Science is science and seat of the pants is very subjective. I am convinced that Peter Ballard has it exactly correct when he says that there should be no force exerted by the upper and lower spring pushing against each other. Contrary to what one would think the two spring rates become cumalative when the springs are acting one against the other. This gives a very high initial spring rate but when the lower spring is freed, either in bump or due to cornering forces, the rate suddenly falls. This is what gives rise to the nodding dog syndrome.
Increasing the rate of the upper spring may simply mask the problem because the the accumulated spring rate will hold for so long that the lower spring seldom become free. This means that if you have an upper spring with 140 pounds per inch and a lower spring of say 200 pounds per inch the actual initial spring rate is 340 pounds per inch. Very rigid ride and a lot of initial roll resistance, but the moment the lower spring becomes free, say with a hard bump mid corner, the car will make a lurch to that side. Sure it can also be controlled with high damper settings but whilst it may be effective on good roads it would hardly be comfortable, and may be a bit disconcerting, on bad roads.
Peter Ballard has opted for more comfort with his spring rates and strongly specifies that the much higher rated rebound spring should be only just kissing the bottom of the sliding hub at static ride height. This strong rebound spring limits the rise of the outside wheel in cornering and reduces body roll. For those that require a more sporting suspension the Ballard designed race kit is available or, alternatively springs from the race kit could be ordered with the standard set up.
Which ever way you go I believe that the springs should be set so that you have a short high rate lower spring that is set, with spacers if necessary, to match the ride height required when it is at full extension, that is to say not under load. The upper spring should be set to support the car at that height.
The SSL (Suplex) kit makes this easy because it has adjusting ring nuts above the spring. It also has the advantage of having an inner rising rate spring that comes into effect in severe bumps and/or really hard cornering.
Having said all that the current range of trad Morgans have much improved springing and the SSL kit, I believe, is actually specified as standard for some markets so it could even be installed from new by the factory.
If you want to read Peter's more complete (and more correct) explanation of this topic it can be found at
https://sites.google.com/site/morganatica/suspension/springs-progressive-rising-rate