Tom(s)/Mark;
While I agree that greensites on many older courses by many architects were built on the "high points" for the reasons we've discussed in here many times before, I think there are some subtle differences that make Flynn standout in this regard.
This is pure conjecture based on observation, but it seems to me that Flynn often located his greens EXACTLY at the top edge of the steepest point of an elevation rise. That is somewhat different than placing a green at the Very Top of the elevation rise, as Ross often did, because in many hilltop Rossian greens you find a somewhat flatter "transition area" just short of the green where a running shot may be accommodated, and where it's unlikely the ball will trickle back 10-50 yards down the fairway.
Flynn, by contrast, seemed to build a lot of greens that sit right on the edge of disaster. Usually, there is additional elevation change and slope through and beyond his green sites...they are not at the very top of a slope, but instead at the top of the steepest point.
I think this accomplished a few other things that might not be entirely obvious. By locating greens this way, aside for the "shot demands" and intensity level it places on the golfer, I think the most important thing Flynn was trying to do was provide target visibility. I find it amazing how much of the green and surrounds Flynn was able to "show" the golfer, even on greens that are benched into steep hillsides.
I think the other reason he might have done that is something Tom Paul mentioned. Flynn loved to build his bunkers into upslopes, for visibility and dramatic purposes, and those steep rises must have been just too attractive for him to pass up.
Going back to contrast where Flynn places his uphill greensites with some others, like Ross, I find that many times Ross's "top of the hill" approach led to a greater degree of blindness for the approach, as well as a less dramatic looking, well-defined target.
I find this difference to be generally true, with exceptions of course, but in any case, I can't think of any architect who created "rollback" situations as much as William Flynn, simply because he seemed more willing than others to place his greensites a little further down an elevation change (to the steepest point) than some of his contemporaries.
I also wonder why we hardly ever see this feature in today's architecture.