Suppression of the activity of arbuscular mycorrhizal fungi (AMF) in unsterile soils, as determined by transfer of 33P from soils to plants, has been recently demonstrated, and soils differ in their degree of suppressiveness. The previous experiments included only one R. irregularis isolate, BEG87. The aim of this study was to analyze the sensitivity to suppression of different cultivated AMF isolates as well of native AMF communities using shoot P-33 content as a proxy of AMF activity in mesh-enclosed patches of unsterile soil. We found that the suppression of a panel of Rhizophagus isolates differed from the suppression of R. irregularis BEG87. Suppression overall correlated with low soil pH and high soil phosphorus, but there were large differences between the isolates. Some Rhizophagus isolates were less sensitive to soil suppression than BEG87, and tolerance to suppression may be a relevant property to consider when choosing future AMF inoculants. For native AMF communities showing high root colonization, we found one example of a community that was not suppressed by its 'home' soil, while BEG87 activity in that soil was completely suppressed. Hence, some AMF communities appear to be reservoirs for inoculants with high tolerance towards suppression.