ArticlePDF Available

Phylogeny and Divergence Times of Lemurs Inferred with Recent and Ancient Fossils in the Tree


Abstract and Figures

Paleontological and neontological systematics seek to answer evolutionary questions with different datasets. Phylogenies inferred for combined extant and extinct taxa provide novel insights into the evolutionary history of life. Primates have an extensive, diverse fossil record and molecular data for living and extinct taxa are rapidly becoming available. We used two models to infer the phylogeny and divergence times for living and fossil primates, the tip-dating (TD) and fossilized birth-death process (FBD). We collected new morphological data, especially on the living and extinct endemic lemurs of Madagascar. We combined the morphological data with published DNA sequences to infer near-complete (88% of lemurs) time-calibrated phylogenies. The results suggest that primates originated around the Cretaceous-Tertiary boundary, slightly earlier than indicated by the fossil record and later than previously inferred from molecular data alone. We infer novel relationships among extinct lemurs, and strong support for relationships that were previously unresolved. Dates inferred with TD were significantly older than those inferred with FBD, most likely related to an assumption of a uniform branching process in the TD compared to a birth-death process assumed in the FBD. This is the first study to combine morphological and DNA sequence data from extinct and extant primates to infer evolutionary relationships and divergence times, and our results shed new light on the tempo of lemur evolution and the efficacy of combined phylogenetic analyses.
Content may be subject to copyright.
TITLE: Phylogeny and divergence times of lemurs inferred with recent and ancient fossils
in the tree
James P. Herrera1,2,3*, Liliana M. Dávalos3,4,5
1Department of Mammalogy, Division of Vertebrate Zoology, American Museum of Natural
History, Central Park West & 79th street, New York NY 10024 USA
2Department of Vertebrate Paleontology, Division of Vertebrate Zoology, American Museum of
Natural History, Central Park West & 79th street, New York NY 10024 USA
3Interdepartmental Doctoral Program in Anthropological Sciences, Department of
Anthropology, Stony Brook University, Stony Brook NY 11794 USA
4Department of Ecology and Evolution, Stony Brook University, Stony Brook NY 11794 USA
5Consortium for Inter - Disciplinary Environmental Research, Stony Brook University, Stony
Brook NY 11794 USA
*Corresponding author contact:
James P. Herrera
Department of Mammalogy, Division of Vertebrate Zoology, American Museum of Natural
History, Central Park West & 79th street, New York NY 10024 USA 1-212-769-5693
© The Author(s) 2016. Published by Oxford University Press, on behalf of the Society of Systematic Biologists. All rights reserved.
For Permissions, please email:
Systematic Biology Advance Access published April 25, 2016
by Liliana Davalos on April 25, 2016 from
Paleontological and neontological systematics seek to answer evolutionary questions with
different datasets. Phylogenies inferred for combined extant and extinct taxa provide novel
insights into the evolutionary history of life. Primates have an extensive, diverse fossil record
and molecular data for living and extinct taxa are rapidly becoming available. We used two
models to infer the phylogeny and divergence times for living and fossil primates, the tip-dating
(TD) and fossilized birth-death process (FBD). We collected new morphological data, especially
on the living and extinct endemic lemurs of Madagascar. We combined the morphological data
with published DNA sequences to infer near-complete (88% of lemurs) time-calibrated
phylogenies. The results suggest that primates originated around the Cretaceous-Tertiary
boundary, slightly earlier than indicated by the fossil record and later than previously inferred
from molecular data alone. We infer novel relationships among extinct lemurs, and strong
support for relationships that were previously unresolved. Dates inferred with TD were
significantly older than those inferred with FBD, most likely related to an assumption of a
uniform branching process in the TD compared to a birth-death process assumed in the FBD.
This is the first study to combine morphological and DNA sequence data from extinct and extant
primates to infer evolutionary relationships and divergence times, and our results shed new light
on the tempo of lemur evolution and the efficacy of combined phylogenetic analyses.
Keywords: total evidence, primatology, Bayesian phylogenetics, calibration, chronogram
by Liliana Davalos on April 25, 2016 from
A primary goal of phylogenetic systematics is discovering and describing species, as well
as placing them in the Tree of Life (Felsenstein 2004). One impediment to this goal is extinction:
more than 90% of species that ever lived are extinct (Novacek and Wheeler 1992).
