In Rees et al. (2014b), we reviewed the current status of environmental DNA (eDNA) to monitor aquatic populations. Our aim was to focus on discussion of methodologies used, application of eDNA analysis as a survey tool in ecology, and to include some innovative ideas for using eDNA in conservation and management.Roussel et al. (2015) claim that analysis of Rees et al. (2014b) and other publications highlights the downsides of the method, and they suggest that some conclusions should be toned down. Many of their arguments were covered in our original paper (Rees et al., 2014b); however, they make the point that modelling approaches should be encouraged, and we fully agree with this suggestion.Roussel et al. (2015) also claim that we neglected to recognize that there are two sources of imperfect detection (at the field level and at the laboratory level). We feel that our review paper implies this point.Synthesis and applications. Roussel et al. (2015) reiterate many of the points made in the original paper but do cover some additional areas that improve the debate on the use of environmental DNA (eDNA). Both the comment (Roussel et al., 2015) and our rebuttal clearly highlight that detailed laboratory protocols and rigorous field sampling design are crucial factors which require sufficient reporting in the literature to allow for experimental comparison and replication. Any development of a new method for eDNA detection should be compared directly with established ‘gold standard’ methods for the detection of the species or habitat under investigation. None of the issues raised in Roussel et al. (2015) would alter our main conclusions.