Razorwit
Yeah, sounds like a clocking thing. I doubt it's 48 to 44.1 because that results in a change of about 1 and a quarter, but it could be a smaller sampling rate discrepancy. I'd drop the word clock and see if it continues. I've gotten some funky word clock stuff from time to time.
Dean
Yep. I used to have an Echo Audio Layla 24 that turned out to have a clock that was just a wee bit off-spec. When I hooked up my old Tascam DM32 via ADAT, I caught the problem: the DM32 has a utility that tells you how much the incoming clock is deviating from the target rate, so when I tried to use the Layla as the master, the DM32 let me know the sample rate was 97% of 44.1k.
So off to the expensive world of proper master clocks I went (and dumped the Layla in favor of RME

)
Back to your OP, Mark - the quickie test is importing the stems back into their respective projects. If they line up on your system, that implies a clocking mismatch between you and the other folks. The next step is taking your Black Lion out of line like Dean suggests. Export stems using the MOTU as the master clock, and send 'em to someone offsite to check. And like Jonesey said, also double check *they* are using a reliable clock set to 44.1....
Black Lion has a good rep, but my first guess would be the problem involves the Micro Clock. The Micro Clock MkIII seems to be a genuine competitor to the higher-end master clocks, but I don't think the earlier versions are quite in the same league. Regardless, whichever device is causing the sample rate deviation, something ain't right somewhere for sure.
Good luck!