PART IV

APPENDIX

A. A POSTSCRIPT ON MYTHOLOGY IN BEOWULF

(1) Beowulf the Scylding and Beowulf son of Ecgtheow

It is now ten years since Prof. Lawrence attacked the mythological theories which, from the time when they were first enunciated by Kemble and elaborated by Müllenhoff, had wielded an authority over Beowulf scholars which was only very rarely disputed^[523].

Whilst in the main I agree with Prof. Lawrence, I believe that there is an element of truth in the theories of Kemble. It would, indeed, be both astonishing and humiliating if we found that a view, accepted for three-quarters of a century by almost every student, had no foundation. What is really remarkable is, not that Kemble should have carried his mythological theory too far, but that, with the limited information at his disposal, he at once saw certain aspects of the truth so clearly.

The mythological theories involve three propositions:

(a) That some, or all, of the supernatural stories told of Beowulf the Geat, son of Ecgtheow (especially the Grendel-struggle and the dragon-struggle), were originally told of Beowulf the Dane, son of Scyld, who can be identified with the Beow or Beaw^[524] of the genealogies.

(b) That this Beow was an ancient "god of agriculture and fertility."

(c) That therefore we can allegorize Grendel and the dragon into culture-myths connected with the "god Beow."

Now (c) would not necessarily follow, even granting (a) and (b); for though a hero of story be an ancient god, many of his most popular adventures may be later accretion. However, these two propositions (a) and (b) would, together, establish a very strong probability that the Grendel-story and the dragon-story were ancient culture-myths, and would entitle to a sympathetic hearing those who had such an interpretation of them to offer.

That Beow is an ancient "god of agriculture and fertility," I believe to be substantially true. We shall see that a great deal of evidence, unknown to Kemble and Müllenhoff, is now forthcoming to show that there was an ancient belief in a corn-spirit Beow: and this Beow, whom we find in the genealogies as son of Scyld or Sceldwa and descendant of Sceaf, is pretty obviously identical with Beowulf, son of Scyld Scefing, in the Prologue of Beowulf.

So far as the Prologue is concerned, there is, then, almost certainly a remote mythological background. But before we can claim that this background extends to the supernatural adventures attributed to Beowulf, son of Ecgtheow, we must prove our proposition (a): that these adventures were once told, not of Beowulf, son of Ecgtheow, but of Beowulf or Beow, son of Scyld.

When it was first suggested, at the very beginning of Beowulf-criticism, that Beowulf was identical with the Beow of the genealogies, it had not been realized that there were in the poem two persons named Beowulf: and thus an anonymous scholar in the Monthly Review of 1816^[525], not knowing that Beowulf the slayer of Grendel is (at any rate in the poem as it stands) distinct from Beowulf, son of Scyld, connected both with Beow, son of Scyld, so initiating a theory which, for almost a century, was accepted as ascertained fact.

Kemble's identification was probably made independently of the work of this early scholar. Unlike him, Kemble, of course, realized that in our poem Beowulf the Dane, son of Scyld, is a person distinct from, is in fact not related to, Beowulf son of Ecgtheow. But he deliberately identified the two: he thought that two distinct traditions concerning the same hero had been amalgamated: in one of these traditions Beowulf may have been represented as son of Scyld, in the other as son of Ecgtheow, precisely as the hero Gunnar or Gunter is in one tradition son of Gifica (Giuki), in another son of Dankrat.

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 chapter begin as No. I. This is the case in the other MS of the Pastoral Care^[531]. Is there, then, such difficulty in the dissertation on the glory of the ancient Danish kings being treated as what, in fact, it is: a prologue or preface; and being, as such, simply left outside the numbering?

Still less can we argue for the identification of our hero, the son of Ecgtheow, with Frotho, and through him with Beow, from the supposed resemblances between the dragon fights of Beowulf and Frotho. Such resemblances have been divined by Sievers, but we have seen that it is the dissimilarity, not the resemblance, of the two dragon fights which is really noteworthy^[532].

To prove that Beow was the original antagonist of Grendel there remains, then, only the mention in the charter of a Grendles mere near a Bēowan hamm^[533]. Now this was not known to Kemble at the time when he formed his theory that the original slayer of Grendel was not Beowulf, but Beow. And if the arguments upon which Kemble based his theory had been at all substantial, this charter would have afforded really valuable support. But the fact that two names occur near each other in a charter cannot confirm any theory, unless that theory has already a real basis of its own.

(2) Beow

Therefore, until some further evidence be discovered, we must regard the belief that the Grendel and the dragon stories were originally myths of Beow, as a theory for which sufficient evidence is not forthcoming.

But note where the theory breaks down. It seems indisputable that Beowulf the Dane, son of Scyld Scefing, is identical with Beo(w) of the genealogies: for Beo(w) is son of Scyld^[534] or Sce(a)ldwa^[535], who is a Scefing. But here we must stop. There is, as we have seen, no evidence that the Grendel or dragon adventures were transferred from him to their present hero, Beowulf the Geat, son of Ecgtheow. It would, of course, be quite possible to accept such transference, and still to reject the mythological interpretation of these adventures, just as it would be possible to believe that Gawain was originally a sun-hero, whilst rejecting the interpretation as a sun-myth of any particular adventure which could be proved to have been once told concerning Gawain. But I do not think we need even concede, as Boer^[536] and Chadwick^[537] do, that adventures have been transferred from Beowulf the Dane to Beowulf the Geat. We have seen that there is no evidence for such transference, however intrinsically likely it may be. Till evidence is forthcoming, it is useless to build upon Kemble's conjecture that Beowulf the Scylding sank into Beowulf the Wægmunding^[538].

But it is due to Kemble to remember that, while he only put this forward as a tentative conjecture, what he was certain about was the identity of Beowulf the Scylding with Beow, and the divinity of these figures. And here all the evidence seems to justify him.

"The divinity of the earlier Beówulf," Kemble wrote, "I hold for indisputable.... Beo or Beow is ... in all probability a god of agriculture and fertility.... It strengthens this view of the case that he is the grandson of Sceáf, manipulus frumenti, with whom he is perhaps in fact identical^[539]."

Whether or no Beow and Sceaf were ever identical, it is certain that Beow (grain) the descendant of Sceaf (sheaf) suggests a corn-myth, some survival from the ancient worship of a corn-spirit.

Now bēow, 'grain, barley,' corresponds to Old Norse bygg, just as, corresponding to O.E. trīewe, we have O.N. tryggr, or corresponding to O.E. glēaw, O.N. glǫggr. Corresponding to the O.E. proper name Bēow, we might expect an O.N. name, the first letters in which would be Bygg(v)-.

And pat he comes, like the catastrophe of the Old Comedy. When Loki strode into the Hall of Ægir, and assailed with clamour and scandal the assembled gods and goddesses, there were present, among the major gods, also Byggvir and his wife Beyla, the servants of Frey, the god of agriculture and fertility. Loki reviles the gods, one after the other: at last he exchanges reproaches with Frey. To see his lord so taunted is more than Byggvir can endure, and he turns to Loki with the words:

Know thou, that were my race such as is that of Ingunar-Frey, and if I had so goodly a seat, finer than marrow would I grind thee, thou crow of ill-omen, and pound thee all to pieces^[540].

Byggvir is evidently no great hero: he draws his ideas from the grinding of the homely hand-mill, with which John Barleycorn has reason to be familiar:

Loki, who has addressed by name all the other gods, his acquaintances of old, professes not to know who is this insignificant being: but his reference to the hand-mill shows that in reality he knows quite well:

What is that little creature that I see, fawning and sneaking and snuffling: ever wilt thou be at the ears of Frey, and chattering at the quern^[542].

Byggvir replies with a dignity which reminds us of the traditional characteristics of Sir John Barleycorn, or Allan O'Maut. For:

Byggvir adopts the same comic-heroic pose:

Byggvir am I named, and all gods and men call me hasty; proud am I, by reason that all the children of Odin are drinking ale together^[544].

But any claims Byggvir may make to be a hero are promptly dismissed by Loki:

Hold thou silence, Byggvir, for never canst thou share food justly among men: thou didst hide among the straw of the hall: they could not find thee, when men were fighting^[545].

Now the taunts of Loki, though we must hope for the credit of Asgard that they are false, are never pointless. And such jibes as Loki addresses to Byggvir would be pointless, if applied to one whom we could think of as in any way like our Beowulf. Later, Beyla, wife of Byggvir, speaks, and is silenced with the words "Hold thy peace—wife thou art of Byggvir." Byggvir must have been a recognized figure of the old mythology^[546], but one differing from the monster-slaying Beow of Müllenhoff's imagination.

Byggvir is a little creature (et lítla), and we have seen above^[547] that Scandinavian scholars have thought that they have discovered this old god in the Pekko who "promoted the growth of barley" among the Finns in the sixteenth century, and who is still worshipped among the Esthonians on the opposite side of the gulf as a three year old child; the form Pekko being derived, it is supposed, from the primitive Norse form *Beggwuz. This is a corner of a very big subject: the discovery, among the Lapps and Finns, of traces of the heathendom of the most ancient Teutonic world, just as Thomsen has taught us to find in the Finnish language traces of Teutonic words in their most antique form.

