What sort of seed grinding at Pleistocene Lake Mungo?

Fullagar et al. (2015) use microwear and residues to argue for seed grinding in the late Pleistocene at Lake Mungo. The ten Zanci/Arumpo grindstone fragments in their study represent only three grindstones when conjoined. Grindstones appear to be rare in these assemblages and are small, hand-sized implements. Traces of starch are low, of uncertain taphonomy and not from known economic taxa. The use-polish may indicate grinding of seeds, but these grindstones were used differently to ethnohistorical seed-grinders. Even if seeds were used, this was probably not a seed-based economy.