Abstract. There are two good reasons for wanting to detect variability concepts in source code: migrating to a product-line development for an existing product, and restructuring a product-line architecture degraded by evolution. Although detecting variability in source code is a common step for the successful adoption of variability-oriented development, there exists no compilation nor comparison of approaches available to attain this task. This paper presents a survey of approaches to detect variability concepts in source code. The survey is organized around variability concepts. For each variability concept there is a list of proposed approaches, and a comparison of these approaches by the investment required (required input), the return obtained (quality of their output), and the technique used. We conclude with a discussion of open issues in the area (variability concepts whose detection has been disregarded, and cost-benefit relation of the approaches).