Biscutate swift Streptoprocne biscutata nests are usually built on protected rocky cliff walls. Birds often renest at the same location. Remains of previous nests may offer information about potential nests and quality of nest‐sites. Here, we experimentally study nest‐site selection to test the hypothesis that information from previous nests is used in current nest‐site selection. We placed old nest material at artificial nest‐sites to test whether new nest‐sites are chosen based on the presence of nesting material. We also tested whether the use of natural nest‐sites is influenced by nesting material by creating two types of natural sites: previously used natural nest‐sites with vestiges of old nests removed and never used natural nest‐sites to which vestiges of old nests were added. In the first experiment, in 139 nest‐use opportunities, 16 artificial nest‐sites were used, all of which included vestiges. In the second experiment, in 91 nest‐use opportunities, four nests were in previously unused but natural locations to which vestiges had been added, 22 nests were in previously used sites without vestiges, and the remaining 65 nests remained unused. Two processes are apparently in action: first, prior experience and memory; second, vestiges indicate where nesting has occurred, possibly useful for first breeding, or for imperfect memory. Previous nesting information may explain why swifts use nesting locations for many years and why new nesting colonies seldom form. This transmission of information suggests that swifts tend to be conservative and nest where previous nesting has occurred.