The Lappish field has proved the most successful hunting ground^[548]: among the Finns, apart from the Thunder-god, connection with Norse beliefs is arguable mainly for a group of gods of fruitfulness^[549]. The cult of these, it is suggested, comes from scattered Scandinavian settlers in Finland, among whom the Finns dwelt, and from whom they learnt the worship of the spirits of the seed and of the spring, just as they learnt more practical lessons. First and foremost among these stands Pekko, whom we know to have been especially the god of barley, and whose connection with Beow or Byggvir (*Beggwuz) is therefore a likely hypothesis enough^[550]. Much less certain is the connection of Sämpsä, the spirit of vegetation, with any Germanic prototype; he may have been a god of the rush-grass^[551] (Germ. simse). Runkoteivas or Rukotivo was certainly the god of rye, and the temptation to derive his name from Old Norse (rugr-tivorr, "rye-god") is great^[552]. But we have not evidence for the worship among Germanic peoples of such a rye-god, as we have in the case of the barley-god Byggvir-Beow. These shadowy heathen gods, however, do give each other a certain measure of mutual support.

And, whether or no Pekko be the same as Byggvir, his worship is interesting as showing how the spirit of vegetation may be honoured among primitive folk. His worshippers, the Setukese, although nominally members of the Greek Orthodox Church, speak their own dialect and often hardly understand that of their Russian priests, but keep their old epic and lyric traditions more than almost any other section of the Finnish-Esthonian race. Pekko, who was honoured among the Finns in the sixteenth century for "promoting the growth of barley," survives among the present-day peasantry around Pskoff, not only as a spirit to be worshipped, but as an actual idol, fashioned out of wax in the form of a child, sometimes of a three year old child. He lives in the corn-bin, but on certain occasions is carried out into the fields. Not everyone can afford the amount of wax necessary for a Pekko—in fact there is usually only one in a village: he lodges in turn with different members of his circle of worshippers. He holds two moveable feasts, on moonlight nights—one in spring, the other in autumn. The wax figure is brought into a lighted room draped in a sheet, there is feasting, with dancing hand in hand, and singing round Pekko. Then they go out to decide who shall keep Pekko for the next year—his host is entitled to special blessing and protection. Pekko is carried out into the field, especially to preside over the sowing^[553].

I doubt whether, in spite of the high authorities which support it, we can as yet feel at all certain about the identification of Beow and Pekko. But I think we can accept with fair certainty the identification of Beow and Byggvir. And we can at any rate use Pekko as a collateral example of the way in which a grain-spirit is regarded. Now in either case we find no support whatever for the supposition that the activities of Beow, the spirit of the barley, could, or would, have been typified under the guise of battles such as those which Beowulf the Geat wages against Grendel, Grendel's mother, and the dragon. In Beowulf the Geat we find much that suggests the hero of folk-tale, overlaid with much that belongs to him as the hero of an heroic poem, but nothing suggestive of a corn-myth. On the other hand, so long as we confine ourselves to Beow and his ancestor Sceaf, we are in touch with this type of myth, however remotely. The way that Sceaf comes over the sea, as recorded by William of Malmesbury, is characteristic. That "Sheaf" should be, in the language of Müllenhoff, "placed in a boat and committed to the winds and waves in the hope that he will return new-born in the spring" is exactly what we might expect, from the analogy of harvest customs and myths of the coming of spring.

In Sætersdale, in Norway, when the ice broke up in the spring, and was driven ashore, the inhabitants used to welcome it by throwing their hats into the air, and shouting "Welcome, Corn-boat." It was a good omen if the "Corn-boats" were driven high and dry up on the land^[554]. The floating of the sheaf on a shield down the Thames at Abingdon^[555] reminds us of the Bulgarian custom, in accordance with which the venerated last sheaf of the harvest was floated down the river^[556]. But every neighbourhood is not provided with convenient rivers, and in many places the last sheaf is merely drenched with water. This is an essential part of the custom of "crying the neck."

The precise ritual of "crying the neck" or "crying the mare" was confined to the west and south-west of England^[557]. But there is no such local limitation about the custom of drenching the last sheaf, or its bearers and escort, with water. This has been recorded, among other places, at Hitchin in Hertfordshire^[558], in Cambridgeshire^[559], Nottinghamshire^[560], Pembrokeshire^[561], Wigtownshire^[562] as well as in Holstein^[563], Westphalia^[564], Prussia^[565], Galicia^[566], Saxon Transsylvania^[567], Roumania^[568] and perhaps in ancient Phrygia^[569].

Now it is true that drenching the last sheaf with water, as a rain charm, is by no means the same thing as floating it down the river, in the expectation that it will come again in the spring. But it shows the same sense of the continued existence of the corn-spirit. That the seed, when sown, should be sprinkled with water as a rain charm (as is done in places) seems obvious and natural enough. But when the last sheaf of the preceding harvest is thus sprinkled, to ensure plenteous rain upon the crops of next year, we detect the same idea of continuity which we find expressed when Sceaf comes to land from over the sea: the spirit embodied in the sheaf of last year's harvest returning, and bringing the renewed power of vegetation.

The voyage of the Abingdonian sheaf on the Thames was conducted upon a shield, and it may be that the "vessel without a rower" in which "Sheaf" came to land was, in the original version, a shield. There would be precedent for this. The shield was known by the puzzling name of "Ull's ship" in Scaldic poetry, presumably because the god Ull used his shield as a boat. Anyway, Scyld came to be closely connected with Sceaf and Beow. In Ethelwerd he is son of the former and father of the latter: but in the Chronicle genealogies five names intervene between Scyld and Sceaf, and the son of Sceaf is Bedwig, or as he is called in one version, Beowi. Bedwig and Beowi are probably derived from Beowius, the Latinized form of Beow. A badly formed o might easily be mistaken for a d, and indeed Beowius appears in forms much more corrupt. In that case it would appear that while some genealogies made Beow the son of Scyld, others made him son of Sceaf, and that the compiler of the pedigree got over the difficulty in the usual way, by adding the one version to the other^[570].

But all this is very hypothetical; and how and when Scyld came to be connected with Sceaf and with Beow we cannot with any certainty say. At any rate we find no trace of such connection in Danish traditions of the primitive King Skjold of the Danes. But we can say, with some certainty, that in Beowulf the Dane, the son of Scyld Scefing, in our poem, we have a figure which is identical with Beow, son of Scyld or of Sceldwa and descendant of Sceaf, in the genealogies, and that this Beow is likely to have been an ancient corn-spirit, parallel to the Scandinavian Byggvir. That amount of mythology probably does underlie the Prologue to Beowulf, though the author would no doubt have been highly scandalized had he suspected that his pattern of a young prince was only a disguised heathen god. But I think that any further attempt to proceed, from this, to mythologize the deeds of Beowulf the Great, is pure conjecture, and probably quite fruitless conjecture.

I ought not to conclude this note without reference to the admirable discussion of this subject by Prof. Björkman in Englische Studien^[571]. This, with the elucidation of other proper names in Beowulf, was destined to be the last big contribution to knowledge made by that ripe and good scholar, whose premature loss we all deplore; and it shows to the full those qualities of wide knowledge and balanced judgment which we have all learnt to admire in him.

B. GRENDEL

It may be helpful to examine the places where the name of Grendel occurs in English charters.

A.D. 708. Grant of land at Abbots Morton, near Alcester, co. Worcester, by Kenred, King of the Mercians, to Evesham (extant in a late copy).

Ǣrest of grindeles pytt on wīðimære; of wīðimære on þæt rēade slōh ... of ðēre dīce on þene blace pōl; of þām pōle æfter long pidele in tō þām mersce; of þām mersce þā æft on grindeles pytt^[572].

The valley of the Piddle Brook is about a mile wide, with hills rising on each side till they reach a height of a couple of hundred feet above the brook. The directions begin in the valley and run "From Grindel's 'pytt' to the willow-mere; from the willow-mere to the red morass"; then from the morass the directions take us up the hill and along the lea, where they continue among the downs till we again make our descent into the valley, "from the ditch to the black pool, from the pool along the Piddle brook to the marsh, and from the marsh back to Grindel's 'pytt.'" In modern English a "pit" is an artificial hole which is generally dry: but the word is simply Latin puteus, "a well," and is used in this sense in the Gospel translations. Here it is a hole, and we may be sure that, with the willow-mere and the red slough on the one side, and the black pool and the marsh on the other, the hole was full of water.

A.D. 739. Grant of land at Creedy, co. Devon, by Æthelheard, King of Wessex, to Bishop Forthhere.

of doddan hrycge on grendeles pyt; of grendeles pytte on ifigbearo (ivy-grove)...^[573].

The spot is near the junction of the rivers Exe and Creedy, with Dartmoor in the distance. The neighbourhood bears uncanny names, Cāines æcer, egesan trēow. If, as has been suggested by Napier and Stevenson, a trace of this pit still survives in the name Pitt farm, the mere must have been in the uplands, about 600 feet above sea level.

A.D. 931. Grant of land at Ham in Wiltshire by Athelstan to his thane Wulfgar. Quoted above, p. 43. It is in this charter that on Bēowan hammes hecgan, on Grendles mere^[574] occur. "Grendel pits or meres" are in most other cases in low-lying marshy country: but this, like (perhaps) the preceding one, is in the uplands—it must have been a lonely mere among the hills, under Inkpen Beacon.

