'We fear that all our speculations will remain summaries of probabilities. No documents are extant to enlighten us; we have only mobile, complex and confused ideas, incarnate in eccentric, often contradictory theories. That this character attaches to such ideas should keep us on guard against framing theories whose symmetry is sometimes their condemnation' ('Daily Chronicle,' December 10, 1897). Nothing excites my own suspicion of my provisional hypothesis more than its symmetry. It really seems to fit the facts, as they appear to me, too neatly. I would suggest, however, that ancient savage sacred hymns, and practices in the mysteries, are really rather of the nature of 'documents;' more so, at least, than the casual observations of some travellers, or the gossip extracted from natives much in contact with Europeans. Supposing that the arguments in this essay met with some acceptance, what effect would they have, if any, on our thoughts about religion? What is their practical tendency? The least dubious effect would be, I hope, to prevent us from accepting the anthropological theory of religion, or any other theory, as a foregone conclusion, I have tried to show how dim is our knowledge, how weak, often, is our evidence, and that, finding among the lowest savages all the elements of all religions already developed in different degrees, we cannot, historically, say that one is earlier than another. This point of priority we can never historically settle. If we met savages with ghosts and no gods, we could not be sure but that they once possessed a God, and forgot him. If we met savages with a God and no ghosts, we could not be historically certain that a higher had not obliterated a lower creed. For these reasons dogmatic decisions about the _origin_ of religion seem unworthy of science. They will appear yet more futile to any student who goes so far with me as to doubt whether the highest gods of the lowest races could be developed, or can be shown to have been developed, by way of the ghost-theory. To him who reaches this point the whole animistic doctrine of ghosts as the one germ of religion will appear to be imperilled. The main practical result, then, will be hesitation about accepting the latest scientific opinion, even when backed by great names, and published in little primers.