The base adjacent to the 3′ end of universal PCR primers targeting the 16S rRNA gene is often variable and apparently biases the microbial community composition as represented by PCR-based surveys. To test this hypothesis, four templates of 44 bases each and two complementary primers (21 bases) were designed to differ only in the bases adjacent to the primers, and their amplification efficiencies were evaluated using quantitative PCR. For extension temperatures of 72°C, 73°C and 74°C, improvement in initial amplification efficiency was observed for templates with guanine or cytosine at the position contiguous to the primers. However, no clear preference was observed when extension temperature was lowered to 70°C. Shortening the primers by one base, so that the variable position was located two base pairs downstream from the primer, attenuated but did not eliminate this bias. A conformational change of the quaternary polymerase – primer – template – dNTP complex upon commencing of polymerization is thought to be a rate-limiting step. A possible explanation for the observed bias is the stabilization of this complex by the adjacent guanine or cytosine. Reducing PCR extension temperature to 70°C minimizes amplification biases caused by variable template-contiguous bases to the 3′ end of universal PCR primers. Next-base nucleotide composition should be taken in consideration in designing primers targeting 16S rRNA or other functional genes used in microbial ecology studies.