Fluorescence in situ hybridization (FISH) has impacted profoundly on our knowledge of the in situ ecophysiology and biodiversity of bacteria in natural communities. However, it has many technical challenges including the possibility of false positives from the binding of probes to non-target rRNA sequences. We show here that probe target sites containing single-base insertions or deletions can lead to false FISH positives, the result of hybridization with a bulge around the missing base. Experimental and in silico data suggest this situation occurs at a surprisingly high frequency. The existence of such sites is not currently considered during most FISH probe design processes. We describe software to identify potential non-target sites resulting from single-base insertions or deletions in rRNA sequences. This software also provides an estimate of the FISH probe hybridization efficiency to these sites.