Site characteristics The study was undertaken in two locations, (i) polder 31 at Dacope, Khulna (22.37°N and 89.30°E) having high salinity, and (ii) in polder 43/1 at Amtali, Barguna (22.02°N and 90.14°E) with medium salinity level environments. The area has two distinct seasons: a rainy season from June to October and a dry season from November to May. Formerly, the soil was very saline because of the intrusion of seawater during the dry season. To increase agricultural production, in the late 1960’s, the area was included in the government’s Coastal Embankment Project (CEP), in which the areas were surrounded by a dykes or embankments (called polders), separating them hydrologically from the main river system and offering protection against tidal floods, salinity intrusion and sedimentation (Islam, 2006). The river water salinity varies year-round. The onset of the rainy season makes river water fresh with an average electrical conductivity mostly below 1.0 dS m-1 from July to December and after that the river water gradually becomes more saline reaching more than 20 dS m-1, by April-May. Therefore, the rainy season is suitable for crop cultivation from the perspective of water salinity, but there is the problem of waterlogging in most coastal regions.
Soil of the Dacope site is silty-clay to clay and salinity varies from 2-3 dS m-1 in July-September and 8-12 dS m-1 in April-May. Most arable land (70-75%) in the Dacope area is medium-high to low land and about 25-30% of total area is medium-high to high land. The cropping system is based on a monoculture of wet season rice. Most of the land is under low-yielding local varieties and some areas are under long-duration modern varieties. On the other hand, soil in the Amtali area is of silty-clay type and soil salinity in dry season is low to medium (6-8 dS m-1) and decreases to low (1-2 dS m-1) in the wet season. Most arable land (80-85%) in the village is ‘medium-high land two’ and about 15-20% of total area is ‘medium-high land-one’ (BBS, 2017). In a normal wet season, most of the land is inundated between 0.3 to 0.6 m and to cope with this water depth, farmers use tall rice seedlings. Beside this, every year one or two depressions form in the Bay of Bengal and most of the coastal zone’s water depths increase about 0.3-0.5 m above the normal flood level.
Varietal adoption and yield assessment The experiment was established in a Randomized Complete Block Design (RCBD) at Dacope, Khulna and Amtali, Barguna for three consecutive years, 2016- 2018. Tested rice varieties were BR23, BRRI dhan53, BRRI dhan54, BRRI dhan62, BRRI dhan66 and BRRI dhan73 along with popular local varieties (BR11, BR23, BRRI dhan34, Boran, Swarnagota, Tepu, Vogon) as a check for 2016 T. aman season based on the availability and suitability of seeds. According to farmers’ preferences, some unsuitable varieties were discarded in 2017 and 2018, while BRRI dhan76 and BRRI dhan77 were added. Individual farmer’s plots were treated as disperse replications for similar varieties. In the Dacope area, the tested varieties were sown on 10-20 July and 25-30 days old seedlings were transplanted on 10-25 August at 20 cm x 20 cm spacing with 3-5 seedlings per hill. In Amtali, Barguna area, the tested varieties were sown on 15-31 July and 25-30 days old seedlings were transplanted on 10-30 August at 20 cm x 20 cm spacing with 3-6 seedlings per hill. Fertilizer was applied at 75/68-12-53-12 kg NPKS ha-1. Nitrogen fertilizer was used at 68 kg ha-1 for short duration (BRRI dhan 62) and 75 kg ha-1 for other HYV rice varieties.
Under farmers’ practice (FP) in the Dacope area, most cultivated modern varieties and seeds were sown on 14-20 July and transplanted using 25-35 days old seedlings randomly with comparatively wider spacing (25-30 cm x 25-30 cm). In FP at the Amtali site, most of the farmers cultivated local variety and seeds were sown on 1-10 July and transplanted using 30-40 days old seedlings randomly with wider spacing (25-30 cm x 25-30 cm). In Dacope area, most of the farmers apply balanced fertilizer doses for modern varieties, but in Amtali area, most of the farmers used only nitrogen fertilizer with a higher dose (73 - 82 kg ha-1) and a few farmers used other fertilizer with lower doses (5- 11 and 7.5-15 kg ha-1 for P and K, respectively) than recommendation. The whole amounts of P, K and S were applied at the time of final land preparation, except N fertilizer. One-third N fertilizer was applied basally and the remaining two-thirds were applied in two equal splits at intervals of 25 and 45 days after transplanting (depends upon the tidal and flood water depth in the field). Rainfall and flood/tidal water were sufficient to meet crop water requirements for rice crops during T. aman. For yield estimation, 5 m2 area was harvested from four corners and one from the middle of each rice field, and the yield was adjusted to 14% moisture content. Others input costs and cultivation practicerelated economic data were collected and analysed by following the standard procedure (Dillon and Hardaker, 1993).
Preference analysis Dacope, Khulna under Polder 31, and Amtali, Barguna under Polder 43/1, were selected purposively for this study as researchers managed trials on different wet season (WS) rice varieties. Preference polls were conducted at maturity stage to identify the most preferred rice varieties among farmers. In the preference analysis (PA), scientists explained to farmers the importance of this activity and stepby-step procedures of voting to identify most desired and the least desired rice varieties. In Dacope, 11 WS rice varieties including four checks were included for preference analysis. Some of the recently released short/medium duration WS rice varieties such as BRRI dhan53, BRRI dhan54, BRRI dhan62, BRRI dhan66 and BRRI dhan73 were considered to represent good potential to enhance cropping intensity in the location and compared withBR10, BR11, BR23 and Baral(a local cultivar). In Amtali, preference analysis was conducted for BRRI dhan39, BRRI dhan53, BRRI dhan54, BRRI dhan62 and BRRI dhan66 against Swarnogota, Vajon, BR11 and BR23.
Each variety had a stake and bag with variety name. All the participants moved to the trial fields and they observed the physical traits of all varieties included in the preference poll. Each participant was provided with four pieces of ballot-paper, two checks and two crosses, and voted for their two most preferred and two least preferred varieties/cultivars. Ballots were colour-coded to classify male farmers, female farmers and scientists/ researchers. Female farmers voted first, followed by male farmers, and then the researchers last. The regional technicians placed one panicle in the ballot bag of each variety and collected all those bags. Votes were tallied and the preference scores were computed in front of all the participants. The two most preferred and two least preferred varieties were announced, and one of the check varieties was voted top in both the sites. The preference score was calculated by the following equation:
Preference score = (Positive votes - Negative votes)/ Sum of positive and negative votes (Paris et al., 2011).
Correlation analysis was performed to test if there are significant correlations between the preference scores of male and female farmers as well as the preference scores of farmers and breeders.