Tri-State Tornado?

Michael Auker

I recently read an interesting book about the Tri-State Tornado, written by Wallace Akin. He stated that the tornado formed in south-east MO in the cold sector of the low pressure system, and actually didn't catch up with the low and associated warm sector until it crossed the Mississippi River. Supposedly, the winds had already switched to north-west at St. Louis before the tornado even formed further south-west. How was this possible? Violent tornadoes don't form behind cold fronts, how could the meteorological conditions have supported this? Furthermore, the low that the tornado was associated with under-went explosive deepening as it moved from NW Arkansas up the Ohio Valley; might the process of bombogenesis have been helped by a meso-scale low associated with the tornado? A situation like that occurred with the March 28, 1984 tornado outbreak, which was also closely associated with a rapidly deepening low pressure system. p.s. A study was done on how meso-scale lows influenced the intensification of the March 28, 1984 low. It was on the web a few years ago, but has since disappeared. I would love to read it again if anyone knows where I can find it?
From what I understand, the supercell that produced the Tri-State tornado formed to the north of the warm front. As the surface low deepened and intense warm air advection continued, the warm front rapidly lifted to the north allowing the supercell to become surface based and produce tornadoes.

An unusual thing about this storm was its close proximity to the surface low. From what I've read, the Tri-State supercell was pretty much located right at the triple point and had the advantage of synoptic-scale inflow, so-to-speak. With such a raging surface cyclone, it's really amazing to me that more convection didn't form and force a transition to a linear or quasi-linear storm mode. Any way, this is all in the realm of conjecture, given the dearth of weather observations in the 1920s.