In this paper the notion of ethnicity is discussed; in particular the problem of its theoretical boundaries and the levels and contexts it may apply to. With this aim an ethnographical case is analyzed which is the network of relationship between two groups ("orang pesisir" and "orang gunung") on Alor island (Eastern Indonesia). The notion of ethnicity (or rather "microethnicity") may seemingly be applied to the Alorese context. As a matter of fact this concept is a useful label for this kind of social dynamics, but the author questions if it is necessary to elaborate the analytical model applied to the case discussed here. Ethnographical cases like this may help to ascertain the theoretical boundaries of the concept of ethnicity, and consequently to narrow its extension and to increase its semantic density. Nevertheless the author argues that this aim is hard to achieve, because "ethnicity" is a notion intrinsecally vague, applied today by anthropologists to too many contexts and levels of social reality.