For all these reasons, but chiefly because of the fact that stories are usually anonymous at first, that names are added later, and that stories naturally crystallise round any famous name, heroic, divine, or human, the process of analysis of names is most precarious and untrustworthy. A story is told of Zeus: Zeus means sky, and the story is interpreted by scholars as a sky myth. The modern interpreter forgets, first, that to the myth-maker sky did not at all mean the same thing as it means to him. Sky meant, not an airy, infinite, radiant vault, but a person, and, most likely, a savage person. Secondly, the interpreter forgets that the tale (say the tale of Zeus, Demeter, and the mutilated Ram) may have been originally anonymous, and only later attributed to Zeus, as unclaimed jests are attributed to Sheridan or Talleyrand. Consequently no heavenly phenomena will be the basis and explanation of the story. If one thing in mythology be certain, it is that myths are always changing masters, that the old tales are always being told with new names. Where, for example, is the value of a philological analysis of the name of Jason? As will be seen in the essay 'A Far-Travelled Tale,' the analysis of the name of Jason is fanciful, precarious, disputed, while the essence of his myth is current in Samoa, Finland, North America, Madagascar, and other lands, where the name was never heard, and where the characters in the story have other names or are anonymous. For these reasons, and others too many to be adduced here, I have ventured to differ from the current opinion that myths must be interpreted chiefly by philological analysis of names. The system adopted here is explained in the first essay, called 'The Method of Folklore.' The name, Folklore, is not a good one, but 'comparative mythology' is usually claimed exclusively by the philological interpreters. The second essay, 'The Bull-Roarer,' is intended to show that certain peculiarities in the Greek mysteries occur also in the mysteries of savages, and that on Greek soil they are survivals of savagery. 'The Myth of Cronus' tries to prove that the first part of the legend is a savage nature-myth, surviving in Greek religion, while the sequel is a set of ideas common to savages. 'Cupid and Psyche' traces another Aryan myth among savage races, and attempts to show that the central incident of the tale may have had its origin in a rule of barbarous etiquette. 'A Far-Travelled Tale' examines a part of the Jason myth. This myth appears neither to be an explanation of natural phenomena (like part of 'The Myth of Cronus'), nor based on a widespread custom (like 'Cupid and Psyche'). The question is asked whether the story may have been diffused by slow filtration from race to race all over the globe, as there seems no reason why it should have been invented separately (as a myth explanatory of natural phenomena or of customs might be) in many different places.