Circa A.D. 957. A list of boundaries near Battersea^[575].

Ðis synd ðā landgemǣre tō Batriceseie. Ǣrst at hēgefre; fram hēgefre to gætenesheale; fram gæteneshæle to gryndeles syllen; fram gryndeles sylle to russemere; fram ryssemere to bælgenham....

All this is low-lying land, just south of the Thames. Hēgefre is on the river; Bælgenham is Balham, co. Surrey. "From Grendel's mire to the rushy mere" harmonizes excellently with what we know of the swampy nature of this district in early times.

A.D. 958. Grant of land at Swinford, on the Stour, co. Stafford, by King Eadred to his thane Burhelm^[576].

Ondlong bæces wið neoþan eostacote; ondlong dīces in grendels-mere; of grendels-mere in stāncōfan; of stāncōfan ondlong dūne on stiran mere....

A.D. 972. Confirmation of lands to Pershore Abbey (Worcester) by King Edgar^[577].

of Grindles bece swā þæt gemǣre ligð....

A.D. 972. Extract from an account of the descent of lands belonging to Westminster, quoting a grant of King Edgar^[578].

andlang hagan to grendeles gatan æfter kincges mearce innan brægentan....

The property described is near Watling Street, between Edgware, Hendon, and the River Brent. It is a low-lying district almost surrounded by the hills of Hampstead, Highgate, Barnet, Mill Hill, Elstree, Bushey Heath and Harrow. The bottom of the basin thus formed must have been a swamp^[579]. What the "gate" may have been it is difficult to say. A foreign scholar has suggested that it may have been a narrow mountain defile or possibly a cave^[580]: but this suggestion could never have been made by anyone who knew the country. The "gate" is likely to have been a channel connecting two meres—or it might have been a narrow piece of land between them—one of those enge ānpaðas which Grendel and his mother had to tread. Anyway, there is nothing exceptional in this use of "gate" in connection with a water-spirit. Necker, on the Continent, also had his "gates." Thus there is a "Neckersgate Mill" near Brussels, and the name "Neckersgate" used also to be applied to a group of houses near by, surrounded by water^[581].

All the other places clearly point to a water-spirit: two meres, two pits, a mire and a beck: for the most part situated in low-lying country which must in Anglo-Saxon times have been swampy. All this harmonizes excellently with the fenfreoðo of Beowulf (l. 851). Of course it does not in the least follow that these places were named after the Grendel of our poem. It may well be that there was in England a current belief in a creature Grendel, dwelling among the swamps. Von Sydow has compared the Yorkshire belief in Peg Powler, or the Lancashire Jenny Greenteeth. But these aquatic monsters are not exactly parallel; for they abide in the water, and are dangerous only to those who attempt to cross it, or at any rate venture too near the bank^[582], whilst Grendel and even his mother are capable of excursions of some distance from their fastness amid the fens.

Of course the mere-haunting Grendel may have been identified only at a comparatively late date with the spirit who struggles with the hero in the house, and flees below the earth in the folk-tale.

At any rate belief in a Grendel, haunting mere and fen, is clearly demonstrable for England—at any rate for the south and west of England: for of these place-names two belong to the London district, one to Wiltshire, one to Devonshire, two to Worcester and one to Stafford. The place-name Grendele in Yorkshire is too doubtful to be of much help. (Domesday Book, I, 302.) It is the modern village Grindale, four miles N.W. of Bridlington. From it, probably, is derived the surname Grindle, Grindall (Bardsley).

Abroad, the nearest parallel is to be found in Transsylvania, where there is a Grändels môr among the Saxons of the Senndorf district, near Bistritz. The Saxons of Transsylvania are supposed to have emigrated from the neighbourhood of the lower Rhine and the Moselle, and there is a Grindelbach in Luxemburg which may possibly be connected with the marsh demon^[583].

Most of the German names in Grindel- or Grendel- are connected with grendel, "a bar," and therefore do not come into consideration here^[584]: but the Transsylvanian "Grendel's marsh^[585]," anyway, reminds us of the English "Grendel's marsh" or "mere" or "pit." Nevertheless, the local story with which the Transsylvanian swamp is connected—that of a peasant who was ploughing with six oxen and was swallowed up in the earth—is such that it requires considerable ingenuity to see any connection between it and the Beowulf-Grendel-tale^[586].

The Anglo-Saxon place-names may throw some light upon the meaning and etymology of "Grendel^[587]." The name has generally been derived from grindan, "to grind"; either directly^[588], because Grendel grinds the bones of those he devours, or indirectly, in the sense of "tormentor^[589]." Others would connect with O.N. grindill, "storm," and perhaps with M.E. gryndel, "angry^[590]."

It has recently been proposed to connect the word with grund, "bottom": for Grendel lives in the mere-grund or grund-wong and his mother is the grund-wyrgin. Erik Rooth, who proposes this etymology, compares the Icelandic grandi, "a sandbank," and the common Low German dialect word grand, "coarse sand^[591]." This brings us back to the root "to grind," for grand, "sand" is simply the product of the grinding of the waves^[592]. Indeed the same explanation has been given of the word "ground^[593]."

However this may be, the new etymology differs from the old in giving Grendel a name derived, not from his grinding or tormenting others, but from his dwelling at the bottom of the lake or marsh^[594]. The name would have a parallel in the Modern English grindle, grundel, German grundel^[595], a fish haunting the bottom of the water.

The Old English place-names, associating Grendel as they do with meres and swamps, seem rather to support this.

As to the Devonshire stream Grendel (now the Grindle or Greendale Brook), it has been suggested that this name is also connected with the root grand, "gravel," "sand." But, so far as I have been able to observe, there is no particular suggestion of sand or gravel about this modest little brook. If we follow the River Clyst from the point where the Grindle flows into it, through two miles of marshy land, to the estuary of the Exe, we shall there find plenty. But it is clear from the charter of 963 that the name was then, as now, restricted to the small brook. I cannot tell why the stream should bear the name, or what, if any, is the connection with the monster Grendel. We can only note that the name is again found attached to water, and, near the junction with the Clyst, to marshy ground.

Anyone who will hunt Grendel through the shires, first on the 6-in. ordnance map, and later on foot, will probably have to agree with the Three Jovial Huntsmen

But, if some conclusions, although scanty, can be drawn from place-names in which the word grendel occurs, nothing can be got from the numerous place-names which have been thought to contain the name Bēow. The clearest of these is the on Bēowan hammes hecgan, which occurs in the Wiltshire charter of 931. But we can learn nothing definite from it: and although there are other instances of strong and weak forms alternating, we cannot even be quite certain that the Beowa here is identical with the Beow of the genealogies^[596].

The other cases, many of which occur in Domesday Book are worthless. Those which point to a weak form may often be derived from the weak noun bēo, "bee": "The Anglo-Saxons set great store by their bees, honey and wax being indispensables to them^[597]."

Bēas brōc, Bēas feld (Bewes feld) occur in charters: but here a connection with bēaw, "horsefly," is possible: for parallels, one has only to consider the long list of places enumerated by Björkman, the names of which are derived from those of beasts, birds, or insects^[598]. And in such a word as Bēolēah, even if the first element be bēow, why may it not be the common noun "barley," and not the name of the hero at all?

No argument can therefore be drawn from such a conjecture as that of Olrik, that Bēas brōc refers to the water into which the last sheaf (representing Beow) was thrown, in accordance with the harvest custom, and in the expectation of the return of the spirit in the coming spring^[599].

C. THE STAGES ABOVE WODEN IN THE WEST-SAXON GENEALOGY

The problems to which this pedigree gives rise are very numerous, and some have been discussed above. There are four which seem to need further discussion.

(I) A "Sceafa" occurs in Widsith as ruling over the Longobards. Of course we cannot be certain that this hero is identical with the Sceaf of the genealogy. Now there is no one in the long list of historic or semi-historic Longobard kings, ruling after the tribe had left Scandinavia, who bears a name at all similar. It seems therefore reasonable to suppose that Sceafa, if he is a genuine Longobard king at all, belongs to the primitive times when the Longobardi or Winnili dwelt in "Scadan," before the historic or semi-historic times with which our extant list deals. And Old English accounts, although making Sceaf an ancestor of the Saxon kings, are unanimous in connecting him with Scani or Scandza.

Some scholars^[600] have seen a serious difficulty in the weak form "Sceafa," as compared with "Sceaf." But we have the exactly parallel cases of Horsa^[601] compared with Hors^[602], and Hrǣdla^[603] compared with Hrǣdel^[604], Hrēðel. Parallel, but not quite so certain, are Sceldwa^[605] and Scyld^[606], Gēata^[607] and Gēat^[608], Bēowa^[609] and Bēaw, Bēo(w)^[610].

I do not think it has ever been doubted that the forms Hors and Horsa, or Hrēðel and Hrǣdla, relate to one and the same person. Prof. Chadwick seems to have little or no doubt as to the identity of Scyld and Sceldwa^[611], or Bēo and Bēowa^[612]. Why then should the identity of Scēaf and Scēafa be denied because one form is strong and the other weak^[613]? We cannot demonstrate the identity of the figure in the genealogies with the figure in Widsith; but little difficulty is occasioned by the weak form.

