This paper offers a new and more refined interpretation of Margaret Cavendish's panpsychist view of nature. Her panpsychism is at the core of her contributions to early modern natural philosophy. According to Cavendish, to explain the order in the natural world we must suppose that all of nature possesses sense and reason. However, so far there has not been sustained attention to the precise structure of her panpsychism. How must a sensing, rational nature be organised to explain the occurrence of natural order? This paper distinguishes three possible structures her panpsychism could take: (1) a top-down centralised organisation on which the whole of nature determines the behaviour of her parts, (2) a semi-centralised organisation, and (3) a distributed model on which organisation is due to coordination among parts of nature. It shows that where (semi-)centralised constructions of panpsychism face systematic problems, a distributed account fits with Cavendish's wider metaphysical commitments. Moreover, distributed panpsychism makes sense of Cavendish's work developmentally, as arising from her earlier atomism.