Sub-populations of 02-tolerant, obligately anaerobic methanogens were enriched with monomethylamine (MMA) under 3 different regimens: Reglmen 1, where oxygenated seawater was reduced by the aerobic bacterial consorts of the methanogens in sealed gas-permeable polycarbonate flasks that apparently prevented the accumulation of hydrogen sulphlde whlle allowing the production of traces of ethane and ethylene; Regimen 2, like 1, but with bacterial reduchon in glass serum bottles that permitted the accun~ulation of hydrogen sulphide formed by sulphate-reduction but not the C2 hydrocarbons; and Regimen 3, an anaerobic mehum chemically reduced by sodium sulphide and cysteine that is used to culture methanogens from anoxic sediment. In Regimens 1 and 2 , methanogenic bacterial consortia (MBC) were initiated by MMA-oxidizing bacteria that formed reduced microzones in which MMA could be cleaved by methylotrophlc methanogens to form methane, which was apparently oxidized by aerobic methanotrophs almost as quickly as it was produced until dissolved oxygen was exhausted. The process was accelerated under Regimen 2, whose enrichments were used for the most probable number estimation of methanogenic particulates which showed that they accumulate in the pycnoche. These methanogenic particulates were quite sensitive to filter concentration. The distribution of the 3 sub-populations of methanogens is then shown at 10 stations along the density gradient of the estuary where water samples were obtained from the surface layer, pycnocline and bottom layer of water. The results were similar for each layer, where Reqmen 2 consistently produced twice the number of methanogenic enrichments as obtained with Regimen 1 or 2. In contrast, there was a marked dfference in the distribution of the methanogen populations down the density gradient of the estuary. Enrichment in Regimen 3, which was very successful up the estuary, decreased in effectiveness with increasing density. Enrichments in Reg~rnen 2 were quite effective throughout the transect of stratified waters, while those in Regimen 1 had a narrower distribution. I conclude that 02-tolerant methanogens that occur throughout the water column and peak in the pycnocline grow in fragile microniches that are reduced by bacteria that either consume oxygen, produce hydrogen sulphide, or both. A 'top-down' working hypothesis is presented that could explain the diversity, nature and distribution of the bacterial components of the methanogenic enrichments of the water column, and how the methanogens in anoxic sedunents lacking aerobic bacterial consorts may be selected from them. I also postulate that the methanogens living in bacterially reduced microniches may be unique both physiologically and taxonomically, and may have redox potential (Eh) requirements less strict than methanogens from anoxic H2S-rich sediments.