Understanding the evolutionary history of species can be improved with knowledge of extinct
taxa (e.g., Pyron 2011, Pyron 2015). Extinct taxa inform us about the mode of character
evolution and transitional forms (Slater et al. 2012; Lihoreau et al. 2015), the timing of species
origin and disappearance (Foote 2000), and species distributions in deep time (Patzkowsky and
Holland 2012). Unfortunately, biased preservation, incomplete specimens, and the lack of
molecular data for comparison to extant species impedes the phylogenetic placement of fossils
(Wiens and Morrill 2011; Sansom 2015). Despite these limitations, fossils can give key insights
into the phylogenetic placements of living and extinct forms (Wiens et al. 2010; Wiens and Tiu
2012; Pattinson et al. 2015).
Combining morphological and molecular datasets, especially including fossils, can
improve phylogenetic inference. Total evidence analyses including extinct taxa have improved
resolution for phylogenetic problems as intractable as the relationships of amniotes (Eernisse and
Kluge 1993), reptiles (Wiens et al. 2010; Reeder et al. 2015), cetaceans (Spaulding et al. 2009),
wasps (Ronquist et al. 2012a), and spiders (Wood et al. 2012). The temporal information
captured by fossils is most commonly used to calibrate nodes in a molecular phylogeny based on
the assumed position of fossil taxa in extant trees (Parham et al. 2011). Uncertainty in assigning
a fossil taxon to nodes in an extant tree may introduce error in divergence time estimation using
node calibration. Further, multiple fossil taxa may be associated with a particular node in an
extant tree and are reduced to a single calibration point (e.g., 45 fossils could be used for only
seven calibration points in Ronquist et al. 2012a). To overcome these limitations, new methods
by Liliana Davalos on April 25, 2016 from
were designed that infer the topology and divergence times of living and extinct species jointly
(Ronquist et al. 2012a) and parameterize the branching process of the phylogeny based on
speciation and extinction rates from the fossil record (Heath et al. 2014). The model assumptions
differ between these two approaches and the effects of these assumptions on results are
becoming clear (e.g., Zhang et al. 2015).
The first method, known as “tip-dating” (hereafter TD), uses total evidence datasets to
model the substitution rate of the molecular and morphological data partitions with extant and
fossil tips in the phylogeny (Pyron 2011; Ronquist et al. 2012a). The TD method assumes a
uniform prior probability on the branching process, such that branching events occur anywhere
along internodes according to the branch lengths inferred from the data (Ronquist et al. 2012b,
Zhang et al. 2015). Estimating the rate of morphological evolution is difficult with available
Markov k-state models, however (Beck and Lee 2014), and if evolutionary rates are biased then
the uniform branching prior may make divergence time estimation sensitive to the time prior. In
contrast to the uniform branching assumption, process-based models such as a birth-death model
are especially suitable when the study group has had non-zero extinction (Condamine et al.
2015). The fossilized birth-death process (hereafter FBD), implements a model with a branching
process prior based on diversification dynamics (speciation and extinction rates) calibrated with
the fossil record (Heath et al. 2014). The utility of fossil dating methods in systematics is evident
from the recent surge in publications using them (e.g., Wood et al. 2012; Slater 2013; Arcila et
al. 2015) but the efficacy and comparability of the methods have only recently been addressed
(Beck and Lee 2014; Grimm et al. 2014, Zhang et al. 2015). In this study, we compare the
divergence time estimates inferred from total evidence datasets using the TD and FBD
by Liliana Davalos on April 25, 2016 from
techniques with extant and extinct primates as an empirical system, focusing on lemurs of
The systematics of fossil and extant primates have been approached from two
perspectives: paleontologists with morphological data and extensive sampling of extinct taxa
(e.g., Seiffert et al. 2010; Ni et al. 2013; Pattinson et al. 2015), and neontologists with molecular
data for nearly all extant species (e.g., Perelman et al. 2011; Springer et al. 2012; Pozzi et al.
2014a,b). Divergence time estimates from molecular data are typically older (60-80 million years
ago, Ma, e.g., Perelman et al. 2011) than the appearance of the earliest true primate fossils ~56
Ma (Beard 2008). This discrepancy may be due to convergent slowdowns in molecular rates
(Steiper and Seiffert 2012), the fossil record not capturing the timing of emergence (dos Reis et
al. 2014a), or limitations of external calibration techniques that cannot use all available fossil
information (Pyron 2011). In this study, we focus on the latter possibility.
