Mr. Haddon has suggested a theory which was printed in the _Proceedings of the British Association_ (1902). On this scheme, at a very early period, groups, by reason of their local environment, would have special varieties of food. Thus, at present, in New Caledonia, the Sea branch of a tribe has cocoa-nuts, fish of all sorts, and so forth, while the Bush branch has bananas, and other commodities, and the Sea and Bush moieties of the tribe meet at markets for purposes of barter. But, in a really primitive state, there will be no cultivation, as there is in New Caledonia. Still, a coast savage might barter crabs for a kangaroo, and, if landed property is acknowledged, owners of plum-trees, or of a spot rich in edible grass-seeds, might trade these away for lobsters and sea-perch.[37] Not having any idea of real cultivation, or of pisciculture (though they may and do have ‘close’ seasons, under tabu), the savages may set about working magic for their specialities in food. Thus it is conceivable that the fishers might come to be named ‘crab-men,’ ‘lobster-men,’ ‘cuttlefish-men,’ by their neighbours, whom they would speak of as ‘grass-men,’ ‘plum-men,’ ‘kangaroo-men,’ and so on. When once these names were accepted (I presume), and were old, and now of unknown or rather forgotten origin, all manner of myths to account for the connection between the groups and their plant and animal names would arise. When the myth declared that the plants and animals were akin to their name-giving creatures, superstitious practices would follow. We have seen two cases in which Australian totem groups averred that they were named totemically after a small species of opossum, and a fish which their ancestors habitually ate. But that is an explanatory myth. Man cannot live on opossums alone, still less on sardines.