A comparison of different methods for isolation of vesicular-arbuscular mycorrhizal (VAM) fungi into open-pot cultures was undertaken as part of a study of the diversity of these fungi. Four different isolation techniques using spores separated from soil, soil trap cultures, root samples, or transplanted seedlings grown in intact soil cores were used to obtain as many fungi as possible from each site. Isolation methods were compared using paired samples from the same locations within natural (savanna, rocky hill, wetland, rainforest) and disturbed (minesite) habitats in a seasonally dry tropical region in the Northern Territory of Australia. There were large differences in (i) the efficiency (rate of increase in mycorrhizal colonisation), (ii) the proportion of successful cultures, (iii) fungal diversity (number of fungal species in each culture) and (iv) specificity (identity of species isolated) between these four procedures. However, the less-efficient procedures generally resulted in a higher proportion of cultures of one fungus, which could be used without further isolation steps. Most species of Scutellospora, Acaulospora and Gigaspora were obtained primarily from field-collected spores, but only 50% of these culture attempts were successful. Spores from these initial cultures produced mycorrhizas much more rapidly and successfully when used to start second-generation cultures. Several species of fungi, rarely recovered as Living spores from field soils, were dominant in many trap cultures started from soil or roots. Most of these fungi were Glomus species, that were first distinguished by colonisation patterns in roots and eventually identified after sporulation in second- or third-generation trap cultures. These experiments demonstrated that glomalean fungi in the habitats sampled belonged to two functional categories, based on whether or not spores were important propagules. The "non-sporulating" fungi were dominant in many trap cultures, which suggests that these fungi had higher total inoculum levels in soils than other fungi. Pot-culturing methods provided additional information on fungal diversity which complemented spore occurrence data obtained using the same soil samples and provided valuable new information about the biology of these fungi.