A two-phase sample design employing large-scale aerial photographs was used to quantify early successional woody and herbaceous plant community structures. Two conventional 35-mm cameras were mounted on a boom and suspended from a helium-filled blimp to obtain low-cost 1:366 scale stereo photographs (1:80 scale prints) of seven experimental vegetation complexes. Estimates of woody crown volume index and herbaceous percent cover were generated for 5 × 5 m plots by calibrating photo measurements to a limited ground-truth sample. The method offered significant increases in estimation precision (>35%) over ground sampling alone, as well as attractive cost advantages (0 to 40%). The highest levels of precision were obtained by measuring entire plots on the photographs. This procedure added approximately 10% to the cost of photo evaluations but resulted in estimates with standard errors that were, on average, 78% smaller than those of ground samples. Simulation trials suggested that more than nine ground-truth sample units per vegetation community type provided only marginal increases in estimation precision. If individual species or species groups of interest are well represented in sample areas, large-scale photographs, employed in a two-phase sample design, can be an effective tool for quantifying and monitoring vegetation community structures in silvicultural and related field studies.