My motivation for writing about the criteria for detecting one form of cryptic female choice, female sperm choice (Birkhead 1998) was twofold: to stimulate discussion (which has occurred) and to draw attention to the fact that novel ideas need to be subject to rigorous testing. There are many different styles of conducting science, one is that it is sufficient to generate novel ideas; at the other end of the continuum is the view, which coincides with my own, that unless an idea has been subject to rigorous testing, its value is limited. To some extent, my conviction that new hypotheses should be subject to testing arises from the impression that in the field of behavioral ecology there is an increasing trend for researchers to feel that it is sufficient to generate interesting ideas, to accept circumstantial evidence, or to report only that evidence which is consistent with a favorite hypothesis. In turn, this change in attitude may be partly a consequence of an increasingly competitive academic climate (e.g., Maddox 1993; Woodward and Goodstein 1996) in which there is little incentive and insufficient time to test ideas properly or replicate the experiments of others, particularly in fields with little impact on the quality of human life. Cryptic female choice is a form of sexual selection and part of behavioral ecology (Krebs and Davies 1997). Behavioral ecology has been a particularly successful area of research and, as is typical of many young sciences, ideas often outstripped data, particularly in its early days. Although most of its practitioners would probably agree that behavioral ecology is now in a phase of more ‘‘normal science’’ (Kuhn 1970) much of the early enthusiasm still remains—especially for novel ideas, recent examples of which include: fluctuating asymmetry, immunocompetence, and cryptic female choice. I regard novel ideas as crucial starting points for progress, but they are not an end in themselves and to make genuine progress we need firm foundations on which to build. A combination of novel ideas, subsequently tested by the most rigorous methods available, including replication, provide the best recipe for success. Behavioral ecologists as a whole have not been especially good about this. A common excuse for the lack of replication for example, apart from the difficulty of obtaining funding for this purpose, is that biology is complex and replication is technically difficult, particularly with studies conducted in the field. However, cryptic female choice, because of its more mechanistic nature, lends itself to careful experimentation and replication more readily than other areas of behavioural ecology, once methods for identifying it have been agreed upon. As soon as we have some sound methodology in place, we will be in a position to fully assess the importance of cryptic female choice in evolution. Here I respond to comments made by three sets of authors in response to my suggestions for how we might seek to identify one form of cryptic female choice: female sperm choice (Birkhead 1998).