Even with the extensive primate fossil record, multiple fossils are often reduced to
calibrations of a single node; for example, 35 fossil taxa were reduced to 14 node calibrations in
Springer et al. (2012). Calibration of the crown primate node has been suggested to be 55-56 Ma
(Wilkinson et al. 2011; Kspeka et al. 2015), despite the fact that multiple fossils which may
represent the first crown primates are known from a range of ages (e.g., Ni et al. 2013; Seiffert et
al. 2015). Among the nodes in the primate tree used for divergence time calibration, the last
common ancestor of Lorisiformes has been calibrated based on two key fossils: Saharagalago
and Karanisia (e.g., Horvath et al. 2008; Chatterjee et al. 2009; Pozzi et al. 2014a, see Fig. 1 for
taxonomy and simplified phylogeny). Dated at ~37 Ma (Seiffert et al. 2003), these two fossils
have only informed a single node – a minimum bound for the divergence between Lorisidae and
Galagidae (Springer et al. 2012; Pozzi et al. 2014a,b). The fossil lorisiforms do not represent the
by Liliana Davalos on April 25, 2016 from
ancestral node themselves, however, because they too share an ancestor with lorises and galagos
in the past (Seiffert et al. 2003). Another limitation to node dating is topological uncertainty. The
position of Karanisia, for example, is not well resolved and it is possibly a stem strepsirrhine,
lemuriform or crown lorisid (Seiffert 2012). Given these caveats, calibrating the lorisiform node
to the dates of the fossils may be biasing divergence time estimates towards the calibration point.
Other fossils have not been informative at all because stem taxa cannot be assigned to a
node for calibration. Plesiadapiformes is possibly a stem primate lineage from the earliest
Paleocene/Eocene (Bloch et al. 2007, but see Beard 1990 for the alternative view that
plesiadapiforms are sister to Dermoptera), and as such it has had no bearing on the dating of the
primate phylogeny because the lineage falls outside the crown group. Eocene crown primates
(e.g., Adapiformes, Omomyiforms) and African stem strepsirrhines such as Djebelemur have not
been informative in divergence time estimation despite their important time periods and
geographic locations because they cannot be assigned to nodes for calibration. The lemurs of
Madagascar are especially intractable with respect to fossil calibration because there are no true
fossil lemurs. There are, however, 17 species of extinct lemurs that are subfossils dating from
400 – 20,000 years ago (Godfrey et al. 2010). Calibrations of lemur divergence times have used
multiple primate and nonprimate outgroups (e.g., Yoder and Yang 2000; Horvath et al. 2008).
Recently published ancient DNA has allowed some of the subfossils to be placed in the tree with
greater precision (Kistler et al. 2015). To close the gap between neontology and paleontology,
we focus on the strepsirrhine primates: Lemuriformes from Madagascar and Lorisiformes from
Africa and Asia. We include 33 extinct primates, focusing on the earliest possible stem and
crown primates, stem strepsirrhines and subfossil lemurs.
by Liliana Davalos on April 25, 2016 from
Lemurs are a monophyletic radiation of primates that diverged from their closest
relatives, the lorisiforms, between 50 and 70 Ma based on node-calibrated molecular divergence
times (Yoder and Yang 2000; Horvath et al. 2008; Fabre et al. 2009; Perelman et al. 2011; Pozzi
et al. 2014a; Kistler et al. 2015). Living lemurs are species-rich (99 species currently recognized,
Schwitzer et al. 2013, IUCN Redlist database, accessed February 28 2015),
in addition to the 17 recently extinct species. It has proven difficult to resolve the lemur
phylogeny using molecular data alone (Yoder 1994; Yoder and Yang 2000; Horvath et al. 2008;
Perelman et al. 2011; Springer et al. 2012). Molecular analyses conflict regarding the placement
of major clades, including the earliest diversification of taxonomic families characterized by
short internodes and long branches (Horvath et al. 2008). The placement of the extinct giant
lemurs in the phylogeny was originally based on the morphometric affinities of the extinct
lemurs to living species (e.g., Jungers et al. 1991; Jungers et al. 1997). Fragments of ancient
mitochondrial DNA (Karanth et al. 2005; Orlando et al. 2008) and, more recently, the entire
mitochondrial genome for five taxa (Kistler et al. 2015) supported or overturned some of these
morphology-based relationships. In this study, we infer near-complete phylogenies of extant and
extinct lemurs and their closest relatives with combined morphological and molecular datasets.
