[Geoff St. Germain]

To measure the tuning frequency you want the mic as close to the driver as possible so that you only measure the output of the driver and not of the PRs in this case.

Blue, part of the problem you are running into with what you are doing is related to the resolution limits of the measurements. The minimas look like a "U" because the resolution isn't high enough to resolve the bottom of the minima so you get a smoothed over look to the measurement. For all three of your measurements the actual minima looks like it is probably near halfway between the two points forming the bottom of the "U". The two bottom graphs show an almost identical tuning frequency, likely within .1 Hz, which makes sense since there is only a 15g difference in the mass added to the PRs. I'd go with either of the bottom ones (335g or 350g) and be happy. It is unlikely you'll be able to hear a difference between the two.

Got any pictures of your enclosure? I've been toying with the idea of picking up an XLS and a couple of the PRs and doing a small subwoofer for the bedroom.