A culture-independent molecular phylogenetic approach was used to survey constituents of microbial communities associated with an aquifer contaminated with hydrocarbons (mainly jet fuel) and chlorinated solvents undergoing intrinsic bioremediation. Samples were obtained from three redox zones: methanogenic, methanogenic-sulfate reducing, and iron or sulfate reducing. Small-subunit rRNA genes were amplified directly from aquifer material DNA by PCR with universally conserved or Bacteria- orArchaea-specific primers and were cloned. A total of 812 clones were screened by restriction fragment length polymorphisms (RFLP), approximately 50% of which were unique. All RFLP types that occurred more than once in the libraries, as well as many of the unique types, were sequenced. A total of 104 (94 bacterial and 10 archaeal) sequence types were determined. Of the 94 bacterial sequence types, 10 have no phylogenetic association with known taxonomic divisions and are phylogenetically grouped in six novel division level groups (candidate divisions WS1 to WS6); 21 belong to four recently described candidate divisions with no cultivated representatives (OP5, OP8, OP10, and OP11); and 63 are phylogenetically associated with 10 well-recognized divisions. The physiology of two particularly abundant sequence types obtained from the methanogenic zone could be inferred from their phylogenetic association with groups of microorganisms with a consistent phenotype. One of these sequence types is associated with the genus Syntrophus; Syntrophus spp. produce energy from the anaerobic oxidation of organic acids, with the production of acetate and hydrogen. The organism represented by the other sequence type is closely related to Methanosaetaspp., which are known to be capable of energy generation only through aceticlastic methanogenesis. We hypothesize, therefore, that the terminal step of hydrocarbon degradation in the methanogenic zone of the aquifer is aceticlastic methanogenesis and that the microorganisms represented by these two sequence types occur in syntrophic association.

We determined factors that affect responses of bacterial isolates and model bacterial communities to the 95 carbon substrates in Biolog microtiter plates. For isolates and communities of three to six bacterial strains, substrate oxidation rates were typically nonlinear and were delayed by dilution of the inoculum. When inoculum density was controlled, patterns of positive and negative responses exhibited by microbial communities to each of the carbon sources were reproducible. Rates and extents of substrate oxidation by the communities were also reproducible but were not simply the sum of those exhibited by community members when tested separately. Replicates of the same model community clustered when analyzed by principalcomponents analysis (PCA), and model communities with different compositions were clearly separated on the first PCA axis, which accounted for >60% of the dataset variation. PCA discrimination among different model communities depended on the extent to which specific substrates were oxidized. However, the substrates interpreted by PCA to be most significant in distinguishing the communities changed with reading time, reflecting the nonlinearity of substrate oxidation rates. Although whole-community substrate utilization profiles were reproducible signatures for a given community, the extent of oxidation of specific substrates and the numbers or activities of microorganisms using those substrates in a given community were not correlated. Replicate soil samples varied significantly in the rate and extent of oxidation of seven tested substrates, suggesting microscale heterogeneity in composition of the soil microbial community. Garland and Mills (2) recently introduced the use of community-level carbon source utilization patterns for comparison of microbial communities from different habitats. They used commercially available microtiter plates that contain 95 carbon substrates (Biolog GN; Biolog, Inc., Hayward, Calif.), which they directly inoculated with environmental samples from freshwater, saltwater, estuarine, and hydroponic solutions, from the rhizosphere of hydroponically grown wheat, and from soils. Similarly, Winding (5) analyzed the whole-community Biolog GN substrate utilization profiles of samples taken from several types of forest soils, as well as from size class fractions within one soil type, and Zak et al. (6) studied the profiles of soil samples from six desert plant communities along an elevational and moisture gradient. In all three studies, multivariate statistical methods were used to establish distinctions among the resulting community substrate utilization profiles and thereby classify the community samples with respect to their carbon source metabolism. Garland and Mills (2) proposed that since separation of community Biolog profiles by multivariate analyses is based on differences in carbon source utilization between samples, it provided ''a functional basis to distinctions among communities.'' All three groups suggested that in addition to establishing ecologically releva...

Samples of Marshall Sandstone, a major source of groundwater with elevated arsenic levels in southeast Michigan, were exposed to bicarbonate ion under controlled chemical conditions. In particular, effects of pH and redox conditions on arsenic release were evaluated. The release of arsenic from the aquifer rock was strongly related to the bicarbonate concentration in the leaching solution.The results obtained suggest that the carbonation of arsenic sulfide minerals, including orpiment (As 2 S 3 ) and realgar (As 2 S 2 ), is an important process in leaching arsenic into groundwater under anaerobic conditions. The arsenocarbonate complexes formed, believed to be As(CO 3 ) 2 -, As(CO 3 )(OH) 2 -, and AsCO 3 + , are stable in groundwater. The reaction of ferrous ion with the thioarsenite from carbonation process can result in the formation of arsenopyrite which is a common mineral in arsenic-rich aquifers.

Measurements of oxidation−reduction potential (E h) and concentrations of dissolved hydrogen (H2) were made in a shallow groundwater system contaminated with solvents and jet fuel to delineate the zonation of redox processes. E h measurements ranged from +69 to −158 mV in a cross section of the contaminated plume and accurately delineated oxic from anoxic groundwater. Plotting measured E h and pH values on an equilibrium stability diagram indicated that Fe(III) reduction was the predominant redox process in the anoxic zone and did not indicate the presence of methanogenesis and sulfate reduction. In contrast, measurements of H2 concentrations indicated that methanogenesis predominated in heavily contaminated sediments near the water table surface (H2 ∼ 7.0 nM) and that the methanogenic zone was surrounded by distinct sulfate-reducing (H2 ∼ 1−4 nM) and Fe(III)-reducing (H2 ∼ 0.1−0.8 nM) zones. The presence of methanogenesis, sulfate reduction, and Fe(III) reduction was confirmed by the distribution of dissolved oxygen, sulfate, Fe(II), and methane in groundwater. These results show that H2 concentrations were more useful for identifying anoxic redox processes than E h measurements in this groundwater system. However, H2-based redox zone delineations are more reliable when H2 concentrations are interpreted in the context of electron-acceptor (oxygen, nitrate, sulfate) availability and the presence of final products [Fe(II), sulfide, methane] of microbial metabolism.