We date the tree with fossil tips and two different models of the branching process. This study is
the first to jointly evaluate the relationships and divergence times of extinct and extant lemurs,
and the results change our interpretation of the mode and tempo of lemur diversification.
The methods follow the schematic given in Figure 2.
Taxonomic Sampling
by Liliana Davalos on April 25, 2016 from
The taxonomy of lemurs has changed with the increasing use of DNA sequences to delimit
many cryptic species that were previously subsumed as single species. The most recent
taxonomic compilation recognizes 97 species of living lemurs (Mittermeier et al. 2010), with two
new species described since then (Rasoloarison et al. 2013; Thiele et al. 2013) for a total of 99
lemur species (IUCN redlist, accessed April 20 2015). Our dataset included 87 living lemurs
(~87.88% of recognized living lemurs), and 14 extinct lemurs (82.35%, Godfrey et al. 2010). We
also included a subset of other primates, including the closest extant relatives of lemurs, the
Lorisiformes (67.85% of 28 IUCN recognized species), and eight haplorhine primates (< 3% of
294 IUCN recognized species). Fossil taxa included the following: four crown and two potential
stem strepsirrhines, five adapiforms, two fossil haplorhines, three early primates of disputed
taxonomy, and three stem primates (Fig. 1, Table 1). The complete data matrix included 148
Morphological Data
For 47 taxa (16 extinct, 31 extant), we collected morphological data de novo from osteological
museum specimens, casts and photographs of original specimens, with multiple specimens
examined when possible to reflect variation and polymorphisms. The sample size per species
varied with the availability of specimens; for example, some species were represented by a single
specimen while others were scored for between five and 10 specimens. We supplemented the
new dataset with data from the literature for 20 fossil taxa and 19 extant taxa (Ni et al. 2013;
Seiffert et al. 2015). The total morphological dataset included 85 taxa.
The starting point for scoring characters was a morphological matrix with 421 characters from
previous studies (OSM, Cartmill 1975; Cartmill 1978; Groves and Eaglen 1988; Tattersall and
by Liliana Davalos on April 25, 2016 from
Schwartz 1974; Tattersall and Schwartz 1991; Yoder 1994; Rasoloarison et al. 2000; Seiffert et
al. 2003; Seiffert et al. 2015). Binary and multi-state characters described cranial and long bone
features such as crests, processes, bony articulations, and foveae, the presence, number and
orientation of foramina. Binary and multi-state dental characters included the presence/absence,
relative orientations and development of teeth, cusps, crests, cristae/ids, conules and cingula/ids.
We included eight quantitative measurements that were size-adjusted by dividing each variable
by the geometric mean of all variables, and then converted to discrete states using gap-coding
(Thiele 1993). Polymorphisms were scored as unique states as in Seiffert et al. (2015) to
incorporate the polymorphic information in the dataset (Wiens 2000). A complete description of
characters and states is given in the Online Supplemental Material (OSM). All characters were
treated as unordered. For the taxa scored de novo, we were able to collect data on 40 – 60% of
the 421 characters, principally cranial and dental characters and postcranial characters of the long
bones. Missing data for each species ranged from <1% to 95% (OSM Table S1).
To test the assumption of character independence in the morphological dataset, we converted
the original species X character data matrix into a pairwise species matrix for each character in
which the values were binary states for the same (1) or different (0) states among pairs of
species. These matrices for each character were concatenated, transposed into a pairwise
character matrix, and the Gower dissimilarities of characters were calculated (using the daisy
function in the cluster package, Maechler et al. 2015, for the R statistical environment, R Core
Team 2014, and code from Dávalos et al. 2014, OSM File S2). Dissimilarity scores of 0 indicate
that the character pair has identical state changes among species; i.e., characters may not be
independent. One character of each pair that had 0 dissimilarity was omitted, choosing the
character that showed the most 0 dissimilarities with other characters in the dataset. Fifty-one
by Liliana Davalos on April 25, 2016 from
characters were found to have identical state distributions among species, suggesting they may
not be independent, and we excluded those characters in a reduced character dataset (OSM, File
Molecular Data
We compiled published molecular sequences from GenBank using the software Geneious
v.7.1.7 (Kearse et al. 2012) or directly from first authors. We selected six protein-coding loci
chosen to maximize overlapping coverage among study species, including two mitochondrial
loci (mtDNA: cytochrome b and NADH dehydrogenase - 4) and four nuclear loci (nDNA:
adenosine A3 receptor, cannabinoid receptor 1, and recombination activating gene 1 and 2) for a
total of 5767 base pairs. The GenBank accession numbers are available in OSM File S4 (data
especially from Yoder et al. 1996; Yoder and Irwin 1999; Pastorini et al. 2001a,b; Murphy et al.
