Researches on local breeds have mainly focused on the scientific and technical activities of genetic gain production and/or maintain genetic variability. The diffusion of the genetic gain used to be taken for granted, or considered as of little importance as the State was subsidizing official breeding schemes. However, diffusion and sustainability of small local-breeding schemes are threatened by current changes in breeding activities and organizations. Diversification of farming and breeding objectives, liberalization of public policies on breeding activities, decrease in public support change the business model of breeding organizations. Local breeds are particularly concerned, as they may be threatened by more competitive and widespread ones. Indeed, the management of the diffusion dimension of breeding activities gets a greater importance. Thus, there is a need for a better understanding of the market of genetic gain and the strategies of its participants. To investigate this issue, we study with quantitative and qualitative data, the way the genetic market works in the case of local dairy sheep breeds in the Western Pyrenees. In this area, the use of artificial insemination (AI) outside nucleus flocks is weak. The diffusion is mainly based on the exchanges of live breeding animals, but the number and substance of the exchanges are unknown. We analyse two types of markets, which are set up: the official sale of breeding animals, organized by the breeding centre; the parallel market of rams' exchanges by mutual agreement between farmers. We find several paradoxical results: the more expensive animals are sold outside of the breeding schemes, while the genetic value is more uncertain; the breeding centre does not find enough buyers for its rams, while there is a shortage of rams in the region; outside the breeding schemes, the parallel market of rams is dominant. We also identify that there is a diversity of prices on the market, which cannot be explained based on the scientific evaluation of animals. We show the existence of a second-hand market of rams. In conclusion, we argue that there are various ways of managing the diffusion of genetic gain, and that the market is only one of this.