(II) Secondly, the absence of the name Scēaf from the oldest MS of the Chronicle (the Parker MS, C.C.C.C. 173) has been made the ground for suggesting that when that MS was written (c. 892) Sceaf had not yet been invented (Möller, Volksepos, 43; Symons in Pauls Grdr. (2), III, 645; Napier, as quoted by Clarke, Sidelights, 125). But Sceaf, and the other names which are omitted from the Parker MS, are found in the other MSS of the Chronicle and the allied pedigrees, which are known to be derived independently from one and the same original. Now, unless the names were older than the Parker MS, they could not appear in so many independent transcripts. For, even though these transcripts are individually later, their agreement takes us back to a period earlier than that of the Parker MS itself^[614].

An examination of the different versions of the genealogy, given on pp. 202-3, above, and of the tree showing the connection between them, on p. 315, will, I think, make this clear.

The versions of the pedigree given in the Parker MS of the Chronicle, in Asser and in Textus Roffensis I, all contain the stages Friþuwald and Friþuwulf. Asser and Roff. I are connected by the note about Gēata: but Roff. I is not derived from that text of Asser which has come down to us, as that text has corrupted Fin and Godwulf into one name and has substituted Seth for Scēaf ["Seth, Saxonice Sceaf": Florence of Worcester]. Roff. I is free from both these corruptions.

Ethelwerd is obviously connected with a type of genealogy giving the stages Friþuwald and Friþuwulf, but differs from all the others in giving no stages between Scyld and Scēf.

None of the other versions contain the names Friþuwald and Friþuwulf. They are closely parallel, but fall into groups showing special peculiarities.

MSS Tib. A. VI and Tib. B. I of the Chronicle show only trifling differences of spelling. The MSS belong respectively to about the years 1000 and 1050, and are both derived from an Abingdon original of about 977^[615].

MS Cott. Tib. B. IV is derived from a copy of the Chronicle sent North about 892^[616].

MS Cott. Tib. B. V and Textus Roffensis II are closely connected, but neither is derived from the other. For Roff. II preserves Teþwa and Hwāla, who are lost in Tib. B. V; Tib. B. V preserves Iterman, who is corrupted in Roff. II. Both Tib. B. V and Roff. II carry the pedigree down to Edgar, mentioning his three sons Ēadweard and Ēadmund and Æþelred æðelingas syndon Ēadgāres suna cyninges. The original therefore apparently belongs to some date before 970, when Edmund died (cf. Stevenson's Asser, 158, note).

Common features of MS Cott. Tib. B. V and Roff. II are (1) Eat(a) for Geat(a), (2) the omission of d from Scealdwa, and (3) the expression se Scēf, "this Scef." Features (1) and (3) are copied in the Icelandic pedigrees. Scealdwa is given correctly there, but the Icelandic transcriber could easily have got it from Scealdwaging above. The Icelandic was, then, ultimately derived either from Tib. B. V or from a version so closely connected as not to be worth distinguishing.

Accordingly Cott. Tib. B. V, Textus Roffensis II, Langfeðgatal and Flateyarbók form one group, pointing to an archetype c. 970.

The pedigrees can accordingly be grouped on the system shown on the opposite page^[617].

(III) Prof. Chadwick, in his Origin of the English Nation, draws wide deductions from the fact that the Danes traced the pedigree of their kings back to Skjold, whilst the West-Saxons included Sceldwa (Scyld) in their royal pedigree:

"Since the Angli and the Danes claimed descent from the same ancestor, there can be no doubt that the bond was believed to be one of blood^[618]."

This belief, Prof. Chadwick thinks, went back to exceedingly early times^[619], and he regards it as well-founded:

"It is true that the Angli of Britain seem never to have included themselves among the Danes, but the reason for this may be that the term Dene (Danir) had not come into use as a collective term before the invasion of Britain^[620]."

Doubtless the fact that the name of a Danish king Scyld or Sceldwa is found in a pedigree of West-Saxon kings, as drawn up at a period certainly not later than 892, points to a belief, at that date, in some kind of a connection. But we have still to ask: How close was the connection supposed to be? And how old is the belief?

Firstly as to the closeness of the connection. Finn also occurs in the pedigree—possibly the Frisian king: Sceaf occurs, possibly, though not certainly, a Longobard king. Noah and Adam occur; are we therefore to suppose that the compiler of the Genealogy believed his kings to be of one blood with the Hebrews? Certainly he did: but only remotely, as common descendants of Noah. And the occurrence of Sceldwa and Sceaf and Finn in the genealogies—granting the identity of these heroes with Skjold of the Danes, Sceafa of the Longobards and Finn of the Frisians, might only prove that the genealogist believed in their common (Germanic) race.

Secondly, how old is the belief? The Anglian genealogies (Northumbrian, Mercian and East Anglian), as reproduced in the Historia Brittonum and in the Vespasian MS, form part of what is doubtless, as is said above, the oldest extant English historical document. But in this document there is no mention of Scyld. Indeed, it contains no pedigree of the West-Saxon kings at all. From whatever cause, the West-Saxon genealogy is not extant from so early a date as are the pedigrees of the Northumbrian, Mercian, East Anglian and Kentish kings^[621]. Still, this may well be a mere accident, and I am not prepared to dispute that the pedigree which traces the West-Saxon kings to Woden dates back, like the other genealogies connecting Old English kings with Woden, to primitive and heathen times. Now the West-Saxon pedigree is found in many forms: some which trace the royal house only to Woden, and some which go beyond Woden and contain a list of names by which Woden is connected with Sceaf, and then with Noah and Adam.

(1) The nucleus of the whole pedigree is to be found in the names between Cynric or Cerdic and Woden. These occur in every version. The pedigree in this, its simplest form, is found twice among the entries in the Chronicle which deal with the events of heathen times, under 552 and 597. These names fall into verse:

Like the mnemonic lists in Widsith, these lines are probably very old. Their object is clearly to connect the founder of the West-Saxon royal house with Woden. Note, that not only do the names alliterate, but the alliteration is perfect. Every line attains double alliteration in the first half, with one alliterating word only in the second half. The lines must go back to times when lists of royal ancestors, both real and imaginary, had to be arranged in correct verse; times when such things were recorded by memory rather than by writing. They are pre-literary, and were doubtless chanted by retainers of the West-Saxon kings in heathen days.

(2) An expanded form of this genealogy occurs in MSS C.C.C.C. 183 and Cotton Tib. B. V. Woden is here furnished with a father Frealaf. We know nothing of any Frealaf as father of the All-Father in heathen days, though Frealaf is found in this capacity in other genealogies written down in the ages after the conversion. Frealaf breaks the correct alliterative system. In both MSS the pedigree is brought down to King Ine (688-726): both MSS are ultimately, no doubt, derived from a list current in the time of that king, that is to say less than a century after the conversion of Wessex.

(3) A further expansion, which Prof. Napier has held on linguistic grounds^[622] to have been written down as early as 750, is incorporated in a genealogical and chronological note regarding the West-Saxon kings, which is extant in many MSS^[623]. In its present form this genealogical note is a recension, under Alfred, of a document coming down to the death of his father Æthelwulf. It traces the pedigree of Æthelwulf to Cerdic, but it keeps this district from the rhythmical nucleus, in which it traces Cerdic to Woden, and no further.

(4) Then, in the Anglo-Saxon Chronicle, under the year 855, the pedigree is given in its most elaborate form. There the genealogy of Æthelwulf is traced in one unbroken series, not merely through Cerdic to Woden, but from Woden through a long line of Woden's ancestors, including Frealaf, Geat, Sceldwa and Sceaf, to Noah and Adam.

It has been noted above^[624] that none of the Chronicle pedigrees stop at Sceaf. The Chronicle, in the stages above Woden, recognizes as stopping places only Geat (Northumbrian pedigree, anno 547) or Adam (West-Saxon pedigree, anno 855).

(5) The Chronicle of Ethelwerd (c. 1000) does, however, stop at Scef^[625]. Now it has been argued that Ethelwerd's pedigree is merely abbreviated from the pedigree in the Anglo-Saxon Chronicle under 855, and that, in making Scef the final stage, and in what he tells us about that hero, Ethelwerd is merely adapting what he had read in Beowulf about Scyld^[626]. But this seems hardly possible. Ethelwerd, it is true, borrows most of his facts from the Chronicle, from Bede, and other known sources: but there are some passages which show that he had access to a source now lost. Ethelwerd was a member of the West-Saxon royal house, and he wrote his Chronicle for a kinswoman, Matilda, in order, as he says, to explain their common stock and race. They were both descended from Æthelwulf, the chronicler being great-great-grandson of Æthelred, and the lady to whom he dedicates his work being great-great-granddaughter of Alfred. So he writes to tell "who and whence were their kin, so far as memory adduces, and our parents have taught us." Accordingly, though he begins his Chronicle with the Creation, the bulk of it is devoted to the deeds of his or Matilda's ancestors. Is it credible that he would have cut out all the stages in their common pedigree between Scyld and Scef, that he would have sacrificed all the ancestors of Scef, thus severing relations with Noah and Adam, and that he would have attributed to Scef the story which in Beowulf is attributed to Scyld, all this simply in order to bring his English pedigree into some harmony with what is told about the Danish pedigree in Beowulf—a poem of which we have no evidence that he had ever heard?

