Abstract
Experiments were conducted in the Agronomy Field Laboratory of the Bangladesh Agricultural University (BAU), Mymensingh, and in farmers' fields in the Old Brhmaputra Floodplain (OBFP) and the Old Meghna Estuarine Floodplain (OMEFP) agro-ecological zones (AEZ) during 1987 and 1988. The experiments were done to study the phytosociology of weeds, weed vegetation in two AEZS, economic aspects of weeding in farmers' fields and to evaluate the cultural, manual, mechanical and chemical methods of weed control to develop integrated weed management for direct-seeded aus rice.
Weed phytosociology study was done in each agroecological zone under micro-environment, that is, in three elevated land- types where aus rice is grown. In each AEZ 100 aus rice fields were randomly selected in each land type for vegetation analysis. The most important weed species in the three land- types of the two AEZS were identified. The co-efficient of similarity of the weed community between land types within an AEZ had a high degree of similarity except in between medium high and lowland where the lowland usually goes under annual flooding for a longer period. The high degree of similarity of weed vegetation between land types was due to similar agro- environmental condition prevailing in different land types. The weed vegetation between the maximum tillering and the reproductive stage had a high degree of dissimilarity due to low moisture at the early growth stage and high moisture or standing water during the reproductive stage of crop growth. This situation changed the weed community from predominated grass to predominated sedge, and broadleaf weeds. The change. of the weed community was also observed in 1988 due to severe flooding. The weed community between land types of the two AEZS had a high degree of dissimilarity, irrespective of the crop growth stages and year of observation due to different agro-environment influenced mainly by annual flooding, crop production technology, cropping pattern, soil soil properties, climatic and other biotic and abioitic factors in the two
zones.
In an agro-economic study conducted in farmers' fields in OBFP and OMEFP zone during 1987 and 1988, it appeared that the weed number and biomass did not vary between two AEZS with an exception in 1988 when weed growth was hampered due to flood- water. Weed infestation did not vary in modern and local varieties. Modern and local varieties tested had similar leaf area indices. Any rice variety, either modern or local, was responsive to weeding, but farmers' weeding practice was inferior to weed free crops in respect of grain yield, weed suppression, and benefit-cost ratio.
Early-sown aus rice was associated with high weed infestation. Optimum or late-sown aus rice had similar influence on weed growth, crop growth, and grain yield. In unweeded plots the degree of land preparation could not control weeds, and gave poor crop growth, and low grain yield. But in weed free and in farmers' weeding plots, reduced and optimum tillages were similar in controlling weeds and gave
high grain yield. The minimum tillage did not perform well in respect of weed control and optimum grain yield.
crop.
The methods of seeding had no influence on weeds and
no
In unweeded plots the increased seed rate had influence on weed growth and crop plants. In weed free and weeded condition high and optimum seed rates had similar influence on weed infestation, crop growth, and grain yields. Low seed rate gave high weed infestation, low crop stand, poor plant characters, and low grain yield.
No difference in weed growth was observed in modern, local improved, and local varieties. All varieties performed better in weed free condition, compared to farmers' weeding. Although high fertilizer gave more weed growth, it gave high yield with MV rice in weed free plots. On the other hand, high fertilizer did not perform better than optimum fertilizer in crop growth and grain yield. Local varieties responded well to low fertilizer doses in a weed free plot. Local improved varieties were not found performing better than a local variety. In unweeded plots fertilizer application had no influence on crop growth and grain yield of any variety.
Pre-emergence application of butachlor and thiobencarb, combined with one hand weeding, controlled weeds and gave grain yield comparable to that of weed free treatment, whereas butachlor and thiobencarb combined with 2 hand weedings, gave higher net returns and benefit-cost ratio than when combined with one hand weeding Similarly, post-emergence application of basagran and post-emergence raking, combined with two hand weedings, gave less cost, more returns and a higher benefit- cost ratio than weed free treatment in a direct-seeded aus rice.
Even though dissimilarity of weed vegetation was observed between the two AEZS, weed free condition was found essential for proper crop growth and grain yield of aus rice. For effective weed control, therefore, integration of proper cultivation technology and weed control methods have been suggested for high net returns in a direct-seeded modern and local aus rice.