If you have access to a good NMR spectrometer (500 MHz or better), you can easily distinguish Mg++ from a Grignard reagent by 25Mg NMR.  I believe the peaks can be integrated to give relative concentrations; if not you can add measured amounts of MgBr2 and obtain a calibration curve.

http://chem.ch.huji.ac.il/nmr/techniques/1d/row3/mg.html