surprised if we find that not only the first-born but every child is stolen by a giant. I make this suggestion, however, only as a possibility.
The fact that Mongán, when three nights old, was taken by his father Manannan to fairyland might reflect a similar idea: the child considered as a god's son belongs to the god (in the Removed Barrenness motive the child was also believed to be god's gift); this would, however, imply a slightly different motive, i.e. a Supernatural Birth, of which we have found some traces in the first branch of the Mabinogi; in such supernatural birth stories the child remains with the human parents, but if we admit that the child is sometimes really taken from them (not given to fosterage), we have to account for the change of such a Mongán story into one of the Mabinogi type. In the Mongán tale we have no traces of such savage ideas (as human sacrifice), and it is most improbable that in the archetype of this tale there ever was anything of this kind; how shall we, then, explain the fact that the supposed Welsh deterioration of the Irish tale has preserved the older, more savage, feature, i.e. the monster stealing the children? or shall we ascribe it to the "monkish redactor" who guessed our modern theories? I hardly think so.
The same difficulties appear in Mr. Gruffydd's association (ibid. pp. 20 ff.) of the tale of Llew Llawgyffes (R.B. 763 ff.; W.B. fo. 195d ff.) with Ir. Aided Conroí, which presupposes a common origin for these two tales. If these words were to mean that this common origin goes back to a very remote period, there would be little difficulty, but Mr. Grufifydd thinks that the Irish story is the original one, and he argues that the identities of both stories prove the Gaelic origin of the Mabinogion. But unfortunately there are not only similarities in both tales, there are also great divergences.
The main divergence is that the Irish tale had originally an External Soul motive (Ériu, vii. pp. 200 ff.), of which the
