as Clough says. If so, as they do not propitiate him, they did not conceive him as a partner in the game of _Do ut des_. But our only witnesses, Messrs. Spencer and Gillen, are extremely reticent about Twanyirika. Nothing is said about his having a superior, and I assume that he has none. It seems to follow that he is a mere Mumbo Jumbo, or bogle, devised by the men to keep the women and children in order. But in South-Eastern Australia (if I may trust Mr. Howitt's evidence, to which Mr. Frazer does not here allude) the counterpart of Twanyirika is a mere servant of a much higher being, everywhere called by names meaning 'our father.' Therefore either 'our father' Baiame, Mungan-ngaur, and the rest, have been developed out of a sportive bugbear like Twanyirika, or Twanyirika (if he really has no superior) is a rudimentary survival of a belief like that in Mungan-ngaur, and his subordinate, Tundun. In the former case Twanyirika, a germ of the more advanced religion of South-Eastern Australia, was not invented as a power behind nature, who might be useful if propitiated, as in Mr. Frazer's theory. In the latter case the Arunta do not represent man prior to religion (as Mr. Frazer holds), but man who has cast off religion. But Mr. Frazer does not seem to notice this dilemma. The evidence for what most people call 'religion' among the Australian natives is so far from scanty that one finds it when looking for other matters, as I am going to show. True, in the following report the religion does not answer to Mr. Frazer's definition, no powerful being is here said to be conciliated or propitiated: he is only said to exist and favour morality. But Mr. Frazer's definition, if pressed, produces the effect of arguing in a vicious circle. His theory asserts that powerful beings are only invented by man, in view of man's tardy discovery that his own magic is powerless. The invented beings are then propitiated, for selfish ends, and that, by the definition, is religion.