Actually, after observation of this forum for 10 years, I can easily conclude that it cannot.

I think the use of taller and shorter gearing is probably the easiest, so I will try to use those terms consistently. Thanks all for your suggestions there.

Hopefully this won't stir the pot too much, but I have attached a graph from a 1978 SAE paper on the early OM617a (turbo) which gives the fuel consumption as a function of RPM. More specifically, it is a graph of Brake Specific Fuel Consumption. The units/dimensions of BSFC can be a bit odd if you haven't used them already, but it is basically fuel consumption rate (grams/hour) divided by how much power the engine is producing (kW). So in the end, the units are presented as g/kWh. Anyway, you can use this to roughly estimate how much fuel economy should change as you drop the RPM due to a differential change (or any gear change). It is not exact because the graph assumes that you are at full load, which probably isn't the case. But hopefully the trends in the info are helpful.

The units for fuel consumption again are g/hour, but that is inversely proportional to mpg at a given highway speed. Due to the nonlinear shape of the curve, it is highly unlikely that a 5% drop in RPM would result in a 5% increase in fuel economy. But you can check it out for your specific RPMs if you like.

And just to clarify, this was for the earliest OM617A. The later engines had various improvements which probably adjusted this graph a little.

Attached Thumbnails

 

I was looking though SAE paper 780633 again, and noticed that I missed the actual BSFC map they published. The graph that I included in my previous post only works for full load. But the graph in this current post is the complete map of BSFC as a function of RPM. So you can use this map to predict fuel consumption changed from gearing changes. Hopefully at least one person out there finds looking at this info interesting.

As an aside, they also include "Road Load" for the 300sd (the focus vehicle of this paper). This is basically how much power is needed to go a certain speed. It is the exact same thing I tried to predict for my 220d in my first post. Unfortunately, they don't explain how that curve was calculated. There are a lot of assumptions that go into that calculation, so it would have been nice if they had documented their assumptions. Or, I guess there is a chance that MB provided that info. It really isn't clear how this graph was created.

Not sure about people's onroad experiences, but this paper suggests that for a given vehicle speed, the turbo version of this engine is actually more fuel efficient than the NA version (and the 240D as well). That is impressive engineering, in my opinion. See 2nd graph attached to this post.

Attached Thumbnails

   

Is the road load indicated...........if so, I missed it?

The fuel consumption table at various power demands is very interesting. Unfortunately, it confirms that the engine does relatively poorly under very low power conditions (road load).

....for a lay person like me, I just based my rear diff ratios' relationship to engine horsepower through engine RPM using the good ole' ratio and proportion equation.

Example: If at 70 MPH, a 2.88 rear diff turns the engine at 3000 RPM.
For a) a 2.47, b) a 3.07, and c) a 2.24 rear diffs, how many RPMs would the engine turn?

a) 2.88/3000 = 2.47/X
X = [2.47 (3000)]/2.88
X = 2573 RPM

b) 2.88/3000 = 3.07/Y
Y = [3.07 (3000)]/2.88
Y = 3198 RPM

c) 2.88/3000 = 2.24/Z
Z = [2.24 (3000)]/2.88
Z = 2333 RPM

Therefore, while the 2.88 rear diff turns the engine 3000 RPM @ 70 mph, a 2.47 rear diff will turn the engine 2573 RPM @ 70 MPH, a 3.07 @ 3198 RPM and a 2.24 @ 2333 RPM.

Considering that all of the above have the same stock OM 617 Turbo and 4-speed automatic trans....

__________________
'85 300D Turbo - CA Version

The torque curve is quite flat at those rpm's, therefore your conclusion of engine speed as a direct correlation to horsepower available speed is fairly accurate.

Don't confuse the available horsepower with the typical horsepower needed for road load.

Road load curve for early 300SD:
In post 48 (the first graph), there is a dotted curve that starts down low on the left side of the graph and ends 2/3 of the way up the right side of the graph (near kW=80). It is associated with the scale on the kW scale on the right side of the graph (although they do a poor job at making all that clear). I haven't looked into drag coefficients or frontal areas for 300SDs, so I am not sure how much this curve would change if you were focusing on a 300D instead.