The angiosperm radiation played a dual role in the diversification of insects and insect pollinators

Abstract

Interactions with angiosperms have been hypothesised to play a crucial role in driving diversification among insects, with a particular emphasis on pollinator insects. However, support for coevolutionary diversification in insect-plant interactions is weak. Macroevolutionary studies of insect and plant diversities support the hypothesis that angiosperms diversified after a peak in insect diversity in the Early Cretaceous. Here, we used the family-level fossil record of insects as a whole, and insect pollinator families in particular, to estimate diversification rates and the role of angiosperms on insect macroevolutionary history using a Bayesian process-based approach. We found that angiosperms played a dual role that changed through time, mitigating insect extinction in the Cretaceous and promoting insect origination in the Cenozoic, which is also recovered for insect pollinator families only. Although insects pollinated gymnosperms before the angiosperm radiation, a radiation of new pollinator lineages began as angiosperm lineages increased, particularly significant after 50 Ma. We also found that global temperature, increases in insect diversity, and spore plants were strongly correlated with origination and extinction rates, suggesting that multiple drivers influenced insect diversification and arguing for the investigation of different explanatory variables in further studies.

Fig. 1. Different pollination modes showing different groups of insects under distinctive patterns of extinction, survival, and origination following the gymnosperm-angiosperm transition.

The fossil pollination cases reviewed in refs. ^,. The periods of Angiosperm Radiation, Cretaceous Terrestrial Revolution, and Angiosperm Terrestrial Revolution are marked. A Diverse fossil community of Cretaceous pollinators and the lifespan of these families. B Representation of the transition from gymnosperm-associated insect families to angiosperms-associated insect families, illustrating examples of known pollinators groups in the different transitional situations. Data extracted from ref. .

Fig. 2. Accumulated diversification of insect families through geological time estimated using the software PyRate and the family-level fossil dataset of ref. compared with the angiosperm relative diversity.

Dark green dotted lines indicate the Permian–Triassic (P–T), Triassic–Jurassic (T–J), and the Cretaceous–Paleogene (K–Pg) mass extinctions. The periods of Angiosperm Radiation, Cretaceous Terrestrial Revolution, and Angiosperm Terrestrial Revolution are marked. A All insect diversity, specifically the pollinator insect diversity, in the left axis; the angiosperm relative diversity from ref. , put in shape in ref. , in the right axis. B Family accumulation through time for five selected insect orders.

Fig. 3. Correlation trends of different analysed drivers for origination (in blue) and extinction (in red) rates on insect diversity for two periods of time: the Angiosperm Terrestrial Revolution timeframe (100–50 Ma), and for the Angiosperm Dominance period (50–0 Ma).

The results for the same drivers analysing only the pollinator insect families, including the possible pollinator lineages. Data used in this representation represent median estimates and the 95% CI are presented in Supplementary Data 1–2. Drivers from top to bottom are Global mean temperature (Temperature), Spore plant relative diversity (Spore plants), Gymnosperm relative diversity (Gymnosperms), Continental fragmentation, Angiosperm relative diversity (Angiosperms), and Insect family richness (Insect richness). Asterisks indicate significant correlations recovered with the MBD model (shrinkage weight >0.5 and 95% CI not overlapping with zero). If any of the dates is out of their corresponding scale, it represents their value inside the box. The x-axes have no unit scale for correlation parameters in the MBD analyses.

Fig. 4. Correlation trends of different analysed drivers for origination (in blue) and extinction (in red) rates on insect diversity divided by five selected orders, for two periods of time: the Angiosperm Terrestrial Revolution timeframe (100–50 Ma), and for the Angiosperm Dominance period (50–0 Ma).

Data used in this representation represent median estimates and the 95% CI is presented in the Supplementary Data 3. Drivers from top to bottom are Global mean temperature (Temperature), Spore plant relative diversity (Spore plants), Gymnosperm relative diversity (Gymnosperms), Continental fragmentation, Angiosperm relative diversity (Angiosperms), and Insect family richness (Insect richness). Asterisks indicate significant correlations recovered with the MBD model (shrinkage weight >0.5 and 95% CI not overlapping with zero). If any of the dates is out of their corresponding scale, it is represented their value inside the box. The x-axes have no unit scale for correlation parameters in the MBD analyses.