Competition experiments on alien weeds with crops: lessons for measuring plant invasion impact?

Can we quantify the impact of invasive species? Here we use the per-plant competitiveness of alien weeds on crops as a model of invasive species impact in general. We reviewed 97 weed-crop competition experiments in 32 papers that included 30 alien weed and 14 crop species. The majority (68.92%) were randomised block designs where the alien weed had been either added (additive experiments) or removed (removal experiments). We propose using the relative competition index to estimate the effect of alien species in all systems, specifying in each case the density and proportion of alien and native plants essayed. We found that the impact of the weed cannot be considered independently of the crop and, thus, we should be cautious in ranking weed species according to their competition effect. A similar situation can be postulated for alien plants interfering with native species. Invaded communities are not random assemblages, and researchers tend to study the most competitive alien plants. We also found that the effect of the weed on crop yield depends on the duration of the interference and the life-history stage of the weed-crop system at which the interaction takes place. We were not able to conduct a more rigorous comparative analysis of the impacts. Such as a meta-analysis. To do this would require some measure of the variation of the competition effect such as standard deviation or standard error, which we found are almost never reported.