This work examines how to adjust the definition of 'dry lightning' in order to optimize the correlation between dry lightning flash count and the climatology of large (>400 km2) lightning‐ignited wildfires over the contiguous United States (CONUS). The National Lightning Detection Network™ and National Centers for Environmental Prediction Stage IV radar‐based, gauge‐adjusted precipitation data are used to form climatic data sets. For a 13 year analysis period over CONUS, a correlation of 0.88 is found between annual totals of wildfires and dry lightning. This optimal correlation is found by defining dry lightning as follows: on a 0.1° hourly grid, a precipitation threshold of no more than 0.3 mm may accumulate during any hour over a period of 3-4 days preceding the flash. Regional optimized definitions vary. When annual totals are analyzed as done here, no clear advantage is found by weighting positive polarity cloud‐to‐ground (+CG) lightning differently than –CG lightning. The high variability of dry lightning relative to the precipitation and lightning from which it is derived suggests it would be an independent and useful climate indicator.