New calculations indicate that 90% of flowering plant species are animal-pollinated


 The number of angiosperms that may depend on animals for pollination has been difficult to estimate because the pollination modes of tens of thousands of species of tall trees, canopy climbers, and epiphytes, especially from the tropics, remain unknown. Ecological research over the past 50 years, however, has established a strong correlation between pollination by wind or water and particular flower traits. Based on a literature search for pollination-ecological data for the families, genera, and species compiled by the Global Biological Information Facility (GBIF) and the World Flora Online (WFO) databases, we estimated the number of species with abiotic pollination and then subtracted it from the total number of taxa. The results show that based on GBIF, 68% of 465 families, 88% of 14,437 genera, and 90% of 332,341 species rely on animal pollination. Based on WFO, 64% of 414 families, 88% of 13772 genera, and 90% of 339,876 species rely on animal pollination. Species exhibiting both pollination modes were identified in about 48 families and 0.1% of species. Finding that 90% of all angiosperm species depend on animals for cross pollination reinforces the importance of prioritizing the conservation of plant-pollinator interactions in both natural and agricultural ecosystems. These results match a previous estimate (87.5% or, after a mathematical correction, 89.5%) obtained by extrapolation from local community studies and also the finding that insects have pollinated flowering plants for most of angiosperm evolutionary history, a consilience of results that supports the validity of these estimates.

The services of animal pollinators that deliver compatible pollen to receptive stigmas are essential for the reproduction of many flowering plants [ 1 , 2 ].The maintenance of plant-pollinator interactions, therefore, is the basis for the function and sustainability of healthy ecosystems.For this reason, there has been widespread concern over the past two decades about the global decline in pollinators and the effects of this loss on pollination services for both wild plants and crops [ 3 , 4 ].One fundamental question relevant to the ongoing pollinator loss remains only partially answered, namely: What proportion of the world's angiosperm species are pollinated by animals?The best estimate so far is that of Ollerton and his colleagues [ 5 ] who inferred that between 85% and 87.5% of the ca.352 0 0 0 species of flowering plants are animal-pollinated.These proportions were obtained by combining mean values of animal-pollinated percentages of flowering plants reported in surveys of pollination systems in various communities.Together, these communities included 3918 plant species, ranging from 15 to 391 species per community ( [ 5 ]: Appendix 1).The higher proportion (87.5%) reflects weighting by species diversity across the three latitudinal belts in which the investigated communities were located (tropical, subtropical, temperate), while the lower value (85%) reflects the unweighted mean.
We use a different approach to estimate the percentage of flowering plants that depend on animals for their pollina-tion.Our estimate is based on publications on pollination modes and the insight that pollination by wind or water, that is, abiotic pollination, is correlated with a suite of floral traits that form the two best-established floral syndromes in pollination ecology [ 1 ].We therefore first calculated the number of families, genera, and species that are wind or water pollinated, modes that can reliably be inferred from a species' morphology, and we then treated the remaining angiosperms as relying on animals for pollination.Separately, we also re-analyzed Ollerton et al. 's [ 5 ] data.
When the data were plotted on the WFO (World Flora Online) taxon list, 103 of 414 families are abiotically pollinated and 64% of families are animal-pollinated.Of 13 772 genera, 1548 are abiotically pollinated, and 90 include species with both biotic and abiotic pollination, which, based on our formula, yields 88% exclusively animal-pollinated genera.Of 339 876 angiosperm species in the WFO database, 32 129 are abiotical ly pol linated, and 267 species utilize both biotic and abiotic pollination.The percentage of animalpollinated flowering species is thus 90% [1 −(32 129 + 267)/339 876] ( Table S2 ; Fig. 1 ).We also present the congruence between GBIF, WFO and APG IV (Angiosperm Phylogeny Group IV) in Table S3 .
After extending and re-analyzing Ollerton et al. 's [ 5 ] data with sample-size weighting, the calculation of the global mean changed from 85% (in the original study) to 88.5% of angiosperms being biotical ly pol linated.After correcting for both sample size and latitudinal diversity patterns, the value changed from 87.5% to 89.5% of angiosperms being biotically pollinated.
Our finding that 90% of floweringplant species are animal-pollinated depends on the corrected inference of the number of wind-or water-pollinated species and the assumed total number of angiosperm species.Neither of these two numbers wi l l ever be final because of progress in plant discovery and field work, and because of changing species concepts.Recent estimates of the number of angiosperm species range from 295 383 to 369 434, in each   S3 ) and the use of the most complete angiosperm name databases GBIF and WFO (accessed in February 2023).Table S4 shows the differences in family circumscription between GBIF, WFO and APG IV.
case based on the best efforts of taxonomists at the Royal Botanic Gardens, Kew [ 6 , 7 ].The total numbers of flowering plant species used in our study, namely 332 341 when using the GBIF or 339 876 when using the WFO, fall between these two values, and our results show that percentages resulting from these two data sources (namely 90%) are identical.
Thirty years ago, it was estimated that 23% (84 of 365) families and 8% (1106 of 13 500) genera are abiotically pollinated [ 8 ], while a slightly more recent estimate gave 20% of families as abiotical ly pol linated [ 9 ].Our estimates are that 33%-36% of families and 11% of genera are pollinated by wind or water or are ambophilous.Our higher percentages of wind-pollinated families and genera may be explained by new insights on wind-pollinated genera as listed in Table S4 and perhaps also by our addition of ambophily to wind pollination.Even so, ambophily may be underestimated (supplementary discussion and Fig. S1 ).
A benchmark study on the question of which percentage of angiosperms might be animal-pollinated is that of Ollerton et al. [ 5 ], who used a communitylevel approach for calculation.Their communities included 3918 species, slightly more than 1% of the estimated number of angiosperms (352 0 0 0 at the time), which resulted in an estimate of 87.5% animal-pollinated angiosperms, while we inferred that 90% of the angiosperms rely on animal pollinators.Even closer correspondence between the results of Ollerton et al. [ 5 ] and our approach resulted when we extended and corrected their data, which yielded 89.5% of angiosperms as animal-pollinated.This degree of consilience suggests that the % of angiosperm species that are biotically pollinated is indeed about 90%.This fits nicely with the finding that insects have pollinated angiosperms for approximately 86% of this plant lineage's evolutionary history [ 10 ] and pinpoints the need for especially the conservation of insects.

Figure 1 .
Figure 1.Percentages of biotic (BP), abiotic (ABP), and ambophilous (Dual) pollination based on the results of our literature search ( TableS3) and the use of the most complete angiosperm name databases GBIF and WFO (accessed in February 2023).TableS4shows the differences in family circumscription between GBIF, WFO and APG IV.