I ran a test over the last month, running ethanol-free gas (or, more precisely, fueling from a pump that claimed it to be ethanol-free) in my FJR for several tanks, then switching back. I MIGHT have gained 1 MPG using "straight" gas in that sequence over the stuff that MIGHT have UP TO 10% ethanol. The straight stuff was rated at higher octane (91) than the 87 I usually run. There was no observable difference in behavior or performance between the two. The "straight" gas ran $0.25/ gallon more than 87 e10.
Some more detail:
This 1 MPG is in the ballpark for the published difference between the energy density (BTU/GAL, e.g.) between gasoline and a 10% (by volume) gasohol. This difference is about 3%. (
http://en.wikipedia.org/wiki/Energy_density). For reference, my FJR right now using (theoretically) e10 is getting 46-47 MPG, given my pattern of normal use.
There are caveats to this experiment:
1. I believe the pattern of use of the bike (sustained highway vs. stop/shutoff, modest length runs vs. short runs) for this experiment was the same throughout. It's possible that there was more variation in this than I realize. So one could argue there is enough experimental error (there's always some) to preclude taking this 1 MPG to be highly accurate.
2. This is summer, when my "e10" gas mileage jumps up in all my vehicles. It has always been more noticeable in my Prius and bikes. It'll almost certainly go back down around late September, like it always does. This seasonal change is about 10%. I've long attributed this to seasonal changes in gas formulations and it might include an e10 in winter and e0 in summer. Or it could be other additives, etc. Who knows?
3. You really do not know what you're putting in your tank. You take it on faith that if the pump says it contains 10% ethanol, it at least has some in it. And if it says ethanol-free, this is the ONLY difference between that fuel and the e10 one buys at the same station.
4. The ethanol-free stuff was rated 91 octane. I know from past experience that running a higher-than required octane in a given engine can actually reduce gas mileage. This, of course, further complicates trying to measure the effect of ethanol on mileage.
5. Filling the tank to the same level every time is a challenge on a FJR. And it happens that the pump available to me with ethanol-free gas is sitting in a spot that is far from level pavement. I did this experiment over several tanks to mitigate this effect, but still.
I'll retry this in a few months, when we're back on the other-season blend, to see what I see.