Along one edge (not always the same one), the brim, and the brim alone, will have considerable gaps between passes.  This, of course, leads to curled prints and hilarity.

The troubleshooting steps I've tried:

* Leveling the bed.  I've done this a dozen times since this problem manifested.  I use a 12" length of 0.3mm feeler gauge stock.  I've grown quite comfortable doing this, thanks to accidentally acquiring a broken Printrbot LC (V1) before the Solidoodle arrived.  I'm confident my bed is as level with respect to the print head as I can get it by feel alone.

* Making sure all of the nuts, bolts, and belts are tight on the X and Y carriages.  There's a small amount (<1mm) of wobble in the X carriage where the rear of the carriage meets the rod, with the teflon block.  This requires four hands to tighten, and that's the best I can do with only two.

* Extrusion rates.  I've managed perfect 0.5mm walls with a Slic3r multiplier of 0.95.  I'm not sure I can get any better than that.  If I up the extrusion rate, it may over-extrude enough to fix the brim, but the walls will then be out of spec. 

* Temperature.  I've printed anywhere between 195C and 199C.  I've left it at 199C for this filament because it seems to be giving me such great results otherwise.

You may be able to see in the attached photos that my walls are decently solid and adhere well, even along the problem brim area.  I initially suspected just simple bed leveling, except that there's no effect on the bottom of the printed piece itself. 

What makes this odd is that, if this were a bed level issue in its entirety, the problem brim area would be consistent with the area of the bed that needs leveling.  If you look at the overall picture of the printed calibration piece, the problem area starts just right of the curve (I'm going anticlockwise here), follows the curve and the short straightaway, and then vanishes abruptly at the turn to the long 50mm straightaway. 

If this were a leveling problem, I'd expect the problem to continue until it's roughly aligned (on the Y axis) with where the problem starts on the opposite end of the piece.  This is most definitely not the case.

Of note is that I currently have my borosilicate glass plate held in place by two binder clips along the very back edge of the bed.  When this problem first manifested, I'd clipped it on the front (to give me bed space back, as the carriage collides with them back there), and the problem brim area was to the front-left edge of the bed. 

As part of the troubleshooting process, I also moved the clips back to where they were (reasoning that if it was broken because they'd moved...).  This only affected where the problem occurs (if it affects it at all, correlation and causation and all that). 

Frankly, I'm out of ideas.  The bed's level, the parts are tight, and my one-bead-thick walls are gorgeous.

So, my question is, what am I missing?

Also, to the person on this forum that invented this particular calibration piece:  THANK YOU.  This is easily the best way I've found to check my printers' (Both the Printrbot and the Solidoodle 4) calibration in every way.