2.1. Chickpea cultivation practices During the 2002–2003 and 2003–2004 chickpea growing seasons, the project targeted five districts in western Bangladesh – Jessore, Jhenaidah, Magura, Faridpur and Rajbari – which were traditional chickpea cultivation areas in the country (Rahman et al., 2000). In the 2004–2005 and 2005–2006 seasons, three additional adjacent districts were included in the programme – Kushtia, Chuadanga and Pabna. The soils of these districts are mostly silty loams formed on recent Gangetic alluvium (Brammer, 1996). The cropping patterns followed by the smallholder farmers of the region are predominantly based on rainy season rice (June–November) with mostly irrigated rice, wheat or vegetables in the post-rainy season; however, there are upland areas where irrigation is not readily accessible and rainfed pulses and oilseeds are grown. Usual farmer practice (FP) for chickpea cultivation in the target region is to give pre-sowing tillage usually by bullock-drawn plough or power tiller after harvest of rainy season rice during November, hand broadcast seed of a ‘‘local’’ variety (often mixed with seed of mustard, linseed or wheat), cover the seed by another ploughing and levelling, grow entirely rainfed, hand weed as required, and use no fertilizer or pesticide. The chickpea ICM package formulated for on-farm evaluation was initially based on earlier studies on BGM management (Pande et al., 2002) and subsequently modified according to findings in on-farm trials conducted during this project period in Bangladesh. It superimposed the following practices on the normal farmer practice: use of a chickpea variety less susceptible to BGM (mainly BARI chola 5, but ICCL 87322 in some cases); delayed sowing, to late November or early December to prevent excessive vegetative growth; from 2003–2004, reduction of the seed rate, from 50 to 37.5 kg ha1, to prevent excessive plant population density, thinning of the crop if required, to maintain an open canopy; need-based foliar application of fungicide (usually Bavistin1 @ 1 kg ha1) to control BGM; integrated management of chickpea pod borer (scouting for eggs and young larvae, placing of perches to encourage predator birds and need-based insecticide spray, according to Harris et al., 2008); application of triple superphosphate (TSP, @ 20 kg P ha1); and from 2004–2005, seed treatment with the fungicide Vitavax2001 to minimize collar rot.
2.2. On-farm trials Various on-farm trials were conducted, in randomized complete block design, to evaluate candidate components of the ICM package. The effects of seed priming (soaking of seed overnight prior to sowing to promote early growth vigour – Musa et al., 2001), intercropping of chickpea with wheat and linseed to reduce pest and disease incidence, and seed treatment with Bavistin1 to control seed-borne BGM were examined. In 2002–2003, the effects of priming and intercropping were tested in 400 m2 plots in five dispersed replications each at three locations and in 2003–2004 intercropping and seed treatment were tested in 267 m2 plots in five dispersed replications each at two locations. No statistically significant advantages of these treatments over the ICM control were obtained and so they were not included in the ICM package, and are not further discussed here; details of these on-farm trials are reported in CLIMA (2006).
To identify means of managing collar rot, on-farm trials were conducted in fields previously known to be infested by collar rot, in three districts in 2004–2005 and 2005–2006. In 2004–2005, treatments comprised: (1) seed treatment with Vitavax-2001; (2) clearing of rice stubble from the plot; (3) delayed sowing; (4) soil incorporation of poultry litter; (5) soil incorporation of mustard oil cake; and (6) a control (ICM without Vitavax-2001). In 2005–2006, treatments comprised: (1) seed treatment with Vitavax-2001; (2) seed treatment with mustard oil; (3) seed inoculation with antifungal microoganisms (Biopharma1); (4) seed treatment with the fungicide Rovral1; (5) soil incorporation of poultry litter; and (6) a control (ICM without Vitavax-2001). Plot size was 225 m2 and there were three dispersed replications at each location in 2004– 2005 and four in 2005–2006; the chickpea variety used was BARI chola 5.
With the objective of obtaining farmer preferences for varieties and traits of chickpea, chickpea Mother Trials (Johansen et al., 2008) were conducted with 11 entries in 2004–2005 and 13 entries in 2005–2006. Six dispersed replications were sown in Rajbari, Jessore and Jhenaidah Districts in 2004–2005, but the three replications in Rajbari were abandoned due to plot damage, and one replication each was sown in Faridpur, Jessore and Jhenaidah Districts in 2005–2006. Plot size was 25 m2 and ICM practice recommended for the particular season was followed.
2.3. On-farm evaluations From the 2002–2003 season, on-farm evaluations (OFE) compared the ICM package with prevailing FP in paired plots of 666 m2 (the average farmer field size in the target region is 1333 m2, which is 1 bigha, the local unit of area measurement). There were 100 such comparisons across five districts in each of 2002–2003 and 2003– 2004, and a further 20 in two new districts in 2004–2005. After receiving training, along with inputs for chickpea cultivation that they do not normally use (seed of improved varieties, TSP, pesticides), farmers implemented OFEs, with regular monitoring from BARI and Department of Agricultural Extension (DAE) personnel. The OFEs were established in clusters of five around a village, within selected blocks of an Upazila (sub-district). Prior to sowing land was ploughed with either tractor-drawn tynes, a power tiller or bullock-drawnmould-board plough. Seed and fertilizer were hand broadcast, the field again ploughed and then levelled. All plots were grown without irrigation. Sowing was done mostly during 20 November to 10 December. Although it was planned to sow the ICM plots after the FP plots, this proved difficult for farmers to implement due to their preoccupation at that time with harvesting of rice and sowing of post-rice crops. Thus, both treatments were invariably sown at the same time, usually at the delayed time recommended for ICM. Crops matured and were harvested during the latter half of March, although the unusually excessive rain in March 2003 delayed maturity by 1–2 weeks in that season. Grain yields were estimated in 5 1 m2 quadrats per plot by DAE Block Supervisors or other project personnel. Yields of ICM and FP plots were compared by paired ‘‘t’’ test applied for each block. Plots that grew poorly due to factors not related to the ICM practice (e.g. grazing damage) were excluded from the analysis.