Latent fingerprint examiners (LPEs) identify distinctive features (minutiae) in fingerprints and consider how rare these observed minutiae are in their decisions about both the value of a fingerprint and whether there is enough correspondence between two fingerprints to support an “identification” or “exclusion” decision. But subjective perceptions about the frequency of events and features tend to be inconsistent and dynamic, which means that variable perceptions of minutia frequency may contribute to inconsistencies in LPEs’ opinions about fingerprint evidence. We surveyed expert LPEs at two time points (NTime 1 = 132; NTime 2 = 99) to establish how rare LPEs believe different minutia types to be and to determine the variation in examiners’ perceptions—both between different examiners and across time for the same examiner. We observed significantly less variation in LPEs’ perceptions of minutia frequency for three minutiae: the two most common minutiae and the minutia perceived to be the least common. We also observed increases in LPEs’ estimates of minutia frequency over time and when they reported recent sightings of the rarest minutiae. LPEs reported frequently using this information in their fingerprint comparison decisions. We propose the use of these consensus-based frequency estimates (until more objective data are available) to increase consistency in LPEs’ use of base rates when examining fingerprint evidence, which may consequently increase the repeatability and reproducibility of decisions made by LPEs.