My guess would be there could be some minor loss of DGM however this could be dependent upon filter size and vacuum pressure. Remembering back to my Analytical Chemistry days, we degassed water samples by applying a vacuum. 

Best bet, compare a bunch of different methods if you can. 

My opinion is that water filtration will result in loss of dissolved gasses as filtering is connected to change of pressure, especially if vacuum is applied. Certainly filtering have influence on dissolved Hg gasses (Hg(0) and DMeHg). It should be checked in filtered and unfiltered water immediately after sampling.

Suspecting  filtrating will result in gas phase loss, I never filter water sample for further DGM or DMHg measurements

I don't know about vacuum filtration. During the U.S. GEOTRACES cruises, we have been filtering water through capsule filters (0.2 µm PES) attached directly to GoFlo bottles with silicone tubing and then another length of tubing to the sample bottle.  Our comparisons of filtered vs. nonfiltered water were few, but we found no significant difference between the filt. and nonfilt. water for dimethylHg (Bowman and Hammerschmidt, 2011, L&O Methods).  Somewhere early on in GEOTRACES we also compared Hg0 and found filtration by this technique to have no significant effect, which is consistent with dimethylHg and Hg0 having similar solubilities.  A laminar flow (no bubbles) is important for minimizing potential degassing of the sample.

Completely agree with Joze, you should check your method before proceeding.

Filter type, and whether it is vacuum aided or pressure filtered are the likeliest sources for introducing bias. The capsule filters are pressure filtered and are filled slowly to wet and fill interior surfaces thoroughly, before closing the vent and obtaining filtered sample, if I remember correctly. Vacuum filtration is on the other hand likely to degas and lose volatile organo mercury species in raw waters. If you would attempt experiments to discover amounts of bias, they should be done with tracer isotope dilution ( if isotope ratio MS available) side by side.In harking back to ways to determine analytical bias from matrix effects, Method of Standard Additions and comparative plots of results could discover limits of accuracy and bias for methodology. 

I agree with the above that using capsule filters (properly) should present very little negative bias, but that vacuum filtration will almost certainly result in some loss of dissolved gasses. Any direct comparison that you can perform between filtered and unfiltered samples would help you make the case for using filtration as a preservation method.

Good luck!

thanks for all the answers. i do not see a reason to filter sea water samples for determination of gaseous species, DGM and DMHg. For now, i do not plan to filter those samples, nor do i plan a comparison. But looking at published literature i noticed large differences of DGM, DMHg and MeHg (MMHg + DMHg) data for same ocean basins. I am wondering if this might be partly due to the fact that some groups filter and others do not.

Chemically, we would not expect that filtration would reduce elemental or dimethyl Mercury.

The issue isn't filtration per say as of course they are not particles. It's the sparing solubility and significant vapor pressure of the analytes. Under vacuum filtration conditions they can partition into the headspace and be evacuated by the vacuum system. That's why using capsule filters under positive pressure (or esp. gentle gravity pressure with no pump cavitation to generate bubbles) is recommended, *if filtration is needed*. If measurements are made immediately, there is no reason to filter. If samples are to be stored, they need to be preserved in a way that preserves the Hg speciation somehow and sometimes filtration is used to remove microbes.