Wayne, how about a few words about the essential DIFFERENCES in how a sand amp's output voltage (does NOT change with load variations) might be expected to INTERACT with the reactive xover components.

Is a tube amp LESS ABLE to deliver extra power into the reactive part of the load, perhaps ENHANCING the electrical effects of the Zobel for instance, in some manner depriving the drivers themselves of power at some freqs, particularly the xover region?
PWK touched on this in one of his patents, where he implied that his xover design EXPLOITED sand amp characteristics specifically. Sam

this is a "voltage source vs. current source" situation I think...wouldn't Spice simulations give different response plots for the overall freq. response of xover filters if one (Vs), rather than the other (Is) was specified?