2001; Pastorini et al. 2002; Pastorini et al. 2003; Andriaholinirina et al. 2006; Louis et al. 2006;
Andriantompohavana et al. 2007; Craul et al. 2007; Lei et al. 2007; Olivieri et al. 2007;
Zaramody et al. 2007; Horvath et al. 2008; Johnson et al. 2008; Orlando et al. 2008; Groeneveld
et al. 2010; Weisrock et al. 2010; Perelman et al. 2011; Rumpler et al. 2011; Springer et al. 2012;
Markolf et al. 2013; Thiele et al. 2013; Pozzi et al. 2014a,b; Kistler et al. 2015). Sequences for
each locus were aligned using amino acid translation alignment in MAFFT (Katoh et al. 2005) as
implemented in Geneious v.7.1.7. Alignments were verified and edited manually as necessary
(OSM File S5 and TreeBASE submission # 17704).
We analyzed three concatenated molecular matrices: 1) the mtDNA loci, 2) the nDNA loci,
and 3) all six loci. For each matrix, the dataset was partitioned to reflect the heterogeneity in
substitution rates within the matrix. Finding optimal partitioning schemes and models of
sequence evolution for multi-locus datasets is an active area of research. One may chose a priori
by Liliana Davalos on April 25, 2016 from
to partition by each gene, each codon position of each gene, or some combination of these
approaches. We conducted searches for the best partitioning scheme using likelihood statistics,
as implemented in PartitionFinder software (Lanfear et al. 2012). We first specified each codon
of each locus and then used the greedy search algorithm to find the partitioning scheme that
maximized the fit of the data to the model while minimizing the number of parameters, using the
Bayesian Information Criterion (BIC) as well as the second-order Akaike information criterion
(AICc) as the measure of model fit. While alternative partitioning approaches are possible, this
method is objective, repeatable, and has been used for tree inference and divergence time
estimation in previous studies (e.g., Condamine et al. 2015, Lambert et al. 2015). The best-fitting
partitioning scheme for the full concatenated dataset was one that included two partitions, each
with their own model of sequence evolution: 1) all nuclear genes, 1st and 2nd codon positions of
cytochrome B, 2nd and 3rd codon positions of NADH dehydrogenase-4 (GTR+G), 2) 3rd codon
position of cytochrome B, 1st codon position of NADH dehydrogenase-4 (SYM+G; OSM Table
S2). This partitioning scheme grouped sites in the molecular matrix into slow- and fast-evolving
partitions, reflecting differences in the substitution rates and probability of state changes (see
Results). Further details on the partitions used for analyses of the mtDNA-only, nDNA-only, and
concatenated molecular dataset with reduced taxon sampling (discussed below, Topology
comparison) are available in the OSM.
Clock Model Comparison
To compare clock models, we used the stepping-stone approach implemented in MrBayes
v3.2.6 to calculate the marginal likelihoods of the data under the strict molecular clock model
and the following relaxed-clock models: Brownian motion (Thorne and Kishino 2002, TK02),
inverse gamma rates (IGR), and Compound Poisson Process (CPP, Ronquist et al. 2012a).
by Liliana Davalos on April 25, 2016 from
Stepping-stone analysis uses Markov chain Monte Carlo (MCMC) to estimate the likelihood of
the given model close to the posterior distribution and at intervals approaching the prior
distribution, and is an accurate measure of marginal likelihoods for model comparison using
Bayes factors (Xie et al. 2010). Models were compared using Bayes factors by taking the
exponent of the difference in marginal log likelihoods between two models. Models with Bayes
factors between 3 and 20 were considered moderately supported over alternate models and
greater than 20 were considered strong support over alternate models (Kass and Raftery 1995).
