Based on observations in the human suggesting improved membranous bone graft take, an experimental study was undertaken in 15 rabbits and 7 monkeys to evaluate the differences in take between membranous and endochondral bone grafts. Using vital stains, serial cephalograms, direct measurements, and point-counting techniques, the grafts were compared. In both types of experimental animals, membranous bone maintained its volume to a significantly greater extent than endochondral bone when autografted in the craniofacial region. The loss of volume with endochondral grafts was at the end of the experiment more than three times that of membranous grafts in the rabbit (65 percent endochondral loss versus 19.5 percent membranous loss) and more than four times volume loss in the monkey (17.2 percent membranous volume loss versus 88 percent endochondral loss). The uptake of vital stains (tetracycline and alizarin) was greater with membranous bone, and point counting demonstrated more living membranous bone (40.5 percent membranous versus 28.1 percent endochondral) at the end of the experiment. These studies confirm the increased resorption of endochondral bone grafts when compared with membranous grafts and substantiate clinical impressions that cranial donor sites are preferable for craniofacial recipient areas when clinically feasible.