Has anyone done studies on the metal composition of 19th century US Mint medals?

Conventional wisdom says that US Mint medals from the 19th century are almost all struck in pure gold or silver. Julian writes (p. XXXIII):

The local coin dealer recently bought an electronic "precious metal verifier". It's not as detailed as a full-blown XRF analysis, but the equipment is much less expensive. You canalso use it without removing medals from their holders (even stapled 2X2s are ok), and it reads past the plating to whatever is inside. It's also FAST - 2-3 seconds per piece if you're goingslowly.

His model has modes for several purities, including .9999, .999, .925, .900, and .800 (Canadian) silver. It doesn't tell you what the purity is, but it does tell you relative to the selectedmode. So for example, a Sterling piece would read "low" in .999 mode, "high" in .900 mode, and right on target in .925 mode. That means that there was really one more mode thatcould turn up, "still silver but less than .800 fine".

He let me use it for an hour or so on a slow afternoon (thanks, Eric!), and I was able to check out a whole bunch of US Mint medals. As expected, almost all of them tested as pure silver... exceptfor the ones that didn't.

Much to my surprise, a group of the smaller medals tested as .800 or lower. Those low-silver pieces included PR-36, PR-40, PR-41, PR-42, and PR-43, as well as CM-2 and CM-43. On the other hand,the similar small-sized PR-26, PR-27, PR-28, PR-29, PR-30, PR-31, PR-32, PR-33, PR-38, and PR-39, and PR-44 all tested .999 as expected.

I always considered those small-sized medals to be a single series, all produced the same way. I'm at a loss to explain these results. Can anyone shine some light on what's going on?

PR-28 and PR-38 (top row) tested as pure silver (.999 fine).
PR-36 and PR-41 (bottom row) tested as low-grade silver (.800 fine or lower).
Actual size for these medals is 19 mm.