Abstract
This study was undertaken to characterize and to estimate the level of adoption of artificial insemination (AI) as well as to identify factors determining its adoption by communal farmers of Nyagatare, Rwanda. This followed the launching of the AI programme in the district several years earlier. A survey of the breeding methods used for small scale dairy cows by farmers from five selected sectors of Nyagatare district of the Eastern Province of Rwanda was carried out by means of a questionnaire. Of the 360 respondents, 16.9% primarily used AI, 63.1% used natural service and 20.0% used both methods for breeding. A significant percentage of respondents residing within 5km of the AI centres (79.2%; P<0.05) adopted AI whilst a significant percentage of respondents residing more than 15km from AI centres (92.5%; P<0.05) preferred natural service. A significant percentage of respondents from big families with 6-7 members were using AI (54.1%; P<0.05), whilst all the respondents from small families with up to 3 members opted for natural service (100% P<0.05). A significant percentage of respondents with secondary education (76.9%; P<0.05) opted for AI whilst a significant percentage of respondents with no formal education (98.5%; P < 0.05) opted for natural service alone. A significant percentage of respondents working in both agriculture and animal husbandry (93.1%; P<0.05) opted for AI whilst a significant percentage of respondents working in crop production and other none animal husbandry related sectors (93.2%; P<0.05) opted for natural service alone. Land size in excess of 5 hectares did not significantly influence adoption of AI. Bellowing (27.9%), vaginal discharge (25.6%) and increased mounting behaviour (22.6%) were used for heat detection. Farmers preferred heat detection in the mornings (85.0%).

Key words: Artificial insemination, heat detection, natural service