To suppose him to have done this, is to make him sacrifice, without any reason, just that part of the pedigree in the Chronicle under 855 which, from all we know of Ethelwerd, was most likely to have interested him: that which connected his race with Noah and Adam. Further, it is to suppose him to have reproduced just those stages in the pedigree which on critical grounds modern scholars can show to be the oldest, and to have modified or rejected just those which on critical grounds modern scholars can show to be later accretion. When Brandl supposes Ethelwerd to have produced his pedigree by comparing together merely the materials which have come down to us to-day, namely Beowulf and the Chronicle, he is, in reality, attributing to him the mind and acumen of a modern critic. An Anglo-Saxon alderman could only have detected and rejected the additions by using some material which has not come down to us. What more natural than that Ethelwerd, who writes as the historian of the West-Saxon royal family, should have known of a family pedigree which traced the line up to Sceaf and his arrival in the boat, and that he should have (rightly) thought this to be more authoritative than the pedigree in the Chronicle under the year 855, which had been expanded from it? Prof. Chadwick, it seems to me, is here quite justified in holding that Ethelwerd had "acquired the genealogy from some unknown source, in a more primitive form than that contained in the Chronicle^[627]."

But, because the source of Ethelwerd's pedigree is more primitive than that contained in the Chronicle under the year 855, it does not follow that it goes back to heathen times. Wessex had been converted more than two centuries earlier.

We are now in a position to make some estimate of the antiquity of Scyld and Sceaf in the West-Saxon pedigree. The nucleus of this pedigree is to be found in the verses connecting Cynric and Cerdic with Woden. (Even as late as Æthelwulf and Alfred this nucleus is often kept distinct from the later, more historic stages connecting Cerdic with living men.) Pedigrees of other royal houses go to Woden, and many stop there; however, in times comparatively early, but yet Christian, we find Woden provided with five ancestors: later, Ethelwerd gives him ten: the Chronicle gives him twenty-five. It is evidently a process of accumulation.

Now, if the name of Scyld had occurred in the portion of the pedigree which traces the West-Saxon kings up to Woden, it would possess sufficient authority to form the basis of an argument. But Scyld, like Heremod, Beaw and Sceaf, occurs in the fantastic development of the pedigree, by which Woden is connected up with Adam and Noah. The fact that these heroes occur above Woden makes it almost incredible that their position in the pedigree can go back to heathen times. Those who believed in Woden as a god can hardly have believed at the same time that he was a descendant of the Danish king Scyld. This difficulty Prof. Chadwick admits: "It is difficult to believe that in heathen times Woden was credited with five generations of ancestors, as in the Frealaf-Geat list." Still less is it credible that he was credited with 25 generations of ancestors, as in the Frealaf-Geat-Sceldwa-Sceaf-Noe-Adam list.

The obvious conclusion seems to me to be that the names above Woden were added in Christian times to the original list, which in heathen times only went back to Woden, and which is still extant in this form. A Christian, rationalizing Woden as a human magician, would have no difficulty in placing him far down the ages, just as Saxo Grammaticus does^[628]. Obviously Noe-Adam must be an addition of Christian times, and the same seems to me to apply to all the other names above Woden, which, though ancient and Germanic, are not therefore ancient and Germanic in the capacity of ancestors of Woden.

And even if these extraordinary ancestors of Woden were really believed in in heathen times, they cannot have been regarded as the special property of any one nation. For it was never claimed that the West-Saxon kings had any unique distinction in tracing their ancestry to Woden, such as would give them a special claim upon Woden's forefathers. How then can the ancient belief (if indeed it were an ancient belief) that Woden was descended from Scyld, King of Denmark, prove that the Anglo-Saxons regarded themselves as specially related to the Danes? For any such relationship derived through Woden must have been shared by all descendants of the All-Father.

Prof. Chadwick avoids this difficulty by supposing that Woden did not originally occur in the pedigree, but is a later insertion^[629]. But how can this be so when, of the two forms in which the West-Saxon pedigree appears, one (and, so far as our evidence goes, much the older one) traces the kings to Woden and stops there. The object of this pedigree is to connect the West-Saxon kings with Woden. The expanded pedigrees, which carry on the line still further, from Woden to Sceldwa, Sceaf and Adam, though very numerous, are all traceable to one, or at most two, sources. It is surely not the right method to regard Woden as an interpolation (though he occurs in that portion of the pedigree which is common to all versions, some of which we can probably trace back to primitive times), and to regard as the original element Scyld and Sceaf (though they form part of the continuation of the pedigree found only in, at most, two families of MSS which we cannot trace back beyond the ninth century).

Besides, there is the strongest external support for Woden in the very place which he occupies in the West-Saxon pedigree. That pedigree is traced in all its texts up to one Baldæg and his father Woden. Those texts which further give Woden's ancestry make him a descendant of Frealaf—they generally make Woden son of Frealaf, though some texts insert an intermediate Frithuwald.

Now the very ancient Northumbrian pedigree also goes up, by a different route, to "Beldæg," and gives him Woden for a father. In some versions (e.g. the Historia Brittonum) the Northumbrian pedigree stops there: in others (e.g. the Vespasian MS) Woden has a father Frealaf. How then can it be argued, contrary to the unanimous evidence of all the dozen or more MSS of the West-Saxon pedigree, that Woden, standing as he does between his proper father and his proper son, is an interpolation? There is no evidence whatsoever to support such an argument, and everything to disprove it.

The fact that Sceaf, Sceldwa and Beaw occur above Woden, that some versions of the pedigree stop at Woden, and that in heathen times presumably all must have stopped when they reached the All-Father, seems to me a fatal argument—not against the antiquity of the legends of Sceaf, Sceldwa, and Beaw, but against the antiquity of these characters in the capacity (given to them in the Anglo-Saxon Chronicle) of ancestors of the West-Saxon kings, and against the vast deduction concerning the origin of the English nation which Prof. Chadwick draws from this supposed antiquity.

(IV) Precisely the same argument—that Sceaf, Sceldwa and Beaw are found above Woden in the pedigree of the English kings, and are not likely to have occupied that place in primitive heathen times, is fatal to the attempt to draw from this pedigree any argument that the myths of these heroes were specially and exclusively Anglo-Saxon. The argument of Müllenhoff and other scholars for an ancient, purely Anglo-Saxon Beowa-myth^[630] falls, therefore, to the ground.

D. EVIDENCE FOR THE DATE OF BEOWULF. THE RELATION OF BEOWULF TO THE CLASSICAL EPIC

A few years ago there was a tendency to exaggerate the value of grammatical forms in fixing the date of Old English poetry, and attempts were made to arrange Old English poems in a chronological series, according to the exact percentage of "early" to "late" forms in each. There has now been a natural reaction against the assumption that, granting certain forms to be archaic, these would necessarily be found in a percentage diminishing exactly according to the dates of composition of the various poems in which they occur. The reaction has now gone to the other extreme, and grammatical facts are in danger of being regarded as not being "in any way valid or helpful indications of dates^[631]."

Schücking^[632], in an elaborate recent monograph on the date of Beowulf, rejects the grammatical evidence as valueless, and proceeds to date the poem about two centuries later than has usually been held, placing its composition at the court of some christianized Scandinavian monarch in England, about 900 A.D.

But it surely does not follow that, because grammatical data have been misused, therefore no use can be made of them. And, if Beowulf was composed about the year 900, from stories current among the Viking settlers, how are we to account for the fact that the proper names in Beowulf are given, not in the Scandinavian forms of the Viking age, nor in corruptions of such forms, but in the correct English forms which we should expect, according to English sound laws, if the names had been brought over in the sixth century, and handed down traditionally^[633]?

For example, King Hygelac no doubt called himself Hugilaikaz. The Chochilaicus of Gregory of Tours is a good—if uncouth—shot at reproducing this name. The name became, in Norse, Hugleikr and in Danish Huglek (Hugletus in Saxo): traditional kings so named are recorded, though it is difficult to find that they have anything in common with the King Hygelac in Beowulf^[634]. Had the name been introduced into England in Viking times, we should expect the Scandinavian form, not Hygelāc^[635].

Even in the rare cases where the character in Beowulf and his Scandinavian equivalent bear names which are not phonologically identical, the difference does not point to any corruption such as might have arisen from borrowing in Viking days^[636]. We have only to contrast the way in which the names of Viking chiefs are recorded in the Anglo-Saxon Chronicle, to be convinced that the Scandinavian stories recorded in Beowulf are due to contact during the age when Britain was being conquered, not during the Viking period three or four centuries later^[637].

And the arguments from literary and political history, which Schücking adduces to prove his late date, seem to me to point in exactly the opposite direction, and to confirm the orthodox view which would place Beowulf nearer 700 than 900.

Schücking urges that, however highly we estimate the civilizing effect of Christianity, it was only in the second half of the seventh century that England was thoroughly permeated by the new faith. Can we expect already, at the beginning of the eighth century, a courtly work, showing, as does Beowulf, such wonderful examples of tact, modesty, unselfishness and magnanimity? And this at the time when King Ceolwulf was forced by his rebellious subjects to take the cowl. For Schücking^[638], following Hodgkin^[639], reminds us how, in the eighth century, out of 15 Northumbrian kings, five were dethroned, five murdered; two abdicated, and only three held the crown to their death; and how at the end of the century Charlemagne called the Northumbrian Angles "a perfidious and perverse nation, murderers of their lords."

