To all these perfectly natural objections Mr. Tylor has replied.[1] Evidence must be collected, sifted, tested, as in any other branch of inquiry.  A writer, ’of course, is bound to use his best judgment as to the trustworthiness of all authors he quotes, and, if possible, to obtain several accounts to certify each point in each locality.’  Mr. Tylor then adduces ‘the test of recurrence,’ of undesigned coincidence in testimony, as Millar had already argued in the last century.[2] If a mediaeval Mahommedan in Tartary, a Jesuit in Brazil, a Wesleyan in Fiji, one may add a police magistrate in Australia, a Presbyterian in Central Africa, a trapper in Canada, agree in describing some analogous rite or myth in these diverse lands and ages, we cannot set down the coincidence to chance or fraud.  ’Now, the most important facts of ethnography are vouched for in this way.’

We may add that even when the ideas of savages are obscure, we can often detect them by analysis of the institutions in which they are expressed.[3]

Thus anthropological, like psychical or any other evidence, must be submitted to conscientious processes of testing and sifting.  Contradictory instances must be hunted for sedulously.  Nothing can be less scientific than to snatch up any traveller’s tale which makes for our theory, and to ignore evidence, perhaps earlier, or later, or better observed, which makes against it.  Yet this, unfortunately, in certain instances (which will be adduced) has been the occasional error of Mr. Huxley and Mr. Spencer.[4] Mr. Spencer opens his ‘Ecclesiastical Institutions’ by the remark that ’the implication [from the reported absence of the ideas of belief in persons born deaf and dumb] is that the religious ideas of civilised men are not innate’ (who says they are?), and this implication Mr. Spencer supports by ’proofs that among various savages religious ideas do not exist.’  ‘Sir John Lubbock has given many of these.’  But it would be well to advise the reader to consult Roskoff’s confutation of Sir John Lubbock, and Mr. Tylor’s masterly statement.[5] Mr. Spencer cited Sir Samuel Baker for savages without even ‘a ray of superstition’ or a trace of worship.  Mr. Tylor, twelve years before Mr. Spencer wrote, had demolished Sir Samuel Baker’s assertion,[6] as regards many tribes, and so shaken it as regards the Latukas, quoted by Mr. Spencer.  The godless Dinkas have ‘a good deity and heaven-dwelling creator,’ carefully recorded years before Sir Samuel’s ‘rash denial.’  We show later that Mr. Spencer, relying on a single isolated sentence in Brough Smyth, omits all his essential information about the Australian Supreme Being; while Mr. Huxley—­overlooking the copious and conclusive evidence as to their ethical religion—­charges the Australians with having merely a non-moral belief in casual spirits.  We have also to show that Mr. Huxley, under the dominance of his theory, and inadvertently, quotes a good authority as saying the precise reverse of what he really does say.