One cannot expect an exact answer to the question “What are the possible structures of space?”, but rough answers to it impact central debates within philosophy of space and time. Recently Gordon Belot has suggested that a rough answer takes the class of metric spaces to represent the possible structures of space. This answer has intuitive appeal, but I argue, focusing on topological characterizations of dimension, examples of prima facie space-like mathematical spaces that have pathological dimension properties, and endorsing a (...) principle of plenitude for possibility, that we get a different rough answer: either the class of metrizable spaces or the class of separable metrizable spaces are possible structures of space. (shrink)