Abstract An area of dry grassland in New Zealand, comprising an equal mixture of native and exotic species, was subject to perturbations of irrigation, fertilization and cessation of grazing. The vegetation response was recorded for 3 years. Total cover, and the contribution of native species to that cover, fluctuated between years even in the control plots. Irrigation increased total cover, but decreased the cover of native species. Fertilization produced the same effects, only less strongly, and also reduced species richness, the loss being in native species. In spite of overall effects of treatments on native and exotic cover, when individual species’ responses to irrigation, fertilization or exclosure were calculated, there was no significant difference between the native and exotic plant guilds. Species differed in their responses, but the native and exotic guilds overlapped. When grouped by morphology, the only significant difference between the responses to perturbation was that forbs and graminoids responded more positively to irrigation than woody and cryptogamic species. The realized responses of the species to the perturbations described here showed little correlation with their physiological responses as determined in previous greenhouse experiments. It is suggested that the realized responses are strongly, and currently unpredictably, influenced by competition from the other species present. Soil nutrients and soil water were both important controls on the community. The relative similarity in the nature of the response to these two factors – nutrients and water – suggests that they affect species in similar ways, possibly because the greater growth rate of the exotic species mediates the short-term response to both. Grazing has less effect on the current community than either nutrients or water, although it may have been historically important in shaping the species pool. From the poor predictability of field responses from morphological guilds or from ecophysiological responses, it is suggested that the ‘functional types’ approach, although conceptually attractive, lacks experimental support in these grasslands. It is concluded that the exotic species have invaded by being pre-adapted to the environment with the same environmental responses as the natives, but with the advantage of generally higher growth rates.