Summary Model denitrifiers convert NO3− to N2, but it appears that a significant fraction of natural populations are truncated, conducting only one or two steps of the pathway. To better understand the diversity of partial denitrifiers in soil and whether discrepancies arise between the presence of known N‐oxide reductase genes and phenotypic features, bacteria able to reduce NO3− to NO2− were isolated from soil, N‐oxide gas products were measured for eight isolates, and six were genome sequenced. Gas phase analyses revealed that two were complete denitrifiers, which genome sequencing corroborated. The remaining six accumulated NO and N2O to varying degrees and genome sequencing of four indicated that two isolates held genes encoding nitrate reductase as the only dissimilatory N‐oxide reductase, one contained genes for both nitrate and nitric oxide reductase, and one had nitrate and nitrite reductase. The results demonstrated that N‐oxide production was not always predicted by the genetic potential and suggested that partial denitrifiers could be readily isolated among soil bacteria. This supported the hypothesis that each N‐oxide reductase could provide a selectable benefit on its own, and therefore, reduction of nitrate to dinitrogen may not be obligatorily linked to complete denitrifiers but instead a consequence of a functionally diverse community.