The 0.1 parts per million was originally or historically established, I believe, based on the kinds of detection limits analytical equipment was capable of. So it would basically be able to detect any pesticide residues at all.
On May 28th, 2007. See this statement in context.