452 1.653 ns ns 0.0663
0.883 1.450 0.958
0.750 0.458 0.625
0.383 0.400 0.850
0.342 0.425 0.367 * ns 0.0718 0.308 0.425 1.083 * * 0.165
Flood depth Duration S.E. $ F. littoralis
Flood depth Duration S.E. $
57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
694
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
R.J. Kent, D.E. Johnson / Crop Protection 20 (2001) 691}694
dry weights. There was a signi"cant interaction between depth and #ood duration (P(0.01) (Table 2); deeper #ooding and draining appears to stimulate growth, while growth was greatest where soil was saturated 4 days out of 7. 3.2. Sown treatments Echinochloa colona (Fig. 1e): Increased depth of #ooding reduced plant number and biomass (P(0.01). Flood duration reduced plant number (P(0.01) but had no signi"cant e!ect on biomass. E. crus-pavonis (Fig. 1f): The experiment shows a clear impact of #ood depth and duration on plant biomass (P(0.01) and emergence (P(0.01) both were lowest where #ooding was deeper of maintained longer. Constant #ooding resulted in the greatest depression of plant growth. Maintaining a constant depth of 2 cm reduced plant growth by over 60% compared to the case where the soil was #ooded for only 2 days out of 7. At this #ood duration, water depth had relatively little e!ect on growth.
4. Discussion The experiment demonstrates substantial di!erences in the response of common lowland rice weeds to #ooding. Flooding soils can be expected to reduce the emergence of some important weeds but encourages the growth of others. The response of S. xlicaulis and E. colona indicates a preference for dry conditions, while H. callifolia was favoured by increased duration and depth of #ooding. For some species, such as F. littoralis, #ooding may have a limited impact on seedling emergence, but it could be employed to check subsequent growth. Pons (1982) reported that submergence to 10}12 cm depth greatly reduced growth and emergence of F. littoralis, while Monochoria vaginalis was not a!ected by #ooding. In our study the presence of standing water encouraged growth in A. prieureana and H. callifolia. The greater problems posed by grassy weeds such as E. crus-pavonis, both in their high competitive ability and tendency to achieve high levels of infestation, mean that the disadvantages of #ooding, by encouraging aquatic weeds, may be outweighed by the potential to control these species. In
undeveloped lowland areas stream diversion or the building of shallow bunds to retain rainfall and surface run-o! may achieve this. It must be acknowledged that many rice farmers in West Africa have limited control over water even in irrigated "elds, which restricts their capacity to use #ooding as a weed control mechanism. Nonetheless, the results of this study show that with only shallow #ooding the growth of one of the most serious weeds can be greatly reduced. Constant #ooding to a depth of 2 cm or more could check the build-up-of E. crus-pavonis. This "nding is similar to that of Smith and Fox (1973), who report that the emergence of E. crus-galli, a close relative of E. crus-pavonis, was reduced by 90% by #ooding to 1.3 cm. A continuous #ood after herbicide application or hand weeding could largely prevent subsequent growth of this weed and reduce the need for further interventions.
Acknowledgements The authors are grateful to Mariko Mariame and Kouadio Benoit for their assistance with the "eld work. This paper is an output from a project partly funded by the UK Department for International Development (DFID).
References Ahmed, N.U., Hoque, Z.M., Khan, A.H., Khan, S.A.A., 1980. E!ect of weed management and water regime on the yield of farmers' boro rice. Int. Rice Res. Newslett. 5 (4), 23. Becker, M., Johnson, D.E., 1999. Rice yield productivity in irrigated systems of the forest zone in Cote d'Ivoire. Field Crops Res 60, 201}208. Misra, A., Tosh, G.C., Nanda, K.C., 1981. E!ects of herbicides and water management regimes on weeds and grain yields of transplanted rice in India. Int. Rice Res. Newslett. 6 (5), 20}21. Oerke, E.C., Dehne, H.W., SchoK nbeck, F., Weber, A., 1994. Crop Production and Crop Protection*Estimated Losses in Major Food and Cash Crops. Elsevier, Amsterdam, p. 808. Pons, T.L., 1982. Factors a!ecting weed seed germination and seedling growth in lowland rice in Indonesia. Weed Res. 22, 155}161. Reddy, B.S., Reddy, S.R., 1999. E!ect of soil and water management on weed dynamics in lowland rice. Indian J. Weed Sci. 31 (3/4), 179}182. Smith, R.J., Fox, W.T., 1973. Soil water and the growth of rice and weeds. Weed Sci. 21 (1), 59}63.
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85