is given as h and i in Codex Salzburgensis 140, we are not entitled to say that the usage in either case is incorrect. (11) The non-occurrence on the Ruthwell Cross of d and t for p. Prof. Cook draws attention to this phenomenon1. Such a usage, however, would be a mistake, and a mistake which actually occurs in one or two later inscriptions (e.g., Falstone, which is cited by Prof. Cook), and which obviously arises from the use by the engraver of a copy in Roman lettering. In earlier days, when the Runic alphabet was in more general use, such an error would naturally not occur. The use of d and th for the dental voiceless spirant and that of c (before t) and ch for the guttural voiceless spirant, are due, of course, to the inadequacy of the Latin alphabet for representing Teutonic sounds which did not occur in the Latin language. Turning now to the Bewcastle column, we cannot but think that Prof. Cook has given insufficient weight to the statements made by Prof. Vietor in his book Die Northumbrischen Runensteine. It is there pointed out that many of the letters, which may appear (p. 14) to be comparatively clear in the photographs, cannot be said to have been engraved on the stone at all, since they were painted over, apparently largely from conjecture, by Maughan, a former vicar of Bewcastle, in 1856. (1) Gessus2. The use of initial g is perfectly regular in the Runic alphabet. A character which is probably the form for j does occur twice in Runic inscriptions (ie., Thornhill and Dover), both times in the word gisl, gil(s), which originally had initial g. It is clear that the letters g and j were wholly confused, because the palatalised spirant g had become j by the middle of the seventh century, and, so far as we can judge, the result on the alphabet was that j was very rarely used. For the age of the change of spirant g to j, we may note the consistent spelling Iaruman (for Gearu-) in early Bede MSS.-which at least suggests that that was the spelling used by the Bishop himself (who died before 670). (2) æft alcfriþu1. If Vietor is right in saying that the sixth letter is really c, and in reading the whole name as Alefripu, we have here a distinct mistake in the use of the Runic letters which could only be accounted for on the 1 Notes on the Ruthwell Cross, p. 387. 2 The Date of the Ruthwell and Bewcastle Crosses, p. 249. 3 It is identical with the letter regularly used (in its later value, a) on certain Northern inscriptions of the transition period-e g., Stentofta and Björketorp. 4 The Date of the Ruthwell and Bewcastle Crosses, pp. 249 f. supposition that the engraver was following a copy in Roman letters, and consequently writing c for h(x). From Vietor's words, however, it appears that the first five letters are by no means certain ('alle beschädigt, aber, wie ich glaube, vorhanden'). If so, we have no need to assume that the engraver has been guilty of any such error. The first element in the name may have been a word which really had final -c1. As regards the form aft to which Prof. Cook attaches so much importance, it is surely more likely that the two have been accidentally omitted, than that the engraver should have made use of a Scandinavian preposition for which there is, as far as we know, no evidence elsewhere in English. In view, however, of Vietor's phrase 'alle beschädigt' we are inclined to regard this reading also with a good deal of scepticism. (3) cynnburug3. With regard to the Svarabhakti in this word, this usage is common in Runic writing between r or l and voiceless spirants from the earliest times-e.g.,worahto (Tune inscription), worohtae (Kirkheaton inscription), wylif (Franks Casket). There is no doubt that before the end of the seventh century final g, not palatalised, was already voiceless (x), though usually preserved in writing. We may refer to such forms as maerh (Epinal 588; marh, Erf., Corp.), duerh (Erf. 1176), haehtis, haehtisse (Corp. 759, 945). Further evidence is afforded by the erroneous use of g for (original) h, e.g., sceptog for sceptlog (Corp. 145), unneg, fegtaþ (Franks Casket), aerigfaerae (Leid. Rid.), Ealghard for Ealh- (Chart. 18 in Sweet's O.E.T.), -leag (frequently in charters). It might be well to note that with regard to the first part of the name cynnburug, Viëtor declares that the fourth letter is really i. The conclusion, therefore, to which we are led is that the language of the inscriptions points to the same period as that indicated by their orthographical characteristics. Certain features, e.g., the regular use of f for b, the occasional appearance of e for i (and ?) in unaccented syllables, and the frequent loss of final n, seem to us-like the form of the letter d-to militate somewhat against a date much earlier than the seventh century. On the other hand the form appile, which seems to be It 1 Ecfripu (for Ecgfripu) would scarcely be impossible. We must note that, if the form Alefripu or Ecfripu is really correct, the final u is an argument for great antiquity. is universally lost in names in Bede's History (except, of course, when those are Latinised). But the letter u really appears to be doubtful. 2 Cf. the form afte in one of the Thornhill inscriptions. + Cf. p. 29. 