We examined net photosynthesis, transpiration, stomatal conductance, and leaf fluctuating asymmetry on two species (Rhus copallinum and Ipomoea pandurata) as indicators of stress at nine sites across a gradient of soil disturbance at Fort Benning, Georgia. There were three sites for each of three disturbance levels. Physical habitat disturbance was caused by activities associated with infantry training, including mechanized elements (tanks and personnel carriers) and foot soldiers. In addition, we examined the influence of prescribed burns and microhabitat effects (within meter‐square quadrats centered about the plant) on these measures of plant stress. Net photosynthesis declined with increasing disturbance in the absence of burning for both species. However, when sites were burned the previous year, net photosynthesis increased with increasing disturbance. Developmental instability in Rhus, as measured by fluctuating asymmetry, also declined with increasing disturbance in the absence of burning but increased with disturbance if sites were burned the previous year. Developmental instability was much less sensitive to burning in Ipomoea and in general was lowest at intermediate disturbance sites. Microenvironmental and microhabitat effects were weakly correlated with measures of plant stress when all sites were combined. However, higher correlations were obtained within site categories, especially when the recent history of prescribed burning was used as a category. Finally, using all of the combined data in a discriminant function analysis allowed us to correctly predict the disturbance level of more than 80% of the plants. Plant stress is responsive to both large‐scale perturbations, such as burning, and microhabitat parameters. Because of this, it is important to include macro‐ and microhabitat parameters when assessing stress. Similarly, we found a combination of developmental and physiological indicators of stress was superior to using them separately.