The Northern Rocky Mountains of Idaho and Montana west of the Continental Divide have a diversity of forest types distributed across a wide range of topography and a long history of forest wildfires. To infer the climate drivers of fire in dry forests over the past several centuries, we reconstructed a history of surface fires from fire scars on trees at 21 sites in this region. While the results of this study are useful for managing fire, we recognize that land managers and other researchers also need site-specific details of fire history that we did not publish with our regional-scale analysis. For example, site-specific fire regimes properties, such as mean fire interval, can be used to understand how fire regime may have changed on a particular site over the past century by determining a Fire Regime Condition Class. Site-specific chronologies of fire can also be related to other aspects of a site's history, such as tree demography and the history of human use of the site. Finally, details of site-specific fire histories and the sites from which they were sampled may be used to extrapolate fire regimes from sampled to similar unsampled areas. The objective of our broader study was to obtain annually accurate histories of surface fire occurrence across Idaho and Montana to identify regional-fire years and infer their climate drivers. To meet this objective, we targeted small sites with many, well-preserved fire scars in dry forests; in other words, those dominated or codominated by ponderosa pine. We obtained annually accurate fire dates by crossdating, which can assign the exact calendar year to tree rings and hence fire scars. Prior to our study, no accurately dated fire histories had been published for this region. Our objective here is to provide site-specific histories of surface fire that can be used for land management or additional research. We report on 20 of the 21 sites we used in our broader study of climate drivers-details of the remaining site were reported elsewhere. In addition, we report on three sites that were sampled in this region for other purposes. For each site, we report both the chronology of surface fires and estimates of time-averaged properties of the fire regime, as well as site location, forest type, and topography. We describe our methods in a single section that applies to all sites but provide details of each site in a separate section of the results.