The branching times of molecular phylogenies allow us to infer speciation and extinction dynamics even when fossils are absent. Troublingly, phylogenetic approaches usually return estimates of zero extinction, conflicting with fossil evidence. Phylogenies and fossils do agree, however, that there are often limits to diversity. Here, we present a general approach to evaluate the likelihood of a phylogeny under a model that accommodates diversity-dependence and extinction. We find, by likelihood maximization, that extinction is estimated most precisely if the rate of increase in the number of lineages in the phylogeny saturates towards the present or first decreases and then increases. We demonstrate the utility and limits of our approach by applying it to the phylogenies for two cases where a fossil record exists (Cetacea and Cenozoic macroperforate planktonic foraminifera) and to three radiations lacking fossil evidence (Dendroica, Plethodon and Heliconius). We propose that the diversity-dependence model with extinction be used as the standard model for macro-evolutionary dynamics because of its biological realism and flexibility.

Ecological change provokes speciation and extinction, but our knowl-1 edge of the interplay among the biotic and abiotic drivers of macroevo-2 lution remains limited. Using the unparalleled fossil record of Ceno-3 zoic macroperforate planktonic foraminifera, we demonstrate that 4 macroevolutionary dynamics depend upon the interaction between 5 species' ecology and the changing climate. This interplay drives 6 diversification, but di↵ers between speciation probability and ex-7 tinction risk: speciation was more strongly shaped by diversity-8 dependence than by climate change, whereas the reverse was true for 9 extinction. Crucially, no single ecology was optimal in all environ-10 ments and species with distinct ecologies had significantly di↵erent 11 probabilities of speciation and extinction. The ensuing macroevolu-12 tionary dynamics depend fundamentally on the ecological structure 13 of species' assemblages. 14The wide-ranging mechanisms that generate and maintain biodiversity have been grouped has long been recognized as fundamental for regulating diversity (6, 7), progress towards 25 understanding their interaction has been slow (2). The fossil record's incompleteness often 26 necessitates temporally and taxonomically coarse paleontological analyses (7-10), while 27 molecular phylogenies are restricted to extant species and therefore o↵er little insight into 28 extinction (11, 12 47The macroperforate clade has diversified from two species that survived the end- 48Cretaceous mass extinction into 32 morphologically distinct species today, though the rise 49 in diversity has been far from smooth ( Fig. 1). The sharpest fall in diversity was during 50 the Eocene -Oligocene transition, when rapid global cooling led to the development of 51 the Antarctic ice cap (15). This suggests that climate change has been important in 52 macroevolution so we approximated the complex, multifacated climate system using the 53 mean and variability of oxygen isotopic composition of deep sea carbonates (15, 18). 54Clade growth (ln( N t+1 Nt ), detrended, N t is the number of species in each 1 million year bin 55 t) was poorly predicted by climate (18, Fig. 2A, Table S1). Models based on diversity-56 dependence only used diversity at the start of each bin and assume a constant limit to niche 57 availability (a taxonomic analogue to a demographic "carrying capacity"), but predictions 58 were similarly poor (Fig. 2B, Table S1). The clade's macroevolutionary dynamics are 59 therefore not well-predicted by either a strictly abiotic ( Fig. 2A) (Fig. 1), with the relative dominance of each group apparently waxing 77 and waning with the changing climate. 78Ecology is more strongly predictive of clade growth than either climate or diversity, but 79 model fit is at best moderate (Fig. 2C, Table S1). Models containing interactions among 80 pairs of these variables are significantly and substantially better, but model support was 81 strong only when species with distinct ecologies were permitted to respond di↵erently 82 5 to changes in diver...

We present a complete phylogeny of macroperforate planktonic foraminifer species of the Cenozoic Era (∼65 million years ago to present). The phylogeny is developed from a large body of palaeontological work that details the evolutionary relationships and stratigraphic (time) distributions of species-level taxa identified from morphology ('morphospecies'). Morphospecies are assigned to morphogroups and ecogroups depending on test morphology and inferred habitat, respectively. Because gradual evolution is well documented in this clade, we have identified many instances of morphospecies intergrading over time, allowing us to eliminate 'pseudospeciation' and 'pseudoextinction' from the record and thereby permit the construction of a more natural phylogeny based on inferred biological lineages. Each cladogenetic event is determined as either budding or bifurcating depending on the pattern of morphological change at the time of branching. This lineage phylogeny provides palaeontologically calibrated ages for each divergence that are entirely independent of molecular data. The tree provides a model system for macroevolutionary studies in the fossil record addressing questions of speciation, extinction, and rates and patterns of evolution.

Planktonic foraminiferal species identification is central to many paleoceanographic studies, from selecting species for geochemical research to elucidating the biotic dynamics of microfossil communities relevant to physical oceanographic processes and interconnected phenomena such as climate change. However, few resources exist to train students in the difficult task of discerning amongst closely related species, resulting in diverging taxonomic schools that differ in species concepts and boundaries. This problem is exacerbated by the limited number of taxonomic experts. Here we document our initial progress toward removing these confounding and/or rate-limiting factors by generating the first extensive image library of modern planktonic foraminifera, providing digital taxonomic training tools and resources, and automating species-level taxonomic identification of planktonic foraminifera via machine learning using convolution neural networks. Experts identified 34,640 images of modern (extant) planktonic foraminifera to the species level. These images are served as species exemplars through the online portal Endless Forams (endlessforams.org) and a taxonomic training portal hosted on the citizen science platform Zooniverse (zooniverse.org/projects/ahsiang/ endless-forams/). A supervised machine learning classifier was then trained with~27,000 images of these identified planktonic foraminifera. The best-performing model provided the correct species name for an image in the validation set 87.4% of the time and included the correct name in its top three guesses 97.7% of the time. Together, these resources provide a rigorous set of training tools in modern planktonic foraminiferal taxonomy and a means of rapidly generating assemblage data via machine learning in future studies for applications such as paleotemperature reconstruction.

The Paleocene-Eocene Thermal Maximum (PETM), ca. 56 Ma, was a major global environmental perturbation attributed to a rapid rise in the concentration of greenhouse gases in the atmosphere. Geochemical records of tropical sea-surface temperatures (SSTs) from the PETM are rare and are typically affected by post-depositional diagenesis. To circumvent this issue, we have analyzed oxygen isotope ratios (d 18 O) of single specimens of exceptionally well-preserved planktonic foraminifera from the PETM in Tanzania (~19°S paleolatitude), which yield extremely low d 18 O, down to <-5‰. After accounting for changes in seawater chemistry and pH, we estimate from the foraminifer d 18 O that tropical SSTs rose by >3 °C during the PETM and may have exceeded 40 °C. Calcareous plankton are absent from a large part of the Tanzania PETM record; extreme environmental change may have temporarily caused foraminiferal exclusion.