The first serious attempt to disentangle the complex character of Melanesian ethnography was made by F. Graebner in 1905[333], followed by G. Friederici, the references to whose work are given above. More recently W. H. R. Rivers[334] has attacked the cultural problem by means of the genealogical method and the results of his investigations are here briefly summarised. He has discovered several remarkable forms of marriage in Melanesia and has deduced others which have existed previously. He argues that the anomalous forms of marriage imply a former dual organisation (_i.e._ a division of the community into two exogamous groups) with matrilineal descent, and he is driven to assume that in early times there was a state of society in which the elders had acquired so predominant a position that they were able to monopolise all the young women. Some of the relationship systems are of great antiquity, and it is evident that changes have taken place due to cultural influences coming in from without. The distribution of kava-drinking and betel-chewing is of great interest. The former occurs all over Polynesia (except Easter Island and New Zealand) and throughout southern Melanesia, including certain Santa Cruz Islands, where it is limited to religious ceremonial. Betel-chewing begins at these islands and extends northwards through New Guinea and Indonesia to India. Kava and betel were introduced into Melanesia by different cultural migrations. The introduction of betel-chewing was relatively late and restricted and may have taken place from Indonesia after the invasion by the Hindus. With it were associated strongly established patrilineal institutions, marriage with a wife of a father's brother, the special sanctity of the skull and the plank-built canoe. The use of pile dwellings is a more constant element of the betel-culture than of the kava-culture. The religious ritual centres round the skulls of ancestors and relatives, and the cult of the skull has taken a direction which gives the heads of enemies an importance equal to that of relatives, hence head-hunting has become the chief object of warfare. The skull of a relative is the symbol--if not the actual abiding place--of the dead, to be honoured and propitiated, while the skulls of enemies act as the means whereby this honour and propitiation are effected.