To study the ecology of ammonia-oxidising bacteria (AOB), quantitative techniques are essential. Real-time PCR assays based on the 16S rRNA or on the structural amoA gene are routinely used. The CTO primer set rooted on the 16S rRNA gene has a number of mismatches with some of the cultures of AOB. To examine if these mismatches have an effect on the outcome of real-time PCR assays, the assay was tested with DNA from a number of closely related isolates of AOB. Standard curves of known amounts of initial DNA were similar among most of the tested cultures of AOB, except for the standard curves of Nitrosomonas strain AL212 and Nitrosospira strain NpAV. Nitrosomonas strain AL212 had 3 mismatches with the CTO primer set. Adaptation of the CTO primer set in order to perfectly match the Nitrosomonas strain AL212 gave a standard curve similar to the majority of the AOB tested. As Nitrosospira strain NpAV has no mismatches with the original CTO primer set, there must be another reason for the less efficient amplification than the sequence itself. Application of an existing sigmoidal mathematical model gave no other results with respect to the standard curves of Nitrosomonas europaea and Nitrosomonas strain AL212, but also demonstrated that primer mismatches can seriously underestimate the initial target concentration. It was concluded that in general correct interpretation of real-time PCR results requires knowledge of the target community composition, in particular of the target sequences of the dominant community members.