Brian Glover-I would assume the longer the leaf bar, the more roll resistance
Actualy the shorter the blade, the more roll resistance.... thought experiment (well , practice too...) shows an infinitely short (nice tautology ,huh?) blade would not bend at all (if the same thickness-always an useful caveat), as links would be acting against each other..... mmmm... very muzzy analogy.... lets start again.
A leaf spring of a given section , will deform (bend) a given amount for each unit of lenght... so double the lenght , it will bend twice as much under the same load... as rate is defined by load/deflection rate will be half, so reducing roll resistance when connected to wheels through link.
And keeping link angle to blade as close as possible to 90deg. keeps rate linear over a reasonable amount of travel (which translates in angular position of this particular double-opposed blade arrangement...imagine it at zero or 180 degrees and it will be infinite , as links again will be pulling against each other , thus stopping all movement notwithstanding the force).
Other angle values can be used to have a progressive or regresive roll resistance... quite interesting, but difficult to implement as related to ride height.... I have used it on rear suspension of car with rear fuel tank location to cater for full-empty tank condition. Ultimately a conventional "C" shaped bar the easiest to use, at a pinch "T" bars, but all other versions depending on packaging or space availiable.
Much more interesting is the Kinetic tipe of bar actuation.. as used on WRC Citroens, and subject of article in RCE some months ago... see their site for more details.
And neat Mustang work around... some good lateral thinking there...interesting way of lowering the roll center on a beam axle.