Hmmm, that fluctuation seems fairly significant.  I wonder if it swings every month.  If so, it probably matters when the actual sample is taken, too. 

Thanks for posting that up!