The use of an ion trap mass spectrometer with three different membrane inlet probes is described. Two methods of removing water from the sample are compared. One is the use of a PTFE-silicone rubber double membrane, PTFE is relatively impermeable to water and so reduces the amount entering with the gas sample (Probe A). The second is the use of a silicone rubber membrane covered long probe, which condenses water out of the sample (Probe B). Response times (100%) for dissolved N2O, O2, Ar and CO2 without He in the chamber vary from between 158 and 684 s with Probe A. For the same probe with He, the response times were between 283 and 551 s. In the gas phase response times were between 99 and 153 s with He and 117 and 122 s without He. Probe B had 100% response of between 122 and 152 s for dissolved gases. Further extension of the probe by 2 m slowed response times as did increasing the ionisation time. Response times for Probe B increased to between 99 and 340 s when ionisation time increased from 1000 to 24,930 microseconds. Plots of output against concentration showed the steepest line of response for the short single membrane covered probe with 1000 microseconds ionisation time. Increasing the ionisation time, extending the probe and the use of a double membrane all reduced the gradient of output against concentration for every gas tested. In an intact sediment core, concentrations of O2, N2O and CO2 rose at the start and the concentration of N2 fell. As the disturbed sediment settled, this was reversed. The initial increase in O2 concentration stimulated respiration and inhibited the final pathway in dentrification producing higher concentrations of N2O and reducing the concentration of N2.