Cryptosporidium parvum is an enteric protozoan parasite of medical and veterinary importance. Dissemination of environmentally resistant oocysts in surface water plays an important role in the epidemiology of cryptospridiosis. Although the polymerase chain reaction (PCR) is a well-established technique and is widely used for detecting microorganisms, it is not routinely applied for monitoring waterborne C. parvum. In order to facilitate the application of PCR to the detection of waterborne C. parvum oocysts, a comparison of published PCR protocols was undertaken and different sample-preparation methods tested. The sensitivity of a one-step PCR method, consisting of 40 temperature cycles, was 10 purified oocysts or fewer than 100 oocysts spiked in raw lake water. The detection limit of two primer pairs, one targeting the ribosomal small subunit and another specific for a C. parvum sequence of unknown function, was approximately ten-fold lower than achieved with a primer pair targeting an oocyst shell protein gene. Three cycles of freezing/thawing were sufficient to expose oocyst DNA and resulted in higher sensitivity than proteinase K digestion, sonication or electroporation. Inhibition of PCR by surface water from different local sources was entirely associated with the soluble fraction of lake water. Membrane filtration was evaluated in bench-scale experiments as a means of removing lake water inhibitors and improving the detection limit of PCR. Using gel and membrane filtration, the molecular size of inhibitory solutes from lake water was estimated to less than 27 kDa.