The enemy release hypothesis is a common explanation for species invasions, suggesting that introduced species benefit from leaving behind natural enemies in the native range. However, any such advantage may attenuate over time. In this study, we test a prediction of this more dynamic enemy release hypothesis: that non-native plant species that became established longer ago exhibit stronger negative feedbacks with the soil. Consistent with declining enemy release over time, we found increasingly negative soil feedbacks for species established longer ago in New Zealand. egative soil feedbacks were also stronger for more widespread species, but weaker for more locally abundant species, suggesting that species accumulate negative interactions as they spread and can be locally regulated by these interactions. We also present data to support the common assumption that relatives have similar impacts on and responses to soil communities. Together, these data highlight the dynamic nature of novel interactions arising from species introductions.