Four test conditions of increasing complexity were used to evaluate the clinical efficacy of amniotic membranes as biologic dressings on donor sites and burn wounds in children. These were the clean-skin donor-site wound, the uncontaminated shallow partial-thickness burn wound, the bed of freshly excised full-thickness wounds, and the granulating surface of colonized burn wounds. The rate of epithelialization under amniotic membranes was the same as that under 5% scarlet red ointment or 0.5% silver nitrate solution dressings. Preservation of a healthy excised wound bed and maintenance of a low bacterial count in contaminated wounds paralleled the experience with human allograft dressings despite technical difficulties and the absence of vascularization of amniotic membrane and its fragile structure. Tentative conclusions are drawn as to the mechanisms by which biologic dressings exert their beneficial effects.