Within twenty years of one another, Bacon and Descartes proposed cosmologies which relied heavily on matter theory. In both, the distribution of matter in the cosmos determined what centers of rotation there were, and rotating bodies were carried around by the motion of an all-encompassing celestial uid in which they were embedded. But the role of matter theory in the two accounts is very different, both in motivation and in the level at which it is active in guiding physical theory. Matter theory in Baconian cosmology stands as a foundational discipline, being virtually constitutive of physical theory, as it had been for natural philosophers from Thales onwards, whereas in Descartes it is subservient to the needs of his optics and his mechanics.Comparison of the two cases shows how the role of matter theory came to be radically modi ed in seventeenth-centur y cosmology.The best-known astronomical theories of the early modern era-those of Copernicus, Brahe, Kepler, Galileo and Newton, for example-offer a model of the cosmos in which solid bodies occupy an otherwise empty space. With Brahe's demolition of the existence of crystalline spheres, the idea that celestial bodies were physically carried-pushed or pulled-around by some solid material structure, such as ice or crystal, was abandoned. 1 The stability of planetary orbits was to be explained in terms of some physical in uence-a combination of gravitational attraction and rectilinear inertia in the most famous case-rather than in terms of something material: mechanics, not matter theory, held the key. But not all early modern cosmologists believed that one could dispense with a material medium, or with matter theory, and a uid version of crystalline 1. On the complex history of the crystalline spheres, and their uid cousins, see Grant (1996).