But surely, at the base of all this argument, lies the same assumption which, as Schücking rightly holds, vitiates so many of the grammatical arguments; the assumption that development must necessarily be in steady and progressive proportion. We may take Penda as a type of the unreclaimed heathen, and Edward the Confessor of the chaste and saintly churchman; but Anglo-Saxon history was by no means a development in steady progression, of diminishing percentages of ruffianism and increasing percentages of saintship.

The knowledge of, and interest in, heathen custom shown in Beowulf, such as the vivid accounts of cremation, would lead us to place it as near heathen times as other data will allow. So much must be granted to the argument of Prof. Chadwick^[640]. But the Christian tone, so far from leading us to place Beowulf late, would also lead us to place it near the time of the conversion. For it is precisely in these times just after the conversion, that we get the most striking instances in all Old English history of that "tact, modesty, generosity, and magnanimity" which Schücking rightly regards as characteristic of Beowulf.

King Oswin (who was slain in 651) was, Bede tells us, handsome, courteous of speech and bearing, bountiful both to great and lowly, beloved of all men for his qualities of mind and body, so that noblemen came from all over England to enter his service—yet of all his endowments gentleness and humility were the chief. We cannot read the description without being reminded of the words of the thegns in praise of the dead Beowulf. Indeed, I doubt if Beowulf would have carried gentleness to those around him quite so far as did Oswin. For Oswin had given to Bishop Aidan an exceptionally fine horse—and Aidan gave it to a beggar who asked alms. The king's mild suggestion that a horse of less value would have been good enough for the beggar, and that the bishop needed a good horse for his own use, drew from the saint the stern question "Is that son of a mare dearer to thee than the Son of God?" The king, who had come from hunting, stood warming himself at the fire, thinking over what had passed; then he suddenly ungirt his sword, gave it to his squire, and throwing himself at the feet of the bishop, promised never again to grudge anything he might give in his charities.

Of course such conduct was exceptional in seventh century Northumbria—it convinced Aidan that the king was too good to live long, as indeed proved to be the case. But it shows that the ideals of courtesy and gentleness shown in Beowulf were by no means beyond the possibility of attainment—were indeed surpassed by a seventh century king. I do not know if they could be so easily paralleled in later Old English times.

And what is true from the point of view of morals is true equally from that of art and learning. In spite of the misfortunes of Northumbrian kings in the eighth century, the first third of that century was "the Golden Age of Anglo-Saxon England^[641]." And not unnaturally, for it had been preceded by half a century during which Northumbria had been free both from internal strife and from invasion. The empire won by Oswiu over Picts and Scots in the North had been lost at the battle of Nectansmere: but that battle had been followed by the twenty years reign of the learned Aldfrid, whose scholarship did not prevent him from nobly retrieving the state of the kingdom^[642], though he could not recover the lost dominions.

Now, whatever we may think of Beowulf as poetry, it is remarkable for its conscious and deliberate art, and for the tone of civilization which pervades it. And this half century was distinguished, above any other period of Old English history, precisely for its art and its civilization. Four and a half centuries later, when the works of great Norman master builders were rising everywhere in the land, the buildings which Bishop Wilfrid had put up during this first period of conversion were still objects of admiration, even for those who had seen the glories of the great Roman basilicas^[643].

Nor is there anything surprising in the fact that this "golden age" was not maintained. On the contrary, it is "in accordance with the phenomena of Saxon history in general, in which seasons of brilliant promise are succeeded by long eras of national eclipse. It is from this point of view quite in accordance with natural likelihood that the age of conversion was one of such stimulus to the artistic powers of the people that a level of effort and achievement was reached which subsequent generations were not able to maintain. The carved crosses and the coins certainly degenerate in artistic value as the centuries pass away, and the fine barbaric gold and encrusted work is early in date^[644]."

Already in the early part of the eighth century signs of decay are to be observed. At the end of his Ecclesiastical History, Bede complains that the times are so full of disturbance that one knows not what to say, or what the end will be. And these fears were justified. A hundred and forty years of turmoil and decay follow, till the civilization of the North and the Midlands was overthrown by the Danes, and York became the uneasy seat of a heathen jarl.

How it should be possible to see in these facts, as contrasted with the Christian and civilized tone of Beowulf, any argument for late date, I cannot see. On the contrary, because of its Christian civilization combined with its still vivid, if perhaps not always quite exact, recollection of heathen customs, we should be inclined to put Beowulf in the early Christian ages.

A further argument put forward for this late date is the old one that the Scandinavian sympathies of Beowulf show it to have been composed for a Scandinavian court, the court, Schücking thinks, of one of the princes who ruled over those portions of England which the Danes had settled^[645]. Of course Schücking is too sound a scholar to revive at this time of day the old fallacy that the Anglo-Saxons ought to have taken no interest in the deeds of any but Anglo-Saxon heroes. But how, he asks, are we to account for such enthusiasm for, such a burning interest in, a people of alien dialect and foreign dynasty, such as the Scyldings of Denmark?

The answer seems to me to be that the enthusiasm of Beowulf is not for the Danish nation as such: on the contrary, Beowulf depicts a situation which is most humiliating to the Danes. For twelve years they have suffered the depredations of Grendel; Hrothgar and his kin have proved helpless: all the Danes have been unequal to the need. Twice at least this is emphasized in the most uncompromising, and indeed insulting, way^[646]. The poet's enthusiasm is not, then, for the Danish race as such, but for the ideal of a great court with its body of retainers. Such retainers are not necessarily native born—rather is it the mark of the great court that it draws men from far and wide to enter the service, whether permanently or temporarily, even as Beowulf came from afar to help the aged Hrothgar in his need.

It is this ideal of personal valour and personal loyalty, rather than of tribal patriotism, which pervades Beowulf, and which certainly suits the known facts of the seventh and early eighth centuries. The bitterest strife in England in the seventh century had been between the two quite new states of Northumbria and Mercia, both equally of Anglian race. Both these states had been built up by a combination of smaller units, and not without violating the old local patriotisms of the diverse elements from which they had been formed. At first, at any rate, no such thing as Northumbrian or Mercian patriotism can have existed. Loyalty was personal, to the king. Neither the kingdom nor the comitatus was homogeneous. We have seen that Bede mentions it as a peculiar honour to a Northumbrian prince that from all parts of England nobles came to enter his service. We must not demand from the seventh or eighth century our ideals of exclusive enthusiasm for the land of one's birth, ideals which make it disreputable for a "mercenary" to sell his sword. The ideal is, on the contrary, loyalty to a prince whose service a warrior voluntarily enters. And the Danish court is depicted as a pattern of such loyalty—before the Scyldings began to work evil^[647], by the treason of Hrothulf.

Further, the fact that the Danish court at Leire had been a heathen one might be matter for regret, but it would not prevent its being praised by an Englishman about 700. For England was then entirely Christian. In the process of conversion no single Christian had, so far as we know, been martyred. There had been no war of religion. If Penda had fought against Oswald, it had been as the king of Mercia against the king of Northumbria. Penda's allies were Christian, and he showed no antipathy to the new faith^[648]. So that at this date there was no reason for men to feel any deep hostility towards a heathendom which had been the faith of their grandfathers, and with which there had never been any embittered conflict.

But in 900 the position was quite different. For more than a generation the country had been engaged in a life-and-death struggle between two warring camps, the "Christian men" and the "heathen men." The "heathen men" were in process of conversion, but were liable to be ever recruited afresh from beyond the sea. It seems highly unlikely that Beowulf could have been written at this date, by some English poet, for the court of a converted Scandinavian prince, with a view perhaps, as Schücking suggests, to educating his children in the English speech. In such a case the one thing likely to be avoided by the English poet, with more than two centuries of Christianity behind him, would surely have been the praise of that Scandinavian heathendom, from which his patron had freed himself, and from which his children were to be weaned. The martyrdom of S. Edmund might have seemed a more appropriate theme^[649]. The tolerant attitude towards heathen customs, and the almost antiquarian interest in them, very justly, as it seems to me, emphasized by Schücking^[650], is surely far more possible in a.d. 700 than in A.D. 900. For between those dates heathendom had ceased to be an antiquarian curiosity, and had become an imminent peril.

If those are right who hold that Beowulf is no purely native growth, but shows influence of the classical epic, then again it is easier to credit such influence about the year 700 than 900. At the earlier date we have scholars like Aldhelm and Bede, both well acquainted with Virgil, yet both interested in vernacular verse. It has been urged, as a reductio ad absurdum of the view which would connect Beowulf with Virgil, that the relation to the Odyssey is more obvious than that to the Æneid. Perhaps, however, some remote and indirect connection even between Beowulf and the Odyssey is not altogether unthinkable, about the year 700. At the end of the seventh century there was a flourishing school of Greek learning in England, under Hadrian and the Greek Archbishop Theodore, both "well read in sacred and in secular literature." In 730 their scholars were still alive, and, Bede tells us, could speak Greek and Latin as correctly as their native tongue. Bede himself knew something about the Iliad and the Odyssey. Not till eight centuries have passed, and we reach Grocyn and Linacre, was it again to be as easy for an Englishman to have a first-hand knowledge of a Greek classic as it was about the year 700. What scholarship had sunk to by the days of Alfred, we know: and we know that all Alfred's patronage did not produce any scholar whom we can think of as in the least degree comparable to Bede.

