Actually, I've given that up. I'm much more confident in my later method (the Rolling article: measuring the temperature, and interpolating between the two extreme scales). The first method, which seemed to be to at least be "in the ballpark" (if not very precise, given the length of that 'cutout') when I did it on my own cars, seemed too difficult for many folks to get repeatable (and therefore reliable) results.

I'd stick with this new method -- I'm much more confident in it (and it's the very same rationale found in Mercedes shop manuals, by the way).