This paper presents the results of a series of experiments attempting to determine the methods of selection, preparation and use of turtle-shells and cattle-scapulas for divination during the late Shang dynasty. These experiments cover a range of topics, including shell and bone selection and preparation, the making of burning-pits, the source and application of heat to produce the divinatory cracks, the court diviners' interpretation of the cracks, why the court historians first wrote the divination-inscription texts on the bone or shell before they were carved, why the divinatory cracks and inscriptions were carved and pigmented, whether or not the bones and shells were softened before preparation and carving, and the use of bronze and jade knives in carving the inscriptions. I have compared the results of my experiments with archaeological data and with the work of other scholars in the field, and have finally established what the complete process of late-Shang oracle-bone divination was. The major conclusions obtained from these experiments include the following: 1) The primary selection criterion for live turtles to be used in divination was that they be good-looking, unscarred individuals with flat, high-quality plastrons. 2) The fifty-eight closely-joined bony plates making up the complete turtle-shell will loosen or fall apart if boiled too long, and are not suited for a softening process before preparation and carving. 3) The backs of the scapulas were flattened in order to facilitate the making of a greater number of burning-pits on the back and the turning of the bone during inscription-carving on the front; this also made them stable for carving and easier to stack. 4) During the making of the burning-pits, the long-oval pits had to be cut deep and straight, as they provided the guide-line for the initial crack during heating. 5) The cracking of the bones and shells required the use of hardwood burning in the form of a hot coal. This was lightly applied to the shallow burning-pit and slowly blown upon to keep it alight for as long as possible, until the crack was produced. 6) Auspicious or inauspicious crack-interpretations were determined by variations in the shape of the horizontal branch of the crack, but these variations were quite likely controlled by the diviners themselves. 7) The cracks and inscriptions were carved and pigmented for the practical purpose of facilitating later reference and confirmation of divinations, and not for aesthetic purposes. 8) The divination-inscription texts were most likely first written on the bone or shell by the historian, then given to specialized artisans for carving. 9) Shell and bone can be softened with vinegar for carving, but this method has its drawbacks. The proper proportions of copper and tin can definitely be used to make a bronze alloy for saws, knives and other tools capable of cutting bone and shell without pre-softening; such tools could especially have been used to make the wide variety of finely-carved late- Shang bone artifacts. This paper represents the first comprehensive inquiry into the actual physical process of oracle-bone divination that has been undertaken in the past eighty-five years of late-Shang research. [ABSTRACT FROM AUTHOR]