Ant assemblages are focal ecological indicators of progress in mine-site restoration, often showing increasing species richness with restoration age. Certain functional groups also behave in predictable ways in response to disturbance and changes in the environment. Whether these ant responses can be applied to other types of restoration and ecosystems is unknown, especially in dynamic environments and where gradients may not be as severe as in mine-site restoration. Ant assemblages would be expected to perform poorly as ecological indicators in dynamic environments because such environs are subject to periodic disturbance of important habitat features. Indeed, periodic disturbance may limit the predictive power of any ecological indicator. In this study, we trapped ants on two separate occasions to compare ant assemblages among four riparian habitat types (Unplanted grassland, Young revegetation, Older revegetation and Mature woodland). These habitat types were assumed to represent progressive stages of restoration. In contrast to the findings of others, species richness was variable among replicate locations of the same habitat type, and did not differ among the four habitat types. Also in contrast to what others have found for functional groups, dolichoderines were equally abundant in all habitat types and did not decrease in abundance with vegetation maturity. While generalized myrmicines and opportunists became more common with maturation of the vegetation, they did not replace dolichoderines as the most common ants. Surprisingly, the relative abundance of Subordinate Camponotini, a functional group considered to be of limited use in discriminating structural types, increased across the restoration gradient. There were also fairly distinct species assemblages associated with unplanted grassland and mature woodland. Communities in revegetated habitats were intermediate of these extremes, suggesting there is a level of predictiveness to their response to revegetation in this system. While species richness and a functional group approach would be of little use in this environment, species composition would provide a useful gauge of restoration progress. Ant species richness and functional group metrics have repeatedly been advocated as ecological indicators. Given our results, we caution against the blind application of metrics that have not been validated in the context in which they are to be applied.