The data needed to derive an accurate estimate of saprobic microfungi are insufficient, incomplete and contradictory. We therefore address issues that will ultimately reveal whether there are 1.5 million global fungal species, which is the generally accepted working estimate. Our data indicates that large numbers of fungi occur on host families, such as Musaceae, host genera such as Nothofagus and individual host species such as Eucalyptus globulus, and that fungi may be specific or recurrent on different plant groups. Recent studies have shown that fungal numbers on hosts may be larger than originally thought as saprobes are organ-specific/-recurrent and changes in fungal communities occur as substrata decays. Other issues, such as the impact of geography, of methodology and of taxonomy are also addressed. There is evidence that fungi on the same host at different locations also differs; site-specific factors and geographic distance may be more important than host/substrate in shaping fungal assemblages. Methodology impacts on estimates of species diversity with many more taxa observed using indirect isolation protocols as compared to direct isolations from leaves. Our understanding of fungal species numbers in speciose genera is important. In some fungal groups accepted species have been reduced to a few species, while in other groups many cryptic species are being uncovered. While we make a number of generalisations from the studies reported here, this review also highlights some of the limitations mycologists currently have to contend with. A large body of knowledge exists for certain groups of microfungi or for microfungi occurring on certain substrata/hosts. However, it is likely that we are drawing conclusions from data that are somewhat biased toward fungi and host/substrata that are of interest to human endeavours. The discrepancy between the numbers of fungi described from only one economically important genus, Eucalyptus, and all the other members of the Myrtaceae is but one example of this bias. By incorporating the large body of work that is already available and adding appropriate complementary studies, we can accelerate our understanding of microfungal diversity and this will eventually lead us to a realistic estimate of global fungal species numbers.