Large-scale multispectral, multitemporal aerial video images were evaluated for speciation of bald-cypress and several species of bottomland hardwoods. Images were acquired with a multispectral video system, including three bandpass filters centered at 550, 800, and 1000 nm, from an altitude of 305 m. The ground-level dimension of the video image pixels was 0.329 m. Images were statistically analyzed with two supervised classification methods (minimum distance and maximum likelihood. The minimum-distance classifier yielded statistically similar results to the maximum likelihood classifier while requiring much less time. Multitemporal imagery increased classification accuracies on the order of 10 percent. Average classification accuracy for individual trees on all plots was 70 percent