The herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) was applied to soils in microcosms, and degradation was monitored after each of five repeated additions. Total DNAs were isolated from soil bacterial communities after each 2,4-D treatment. The DNA samples were analyzed on slot blots and Southern blots by using a tfdA gene probe subcloned from plasmid pJP4 and a Spa probe derived from a different 2,4-D-degrading isolate, a Sphingomonas paucimobilis strain. 2,4-D applied to soil was quickly degraded by indigenous microbial populations. As determined by slot blot analyses of DNA from a Michigan soil, the increase in hybridization signal in response to 2,4-D treatments was greater with the Spa probe than with the tfdA probe. In contrast, the DNA from a Saskatchewan soil exhibited an increase in hybridization signal with the tfdA probe. This indicated that a population with 2,4-D-degradative gene sequences different from the tfdA gene sequence was dominant in the Michigan site, but not in the Saskatchewan site. A Southern blot analysis of DNA from Michigan soil showed that the dominant 2,4-D-degrading population was S. paucimobilis 1443. A less dominant 2,4-D-degrading population was detected with the tfdA probe; further analysis revealed that this population was a Pseudomonas pickettii 712. These gene probe analyses revealed that an important population carrying out 2,4-D degradation was not detected when the canonical tfdA gene probe was used. After a series of new strains were isolated, we identified a probe to detect and identify the dominant members of this new group.