3 Ib. pp. 255-6. reasonably safe, and the usual retention of œ and i decidedly favour the eighth century rather than the ninth. Under no circumstances, however, can we credit the supposition that the inscriptions were written later than the ninth century, which on historical grounds practically means that they must date from before 867. The linguistic evidence which we have just summarised comes almost wholly from the Ruthwell Cross; but we are by no means satisfied that the Bewcastle inscription really contains any forms which point to a different date1. M. D. FORBES. CAMBRIDGE. 1 Since this article was sent to press, a further monograph on the subject of the Ruthwell and Bewcastle Crosses has come to our notice. In The Runic Roods of Ruthwell and Bewcastle (John Smith and Son, Glasgow, 1914) Dr King Hewison ascribes the two monuments to the tenth century and to St Dunstan in particular. We have examined his arguments, but see no reason to modify the views expressed above. Professor Blackburn in Modern Philology, IX, 555-66, argues that the 'confused order of events' in the story of the slaying of Grendel's mother is due to the shifting of a leaf. Professor Blackburn opens his argument thus: The story of the slaying of Grendel's mother, apart from the preparations at the beginning and the rejoicing at the end, may be analyzed as follows: i. Vss. 1497-1569. The fight and the death of the monster. (In the cave.) ii. Vss. 1570-90. Beowulf explores the den, finds Grendel dead, and strikes off his head. (In the cave.) iii. Vss. 1591-1605 a. The Danes wait until they are convinced that Beowulf has perished; they then return home, but the Geats, though hopeless, remain longer. (On the cliff.) iv. Vss. 1605 b-17. The sword-blade melts in the blood of Grendel, but Beowulf takes the hilt and the head as trophies. (In the cave.) V. Vss. 1619-43. Beowulf comes up from the den and swims to land; the Geats meet him and escort him back to Hrothgar's Hall. The purpose of this paper is to show the probability that a part of the story has been misplaced, to suggest a way in which this may have happened, to discuss the natural results of such an accident, and perhaps thus to contribute something to the method of discovering and correcting such faults elsewhere. I assume that originally the poet finished the account of the fight in the cave as described in I, II, and IV before telling what took place meanwhile on the cliff as related in III and V, and that the shift of scene from cave to cliff and back again, which we find in the poem as it has reached us, is the result of an exchange of the position of III and IV. This method of exposition has one serious drawback. The story is summarised by a critic who regards it as confused. Starting with this feeling with regard to his text, such a critic is likely to show, in his summary, a want of sympathy towards the narrative which he is summarising. The summary may show a break in the continuity of the story,' but the question has still to be faced, 'Is this break due to the original story-teller or to the summariser'? There certainly appears to be in Professor Blackburn's summary 'a break in continuity' between III, where we are told that the Danes return home, and the earlier sections I and II, where the slaying or mutilation of Grendel and his mother is narrated. But this 'break in continuity' is due to the simple fact that Professor Blackburn in his summary has omitted the connecting link. The slaughter or mutilation. of the monsters in the cave leads to the water being reddened with blood, and it is the sight of this blood that leads the Danes on the cliff to conclude that Beowulf is dead, and to return home. The omission of this connecting link is serious, for the effect produced upon the mind of the watchers by the sight of the bloodstained water, so far from being negligible, is an essential and persisting feature of the story. It will strike most readers as strange that Professor Blackburn, in dealing with the story of the fight in the underground cave, should have nothing to say concerning the order of events in the Grettis Saga. Beowulf and the Grettis Saga are here closely parallel, and, what is strictly to the point, they agree in giving the story in the form which Professor Blackburn regards as dislocated. The main features of resemblance between Beowulf and the Grettis Saga in this portion of the tale are these: The hero dives into the water, comes into a cave, and there fights with a monster whose blood he spills. [Beowulf also mutilates the dead body of an earlier adversary, Grendel.] The watcher above [in Beowulf some of the watchers above], seeing the water stained with blood, concludes that the hero is slain, abandons his watch and goes home. Meantime the hero surveys the cave; he swims back with certain trophies. Now clearly the teller of a story like this may treat his subject in one of two ways. He may tell the story as it has been outlined above, that is to say as it is told in Beowulf and the Grettis Saga, thus: (i) The hero slays the monster in the cave. (ii) The watcher or watchers on the cliff, seeing the blood-stained water, depart. (iii) Meantime the hero collects his trophies and swims back to the cliff. Or the story-teller may prefer the following order : (i) The hero slays the monster in the cave. (ii) He collects his trophies and swims back, but (iii) He finds the watchers on the cliff gone, because in the meantime they, seeing the blood-stained water, had departed. To assume that the story ought only to be told in the second of these ways appears to me quite wrong. The first way seems quite as |