Over half of all described species are insects, but until recently our understanding of the reasons for this diversity was based on very little macroevolutionary evidence. Here I summarize the hypotheses that have been posed, tests of these hypotheses and their results, and hence identify gaps in knowledge for future researchers to pursue. I focus on inferences from the following sources: (i) the fossil record, normally at family level, and (ii) insect phylogenies, sometimes combined with: (iii) the species richness of insect higher taxa, and (iv) current extinction risks.
There is evidence that the species richness of insects has been enhanced by: (i) their relative age, giving time for diversification to take place; (ii) low extinction rates. There is little evidence that rates of origination have generally been high or that there are limits on numbers of species. However, the evidence on macroevolutionary rates is not yet so extensive or coherent as to present unequivocal messages.
As regards morphological, ecological, or behavioural hypotheses, there is evidence that diversity has been enhanced by (iii) flight or properties resulting from it like enhanced dispersal, (iv) wing folding, and (v) complete metamorphosis. However, in all these cases the evidence is somewhat equivocal, either because of statistical issues or because evidence from different sources is conflicting.
There is extensive evidence that diversity is affected by (vi) the ecological niche. Comparative studies indicate that phytophagy generally increases net diversification rates, and reduces extinction risk. However, niche specialization is also associated with an increase in extinction risk. Small body size (vii) is often associated with low extinction risk in comparative studies, but as yet there is no solid evidence that it consistently enhances net rates of diversification. Mouthpart diversity (viii) has generally increased over time in the insects, but cannot explain the apparent great increase in diversity seen in the Cretaceous and Tertiary. Sexual selection and sexual conflict (ix) are two processes that are widespread in insects, and there is comparative evidence linking both to increased diversification. Although some comparative evidence links tropical distributions (x) to increased rates of diversification, the extent to which latitudinal richness gradients are unusual in insects is equivocal.
There is little to no direct evidence from fossils and phylogenies that insect diversity has generally been affected by (i) sensory- or neuro-sophistication, (ii) population size or density, (iii) generation time or fecundity, (iv) the presence of an exoskeleton or cuticle, (v) segmentation or appendage diversity, (vi) adaptability or genetic versatility, though all of these remain plausible hypotheses awaiting further tests. The data suggest that the insect body ground plan itself had no direct effect on insect diversity.
Thus, whilst studies to date have given substantial understanding, substantial gaps still remain. Future challenges include: (i) interpreting conflicting messages from different sources of data; (ii) rating the importance of different hypotheses that are statistically supported; (iii) linking specific proximate to specific ultimate explanations and vice versa; and (iv) understanding how different ultimate hypotheses might be dependent on each other.