Mike -- just in the interest of troubleshooting this in this thread, at Danny's suggestion, I tried the following.
1) Swapping the left and right tweeters, to see if the dip followed the tweeter or if it stayed with the speaker. Measurements would indicate that the dip indeed followed the tweeter.
2) With original configuration of tweeters (L on L speaker and R on R speaker), I disconnected the mid-bass drivers and measured to see if the dip was still with the R tweeter or if there was a tweeter/mid-bass interaction. Measurements would indicate that the dip is apparent in both configurations.
Green and red traces are right speaker (with original CD installed on right V). Blue and orange are left speaker.
Other than the dip, the two channels look pretty closely matched. I'd assume that the peaks and dips > 4kHz or so are highly dependent on microphone position; the delta I see between L and R speakers is that 1.8 kHz dip alone.
I'll report back anything else we figure out.