6 MSS: A (32r) Marcabruns, E (151-152) Marcabru, I (119v) Marcabrus, K(105v) Marcabrus, a1 (305-306) Marcabrus, d (306r-v) Marcabrus.

Analysis of the manuscripts

The MS tradition is uniform. All the MSS have irregularities at the rhyme in 5 -ida for -ina, and 17 -anas rather than -ana, where -ina would in fact be expected. This failure to alternate the ‘cd’ rhymes regularly, coupled with syntactical difficulties and problems of interpretation in 15–13 (see the note), suggest that some six lines of the poem were lost at an early stage of the written tradition (compare XIV). The imperfections at the rhyme in lines 5 and 17 may also support this, although early troubadour poetry and Marcabru’s practice in particular may not have aimed at total regularity of rhyme (see our Introduction), Rather than assuming the loss of one complete stanza between II and III, or the misplacing of III, which Pillet (‘Zum Texte’, p. 17) proposed should follow VI, it seems more likely that the first four lines of stanza III were joined to the last two of the fo lowing stanza, the intervening six lines being lost altogether (see Lewent, ‘Beiträge’, p. 440), This suggestion of a faulty archetype is supported by the readings of line 13 (see note).

Ea1 are united in error in 26 (senhor), 33 (d’ missing) and 35 (mas); together with their other common variants (3, 9, 13, 20 (see note), 29, 30), this suggests the division between Ea1 and AIK, However, the agreement of AIKa1 in 22, 32 and 38 (see the notes) may suggest that the source of a1 had been ‘contaminated’ by an antecedent of AIK and E then represents a separate (if imperfect) strand of the tradition. Despite these isolated readings (see also 7, 8 and 26), E’s many isolated errors (2 (repeated rhyme, compare 22), 17, 27, 28, 31, 37) make it unsuitable to be used as base here. The isolated errors in a1 (13, 18, 23, 26, 35, 36) are not compensated for by any individual and clearly ‘right’ reading.