I also should add, they way I've seen this come up before is Williams and others used similar 1.5V marker lamp "regulation" using bridge rectifiers (4 diodes) shorted + and - with the 1.5V bulbs across the ~ AC legs of the rectifier.
The problem came when someone put LED bulbs into the main original incandescent sockets that form the main load. Back at no load (minimal current LED bulbs) results in less voltage drop across the diodes and no incandescent marker lights.
Again, this is like a peak regulator. The voltage output on the markers can rise up to a "limit" and that rise only happens when there is sufficient main load current through the diode to push the diode into the "knee" of the voltage drop curve.
