Anima mea liquefacta est is the only motet among the group of motetti missales that is transmitted in a large number of sources, namely eight manuscripts of different provenance. Six of the eight sources transmit the music in its entirety. CH-SGs 463 includes only the Cantus and Altus voices; in the source in Washington (US-Wc M2.1.M6) Cantus und Tenor are complete, whereas Altus und Bassus were copied only until m. 46. The Washington manuscript as well as both Florentine sources do not include the text (just an incipit in I-Fn XIX.178 and US-Wc M2.1.M6; no incipit in I-Fn BR 229). Librone 1 is the earliest source of the motet and it preserves a good version of the work. In two cases, however, a semibrevis is notated instead of a brevis in the Altus (at mm. 58 and 63), an inaccuracy shared by the Silesia source (and the first also by I-Fn XIX.178). However, a thin barline traced after the semibrevis in I-Mfd 1, in both cases with fermata, makes clear that the musical phrase ends at that point (on ‘Filiae Ierusalem’). At mm. 58–59 the final note of the Cantus and Tenor cadence is split into two semibreves, the second one with fermata. It seems that this notation indicated a delayed pronunciation of the last syllable ‘-lem’ in these two voices (as maintained in the edition), probably intending to create a special sound effect (similar to m. 32, although there Altus and Bassus sing two different notes). It is not entirely clear, however, if this might have occurred instead as the result of a flexible notational practice or of inaccurate copying, even if other similar passages in the motet in the Milanese version seem to point to a specific intention. The later version B-Br 228 (as well as CH-SGs 463 and US-Wc M2.1.M6) shows a ‘regular’ ending with breves in all voices. Since the surviving manuscripts display different solutions with semibreves or breves in one or more voices at that very point (mm. 58–59), it cannot be excluded that the reading of B-Br 228 was possibly the reading of the ancestor for one or more of the other sources.

Considering the transmission on the whole, rhythmic substitutions, variants, or errors occurring in all the sources suggest that direct copying of one from the other is quite unlikely. The Florentine sources share a good deal of small variants (mostly rhythmic substitutions), suggesting evident proximity, which is, however, attenuated by some diverging readings, such as, for instance, the addition of a semibrevis in all relevant voices at m. 64 in I-Fn BR 229. I-Fn XIX.178 is closer to I-Mfd 1 than I-Fn BR 229 (displaying for instance, the fermata signs at the end of ‘Ierusalem’ as in I-Mfd 1), although the accidentals accurately notated in the Milanese sources are mostly absent here; moreover, in I-Fn XIX.178 a few unique errors occur.

Relatively close to the Milanese transmission is E-Bbc M 454, the Barcelona manuscript, which contains some rhythmic substitutions, several ligatures not notated in I-Mfd 1, and almost no additional accidentals. This manuscript is unique in the addition of ‘Amen’ at the end of the motet, set to additional music, which prolongs the piece for a few measures. This addition seems to point to a later origin of the Barcelona version compared to the one transmitted in I-Mfd 1, even if it cannot be established in which context this variant originated and to whom it is to be ascribed. The Silesian manuscript shows some differences in the text underlay, some errors, and small variants compared to the Milanese transmission, with which it shares the reading of the Altus at mm. 58–59. The incomplete Washington manuscript has several rhythmic substitutions and small variants that set it apart from the other sources.

The version handed down in B-Br 228, which stands out for some major variants, requires specific discussion. At m. 32 the delayed singing of the syllable ‘-hi’ of ‘mihi’ in two voices is avoided; no fermata signs are notated. At mm. 46–47 the passage of Altus and Bassus is rewritten so that the ending on ‘me’ coincides with the new measure, namely on a brevis with fermata; breves in Cantus and Tenor provided with fermatas are to be sung as longas in order to preserve a correct polyphonic structure. The next major variant affects the passage at mm. 54–59. The brevis in Cantus and Tenor at m. 54 causes a shift of a semibrevis in the melodic lines on ‘Filiae Ierusalem’, requiring adjustments in the counterpoint and notably a new passage in the Altus line. As a result of this re-elaboration the phrase ends in all voices in a full cadence with breves as final notes. The delayed singing of Altus and Bassus on the final syllable ‘-lem’ is then eliminated. The last remarkable variant involves m. 63, where a longer fermata on the cadence on G is avoided. Fermatas are in fact absent and the Altus melody is slightly modified in order to create a connecting line with the final phrase (more appropriate from a semantic point of view and also leading to a final chord in a new measure). In the version of B-Br 228, therefore, some articulation points of the motet are ‘regularized’, always discarding the delayed singing of the final syllable by two of the voices. The late incomplete source CH-SGs 463 shares these main variants with B-Br 228 and clearly belongs to the same line of transmission.

Considering the dating of B-Br 228, it can be assumed that these changes happened at a later point compared to the Milanese transmission. In the light of the Flemish provenance of the manuscript, they might represent variants going back to the composer. However, there is no evidence confirming this hypothesis and the version could be the result of external intervention as well.

The edition is based on I-Mfd 1, which, according to current knowledge, is the earliest source of the motet and the only one to transmit it within a motet cycle. On I-Mfd 1 as source of repertory from the Sforza court, to which Weerbeke belonged, see the Introduction to the present edition.