Of course such duplication as Kemble assumed is conceivable. Kemble might have instanced the way in which one and the same hero reappears in the pages of Saxo Grammaticus, with somewhat different parentage or surroundings, as if he were a quite different person. The Lives of the Two Offas present another parallel: the adventures of the elder Offa have been transferred to the younger, so that, along with much that is historical or semi-historical, we have much in the Life of Offa II that is simply borrowed from the story of Offa I. In the same way it is conceivable that reminiscences of the mythical adventures of the elder Beowulf (Beow) might have been mingled with the history of the acts of the younger Beowulf, king of the Geatas. A guarantee of the intrinsic reasonableness of this theory lies in the fact that recently it has been put forward again by Dr Henry Bradley. But it is not enough that a theory should be conceivable, and be supported by great names. I cannot see that there is any positive evidence for it at all.

The arguments produced by Kemble are not such as to carry conviction at the present day. The fact that Beowulf the Geat, son of Ecgtheow, "is represented throughout as a protecting and redeeming being" does not necessarily mean that we must look for some god or demigod of the old mythology—Frey or Sceaf or Beow—with whom we can identify him. This characteristic is strongly present in many Old English monarchs and magnates of historic, Christian, times: Oswald or Alfred or Byrhtnoth. Indeed, it might with much plausibility be argued that we are to see in this "protecting" character

of the hero evidence of Christian rather than of heathen influence[[526]].

Nor can we argue anything from the absence of any historic record of a king Beowulf of the Geatas; our records are too scanty to admit of argument from silence: and were such argument valid, it would only prove Beowulf fictitious, not mythological—no more necessarily an ancient god than Tom Jones or Mr Pickwick.

There remains the argument of Dr Bradley. He points out that

"The poem is divided into numbered sections, the length of which was probably determined by the size of the pieces of parchment of which an earlier exemplar consisted. Now the first fifty-two lines, which are concerned with Scyld and his son Beowulf, stand outside this numbering. It may reasonably be inferred that there once existed a written text of the poem that did not include these lines. Their substance, however, is clearly ancient. Many difficulties will be obviated if we may suppose that this passage is the beginning of a different poem, the hero of which was not Beowulf the son of Ecgtheow, but his Danish namesake[[527]]."

In this Bradley sees support for the view that "there were circulated in England two rival poetic versions of the story of the encounters with supernatural beings: the one referring them to Beowulf the Dane" [of this the Prologue to our extant poem would be the only surviving portion, whilst] "the other (represented by the existing poem) attached them to the legend of the son of Ecgtheow."

But surely many objections have to be met. Firstly, as Dr Bradley admits, the mention of Beowulf the Dane is not confined to the Prologue; this earlier Beowulf "is mentioned at the beginning of the first numbered section" and consequently Dr Bradley has to suppose that "the opening lines of this section have undergone alteration in order to bring them into connection with the prefixed matter." And why should we assume that the "passus" of Beowulf correspond to pieces of

parchment of various sizes of which an earlier exemplar consisted? These "passus" vary in length from 43 lines to 142, a disproportion by no means extraordinary for the sections of one and the same poem, but very awkward for the pages of one and the same book, however roughly constructed. One of the "passus" is just twice the average length, and 30 lines longer than the one which comes next to it in size. Ought we to assume that an artificer would have made his book clumsy by putting in this one disproportionate page, when, by cutting it in two, he could have got two pages of just about the size he wanted? Besides, the different "passus" do not seem to me to show signs of having been caused by such mechanical reasons as the dimensions of the parchment upon which they were written. On the contrary, the 42 places where sections begin and end almost all come where a reader might reasonably be expected to pause: 16 at the beginning or end of a speech: 18 others at a point where the narrative is resumed after some digression or general remark. Only eight remain, and even with these, there is generally some pause in the narrative at the point indicated. In only two instances does a "passus" end at a flagrantly inappropriate spot; in one of these there is strong reason to suppose that the scribe may have caused the trouble by beginning with a capital where he had no business to have done so[[528]]. Generally, there seems to be some principle governing the division of chapter from chapter, even though this be not made as a modern would have made it. But, if so, is there anything extraordinary in the first chapter, which deals with events three generations earlier than those of the body of the poem, being allowed to stand outside the numbering, as a kind of prologue?

The idea of a preface or prologue was quite familiar in Old English times. The oldest MSS[[529]] of Bede's History have, at the end of the preface, Explicit praefatio incipiunt capitula. So we have in one of the two oldest MSS[[530]] of the Pastoral Care "Ðis is seo forespræc." On the other hand, the prologue or preface might be left without any heading or colophon, and the next