In the past, fixed prostheses were believed to be more efficient implant-supported devices than removable types for edentulous patients. However, this hypothesis was never properly tested. Therefore, a within-subject crossover clinical trial was designed in which 145 completely edentulous subjects were tested wearing implant-supported mandibular fixed prostheses and long-bar overdentures. Eight subjects received the fixed appliance first and seven the removable type. The patients' perceptions of various characteristics of the implant-supported prostheses were measured after a minimum of two months' adaptation. Mandibular movements and jaw muscle electromyographic activity were recorded while the patients chewed five standard-sized test foods: bread, apple, hard cheese, sausage, and raw carrot. The prostheses were then changed, worn for the same period of adaptation, and the procedures repeated. There were three test sessions per prosthesis, and each included five trials per food. The measurements were repeated three times at one-week intervals. Mastication time was found to be shorter for three foods (bread, cheese, and sausage) when subjects wore the long-bar overdenture. The vertical amplitude of the masticatory strokes was significantly less with the overdenture for all foods except carrot. Cycle duration was significantly longer with the overdenture for sausage and carrot. Contrary to what might be expected, the long-bar overdenture appears to be no less efficient than the fixed prosthesis. Furthermore, these data suggest that patients are capable of adapting their masticatory movements to the characteristics of the two prostheses.