Geographical and life-history traits associated with low and high species richness across angiosperm families

Abstract

Introduction: The phenomenal expansion of angiosperms has prompted many investigations into the factors driving their diversification, but there remain significant gaps in our understanding of flowering plant species diversity.

Methods: Using the crown age of families from five studies, we used a maximum likelihood approach to classify families as having poor, predicted or high species richness (SR) using strict consensus criteria. Using these categories, we looked for associations between family SR and i) the presence of an inferred familial ancestral polyploidization event, ii) 23 life history and floral traits compiled from previously published datasets and papers, and iii) sexual system (dioecy) or genetically determined self-incompatibility (SI) mating system using an updated version of our own database and iv) geographic distribution using a new database describing the global distribution of plant species/families across realms and biomes and inferred range.

Results: We find that more than a third of angiosperm families (65%) had predicted SR, a large proportion (30.2%) were species poor, while few (4.8%) had high SR. Families with poor SR were less likely to have undergone an ancestral polyploidization event, exhibited deficits in diverse traits, and were more likely to have unknown breeding systems and to be found in only one or few biomes and realms, especially the Afrotropics or Australasia. On the other hand, families with high SR were more likely to have animal mediated pollination or dispersal, are enriched for epiphytes and taxa with an annual life history, and were more likely to harbour sporophytic SI systems. Mapping the global distribution of georeferenced taxa by their family DR, we find evidence of regions dominated by taxa from lineages with high vs low SR.

Discussion: These results are discussed within the context of the literature describing "depauperons" and the factors contributing to low and high biodiversity in angiosperm clades.

Keywords: depauperons; diversification rates; floral traits; growth habit; life-history traits; self-incompatibility; whole genome duplication event.

Figure 1

The maximum likelihood estimate, MLE (diamonds), and minimum and maximum confidence interval for the net diversification rate, r, and the speciation rate, λ, estimated using the method of Bokma (2003) based on the species richness of families given by APG IV and the crown age of families estimated by Hernández-Hernández and Wiens (2020); Wikström et al. (2001); Bell et al. (2010); Li et al. (2019) , and Ramírez-Barahona et al. (2020).

Figure 2

(A–C) Venn diagrams showing the levels of agreement between classification of angiosperm families with low (A), expected (B), or high (C) levels of species richness as inferred from five datasets. The strict consensus is given in the center. (D) Violin plot of the median age of angiosperm families across the five phylogenetic studies employed in the analysis by family classification category (poor, predicted, or rich) based on strict consensus criterion. The distribution of the median crown ages was not significantly different for families with poor, predicted, or high SR using the Kolmogorov–Smirnov test.

Figure 3

Families categorized as having high, predicted, or low SR (left panel) or DR (right panel) by family crown age based on the datasets from Wikström et al. (2001); Hernández-Hernández and Wiens (2020), and Li et al. (2019).

Figure 4

Ultrametric molecular phylogeny of angiosperm families based on Hernández-Hernández and Wiens (2020). The family name and external branches are colored by the inferred species richness status, estimated using the MLE modeling described in the methods. The root of the phylogeny (Amborellaceae) is 132 mybp; concentric circles are placed every 14 mybp to the present. (A) Families with poor species richness based on majority consensus (4/5 studies in agreement) criterion. (B) Families with high species richness based on majority consensus (4/5 studies in agreement) criterion. Solid lines indicate families for which the SR category was declared by strict consensus, and dotted lines denote families for which the SR category was declared based on majority consensus.

Figure 5

Distribution of nine traits across families with poor, predicted, or high species richness based on strict consensus. Differences in the proportion of species per family between SR categories were analyzed using a Kruskal–Wallis, non-parametric rank test, at an FDR = 0.1. The remaining 11 traits with non-significant differences among SR categories are shown in the Supplementary Data File 3 ( Figure S2 ).

Figure 6

(A) The mean diversification rate (DR) of georeferenced taxa from families with poor SR (top) or high SR (bottom). The colour of each pixel on the map represents the global mean DR of the families represented by all taxa georeferenced to that location. The corresponding figure for families with predicted SR is in the supplementary data ( Data File 3 , Figure S3 ). (B) Global distribution of taxa in families with poor, predicted and high species richness. The geographic distribution of the SR of taxa in families classified as having poor (top left), predicted (middle left) or high (bottom left) SR. Intensity of red scale indicates the number of taxa/family in that geographic region.