New techniques to explore microbial diversity have led to resurgent interest in prospecting for natural products (bioprospecting or biodiscovery). Although many bioprospecting projects may share little in common at first glance, the vast majority share one particular challenge. Their targets are rare to very rare members of complex natural assemblages. Despite the advances made in bringing new organisms into cultivation and application of culture-independent techniques to isolation of novel genes there remain systematic biases against relatively rare organisms with specific growth requirements. These can frequently be overcome by application of multidisciplinary approaches that take into account principles of evolutionary ecology. Our experiences with prospecting for soluble di-iron monooxygenases (SDIMO) indicate that conventional approaches to organism isolation and metagenomic cloning systematically under-sample diversity in this enzyme family. This reflects that SDIMO-containing organisms are typically relatively low-abundance members of natural assemblages (thus biased against by direct cloning) and SDIMOs have discrete physiological roles in each organism (thus are not amenable to generic enrichment culture strategies). We have sought to overcome this by a PCR-based survey of gene diversity to guide evaluation of subsequent culture or cloning studies. A surprising outcome of this survey was that conventional PCR approaches using degenerate primers also systematically under-sampled diversity, but nested PCR strategies revealed unprecedented diversity. We conclude that many PCR-based gene-prospecting studies are likely to have under-estimated the impact of target:competitor ratios on their success.