"gotta a liquid mercury roll bar in the trunk"

I've restored a lot of GM F body OT's ragtops (69) and they had "cocktail shakers" in the corners, front and rear. Same principal, reducing harmonics, increasing handling in sub-framed convertibles. A "technology" that didn't last long and has never been repeated that I know of. Weird Science.

 

Those things made a big splash back in the 60s, but I never could see how they could make any difference in a car's handling beyond the placement of ballast.
One interesting application of mercury was/still is a sealed steel cylinder filled with mercury placed in the stock of competition shotguns used in skeet or trap shooting. The purpose is to reduce recoil The cylinder is placed inside the stock of the gun in a hole covered by the buttplate. The device is called a "dead mule"
In competition skeet or trap shooting, for a shooter firing all 4 gauges, a meet can consist of firing 400 or more shells, and anything that helps any at all in recoil reduction is a welcome device.

 

It sounds like the inertia dampers fitted to the Citroën 2CV when it first appeared in 1948, and for a large part of subsequent production, until affordable telescopic dampers which could work in a horizontal orientation eventually became available. Dampers mounted conventionally would have partly defeated the object of the 2CV's clever suspension.

The inertia damper was basically a weight supported on a spring inside a fluid-filled canister, with valving to resist the movement of the weight - or rather the movement of the canister relative to the weight. A canister was attached to the extreme end of each suspension arm, and to nothing else. It worked after a fashion, in terms of 1948 expectations of damping; getting it to work really well would have added too much weight.

The supposed anti-roll devices described in this thread seem to be no more than lateral inertia dampers, and would thus suffer from the same limitations.

If the effect is important (I can't see it) it would make more sense to use the entire sprung mass of the car as the inertia mass by building a tiny bit of lateral compliance into the suspension system, and then damping that amply. It's the sort of thing the OEMs like to do, so they can say that they'd spent gazillions and gazillions sorting out NVH, and that you and I ought therefore not to imagine that we could do better.