So that from the point of view of its close touch with heathendom, its tolerance for heathen customs, its Christian magnanimity and gentleness, its conscious art, and its learned tone, all historic and artistic analogy would lead us to place Beowulf in the great age—the age of Bede.

This has brought us to another question—more interesting to many than the mere question of date. Are we to suppose any direct connection between the classical and the Old English epic?

As nations pass through their "Heroic Age," similar social conditions will necessarily be reflected by many similarities in their poetry. In heroic lays like Finnsburg or Hildebrand or the Norse poems, phrases and situations may occur which remind us of phrases and situations in the Iliad, without affording any ground for supposing classical influence direct or indirect.

But there is much more in Beowulf than mere accidental coincidence of phrase or situation.

A simple-minded romancer would have made the Æneid a biography of Æneas from the cradle to the grave. Not so Virgil. The story begins with mention of Carthage. Æneas then comes on the scene. At a banquet he tells to Dido his earlier adventures. Just so Beowulf begins, not with the birth of Beowulf and his boyhood, but with Heorot. Beowulf arrives. At the banquet, in reply to Unferth, he narrates his earlier adventures. The Beowulf-poet is not content merely to tell us that there was minstrelsy at the feast, but like Virgil or Homer, he must give an account of what was sung. The epic style leads often to almost verbal similarities. Jupiter consoling Hercules for the loss of the son of his host says:

In the same spirit and almost in the same words does Beowulf console Hrothgar for the loss of his friend:

On the other hand, though we are often struck by the likeness in spirit and in plan, it must be allowed that there is no tangible or conclusive proof of borrowing^[652]. But the influence may have been none the less effective for being indirect: nor is it quite certain that the author, had he known his Virgil, would necessarily have left traces of direct borrowing. For the deep Christian feeling, which has given to Beowulf its almost prudish propriety and its edifying tone, is manifested by no direct and dogmatic reference to Christian personages or doctrines.

I sympathize with Prof. Chadwick's feeling that a man who knew Virgil would not have disguised his knowledge, and would probably have lacked both inclination and ability to compose such a poem as Beowulf^[653]. But does not this feeling rest largely upon the analogy of other races and ages? Is it borne out by such known facts as we can gather about this period? The reticence of Beowulf with reference to Christianity does not harmonize with one's preconceived ideas; and Bishop Aldhelm gives us an even greater surprise. Let anyone read, or try to read, Aldhelm's Epistola ad Acircium, sive liber de septenario et de metris. Let him then ask himself "Is it possible that this learned pedant can also have been the author of English poems which King Alfred—surely no mean judge—thought best of all he knew?" These poems may of course have been educated and learned in tone. But we have the authority of King Alfred for the fact that Aldhelm used to perform at the cross roads as a common minstrel, and that he could hold his audiences with such success that they resorted to him again and again^[654]. Only after he had made himself popular by several performances did he attempt to weave edifying matter into his verse. And the popular, secular poetry of Aldhelm, his carmen triviale, remained current among the common people for centuries. Nor was Aldhelm's classical knowledge of late growth, something superimposed upon an earlier love of popular poetry, for he had studied under Hadrian as a boy^[655]. Later we are told that King Ine imported two Greek teachers from Athens for the help of Aldhelm and his school^[656]; this may be exaggeration.

Everything seems to show that about 700 an atmosphere existed in England which might easily have led a scholarly Englishman, acquainted with the old lays, to have set to work to compose an epic. Even so venerable a person as Bede, during his last illness, uttered his last teaching not, as we should expect on a priori grounds, in Latin hexameters, but in English metre. The evidence for this is conclusive^[657]. But, at a later date, Alcuin would surely have condemned the minstrelsy of Aldhelm^[658]. Even King Alfred seems to have felt that it needed some apology. It would have rendered Aldhelm liable to severe censure under the Laws of King Edgar^[659]; and Dunstan's biographer indignantly denies the charge brought against his hero of having learnt the heathen songs of his forefathers^[660].

The evidence is not as plentiful as we might wish, but it rather suggests that the chasm between secular poetry and ecclesiastical learning was more easily bridged in the first generations after the conversion than was the case later.

But, however that may be, it assuredly does not give any grounds for abandoning the old view, based largely upon grammatical and metrical considerations, which would make Beowulf a product of the early eighth century, and substituting for it a theory which would make our poem a product of mixed Saxon and Danish society in the early tenth century.

E. THE "JUTE-QUESTION" REOPENED

The view that the Geatas of Beowulf are the Jutes (Iuti, Iutae) of Bede (i.e. the tribe which colonized Kent, the Isle of Wight and Hampshire) has been held by many eminent scholars. It was dealt with only briefly above (pp. 8-9) because I thought the theory was now recognized as being no longer tenable. Lately, however, it has been maintained with conviction and ability by two Danish scholars, Schütte and Kier. It therefore becomes necessary once more to reopen the question, now that the only elaborate discussion of it in the English language favours the "Jute-theory," especially as Axel Olrik gave the support of his great name to the view that "the question is still open^[661]" and that "the last word has not been said concerning the nationality of the Geatas^[662]."

As in most controversies, a number of rather irrelevant side issues have been introduced^[663], so that from mere weariness students are sometimes inclined to leave the problem undecided. Yet the interpretation of the opening chapters of Scandinavian history turns upon it.

Supporters of the "Jute-theory" have seldom approached the subject from the point of view of Old English. Bugge^[664] perhaps did so: but the "Jute-theory" has been held chiefly by students of Scandinavian history, literature or geography, like Fahlbeck^[665], Steenstrup^[666], Gering^[667], Olrik^[668], Schütte^[669] and Kier^[670]. But, now that the laws of Old English sound-change have been clearly defined, it seldom happens that anyone who approaches the subject primarily as a student of the Anglo-Saxon language holds the view that the Geatas are Jutes.

And this is naturally so: for, from the point of view of language, the question is not disputable. The Gēatas phonologically are the Gautar (the modern Götar of Southern Sweden). It is admitted that the words are identical^[671]. And, equally, it is admitted that the word Gēatas cannot be identical with the word Iuti, Iutae, used by Bede as the name of the Jutes who colonized Kent^[671]. Bede's Iuti, Iutae, on the contrary, would correspond to a presumed Old English *Īuti or *Īutan^[672], current in his time in Northumbria. This in later Northumbrian would become Īote, Īotan (though the form Īute, Īutan might also survive). The dialect forms which we should expect (and which we find in the genitive and dative) corresponding to this would be: Mercian, Ēote, Ēotan; Late West-Saxon, Ȳte, Ȳtan (through an intermediate Early West-Saxon *Īete, *Īetan, which is not recorded).

If, then, the word Gēatas came to supplant the correct form Īote, Īotan (or its Mercian and West-Saxon equivalents Ēote, Ēotan, Ȳte, Ȳtan), this can only have been the result of confusion. Such confusion is, on abstract grounds, conceivable: it is always possible that the name of one tribe may come to be attached to another. "Scot" has ceased to mean "Irishman," and has come to mean "North Briton"; and there is no intrinsic impossibility in the word Gēatas having been transferred by Englishmen, from the half-forgotten Gautar, to the Jutes, and having driven out the correct name of the latter, Īote, Īotan. For example, there might have been an exiled Geatic family among the Jutish invaders, which might have become so prominent as to cause the name Gēatas to supplant the correct Īote, Ēote, etc. But, whoever the Geatas may have been, Beowulf is their chief early record: indeed, almost all we know of their earliest history is derived from Beowulf. In Beowulf, therefore, if anywhere, the old names and traditions should be remembered. The word Gēat occurs some 50 times in the poem. The poet obviously wishes to use other synonyms, for the sake of variety and alliteration: hence we get Weder-Gēatas, Wederas, Sǣ-Gēatas, Gūð-Gēatas. Now, if these Geatas are the Jutes, how comes it that the poet never calls them such, never speaks of them under the correct tribal name of Ēote, etc., although this was the current name at the time Beowulf was written, and indeed for centuries later?

For, demonstrably, the form Ēote, etc., was recognized as the name of the Jutes till at least the twelfth century. Then it died out of current speech, and only Bede's Latin Iuti (and the modern "Jute" derived therefrom) remained as terms used by the historians. The evidence is conclusive:

(a) Bede, writing about the time when Beowulf, in its present form, is supposed to have been composed, uses Iuti, Iutae, corresponding to a presumed contemporary Northumbrian *Īuti, *Īutan.

(b) In the O.E. translation of Bede, made in Mercia perhaps two centuries after Bede's time, we do indeed in one place find "Geata," "Geatum" used to translate "Iutarum," "Iutis," instead of the correctly corresponding Mercian form "Eota," "Eotum." Only two MSS are extant at this point. But since both agree, and since they belong to different types, it is probable that "Geata" here is no mere copyist's error, but is due to the translator himself^[673]. But, later, when the translator has to render Bede's "Iutorum," he gives, not "Geata," but the correct Mercian "Eota." There can be no possible doubt here, for five MSS are extant at this point, and all give the correct form—four in the Mercian, "Eota," whilst one gives the West-Saxon equivalent, "Ytena."

