Whoa guys!
The 6 lbs. per horsepower is correct!
More to follow!
That's another of those "as far as it goes" kind of things Chip. Once a rolling mass starts rolling it's not the increased weight in isolation that uses up greater amounts of energy (in our case measured horsepower). Remember, the actual horsepower output of the engine will (for the purposes of the measures we're looking at) stay the same. What we're actually taking about is how much of it goes to parasitic losses like drag due to increased weight. And that takes us back to the beginning.
Weight, in isolation, won't give a static loss of some amount of horsepower (transferrable). Weight will increase drag. That's actually what we're talking about here. I've never seen aero drag numbers for an Electra Glide so can't calculate for you what the amount of increased drag will be for increased weights at differetn speeds. But it's the drag that causes the "performance" loss.
The effect of drag will go up (exponentially) as speed increases. So the performance drag on a rolling mass (our bikes) of 10 pounds at 20 mph will be less than the loss due to that same 10 pounds at 80 mph unless something else changes aerodynamically between those speeds to change the amount of drag. But since our bikes don't have a variable sweep wing to change aerodynamics over a speed range like, say, the old F14 our drag only gets worse as we go faster.
So, yeah, both directly and indirectly weight "costs" horsepower. After all, it takes more push to get something going that weighs 900 pounds than it does soemthing that weighs 800 pounds. So that part is only obvious. But the greater loss at speed will be drag and that will increase more the faster we go. The curve may be smooth but the number won't be static.