Poly(ethylene terephthalate) (PET) bottles used in commercially available m
ineral water were compared to Teflon bottles for sampling and storage of ri
ver water samples prior to mercury determination. The metal concentration w
as in the range 0.3-2.0 ng L-1 for the reactive species, and in the range 1
-9 ng L-1 for total mercury. A paired t-test showed no significant differen
ce in the results for both reactive (p = 0.011) and total Hg (p = 0.024). S
torage time was studied for PET bottles, using synthetic samples spiked wit
h 10 ng L-1 Hg2+ ions, using two different types of preservation: (a) by fr
eezing at -18 degrees C, and (b) by the addition of BrCl solution. Both pro
cedures yielded the same results up to the 14th day of storage (ANOVA, p =
0.72 for BrCl and p = 0.12 for freezing). The freezing procedure proved to
be more suitable for longer storage times, i.e., up to 40 d (significance l
evel of 0.03, ANOVA).