The microtrack has no other memory in the device large enough to hold the recording so whenever it's recording it would have to be writing to the card. Of course I guess there could be a small buffer where a little bit could be lost in the event of an abruptly stopped recording, but it woudl be a small part of what you were recording.
But why would it 'have' to be writing to the card - it might've been in some sort of record-pause mode - since file recovery programs on my computer at home didn't appear to be able to find any lost data or files, it wasn't clear if anything had actually been written.
But to answer my own question, I tried the card in my fairly new WinXP computer at work, and wahey - R-Studio found the missing file and was able to recover it with the correct filename and all. So with the costs of getting to the gig, and then buying file recovery software, not the cheapest recording I've ever made.
And the other night when I had the same situation again with manual file split, waiting 'til the first file had finished writing before the second button press meant it worked perfectly, albeit with the process now taking a couple of seconds longer, so I'm back to being happy with my MT at least.