Classical atoms—“part-less, ontologically irreducible simples” as the conference flyer puts it—are not the atoms of modern chemistry and analogies with the latter can be construed in various ways. They have figured in the historical development of concepts of chemical affinity but without, as Alan Chalmers and I have independently argued, making any significant contribution to empirically justified theories. A purely combinatorial conception of the formation of compounds by juxtaposing atoms is associated with Daltonian atomism. I review the merits of this idea as a solution to problems posed by developments in early nineteenth century chemistry and go on to suggest that subsequent developments in chemistry are as much in agreement with ideas associated with Aristotle as with the ancient atomists, if not more so. The basic issues can be understood without getting involved in detailed chemistry and I hope what I have to say will be of interest to philosophers concerned with the metaphysics of matter but with no special knowledge of chemistry.