Since the discovery of free neutral fat in the blood vessels following injury, the subject of fat embolism has been widely investigated clinically and experimentally. As may so easily happen when an apparently direct and unequivocal cause-effect relationship is presented, the search for other causes has been overlooked; and this, in spite of the fact that the more familiar we become with fat embolism, the less unequivocal the commonly accepted etiology appears. This article is the record of an attack on this unsettled problem from a new direction. A detailed review of the literature of fat embolism need not be given here. It has been thoroughly covered in English by Warthin (19131) and in German by Landois (19232), Paul and Windholz (19253) and others. We need only summarize those facts on which complete agreement exists among most authors, and in somewhat more detail discuss those observations that