Although Liszt played the piece in Berlin in 1843, he left the manuscript in a slightly unfinished state and certainly did not commence the meticulous preparation of it for his printer (he usually marked up important printing details in coloured crayon, and dynamics, tempo changes, pedal marks and other performance directions in coloured ink). The manuscript shows his method of composition of such a fantasy quite clearly: the broad outlines, such as the new themes or sections, are commenced on new pages, and space is left to concoct transition material. The introduction, the two pieces from Figaro, the Don Giovanni minuet and the coda are each in self-contained parts of the manuscript. On several occasions Liszt has marked some repeated figures for possible excision by placing the fragments in brackets with a question mark over them. These tiny tightenings of the structure add up to just a few bars, and they are observed here. The penultimate section breaks off in the middle of some transitional material which does not join happily to the coda. For this performance the sequence has been extended by one bar and two bars derived from earlier material have been added to make the join. The very end of the work is also missing (Liszt may well have improvised both of these sections in performance), but it is clear that the end is fast approaching and that it conforms closely to the closing bars of Act I of Figaro. So the completion of the coda required the addition of a further fifteen bars, making the whole work of 647 bars. It can be seen, therefore, that very little was required to make performable this grand piece, and it remains a mystery quite why Busoni suppressed so much of the work in his edition.
from notes by Leslie Howard © 1994