Sampling populations without clear administrative boundaries is a major issue in social sciences. Some solutions have, however, been proposed in order to gather reliable data on this kind of populations. The Respondant Driven Sampling (RDS) undoubtedly figures as the most widespread of alternative sampling techniques. The main goal of this article is to discuss the strengths and the weaknesses of this procedure through our field work experience on Swiss “ordinary musicians.” The theoretical, methodological and statistical tools involved in this kind of procedure will be presented first. Secondly, the article turns to discuss the originality and robustness of data gathered by the RDS method as compared with other more common procedure (like online surveys) in research on “rare” but not necessarily “hard to reach” populations. Finally, the article evaluates the limits of the sample by analysing the structuration of the whole network of musical collaborations.