Sorry if I rambled on a bit and rather missed your point. Except for the last three paragraphs.
But here is another idea of how to estimate much void space there might be inside a seated primer without the use of water:
Model a large rifle primer as a cylinder, made of brass. Density of brass is about 8.5 g/cc (depending on alloy composition).
Diameter say 5.4 mm, length say 3.2 mm. If that was solid brass it would weigh about 623 mg. Of course you could measure your own primers instead of using this estimate.
Weigh an actual fired primer. Subtract that from 623 mg. The difference is the mass of brass that represents the empty space inside. Correct that for brass density (i.e. divide by 8.5). The result is the volume of air space inside. I.e. the same as water weight. Since 1 mm^3 of water weighs 1 mg.
View attachment 252379
Or, lets turn it around:
You have measured 35mg of water capacity, though that is for the cup only, not including the anvil. So, that, to me, suggests that you might discover that if you simply weighed a fired primer without the anvil it would come in at around 326 mg. If so, you might conclude that fired primers are about 50% hollow inside, including the firing pin indentation. but not including the anvil.
For your anvil-in figure of 25mg instead, I estimate the fired primer would weigh about 411mg. 34% hollow.
What do yours weigh ?
Somewhere earlier, or maybe on another thread, I made a SWAG that the figure for a complete fired primer might be about 30%. I think that might be about right.
Is this important ? Simply do a sensitivity analysis. Run your simulator of choice with, and without, correction for void space inside the primer. Let us know how the Pmax and muzzle velocity changes, and whether you think it is likely to make a real world difference.