After a study of temperature and humidity soundings in Florida during the summer of 1946 failed to produce an explanation for the occurrence or nonoccurrence of thunderstorms, the theory was proposed that large-scale horizontal convergence in the low levels was the necessary condition for thunderstorm convection. Traveling or semipermanent synoptic features of the region producing this convergence were found to be too rare or too remote to account for the almost daily thunderstorm activity of the interior of the Florida peninsula. Low-level horizontal convergence caused by the afternoon sea breezes entering the peninsula from both sides was found to be the most rational explanation of the thunderstorms. Using the Bellamy “triangle method” applied to pilot-balloon stations it was found that low-level convergence developed practically every afternoon during the summer months as a result of this diurnal sea-to-land circulation. The effects extended to an altitude of 3000 or 4000 ft. A comparison with adjacent regions showed that the double sea-breeze effect of the peninsula produces much greater and more consistent convergence than that found along the northern Gulf Coast, thus accounting for the very striking maximum frequency of thunderstorms in the interior of the peninsula.