Study Site: The study was conducted at Bolipara union in Thanchi upazilla (sub-district) of Bandarban district, Bangladesh over a period of five months from October 2005 to February 2006. Thanchi upazilla occupies an area of 1020.82 sq km including 680.55 sq km forest area. It lies between 21° 15' and 21° 57' north latitude and between 92° 20' and 92° 41' east longitude (BBS, 1992). Thanchi, about 55 km southeast of Bandarban is the remotest and perhaps, one of the most inaccessible upazillas of Bangladesh (Haque, 2000b). 90% of the area is hilly; rivers and marshes make up 4% and only 6% is suitable for intensive agriculture (Khisa, 1998). The upazilla consists of 4 unions, 12 mauzas and 82 villages. In the upazilla, there are 2885 households of which 2661 are tribal. 98% of the main houses are made of straw/bamboo, the remaining 2% being of a combination of different materials including cement. The upazilla has a total population of 18,000 of which Mro number 3,738 (some 20%). Methods Multi-stage, partly random sampling was used in the study. Out of seven upazillas of Bandarban district, the Mro are concentrated in the hilly areas of Thanchi, Alikadam, Lama, Ruma and Naikhyangchari upazillas (Drong, 2001). A list of these five upazillas was arranged alphabetically and Thanchi was selected randomly from the list. Mostly the Mro inhabit at Balipara union of this upazilla. So this union was selected purposively. A list of the Mro hamlets was collected from the office of Caritas Bangladesh, an international NGO working locally for the development of the Mro tribe. Three hamlets out of 17 were finally chosen at random from the list. From each of the three hamlets, 12 households were selected randomly. Thus a total of 36 households were selected from the study area. A preliminary discussion was carried out at the local office of ‘Caritas Bangladesh’ with its officials, members of the local Union Council, the headmen of the respective hamlets (locally called Karbari) and the members of farmer groups (beneficiaries of Caritas) before the household sampling began. The objective of the discussion was to provide information on the work intended and to select research assistants. Four assistants were recruited of whom one was from Caritas Bangladesh and the remaining were the headmen (Karbaris) of the three respective hamlets. Interview techniques were employed to elicit information and IK on the management of shifting cultivation; daily life activities; collection, processing and management of forest produces and other resources. An open-ended semistructured questionnaire was used for the survey. Since this research was mainly qualitative, emphasis was given to information concerning the utilization of natural resources in the community’s own way. The main questions covered: what resources do they collect from forest, how do they process the raw products, how do they manage their main crop production system (shifting cultivation) and so on. Side by side, physical observations of practices embodying some IK were also made by the first author when living in the community. Household heads were the main respondents in the interviews, helped by other members of the family when necessary. Ambiguities arising from the use of local terms were clarified by the research assistants. The information gathered was arranged systematically and documented as ‘qualitative descriptions’.