Culture-dependent DGGE (CD-DGGE) fingerprinting 16S rRNA gene can be used to characterize mixed bacterial communities recovered on agar plates; however, few studies have systematically evaluated the utility of this approach for soil populations. Using R2A agar as a growth medium, CD-DGGE analysis of 16S rDNA resulted in clear banding patterns of sufficient complexity (16-32 major bands) to investigate differences in microbial communities in a silt loam soil (fine loamy, mixed, superactive, mesic, Oxyaquic Hapludalfs). There were significant differences in the DGGE profiles between the 10^-3 and 10^-6 soil dilutions (R=0.409, P < 0.05), but no significant differences between the 10^-4 and 10^-5 dilutions (R=0.250, P < 0.07) as determined by ANOSIM, a non-parametric multivariate test. Replicate CD-DGGE profiles from the least dilute plates were more similar (72-77%) than those from the most dilute plates (51-61%). Use of different culture media resulted in distinct community fingerprints (average similarity 44%; R=0.981, P < 0.001) as did pasteurization of soil and anaerobic incubation of culture plates (average similarity 29%; R=1.000, P < 0.001). Bacteria growing on low-nutrient media (R2A and oil agar) were highly similar, but were dissimilar to bacteria growing on media designed for the culture of copiotrophs.  We compared culture-independent DGGE (CI-DGGE) to CD-DGGE. When all communities cultured using different media and incubation conditions were included in the comparison the number of bands detected unique to CD analysis (40) was comparable to that found only in the CI profile (37).  The number of bands shared between both methods was 39. The ability of CI and CD methods to resolve distinct members of the soil bacterial community demonstrates that a more comprehensive profile of soil bacterial communities is achieved when the two methods are used in combination.