is out of spells that prayer has originated. Mercury raised to a high temperature becomes red precipitate; and red precipitate exposed to a still greater heat becomes mercury again. Spells may be the origin of prayers, if prayers show a tendency to relapse into spells. That possibility fits in either with the theory that magic preceded religion or still more exactly with the theory that religion simply is magic raised, so to speak, to a higher moral temperature. We have therefore to consider the possibility that the process of evolution has been from spell to prayer (R. R. Marett, _Folk-Lore_ XV, 2, pp. 132-166); and let us begin the consideration by observing that the reverse passage--from prayer to spell--is only possible on the condition that religion evaporates entirely in the process. The prayer does not become a charm until the {152} religion has disappeared entirely from it: a charm therefore is that in which no religion is, and out of which consequently no religion can be extracted. If then, _per impossibile_, it could be demonstrated that there was a period in the history of mankind, when charms and magic existed, and religion was utterly unknown; if it be argued that the spirit of religion, when at length it breathed upon mankind, transformed spells into prayers--still all that would then be maintained is that spoken formulæ which were spells were followed by other formulæ which are the very opposite of spells. Must we not, however, go one step further and admit that one and the same form of words may be prayer and religion when breathed in one spirit, and vain repetition and mere magic when uttered in another? Let us admit that the difference between prayer and spell lies in the difference of the spirit inspiring them; and then we shall see that the difference is essential, fundamental, as little to be ignored as it is impossible to bridge.