This is not the bias spreader.
The track which goes to the output transistor goes to its collector and also to the collector of the driver, the two clearly arranged as a Darlington pair, so this track has to be a supply rail.
The other end of the diode goes through R405 (left channel) to the collector of one of the input transistors and presumably the base of the gain stage. I will have a look what's going there, it's probably some kind of clamping, protection, anti-saturation or whatnot.
edit
Looking at the markings, these green transistors appear to be PNP (NPN is in the center of the board), so it has to be the negative rail. The diode is mounted in forward direction from the VAS (TR405) input to the rail, in parallel with the VAS and its 2.7Ω emitter resistor R413. What's going on?
edit
Maybe they wanted to make the VAS a "precision" current mirror, with current gain exactly defined by the ratio of R413 to R405, so D401 ought to be a single diode. Too bad that the ratio of those resistors appear to be 1800
D403, the similar diode used in the bias generator, doesn't generate the whole bias alone. It's in series with TR407, presumably a VBE mulitplier. Quite possibly a single diode too.