their relation to the normal verses, the way in which they arose, and particularly with regard to the number of beats to be assigned to them, are various. Some scholars deny that Schwellverse are essentially different from normal verses, and look on them as normal verses with extra 'filling' (Kauffmann, Kögel, Möller-Heusler, Fuhr, Franck). Others add a piece to the end of a normal verse to obtain a Schwellvers (Vetter). Others add at the beginning of a verse (Sievers, Foster, Cremer) or in the caesura, i.e. at the end of the first half-line and at the beginning of the second (Luick, Heath), or they slide two normal verses into one another (Luick, and later Sievers and Schipper). In the same way, too, there is no uniformity of opinion with regard to the number of beats to be assigned to Schwellverse. Kauffmann assumes two, others assume three (Vetter, Schipper, Sievers, Luick), four (Kögel, Franck, Heusler), five (Hirt), six (Schubert, Trautmann, Heath) and even eight (Heath for some Schwellverse). Compare Engl. Stud. 21, 337 ff.

§ 88. The Schwellverse in Beowulf.

In order to arrive at a sure conclusion amid this conflict of opinions we must here, too, proceed from an analysis of the Schwellverse which we have. I have, therefore, arranged according to their 'filling' and printed all the OE. Schwellverse (Engl. Stud. 21, 356-375). Here, as in Stud. z. g. A. 2, 83 follow the Schwellverse of Beowulf (1163-68,

1705-7, 2995 f.) arranged according to types, recognized by their endings:

Type 5* þa gyt was hiera | sib ætgædere ne porfte him þa | lean od-wītan

1164

2994

From the above we see that Sievers was right when he said (Altgerm. Metrik § 94): "Nearly every Schwellvers has at the end a piece, equivalent to a normal verse". These pieces at the end of the verse are, in fact, in all respects like those subforms of the normal verses, characterized above

(§ 62 ff.). A-verses, especially of the simplest form 1, are most frequent. But the portion preceding a normal verse is not, as Sievers and others assume, a 'foot' or 'bar', which can be exactly defined. As we see from the examples above the number of syllables ranges from two to five, and there is great variety in the grammatical structure and stress. We must, therefore, give up any attempt to assign to the initial portion of Schwellverse a definite number of beats. We must look on it simply as 'extended anacrusis', which for the real verse and the number of its beats is of no further importance, even though in the first half-line this 'anacrusis' shares in the alliteration in order to knit the verse more closely. For a fuller discussion see Engl. Stud. 23, 378 ff.

In looking on the Schwellverse as normal verses with extended anacrusis no fundamental difference between Schwellverse and normal verses, as assumed by Sievers, Luick, Schubert, Trautmann etc., is recognized. Even in normal verses monosyllabic anacrusis is admissible. Therefore, as we see from the texts, a transition from normal verses to Schwellverse and vice versa is possible at any moment, even within the line, without any disturbance of the rhythm. "In order to express a mood of solemn exaltation the poet interrupts the calm flow of the short alliterative verses and swells the music of his verse by introducing a fuller anacrusis." (Engl. Stud. 23, 381 f.)

*

§ 90. Recitation of the Schwellverse. With regard to the effect produced by their recitation the Schwellverse may be compared with the Latin Psalms. For a fuller discussion see my articles in Stud. z. g. All. 2, 84 and Engl. Stud. 23, 382. The verses of the Psalms are of very various lengths (e.g. Miserere mei Deus secundum magnam misericordiam tuam | Vigilavi* et factus sum sicut passer solitarius in tecto etc.), and yet, whether short or long, they are all recited similarly, for "the melodious formulae, called mediatio and finalis, fall only on the conclusions of the half-lines", whilst in the preceding part of the verse "the voice remains in the same tone, viz. in the dominant of the tone" (v. Liliencron in Paul's Grundriss II, 2, 307).

By alliteration (which adorns and connects the verses) we mean that the words with the strongest stress in each pair of half-lines have the same initial sound. Grimm gives the following verdict on alliteration (Schipper, Engl. Metr. I, 37): "Alliteration was in itself more delicate and nobler, because it assumed a finer sensitiveness of the ear, attached itself to the metrical movement owing to its connection with the beat, and caused less irritation because of the greater freedom, with which it could be used. And here was the cause of its disuse: people felt the need of a more

effective similarity of sound, which should at the same time arouse the attention more by its unalterable position at the end of the line."

$92. Simple Alliteration.

In the second half-line only one syllable can alliterate, and the alliteration must take place as far as possible at the beginning of the verse. Thus in A, D, E, D1 verses, which begin with words capable of alliterating, the alliteration is on the first member (ellen fremedon, blad wide sprang, weord-myndum pah, wine Scyldinga). In B and C verses, which begin with weakly stressed words, the alliteration is on the second member (him on bearme læg, on bearm scipes).

In the first half-line, too, simple alliteration is sufficient. Here also it is generally at the beginning of the verse (lange hwile, heal-arna mæst, cwēn Hrōdgāres him pa Scyld gewāt, ofer hronrāde). Only in the so-called A3 verses (§ 68), which begin with more weakly stressed words, the alliteration is on the third member (hi hine pā œtbaron, pa was on burgum). Occasionally also in C3 verses (was min fæder 262, gesloh pin fæder 459).

§ 93. Double Alliteration.

In the first half-line, however, two sounds may alliterate, viz. the initial sounds of the first two most strongly stressed words. In type A the first