A headspace solid-phase microextraction (SPME) and gas chromatography-mass spectrometry (GC-MS) method has been reevaluated for the determination of methyl tert-butyl ether (MTBE) in water. Factors affecting method performance were examined. The internal standard, deuterated MTBE, is necessary because it allows normalization of the calibration since the deuterated MTBE and MTBE experience the same extraction efficiency. The method detection limit was 6.6 ng/l, with a linear range up to 160 microg/l. SPME extraction efficiency and detection sensitivity for MTBE decreased in the presence of gasoline matrix. However, reliable measurement of MTBE was possible in the background matrix concentration of 0.97 mg/l gasoline using deuterated internal standard. Reproducibility was 4.5% at the spiking level of 47 ng/l. The average recoveries of MTBE in spiked fountain water, tap water, and snow at 60 ng/l and in spiked rainwater at 580 ng/l ranged from 101 to 105%. The method was successfully applied to the analysis of trace levels of MTBE in various water samples, including rain and snow samples collected in Albany, NY, USA.