I had a 175w /DX MV lamp made by GE. The starting probe in it didn't function, it just glowed dim pink. Due to this, I ran it up on a 250w PSMH ballast, and it worked. It got really bright. I then did a restrike. There was arcing going on in the outer bulb, and it was really pretty. Then, suddenly, there was an intense incandescent flash. I looked in and see one of the arctube leads had melted or vaporized and was glowing white-hot . It was pretty cool, but also scary.
What causes this?
Photo:
http://www.lighting-gallery.net/gallery/displayimage.php?album=lastup&cat=0&pos=0&pid=111556

The outer envelope of mercury lamps is not under vacuum like metal halide, HPS and other HV ignitor-started lamps. Its usually filled with a gas, and therefore an arc can establish in the outer before the burner is cool enough to restrike. Since all mercury lamps are supposed to be operated probe start (started by mains voltage), this is never a problem. If you want to run mercury lamps off that M137 ballast in the future, disconnect the ignitor. The only exception would be those pulse-start EYE "moon pulse" mercury lamps which are basically just a mercury arc tube in a PSMH vacuum envelope.   

The pulse start lamps need HV pulses (in the order of few kV) to ignite the arc in the burners, so the ballast generate these until it senses the arc becoming present (well, it senses the voltage - arc present => below 150V, arc not yet present, above 200V with the pulse MH). This means all the equipment is designed to handle that high voltage, including the lamp internals.

The MV's are designed to have the arc ignited by the auxiliary probe on just the 230..300VAC ballast OCV, so their design is made to handle just that. That means the high voltage from the ignitor may be above what the structures in the outer are capable to handle, so actually ignite the arc there.

With these, the arc always ignites oon the place with the lowest breakdown voltage.
With the first time turn ON, it was inside of the main arctube, so in fact it struck the arc where the arc is supposed to be, so everything worked well.
When you have tried to restart the hot lamp, the hjigh pressure inside of the hot arc tube made the breakdown voltage there high, so the next place becomes somewhere between the support wires in the outer and the ballast struck the arc there. And because these structures are not designed to carry an arc, they have failed (the arc root heated them up, so the incandescence, till they have melted)...

The pulse start lamps are designed with their internal structures so, the breakdown would need way higher voltage than the few kV coming from the ballast, just because the attempt to hot start the lamp is a thing to expect during the normal life, but the MV is not supposed to be exposed to such high voltages.

So you have made a good practical experience to really not connect any probe start lamp on any pulse start ballast with the ignitor still connected.

And side note: The 250W ballast will severely overdrive the 175W lamp (if I remember well, 250W pulse start is rated for 3A, 175W MV for just 1.8A)

I knew that it would overdrive it. Because of that, I didn't run it for too long.
@Medved, thatnk you for your detailed explanation.