Clade age and species richness are decoupled across the eukaryotic tree of life

Abstract

Explaining the dramatic variation in species richness across the tree of life remains a key challenge in evolutionary biology. At the largest phylogenetic scales, the extreme heterogeneity in species richness observed among different groups of organisms is almost certainly a function of many complex and interdependent factors. However, the most fundamental expectation in macroevolutionary studies is simply that species richness in extant clades should be correlated with clade age: all things being equal, older clades will have had more time for diversity to accumulate than younger clades. Here, we test the relationship between stem clade age and species richness across 1,397 major clades of multicellular eukaryotes that collectively account for more than 1.2 million described species. We find no evidence that clade age predicts species richness at this scale. We demonstrate that this decoupling of age and richness is unlikely to result from variation in net diversification rates among clades. At the largest phylogenetic scales, contemporary patterns of species richness are inconsistent with unbounded diversity increase through time. These results imply that a fundamentally different interpretative paradigm may be needed in the study of phylogenetic diversity patterns in many groups of organisms.

Figure 1. Phylogenetic distribution of species richness across the eukaryotic tree of life.

(A) Time-calibrated tree of 1,397 clades of multicellular eukaryotes; length of gray bars indicates relative log-transformed species richness of each group. (B) Total species richness of major groups. Clade colors in (A) correspond to names in (B).

Figure 2. Clade age and species richness are unrelated across 1,397 clades of multicellular eukaryotes.

(A) Relationship between log(richness) and clade age (PGLS β = 0.0008, p = 0.66). (B) Same relationship as (A), but fitted model is projected onto logarithmic timescale to better visualize the relationship among age and richness for younger clades. The regression coefficient β represents the change in log-transformed diversity per million years.

Figure 3. Relationships between age and richness within 12 major taxonomic groups for which dense subclade sampling was available as part of the timetree project .

Lines represent fitted PGLS relationships between log(richness) and clade age. Beetles show a significant age-diversity relationship (β = 0.017, p = 0.004). However, all slopes are less than expected under both relaxed-rate and phylogenetic-rate models of among-clade heterogeneity in net diversification rates (Table 1).

Figure 4. Distributions of rank-order correlations between clade age and species richness predicted under MEDUSA model of rate variation for 12 major taxonomic groups.

Vertical red lines show the observed correlation for each group. Observed correlations are significantly lower than the corresponding model-predicted value for 10 of the 12 groups. The high variance of the MEDUSA-predicted distributions for gymnosperms and actinopterygiians is largely explained by the small number of clades (N = 12) available for those groups (Figure S5).