2.1 Diagnostic trial, 2001/02 An experiment using a subtractive experimental design with the following treatments was conducted: 1.Control; no seed treatment or minor element addition. 2.Seed-treated control; seed priming (soaking of seed in water for 8 h overnight prior to sowing), inoculation with Rhizobium (specific for chickpea, obtained from Bangladesh Institute of Nuclear Agriculture [BINA]). 3.Full-nutrient control (All); Treatment 2 with the following elements added as Reagent grade salts to the soil (kg ha–1): 1.0 B + 0.5 Mo + 5 Zn + 20 S. 4.Treatment 2 with Mo + Zn + S; i.e. All – B. 5.Treatment 2 with B + Zn + S; i.e. All – Mo. 6.Treatment 2 with B + Mo + S; i.e. All – Zn. 7.Treatment 2 with B + Mo + Zn; i.e. All – S. The experiment was done in a farmer’s field at Chabbishnagar Village, Godagari Upazila (= subdistrict), Rajshahi District, HBT (southern part), Bangladesh. A complete randomized block design was used with four replications. Plot size was 1.8 m × 2.0 m, with six rows of chickpea variety BARI chola 5 sown 30 cm apart and 2 m long. The land was plowed and leveled on November 16, 2001 and nutrients added and seeds sown on the following day. Furrows were opened to 6–8 cm depth and nutrient salts added evenly throughout the furrows of a plot, according to treatment. Nutrient salts were mixed with river sand to aid even distribution. Phosphorus was also added to each plot in the furrows at 20 kg P ha–1, as triple superphosphate (TSP). Furrows were then partially filled with soil, to 4–5 cm depth, to prevent direct contact of seed with fertilizer and hence possible nutrient toxicity, and seeds placed at 5 cm intervals within a row. The furrows were then immediately covered and spray-irrigated, to ensure even germination. This experiment was attempted in the previous two seasons, but it had to be abandoned due to uneven emergence because of rapid drying of the soil surface once furrows were opened. Careful planting in rows in small plots was considered necessary in order to reduce variability to the extent that treatment differences could be detected. Soil samples were taken before planting at 0–10, 0–15, and 15–30 cm, with one composite sample per replication. Samples were taken at both 0–10 and 0–15 cm depths, because 0–15 cm is the recommended sampling depth for Bangladesh, but for most HBT soils, the depth to the plough pan is only around 10 cm, and nodulation of chickpea occurs within this layer. The samples were analyzed by the Soil Resources Development Institute (SRDI), Shyampur, Bangladesh for pH, organic carbon, total N, and available P. The crop was grown rainfed, with only 16 mm rainfall falling on the crop during January–February 2002. Plots were thinned to give a within-row spacing equivalent to 10 cm at 20–30 d after sowing. No weeding was necessary. The crop was sprayed with Helicoverpa nuclear polyhedrosis virus (HNPV) twice in January 2002 to control pod borer (Musa et al., 2007), which proved effective as there was negligible pod damage. A visual ranking of vegetative growth in each plot was made on February 2, 2002. The plot with the least growth was designated “1” and that with most growth designated “5”. A plot intermediate between these was designated “3”. A plot intermediate between “1” and “3” was designated “2” and one between “3” and “5” designated “4”. All other plots were ranked in relation to these reference plots. Due to the small plot size, plant samples were not cut so as to calibrate rank with dry mass, and thus, only rankings were analyzed. Plots were harvested at maturity on March 22, 2002. The central two rows of each plot were harvested, to avoid border effects, and grain separated from aboveground residue and all samples sun-dried. Four adjacent plots along one edge of the experiment nearest a footpath had been damaged by passing humans and grazing animals. These included two full nutrient control plots (Treatment 3) and one each of “minus B” (Treatment 4) and “minus Zn” (Treatment 6). Thus, these plots were excluded from the analysis of variance by using a missing-plot technique (Steel and Torrie, 1980; pp. 209–213).
2.2 Multilocation Mo experiments, 2002/03 In order to determine how widespread Mo deficiency is across the HBT, and if a Rhizobium response can be found in the presence of Mo, on-farm trials were done in the northern (Porsha), central (Gomostapur), and central-southern (Amnura) regions of the HBT. Treatments were applied as follows: Control: recommended agronomic practice for chickpea (Musa and Johansen, 2003), viz. variety BARI chola 5 at a seed rate of 50 kg ha–1, seed priming, 20 kg P ha–1 as TSP, hand broadcasting, cross-ways plowing, grown rainfed, IPM for pod borer, etc. Soil Mo: same as the control, but Mo added as sodium molybdate, at the rate of 500 g Mo ha–1, mixed with river sand and hand broadcast prior to plowing. Soil Mo + Rh: same as the control, but with Mo added to the soil and Rhizobium inoculation with lime pelleting after coating seed with sticker and inoculum.