decided later that the first letter was i not ƒ and came to the conclusion that the dotted cross was not a rune but a sign of punctuation, cf. Die Runenschrift, p. 166 ft. Bugge reads: Ana hahaisla inir frawaradar='Over Hahaisl (satte) Ini (og) Frawarad (Stenen),' cf. Norges Indskrifter med de ældre Runer, p. 233. For objections to these readings see Burg, Die älteren Nordischen Runeninschriften, p. 106 etc.

There are only two of the runes about which there can be any question. The picture of 1750 shows as the first letter of the lower line. In 1862 Säve found only | but reported the stone as more or less injured at the left, it having been long used as a step at the well of the Hageby rectory. Most scholars now read f, which is almost certainly correct. The dotted cross is very interesting. In Philologische Studien (Festgabe für Eduard Sievers, p. 18) I showed that the various forms of the rune for ng are only different ligatures of the runes for n and g. I can now be more definite. The combination of these two consonants was one of the most common ones in Germanic. Moreover, the nasal was really and occurred only in this combination. Hence it is not strange that its written form drifted away from that of n, and thus an individual spelling for the group arose. The runes for n and g were but slightly different crosses and easily became alike, just as vy became yy in Greek. They were thus regarded as a sort of double letter for the group ng and developed differently according as they were written next to one another, above one another, or over one another :

That the vertical line should on stone be displaced by two dots was natural; for the vertical line is apt to break up the stone where it crosses the two oblique lines.

Wimmer was right in beginning with the lower line and reading up as well as to the left. The position of the rune for the final above the upper line is sufficient to show that the lowest line is the first of the inscription. This early retrograde legend is one of the many blocks in the way of Wimmer's derivation of the runes from the Latin alphabet. I read as follows in normal order :

FREPEREAFY FXFHFHFIS NEXITY
Frawara a R ana ha hais langiniR

The guard

watched

long.

In all probability, the guard was himself the artist and cut. his picture in the rock to while away the time that hung heavy on his hands. We think at once of the mounted guard that met Beowulf and his men as they landed on the Danish coast (Beowulf, 229, etc.). The importance of the inscription will appear from a consideration of the forms of the words, especially the last two:

frāwaraðar is a compound of fră (older fram) 'fore,' and waraðar ‘guard,' compare OE. foreweard 'outpost,' 'Vorposten,' 'advance guard.' The idea (1) 'in front,' 'forward,' on ahead' (also 'advanced,' 'very') is prone to change into that of (2) 'away,' 'off' (and later loss,' 'destruction,' 'wrong'); and a comparative or superlative or some other form of the word or an entirely different word assumes the earlier meaning. In Greek we have pó with the earlier meaning intact; in Germanic, fra- (OE. for-, Ger. ver-) has gone over to the second meaning. In Germanic the old meaning was taken up (1) by fora (OE. and Eng. fore, Ger. vor; cf. Skt. purá and purás, and Greek Tápos, which was used of time, not place) and (2) by the superlative fram (cf, Greek Tрópos 'foremost'). Compare the many Old-English compounds having for- (=fra-) in the second meaning with the later ones having fore- in the earlier meaning. With Skt. prá 'forward' and prá-svādas 'very pleasing,' and OE. for-heard 'very hard,' compare OE. fram 'forward,' 'bold,'

and Gothic fram farther,' fram-aldrs 'advanced in age,' very old.' But already in Germanic, fram itself changed from the first to the second meaning, and the new comparative Gothic framis, ON. fram(m) (<fram®, Sievers, PB. 15, p. 405 N, Noreen, AI. Gr.3, § 217, A,) assumed the meaning 'farther on,' forward.' In Pr. ON. fräwaraðar 'outpost,' 'advance guard,' 'Vorposten,' we have just such a compound of the earlier fra(m) as the later Old Norse shows of fram(m) in fram-bryggja ‘forward gangway,' fram-kirkja ‘fore-church,' 'nave,' fram-allari, fram-tønn, fram-bogr, &c. Compare the older compound fra-saga 'Vortrag,' 'story,' with the later fram-saga 'pleading,' 'delivery' (in court); also the earlier OE. for- (=fra) in for-heard 'very hard,' for-strang 'very strong,' but the later fore- in fore-halig 'very holy.'-The second part of the compound fräwaradar is common in Germanic, especially in compounds, compare Gothic -wards, ON. -varðr in names (the independent word is in Old Norse an u-stem: vordr), OHG. -wart, OE. weard and -weard.

anahahais is the reduplicated preterit of the verb anahaisan 'keep watch.' ana- has the value of German auf- in aufpassen 'keep watch;' compare Gothic analagjan 'auflegen,' anatimrjan 'aufbauen,' etc. haisan has the stem Gc. hais, IE. kois, seen in Paelignian coisatens 'curaverunt,' OldLatin coiro, Latin curo, Umbrian kuraia 'curet,' all meaning 'take care of,' 'watch,' Brugmann, r2, § 874. (I shall treat this stem and its other representatives on another occasion.) No such reduplicated preterits have thus far been found in Primitive Old Norse (Streitberg, Urgerm. Gr., p. 327, 2), but I have recently discovered several, including two cases of hahait Gothic hathait. The a of ha- is evidently the sign for the obscure vowel, represented by ai in Gothic. Compare also the second (epenthetic) a of waraðar.

=

langinir the Old Norse adverbial genitive lengi 'long,' 'for a long time,' really the genitive of the in- abstract, Gothic laggeins (Urgerm. Gr., p. 255, 259), with the ending -in intact.

GEORGE HEMPL.

IX.-GERMANIC ELEMENTS IN THE STORY OF KING HORN.

To the mass of romances current during the Middle English period of our literature, the contribution of purely Germanic tradition was a relatively meagre one. The spirit which had produced the earlier epic was at this time extinct. A solitary offshoot of the earlier epic seems to have survived in the story of the dragon-killing Wade with his famous boat, Guingelot. But even this story is lost to us save in occasional references, and from these we must infer that all definite idea of its origin was lost, since it is associated, now with Weyland, now with Horn and Havelok, now with Launcelot. To these earlier tales, such as those of Beowulf and possibly of Wade, having a popular, epic origin, succeeded in the Middle English period a mass of tales and romances of the most diverse origin imaginable. Even in the popular romances of Guy of Warwick and Bevis of Hampton, which are supposed to contain a kernel of genuine English tradition, the original story is almost lost amid the mass of mythical, imaginary, or purely conventional matter later added. The historical events in the lives of Waldef and Hereward are embellished with much of the conventional romantic matter, and the late romance of Richard Coeur de Lion consists very largely of the purely conventional.

The stories of Havelok and Horn, which are supposed to have been among the first products of this second growth of story, seem to preserve more than the other later romances, their primitive traits and are hence usually classed as English, or Germanic, in origin. I have undertaken in this paper to distinguish if possible some of the peculiarly Germanic elements in the story of King Horn.

The story of Horn, it is generally believed, had its origin in the turbulent times of the Danish invasions, though the