Birdbeard,
I had to weigh in and couldn't agree more.
Glycogen, or fat burning (as another poster pointed out, there's more than one way to word this problem) is so often a limiting factor. This must be solved before you can run to the point that your aerobic fitness should theoretically allow. And you're right, Canova would agree. A 40k long run is fairly standard for his guys. Mosop did one -and hammered it, setting the Course Record on a loop Renato often uses - 4 weeks and 3 days out from 2011 Boston. He had done a 35k with an impressively fast finish earlier in the cycle. The old standard 20 miler just doesn't do it. You're not hitting the wall during that run and I never have that soon (23 is common in my experience) in the races where I did blow it.
Even the 2-5k interval type long runs Canova prescribes can be up to 20 miles or so. Mosop does 8X3k with 1k float (the kind of thing a couple posters have mentioned) 3 weeks out. This is 19+ miles, and including the warm-up, this makes it into what most would consider 'long run' territory.
I see someone linked MPR's thread. I know some of you dig this, and I don't disagree with much there. The 24 should indeed, be considered a 'normal' long run distance in my mind. The 10 + 10 workout is great of course. My only reservation is that he throws in a couple 10k-half marathon specific workouts too. (This is all from memory - I didn't click on it just now, but the way, so I have indeed paid attention to that thread.) These are fine for turnover, as some say, or just to not lose touch with your full spectrum of speeds, but I would modify MPR's advice in the following way: stress the longer stuff, downplay the shorter and faster. It's after 20 when the problems usually begin. One can train to that effect by running 15 miles moderate and then doing 6X1 mile. It's just that although doing those intervals fresh helps neuromuscular fitness and speed, it doesn't help much when you're 2 miles from the finish and have to walk!
Look at this:
https://www.letsrun.com/forum/flat_read.php?thread=10339274
I linked to some of Nate Jenkins' writing in there. He reminds me of myself or many of us amateurs in the sense that, while faster, he is not so much so that it seems he is from another species. Don't read that wrong - I can no longer go under 3 hours at my age, so I'm not bragging - it's just that I find him relatable. It was hard for him to get good and he seemed to gain an understanding of how to do it. He very strongly believes that glycogen is the primary limiting factor for so many. This often includes track stars who never seem to put up marathon performances close to an equivalent of their 5k PRs.
Another point. It is very hard to do really good marathon specific long runs. It takes a fair amount of training to get to where you can put the hammer down as you go past the 20-mile-mark in your long run. The standard 20 miler will, of course, make 24 easier the firs time you're doing it. I mentioned a few things in the linked thread that I do semi-regularly that are close enough to the most specific stuff that the latter are accessible.
One last point. Something that is 22-25 miles long with 15-17 miles of hard running will take a lot out of you. I would record once every 2 weeks or slightly more often for something like best workouts mentioned here or the ones I lay out in that other thread. Mosop took 5 easy days after his hard 40, and the next hard one had 'only' 10 miles worth of fast running, Special blocks and the like often occur 10 or 11 days apart at the most frequent even among young superstars such as Canova coaches.
Anyway, great thread!