There has been talk, without proof, for decades of "detonation," where a small charge is somehow magically distributed in a case such that most powder is at the front and rear of the case. When the primer goes off, both charges ignited and the two shock waves meet and form a much larger shock wave that blows up the gun. Of course, the low powder charge does not have the chemical energy needed to produce that pressure, but the idea is still around--despite test labs trying to duplicate for decades.
296/H110 are very difficult to ignite and pressure must be high enough. A light load will not produce enough pressure to ignite the whole mass of powder and you can get a semi-burned lump of powder stuck in the barrel (along with a stuck bullet). Thus, a magnum primer is required and enough powder to produce enough pressure fast enough to get full ignition.
What does happen is that at a velocity below 550fps, the bullet can get stuck. The very light charge is too light to maintain pressure and can not produce the pressure needed to push the bullet out of the barrel. Just the force of the primer is enough to push a bullet in a semi-auto pistol any where from half way out the barrel to actually eject the bullet (I used to do this in my 1911s to recover a spent bullet).
Thus, dropping 0.5gn below the starting load is NOT a recipe for disaster, unless it takes the load below 500-550 fps. The 10% reduction from max shows how totally arbitrary the start load is. The decades of very light target loads in .45 Auto and .38 Special for Bullseye competition further disproves all the concern about going below the starting load.
However, if you don't know how to feel a squib load, don't experiment with really reduced loads.
For most purposes, you will find the most accurate load somewhere between start and max and will never need to deviate beyond those limits.
Very sad indeed as these were some of my favorite bullets!! 
