3 MSS: A (31r) Marcabruns, C (258v-259r) Bemat de Uenzac (C Reg. Bemat de uenzac), a1 (294-295) Marchabras with Gothic ‘M’

Analysis of the manuscripts

Aa1’s versification is irregular and C appears to have intervened to correct it (see ‘Versification’). Archetype errors appear to lie behind the readings in 24, 29 and 32; in 24 and 29 (loosely) Aal appear joined in error with C again innovating. A has isolated errors in 10, 13, 14, 24, 29, 30, 35, 38, 50 and 52 and an isolated reading against the agreement of Ca1 in 12; in 42 it seems best. C has isolated errors in 4, 16, 17, 32, 36 (inflexion only), and isolated readings, sometimes apparently smoothing out difficulties, against the agreement of Aax in 3, 9, 18, 20, 22, and so on; 17 and 39 are probably interventions, a1 has numerous isolated errors in 1, 7, 10, 12, 17 and so on, and is also isolated against the agreement of AC in 7; it has the only correct reading in 8 and 14, and is best in 24 (t). A is the least bad source, C having undergone extensive intervention and a% being too error-prone. While adopting A as base we reject some of its isolated readings (10, 13, 14) as suspect.

Versification

Frank, Repértoire, 737.1: a7’ b7’ b7’ c7 d7’ c7 d7’. Seven coblas unissonans and one tornada of 3 lines. Irregular rhymes occur in A 1, 10, 13, 22, 29, 38, 39; C32; ax 1, 19, 29, 39. (a1 9 and C 37 appear to be graphies.) A 13 and 38, C 32 and a1 19 are certainly or almost certainly errors. Given, that Aa1 agree on all the others except 10, where C supports a1 A 10 is probably an error. In 17, C appears to have innovated to correct the rhyme, but at the expense of phonology. This encourages the suspicion that it may have intervened in some or all of the remaining cases, most obviously in 39 where it alone repeats the rhyme-word of 18 (with the same sense). This leaves the question of the Aa1 rhymes in -isa in 1, 22 and 29.