Now the Gēata-passage in the Bede translation is the chief piece of evidence which those who would explain the Geatas of Beowulf as "Jutes" can call: and it does not, in fact, much help them. What they have to prove is that the Beowulf-poet could consistently and invariably have used Gēatas in the place of Ēote. To produce an instance in which the two terms are both used by the same translator is very little use, when what has to be proved is that the one term had already, at a much earlier period, entirely ousted the other.

All our other evidence is for the invariable use of the correct form Īote, Īotan, etc. in Old English.

(c) The passage from Bede was again translated, and inserted into a copy of the Anglo-Saxon Chronicle, which was sent quite early to one of the great abbeys of Northumbria^[674]. In this, "Iutis, Iutarum" is represented by the correct Northumbrian equivalent, "Iutum," "Iotum"; "Iutna."

(d) This Northumbrian Chronicle, or a transcript of it, subsequently came South, to Canterbury. There, roughly about the year 1100, it was used to interpolate an Early West-Saxon copy of the Chronicle. Surely at Canterbury, the capital of the old Jutish kingdom, people must have known the correct form of the Jutish name, whether Gēatas or Īote. We find the forms "Iotum," "Iutum"; "Iutna."

(e) Corresponding to this Northumbrian (and Kentish) form Īote, Mercian Ēote, the Late West-Saxon form should be Ȳte. Now MS Corpus Christi College, Cambridge, 41, gives us "the Wessex version of the English Bede" and is written by a scribe who knew the Hampshire district^[675]. In this MS the "Eota" of the Mercian original has been transcribed as "Ytena," "Eotum" as "Ytum," showing that the scribe understood the tribal name and its equivalent correctly. This was about the time of the Norman Conquest, but the name continued to be understood till the early twelfth century at least. For Florence of Worcester records that William Rufus was slain in Noua Foresta quae lingua Anglorum Ytene nuncupatur; and in another place he speaks of the same event as happening in prouincia Jutarum in Noua Foresta^[676], which shows that Florence understood that "Ytene" was Ȳtena land, "the province of the Jutes."

It comes, then, to this. The "Jute-hypothesis" postulates not only that, at the time Beowulf was composed, Gēatas had come to mean "Jutes," but also that it had so completely ousted the correct old name Īuti, Īote, Ēote, Ȳte, that none of the latter terms are ever used in the poem as synonyms for Beowulf's people^[677]. Yet all the evidence shows that Īuti etc. was the recognized name when Bede wrote, and we have evidence at intervals showing that it was so understood till four centuries later. But not only was Īuti, Īote never superseded in O.E. times; there is no real evidence that Gēatas was ever generally used to signify "Jutes." The fact that one translator in one passage (writing probably some two centuries after Beowulf was composed) uses "Geata," "Geatum," where he should have used "Eota," "Eotum," does not prove the misnomer to have been general—especially when the same translator subsequently uses the correct form "Eota."

I do not think sufficient importance has been attached to what seems (to me) the vital argument against the "Jute-theory." It is not merely that Gēatas is the exact phonological equivalent of Gautar (Götar) and cannot be equivalent to Bede's Iuti. This difficulty may be got over by the assumption that somehow the Iuti, or some of them, had adopted the name Gēatas: and we are not in a position to disprove such assumption. But the advocates of the "Jute-theory" have further to assume that, at the date when Beowulf was written, the correct name Iuti (Northumbrian Īote, Mercian Ēote, West-Saxon Ȳte) must have so passed into disuse that it could not be once used as a synonym for Beowulf's people, by our synonym-hunting poet. And this assumption we are in a position to disprove.

The Jute-theory would therefore still be untenable on the ground of the name, even though it were laboriously proved that, from the historical and geographical standpoint, there was more to be said for it than had hitherto been recognized. But even this has not been proved: quite the reverse. As I have tried to show above, historical and geographical considerations, though in themselves not absolutely conclusive, point emphatically to an identification with the Götar, rather than with the Jutes^[678].

The relations of Beowulf and the Geatas with the kings of Denmark and of Sweden are the constant topic of the poem. Now the land of the Götar was situated between Denmark and Sweden. But if the Geatas be Jutes, their neighbours were the Danes on the east and the Angles on the south; farther away, across the Cattegat lay the Götar, and beyond these the Swedes. If the Geatas be Jutes, why should their immediate neighbours, the Angles, never appear in Beowulf as having any dealings with them? And why, above all, should the Götar never be mentioned, whilst the Swedes, far to the north, play so large a part? Even if Swedes and Götar had at this time been under one king, the Götar could not have been thus ignored, seeing that, owing to their position, the brunt of the fighting must have fallen on them^[679]. But we know that the Götar were independent. The strictly contemporary evidence of Procopius shows quite conclusively that they were one of the strongest of the Scandinavian kingdoms^[680]. How then could warfare be carried on for three generations between Jutes and Swedes without concerning the Götar, whose territory lay in between?

Again, in the "Catalogue of Kings" in Widsith, the Swedes are named with their famous king Ongentheow. The Jutes (Ȳte) are also mentioned, with their king. And their king is not Hrethel, Hæthcyn, Hygelac or Heardred, but a certain Gefwulf, whose name does not even alliterate with that of any known king of the Geatas^[681].

Again, in the (certainly very early) Book on Monsters, Hygelac is described as Huiglaucus qui imperavit Getis. Now Getis can mean Götar^[682], but can hardly mean Jutes.

The geographical case against the identification of Geatas and Götar depends upon the assumption that the western sea-coast of the Götar in ancient times must have coincided with that of West Gothland (Vestra-Götland) in mediæval and modern times. Now as this coast consists merely of a small strip south of the river Götaelv, it is argued that the Götar could not be the maritime Geatas of Beowulf, capable of undertaking a Viking raid to the mouth of the Rhine. But the assumption that the frontiers of the Götar about A.D. 500 were the same as they were a thousand years later, is not only improbable on a priori grounds, but, as Schück has shown^[683], can be definitely disproved. Adam of Bremen, writing in the eleventh century, speaks of the river Gothelba (Götaelv) as running through the midst of the peoples of the Götar. And the obvious connection between the name of the river and the name of the people seems to make it certain that Adam is right, and that the original Götar must have dwelt around the river Götaelv. But, if so, then they were a maritime folk: for the river Götaelv is merely the outlet which connects Lake Wener with the sea, running a course almost parallel with the shore and nowhere very distant from it^[684]. But even when Adam wrote, the Götar to the north of the river had long been politically subject to Norway^[685]: and the Heimskringla tells us how this happened.

Harold Fairhair, King of Norway (a contemporary of King Alfred), attacked them: they had staked the river Götaelv against him, but he moored his ships to the stakes^[686] and harried on either shore: he fought far and wide in the country, had many battles on either side of the river, and finally slew the leader of the Götar, Hrani Gauzki (the Götlander). Then he annexed to Norway all the land north of the river and west of Lake Wener. Thenceforward the Götaelv was the boundary between Norway and West Gothland, though the country ultimately became Swedish, as it now is. But it is abundantly clear from the Heimskringla that Harold regarded as hostile all the territory north of the Götaelv, and between Lake Wener and the sea^[687] (the old Ránriki and the modern Bohuslän).

But, if so, then the objection that the Götar are not a sufficiently maritime people becomes untenable. For precisely to this region belong the earliest records of maritime warfare to be found in the north of Europe, possibly the earliest in Europe. The smooth rocks of Bohuslän are covered with incised pictures of the Bronze age: and the favourite subject of these is ships and naval encounters. About 120 different pictures of ships and sea fights are reproduced by one scholar alone^[688]. And at the present day this province of Göteborg and Bohus is the most important centre in Sweden both of fishery and shipping. Indeed, more than one quarter of the total tonnage of the modern Swedish mercantile marine comes from this comparatively tiny strip of coast^[689].

It is surely quite absurd to urge that the men of this coast could not have harried the Frisians in the manner in which Hygelac is represented as doing. And surely it is equally absurd to urge that the people of this coast would not have had to fear a return attack from the Frisians, after the downfall of their own kings. The Frisians seem to have been "the chief channel of communication between the North and West of Europe^[690]" before the rise of the Scandinavian Vikings, and to have been supreme in the North Sea. The Franks were of course a land power, but the Franks, when in alliance with the Frisians, were by no means helpless at sea. Gregory of Tours tells us that they overthrew Hygelac on land, and then in a sea fight annihilated his fleet. Now the poet says that the Geatas may expect war when the Franks and Frisians hear of Beowulf's fall. The objection that, because they feared the Franks, the Geatas must have been reachable by land, depends upon leaving the "and Frisians" out of consideration.

"Now we may look for a time of war" says the messenger "when the fall of our king is known among the Franks and Frisians": then he gives a brief account of the raid upon the land of the Frisians and concludes: "Ever since then has the favour of the Merovingian king been denied us^[691]." What is there in this to indicate whether the raiders came from Jutland, or from the coast of the Götar across the Cattegat, 50 miles further off? The messenger goes on to anticipate hostility from the Swedes^[692]. To this, at any rate, the Götar were more exposed than the Jutes. Further, he concludes by anticipating the utter overthrow of the Geatas^[693]: and the poet expressly tells us that these forebodings were justified^[694]. There must therefore be a reference to some famous national catastrophe. Now the Götar did lose their independence, and were incorporated into the Swedish kingdom. When did the Jutes suffer any similar downfall at the hands of either Frisians, Franks, or Swedes?