It is my aim in this paper to show by comparative study of the forms of subsection terms and cognate section terms that Elkin was substantially correct in his view of the origin and diffusion of subsections. In particular I hope to show:(1) that the area of origin of the subsection system can be located fairly precisely on the basis of linguistic evidence in the area just north of the lower Victoria River, Northern Territory;(2) that the eight subsection terms most generally used arose through the amalgamation of the two sets of four terms of two pre-existing section systems, one originating in the Pilbara district of Western Aus-tralia (which I shall call the Northwestern section system), and the other originating in the so-called Top End'of the Northern Territory, south of Darwin (which I shall call the Northcentral section system);(3) I also speculate that the amalgamation of the two section systems into a new and qualitatively different subsection system could have taken place in the first instance through the practice of a type of marriage exchange whereby men of only one equivalent patrimoiety in each of the groups possessing the Western and Northcentral section systems married women of the opposite moiety of the other group, while men of the other moiety continued for some time to marry women of the same group (ie the group possessing the same section terms). It will be shown that this arrangement produced all the main features of a sub-section system, which was then adopted as a classification system generally irrespective of the form of marriage exchange practised. 1