And do the crystals grow as mirror images of each other, like Pasteur observed for tartrate crystals?
I believe to understand that most compounds racemates would make a precise 1:1 stacking of right and left upon crystallization. The compounds that separate spontaneously are very uncommon and useful, because these serve to separate the others. That is, if you have a right amine, its salts crystallize more or less easily with the desired right or left acid (boths salts aren't enantiomers) - and if you have the right acid, differing salt crystals lets you obtain the right or left amine.
So from the exceptional tartrate, keep the pure right or left tartric acid, use it to purify the desired amine that wouldn't separate otherwise.
I could be horribly wrong, so please take with caution.
Probably a too big effort for a single compound... I hope the Czochralski process will improve crystallization of chemical compounds as well, especially for difficult operations like chiral separation:http://www.chemicalforums.com/index.php?topic=65386.0