The latest test backbone contains some new ids for names that are still exactly the same as in the last backbone. This happens because it was formerly a synonym and is now regarded as accepted in the new backbone.
I would take this as a bug and rather issue the same, previous taxon key if the name is unchanged, has the same rank and is still within the same kingdom.
Example bird Eolophus roseicapilla (248.000 occurrences linked):
NEW: 2479598; SPECIES; Eolophus roseicapilla (Vieillot, 1817); ACCEPTED
OLD: 2479601; SPECIES; Eolophus roseicapillus (Vieillot, 1817); SYNONYM