caused by those verses, in which one must put a beat on a prefix, or give two beats to a disyllabic word, with a short root-syllable, at the end of the verse, in order to scan with four beats. Verses such as: land gesawon, him pa Scyld gewāt, fyrst ford gewāt, sã-bāt gesæt, gegrētte pā, gehroden golde, in gear-dagum, on bearm scipes, leof land-fruma, swutol sang scopes, gud-rinc monig. Now it is certainly not an accident that except in the case of the fairly rare types gegrētte pā and gehroden golde - a long and strongly stressed syllable, which generally alliterates, always precedes a prefix which has a beat or a disyllabic word, with a short root syllable (standing at the end of a verse), which has two beats; whilst in those cases, in which a prefix counts as a thesis and two short syllables at the end of a verse have only one beat, this prefix or these two syllables are preceded by a more weakly stressed or unstressed inflexional or derivative syllable. Accordingly I was able to establish the following rules in my Studien zum germanischen Alliterationsvers I, Kritik der bisherigen Theorien, Berlin 1894, § 27. 56. § 51. Law stating when Prefixes may have a Beat. 1. A monosyllabic prefix (ge-, be-, a-, on-, for etc.) may have a beat when it directly follows a long, strongly-stressed syllable (independent word or second part of a compound), e.g. land gèsawon, him pa Scýld gèwat, þæt fram hám gèfrægn, fyrst fórð gèwat să-bát gèsæt, word-hórd onleac etc., but it is counted as a thesis when it follows a more weakly stressed inflexional or derivate syllable or a weakly stressed monosyllable, e.g. folcùm gefrage, he pas frōfrè gebād, atol ÿdà geswing, ædelingà gedriht, worold-arè forgeaf sōd is gecided, ic pat gehÿre etc. - NOTE. In the first case Sievers also counts the prefix by itself as a member of the verse, in the latter case only together with the preceding inflexional syllable or more weakly stressed word. At the beginning of the verse in types B and C a prefix directly before the haupthebung can also have a weak beat, e.g. gègrétte pā, àlédon pā gèmýne mārđo, gèhróden golde etc. Before the first beat of the types A and D1 it is only an anacrusis: gewát him pa to warode, a rás pā së rica, ge sagd sōdlice etc. § 52. Law stating when a Disyllabic Word with a short root-syllable at the end of a verse may have two Beats. 2. At the end of a verse a word, consisting of two short syllables, has two beats, when it is immediately preceded by a long strongly stressed syllable (independent word or second part of a compound), so that a foot of three members with descending stress results, e.g. in geár-dàgùm, on béor-sèlè, on béarm scìpès, lēof lánd-frùmà, māre méarc-stapà gud-rinc monig, mago-dríht micèl, hring-nét bèràn etc. It is counted as only one beat, however, i.e. 'it is slurred', when it follows a more weakly stressed inflexional or derivative syllable or prefix, e.g. wid Grendles grýre, sõhte holdne wine, heal-pegnes héte, scadu-helma gescéapu, wīg-spēda gewíofu etc. Also it is to be looked on as containing one beat at the beginning of or within the verse, e.g. flóta stille bad, fæder alwalda, folc-stède frætwan etc. Only in a few rare verses of type E, such as béag-hròdèn cwēn, Nórđ-Dènùm stōd can two short syllables within the verse in the second half-line be counted as two beats. Sievers has recognized the same difference in the value of two short syllables at the end of a verse according as they follow a strongly stressed or a more weakly stressed syllable, when in the 'shortened' types C and D1 and in the A2k verses he counts the two short syllables as two members of the verse, e.g. in geār-dagum, on bearm scipes, leof land-fruma, gud-rinc monig, but at the end of B and E verses only as one member, as a 'resolved arsis': purh Grendles gryre, heal-pegnes hete. On a closer examination we generally find that in older poetry, e.g. in Beowulf, where a word of this kind at the end of the verse must contain two beats or members of the verse, the second syllable is one which was stressed at a former period or was a long inflexional or derivative syllable, which perhaps still preserved something of its old characteristic sound, e.g. scipes, sele d. sg., dagum, fruma, cuman, cumen, stigon, micel, monig, fæder, whilst where there is 'resolved stress' of the hebung at the end of types B, D2, E and 'resolved stress' of the nebenhebung in the first foot of the enlarged type A2 there may be only two quite short syllables, e.g. nom. or acc. of i- and u-stems: gryre, hete, sele, stede, wracu, searo, fela or pronouns as hine, pone, (ge)hwone; cf. Kaluza, Zur Betonungs- und Verslehre des Altenglischen. Festschrift für O. Schade (1896), pp. 120-131. We must also notice that the two short syllables at the end of C and D1 verses are not haupthebung and nebenhebung, but only a stronger and weaker nebenhebung of a foot of three members: in gear-dagùm, òn béarm scìpès, léof land-frùmà, gúð-rínc mồnìg. § 53. The Application of these two Laws. The correctness of these two rules with regard to the different value of prefixes and of two short syllables at the end of a verse according to the quality of the preceding syllable, is proved by the simple fact that, by means of these laws and those other laws relating to stress, which hold also for the later Germanic rimed verse of Otfrid and Lazamon, according to which, for example, disyllabic words with a long root-syllable or trisyllabic words with a short root-syllable form two beats. of a verse, and trisyllabic words with a long rootsyllable and middle syllable form three beats of a verse, we can read all the OE. alliterative verses handed down to us with four beats. Verses, which according to the above rules would have five beats, e.g. þæt fram hám gèfrúnòn, sæ-bát gèsætòn, wórdhórd inlicàn mórgèn-swég micèl etc. do not occur in OE. apart from text-corruptions, nor such, which according to these rules would have only three or even two beats, e.g. frófrè gebád, gódè forgéald, árè forgéaf - mícel mágo-drìht, mónig médu-bènc, fáran wóldè, hélmàs béran, hýràn scúlon fáran sculon etc. But "what other test have we of the worth or worthlessness of a scientific hypothesis than the greater or less degree to which it may be applied, to which it regularly explains individual facts?" (Sievers, Altg. Metr., p. VIII.) § 54. Compatibility of the Four-Beat Theory with Sievers' Types. The assumption of four beats for the old alliterative verse is, as Martin has shown (ZfdPh. 22, 468. 27, 120 f.), not at all incompatible with Sievers' types. We have only to substitute 'beat' for 'member' ('Glied'), and thus to mark Sievers' 'senkungen' as weaker beats. Then, by combining stronger and weaker beats to form feet with two or three beats, we get Sievers' same types in my modified grouping (§ 40): I. (2+2) Type A: 2×2× lángè | hwílè. II. (1+2+1(a) Type B: 11 þùrh | mínè | hánd. b) Type D2: 1צ blæd|wídè|spráng. |