As we shall see, Gilbert continued in the same tradition, but his interpretation of form and formal cause was much more anthropomorphic than that of his predecessors. Gilbert began his _De magnete_ by expounding the natural history of that portion of the earth with which we are familiar.[44] Having declared the origin and nature of the loadstone, we hold it needful first to give the history of iron also ... before we come to the explication of difficulties connected with the loadstone ... we shall better understand what iron is when we shall have developed ... what are the causes and the matter of metals ... His treatment of the origin of minerals and rocks agreed in the main with that of Aristotle,[45] but he departed somewhat from the peripatetic doctrine of the four elements of fire, air, water, and earth.[46] Instead, he replaced them by a pair of elements.[47] (If the rejection of the four Aristotelian elements were clearer, one might consider this a part of his rejection of the geocentric universe but he did not define his position sufficiently.)[48] [41] Richard Hooker. _Of the laws of ecclesiastical polity_, bk. 1, ch. 3, sect. 4 (_Works_, Oxford, Clarendon Press, 1865, vol. 1, p. 157) [42] Francis Bacon, _De augmentis scientiarum_, bk. 3, ch. 4, in _Works_, ed. J. Spedding, R. L. Ellis, and D. D. Heath, Boston, n.d. (1900?), vol. 2, p. 267. [43] _The poems of John Donne_, ed. H. J. C. Grierson, London, Oxford University Press, 1933, p. 175 ("To the Countesse of Bedford, On New Yeares Day"). [44] M: pp. 33, 34. [45] M: pp. 34, 35. Aristotle, _Works_, ed. W. D. Ross, Oxford, 1908--1952, vol. 2, _De generatione et corruptione_, translated by H. H. Joachim, 1930, vol. 3, _Meteorologica_, translated by E. W. Webster, 1931. [46] M: pp. 34, 35, 64, 65, 69, 81. Dr. H. Guerlac has kindly brought to my attention the similarity between the explanation given in Gilbert and that given in the _Meteorologica_, bk. 3, ch. 6. p. 378. [47] M: p. 83. [48] A statement of the relation between Aristotle's four elements and place can be found in Maier, _op. cit._ (footnote 17), pp. 143-182.