A study (Aguust 1990 to July 1991) of profiles of dissovled CH4 concentrations, diffusive flux of CH4, and CH4 production rates of 45 sites in the Lake Agassiz Peatlands in northern Minnesota shows that dissolved CH4 deep in the peat (>1 m depth) mobilized easily to the vadose zone. During August 1990 the dissolved CH4 concentrations at some depths at some sites were supersaturated with respect to one atmosphere partial pressure of CH4. At one site (2.5 m depth) the concentration of dissolved CH4 in the peat pore-water was 140 mg L-1. In July 1991, at no site did the concentration of dissolved CH4 in the peat pore water exceed 40 mg L-1 in the peat profile. The average calculated diffusive flux of CH4 decreased from 95 to 45 mg m-2 d-1 between 1990 and 1991. Gaseous CH4 was more in evidence in 1990 than in 1991. In 1990, CH4 at many depths bubbled vigorously when peat pore water was sampled. At some sites there was sufficient pore pressure to eject slugs of water forcibly from piezometers. Similarly, dissolved inorganic carbon (DIC) consisting of H2CO3, CO2, HCO-3, and CO2-3 decreased between the sampling times from an average for both bogs and fens in 1990 of 62 mg C L-1 to 38 mg C L-1 in 1991. A dynamic mechanism must exist which traps CH4 deep in the peak column allowing gaseous CH4 to build up, increasing dissolved CH4. Other times, CH4 passes freely from deep peat to the vadose zone. We suggest as a hypothesis that a confining layer of trapped CH4 bubbles forms at depth in the peat, trapping gaseous CH4. The duration of the ''bubble confining layer'' is uncertain. We propose two hypotheses. (1) The confining layer is usually present and deterioriates after a major climatic event such as a drought, or (2) the confingin layer forms and collapses seasonally with seasonal variations in the water table elevation. ¿ American Geophysical Union 1995