-what's the diff between the two stepped outputs, one seems to have a much bigger effect than the other?
you're speaking of the differential Sample and Hold circuit which is applied to each pair of barres.  It actually does Sample and Hold twice, for two barres, as follows:
-when A is greater than B, trigger a step sample of B
-when B is greater than A, trigger a step sample of A
Thus, making these work is all about your technique of approaching the pairs of barres.  If you find that inward, palm-style pressure changes one of the stepped outputs, then you will find that a different kind of pressure, called "outward pressure" will affect the other one greater.  This is accomplished by tapping the barres, either ping pong or at the same time, which is like a short inward spike and then a long outward decay, which comes accross (in Calculus) as more area underneathe the outward decay, thus "outward pressure".  Experiment with different ways of playing the barres, trying to create differential articulations, and you will find that these "stepped outputs" really come to life.

-is there a way (or easyish mod) to bring down the input levels to the stuber when putting cocosquares to the brass screws, as with touch-playing?
put resistor in alligator cord, up to 10m, but definitely 1m would make them nice and soft. The series resistor has been experimented with before, for example on the Buchla music easel.

Is the any internal default routing between any of the four sections, via the Connex or otherwise, or is it easier just to link the output of one to the next with a short 1/4" jack?

no default routing.  make it default with a nice little resistor between connex (470k will make it not compression saturated), or do, as you said, a 1/4" patch between output of one to input of another, ideally, through a mixer so you can moderate/modulate the volume and explore the saturation curvature of the tube-type inputs; i.e. as patched it will be a little on the hot side, and you need to turn it down between output and input, and this could even be done with a passive resistor attenuator (or pot) embedded in the cord (it does not need be powered).

When it arrived the Connex pins were covered with that antistatic silver plastic - Should I take equal precautions against static? Thailand has almost none from nylon carpets etc due to no carpets and lots of heat and humidity, only lightning which would fry ANY box if it were hit. Most likely me too :-) - but not so, some of the dry and cold countries I am travelling to (Finland etc). Should I worry or not?
I was just being nice karmatically by wrapping it so.  Technically, Ciat-Lonbarde advises that you not touch any nodes with anything except the tip of an alligator clip, held through insulation.  But some people of course touch the nodes.  And worse people experiment by connecting high voltage probes to the nodes in an attempt to waste their machine.  I cannot control this.  Nylon, it's true, is actually low on static, much less than styrene.  The cocostuber's cocosquare outs are 5 volt nodes, which are not on the other machines.  They are, however protected by an 100k resistor plus the input protection of 4000 series logic (diode to ground and diode to 5v).  Furthermore, the 5v bus is protected with a TVP (not texturized vegetable protein, but transient voltage protector), which should be able to blot out serious machine-life-threatening static, with its "crobar action".

My first attempt to come to grips with the Cocostuber function, revealed more effect from contact with various brass touch contact screws, but less from Barre pressure than for example the Barres on the truly wondrous Tetrazzi which are brilliant. So far I have not be able to detect much "pressure effect" despite using both hands on both Barres. Could you advise me what I should do to improve this?
you have to patch the barre representation section to the modulations section.  For example, take a steppe signal and connect it to SP.AF of a coco section.

I understand that without the screened graphics (at my request) the working of the brass Connex screw pins is less clear than normal, and I plan to rectify this by adding your design in gold leaf (not shorting any pins obviously), but is there any significant difference in operation between the sets of three pins that flip the modes of the four sections, and the remaining four pairs? Of course audio singnals can be used to trigger the flips to good effect, but if I only wanted to work on the audio side, would sticking to the four pairs be useful (maybe for adding external signals or effects)?
yes. these are the CV inputs.  The other pins, again, are "Barre Representations", i.e. CV out from the expression barres.

The function of the 2 pairs of contact screws of the ends of the unit, seem to be quite useful and important, and I notice one set has a resistor between them. My question about these is two fold: a) Would different value resistors change the sound on either pair? and b)not having an indigenous thumb on the outer side of my wrists, could I add some kind of copper, brass or gold contact panel, connected to these screws externally in some place where my present (inner mounted ,-) thumbs could reach them more comfortably?