Dynamic Range and noise; APS-C vs. 35mm Full Frame

I've used identically framed Color Checker Classic shots taken at 42MP FF and 18MP APS_C crop mode using my Sony a7RII and Sony FE70-300G lens to show noise levels at a suitably illustrative grey. Developed in ACR & edited in PS CC 2018.

Both images viewed at 100% have the same noise characteristics, 100% crop for 100% crop but one image is 42MP and the other 18MP. I down-sampled the FF image from 7952x5304 to 5168x3448 to match the pixel size of the crop image.
These two images are still exposed the same but are now the same size in pixels. The original FF image now has noise structure that is markedly finer than the original crop image. An original 42MP gray section at 100% is outlined in GREEN.
Outlined in RED is a 100% un-sampled dark gray section from the crop-mode shot and superimposed over a larger DOWN-SAMPLED section of that same gray from what was originally the FF shot.
To the left is the framing of that Color Checker Classic but is obviously massively down-sampled and rotated counter-clockwise 90 degrees to fit the web page.

This whole experiment revolves around two premises:

1. Dynamic Range is often measured from the brightest white (defined by clipping) to what is actually a very, very dark gray with a specific amount of noise. Less noise and that black is ‘pushed’ further down giving a greater effective DR.

2. The measurement of noise is (in this case) dependant on the normalization of image size. This normalization proves to have a positive relationship to sensor area at any given technology level.

As an aside, modern sensor design has gotten to the point that smaller sensels don’t seem to be effectively less efficient over a given area like they used to be.
Normalized for size and at modestly high ISOs, the a7RII does very well even against the large sensel low-light juggernaut a7SII.

Recapping: The two shots do produce similar images but at different sizes. While I down-resed the FF images, I could up-res the crop-mode shot to 42MP and that would have made the noise look worse.
Either way the FF shot has less apparent noise per total image and given how DR is typically measured it thus has greater dynamic range.