Chimeras are a frequent artifact in polymerase chain reaction and could be the underlying causes of erroneous taxonomic identifications, overestimated microbial diversity, and spurious sequences. However, little is known about the regional effects on chimera formation. Therefore, we investigated the chimera formation rates in different regions of phylogenetically important biomarker genes to test the regional effects on chimera formation. An empirical study of chimera formation rates was performed using the Roche GSFLX™ system with sequences of the V1/V2/V3 and V4/V5 regions of the 16S rRNA gene and sequences of the nifH gene from a mock microbial community. The chimera formation rates for the 16S V1/V2/V3 region, V4/V5 region, and nifH gene were 22.1-38.5%, 3.68-3.88%, and 0.31-0.98%, respectively. Some amplicons from the V1/V2/V3 regions were shorter than the typical length (∼7-31%), reflecting incomplete extension. In the V1/V2/V3 and V4/V5 regions, conserved and hypervariable regions were identified. Chimeric hot spots were located in parts of conserved regions near the ends of the amplicons. The 16S V1/V2/V3 region had the highest chimera formation rate, likely because of long template lengths and incomplete extension. The amplicons of the nifH gene had the lowest frequency of chimera formation most likely because of variations in their wobble positions in triplet codons. Our results suggest that the main reasons for chimera formation are sequence similarity and premature termination of DNA extension near primer regions. Other housekeeping genes can be a good substitute for 16S rRNA genes in molecular microbial studies to reduce the effects of chimera formation.