I asked Gary Wheeler, former chassis engineer/designer for Traction Master and then Dan Gurney's AAR team and this is what he had to say on the subject...
That's a very interesting question and the answer is a kinematic subtlety. The answer is: the two methods of utilizing a Watts Linkage are effectively the same. (After all, straight-line motion between axle housing and chassis - - - is straight-line motion between axle housing and chassis no matter how it's achieved.) As it turns out, straight-line motion is great when the car is on a straightaway. But during cornering, straight line motion isn't as good as a LONG Panhard Rod. Why? Well, while cornering the chassis has a roll-angle. For clarity, let's say it's softly sprung and has a rather large roll angle of 4 degrees. That means any rear axle movement relative to the chassis will rise or fall along that same 4 degree angle with respect to the horizontal road surface. THAT forceably yanks both rear tire contact patches laterally a tiny fraction of an inch - which is undesirable because any lateral motion of a tire's footprint caused by suspension geometry is immediately subtractive from rear-axle cornering power.
Yanking the rear tire footprints laterally even a tiny bit is no problem on the straights, but it's the last thing you want during cornering, simply because IF you're cornering at 10/10ths and the car is perfectly balanced, a momentary degradation of rear axle cornering power will momentarily move the car toward oversteer. The driver will perceive the car as needlessly twitchy when encountering tiny bumps in the corners. And there are ALWAYS tiny bumps in the corners.
Short & highly-angled Panhard Bars are obviously deadly. Although LONG horizontal Panhard Bars aren't perfect either, a careful layout of the geometry of a long-bar, low Roll Center Panhard Bar installation will reveal that during bump or rebound travel, it will actually move the rear tires laterally during body roll slightly LESS than a perfectly straight-line travel Watts or Roberts link. THAT, not simplicity or weight, is why I used a Panhard Bar the AAR 'Cudas. (I consulted a bit to Mark Donohue in 1968-69 and that's precisely why the Penske Camaros always used Panhard Rods.)
Parnelli and Follmer's factory Mustang engineers didn't catch this subtlety and used Watts links. They won the 1970 TA championship because of our unreliable engines, but if memory serves me right, we ended up with more lap records. One thing few people appreciate to this day - is that our works Cudas weighed 3200 lbs. yet Gurney broke all his own previous A-production works Cobra lap records at every track in the country - by approx. two seconds per lap! This is stunning, considering both the Cobra & 'Cuda had Gurney as the driver, nearly identical 540 horsepower, similar tire footprints, yet the 'Cuda was exactly 1,000 lbs heavier than the Cobra! (3200 vs 2200 lbs.) It's interesting to consider how well Dan's works 'Cuda had to be handling for him to hot lap two seconds faster than his own works Cobra lap records - in a pony car weighing nearly 50% heavier!!
