Take a look at Steve's (aeroexperiments) older vids when he was experimenting with small drag 'chutes near his wingtips. You can see the effects of the tip vorticies generated at high and lower lift coefficients. At high Cl's (slow airspeeds) the drogues are rotating the way you'd initially think they would, rotation from the bottom to the top surfaces, indicating upward tip-lift. But at a certain Cl (lower), they would stop their rotations, then at still-lower Cl's (higher airspeeds), they would rotate the opposite direction (rotation from top to bottom surfaces), indicating downward tip-lift.
Because a flexy has a bunch of aeroelasticity, I figure that the extensions are reducing tip vorticies, not merely by their shape and span-extension to the wing alone, but also because they add more moment behind the local neutral moment of the wing at the tip, thus deflecting the tip more,
increasing the wash-out twist, and modifying the lift distribution about the whole wing to more of an exaggerated bell-curve (more lift at the root, less lift nearer the tips) at high Cl's (much to Prandtl's supposed second theory),
theoretically further reducing tip drag and overall induced drag. However, I don't completely subscribe to Bower's ideas. By modifying the lift curves, the root now has a greater portion of the lift distribution, thus there is more leading and trailing, vertical flow deflection (action/reaction) near the root, creating more induced drag, whereas the mid wing and tips are doing less work, and almost merely along for the ride (not really, but closer to that condition than without the tips).
At lower Cl's, again the tip extensions are creating a larger local aerodynamic pitch moment to the aeroelastic wing, and possibly help to
reduce geometric wash-out (reducing the negative lift dist at the tips and flattening the whole lift-dist curve), allowing less overall induced drag due to the inefficiencies of the typical, highly-twisted (washed-out) wing platform at low Cl's. I would guess there is a limit to that reduced twisting at low Cl's.... too much local pitch-moment-induced deflection from the original design goal and it would hinder the nice, humane, airspeed-limiting characteristics of the glider (more prone to tucking?).
At least this is my theory at this point in time. I'm sure Steve will set me straight.