The Ethiopian region is characterised by a wide range of agro-climatic conditions, which accounted for the enormous resources of agro-biodiversity that exist in the country. The most important of these resources is the immense genetic diversity of the various crop plants in the country. Of these, one of the most on farm genetically diverse crops is sorghum. Since the advent of formal breeding in particular after green revolution, genetic diversity of most crops has been threatened worldwide. In order to assess on farm genetic erosion (GE), various research methodologies were employed. These were focused group interviews with 360 farmers, on farm monitoring and participation with 120 farmers, key informant interviews with 60 farmers and development agents, and semi-structured interviews with 250 farmers. Besides, diversity fairs were done with over 1200 farmers. Notwithstanding the complexity of assessing GE, it was assessed by various methods; namely, by temporal method (comparing 1960 and 2000 collections), area method, and semi-structured interview method at individual, community or wereda level and causes of varietal loss from other various perspectives. Farmers perceived GE as the reduced importance of the variety as indicated by lower proportion in the varietal portfolio. The five most important factors for varietal loss at individual farmers' level were reduced benefit from the varieties, drought, Khat expansion, reduced land size and introduction of other food crops respectively. GE was not affected by wealth groups and ecological regions. Farmers do not make simple replacement as a strategic mechanism for genetic resources management. GE at regional level was quantified by temporal and spatial method. There was a complementation not rivalry between farmer varieties (FVs) and improved varieties (IVs). The whole process of GE is explained by three models, namely: Bioecogeographic enhanced genetic erosion model, Farmer induced genetic erosion model and Farmer-cum-bioecogeographic genetic erosion model. As aforementioned, sorghum genetic erosion behaviour is completely different from other food crops such as tetraploid wheat. The prediction in the late seventies that complete erosion of FVs by IVs by the end of the eighties, the principle of GE that competition between IVs and FVs, favours the former and results in the replacement of the latter is not valid in the context of sorghum in Ethiopia. Hence, maintenance of the on farm genetic diversity of sorghum is a reality but GE is rhetoric.