> Like defies imagination (as Bergstrom said).

Could there be another source for this?

Hasn’t uniformity in test / detection standards been a source in region to region variability in some instances?