Very good information Saul.
The diodes are not likely to be VR2 and VR3, on the right side of the A17 schematic, based on their connection to -15 Volts. Do the other two boards look identical to this defective board, in this area? The fact that HP revised this differential amplifier and added these diodes, together with +/-14.2 Volts on the "+/-12" Volt rails adds confirmation that these are indeed +/- 15 Volt rails.
The diode test on Q17 certainly is conclusive. The diode test on this unknown diode may be influenced by the shorted Q17. Can you see if this diode connects in some way to Q17, perhaps through a low value resistor? It would be good to unsolder one leg of the suspect diode and try the diode test again.
I don't see the mystery Q13 (on the A3 parts layout page) on the A3 schematic, either. So far, this manual has not been very helpful. More searching has not turned up different scans of the manual.
Just now saw your new post - it would appear that someone has "repaired" this board. I "think" some people do this just to help us sharpen out troubleshooting skills.
For a quick fix, a 2N3906 should work. The bandwidth of the amplifier may suffer, but at least we will know more.
RF+ Tech