We ran stepping-stone analyses for 50 steps of 2.5 million generations each, sampling every
2,500 generations and discarding the first step and first 10% of each subsequent step as burn-in.
Phylogenetic Inference
We jointly inferred the phylogeny and divergence times by conducting Bayesian analyses of
the total evidence dataset using MrBayes v3.2.6 (Ronquist et al. 2012a; Ronquist et al. 2012b).
Additional unconstrained (i.e., non-clock) analyses were conducted on the mtDNA, nDNA,
morphology and total evidence datasets to investigate the phylogenetic signal in each dataset,
and we compared trees inferred from each dataset based on their marginal likelihoods from
stepping-stone analyses (see OSM). With the total evidence datasets, we used both the TD and
FBD approaches to estimate the divergence times calibrated with the dates of fossil taxa included
in the dataset. The TD analysis parameterizes the substitution rate of the morphological data
partition as well as the molecular data and assumes the probability of branching events is
uniform (Ronquist et al. 2012a). The FBD analysis estimates speciation, extinction, and
preservation parameters from the fossil data to calibrate the diversification rate of the tree and
parameterize the branching process (Heath et al. 2014). In the original implementation of the
FBD method, the taxonomic association of fossils to living clades is specified a priori, similar to
by Liliana Davalos on April 25, 2016 from
node dating. In MrBayes v3.2.3 and more recent versions, the phylogenetic position of fossils is
inferred from the morphological data partition. The fossil dates were taken from the literature
(Table 1). For both TD and FBD analyses, we used a uniform prior between the minimum and
maximum fossil ages when these dates were available. For the subfossil lemurs, recent dates
(i.e., Holocene) are available but setting maximum bounds is not possible. We therefore used a
fixed prior because only a point estimate was available with comparatively narrow confidence
intervals (a few hundred years, compared to millions of years for other fossil taxa). Dates were
first taken from the Paleobiology database (Behrensmeyer and Turner accessed 2015) and
verified with primary and secondary literature, especially Hartwig (2002) and references therein,
and Godfrey et al. (2010). To evaluate the effects of having a distribution for the calibration
priors on divergence time estimates, we ran two FBD analyses: one with the age-range
distributions from Behrensmeyer and Turner (accessed 2015, Table 1) and one analysis with only
fixed point estimates on divergence dates (the midpoint of the age ranges). Here we focus on the
results with distributions on age calibration priors (results from fixed date analyses were similar
and are discussed in OSM, see Fig. S3-5 and TreeBASE submission # 17704). We set
Purgatorius as the outgroup because it is the earliest known possible stem primate or stem
euarchontan (Hartwig 2002; Rose 2006).
Model Specifications and Diagnostics
The model of evolution for each data partition was specified a priori using the results of
the optimal model tests for the molecular dataset and the Markov-k model of morphological
evolution (standard variable model, Lewis 2001). For the morphological partition, we used the
ascertainment bias correction implemented in MrBayes, such that constant characters were
removed, while variable and autapomorphic characters were retained. Shapes of the gamma
by Liliana Davalos on April 25, 2016 from
distributions of rate variation among characters, substitution rates and state frequencies were
unlinked among data partitions. We ran two independent Metropolis coupled MCMC (MC3)
searches with four chains for 60 - 70 million generations, sampling every 5,000 generations.
Three chains were heated (temperature = 0.01) and one was cold which recorded the model
parameters. We used the following prior parameter settings: variable rate prior, uniform
branching prior for TD and birth-death process prior for FBD, TK02 relaxed clock model with
values chosen from an exponential distribution with a rate parameter of 0.1, and a gamma-
distributed clock rate. This latter prior defines the prior probability of the evolutionary rate
parameter, and dating analyses are sensitive to the clock rate prior, especially when multiple data
partitions are defined (dos Reis et al. 2014b). We adjusted the gamma distribution according to
the number of data partitions to approximate an independent identically distributed prior by
dividing the initial prior shape and rate parameters (2 and 4, respectively) by the number of
partitions, such that the shape parameter was 0.666 and the rate parameter was 1.33 (following
dos Reis et al. 2014b). This prior placed the highest probabilities on substitution rates in the
range of 1X10-2 to 1X10-3 substitutions/site/million years, in line with previous studies of
primate molecular evolution (Yoder and Yang 2000; Yang 2008). The FBD analysis included
additional parameters with the following prior settings: exponentially distributed speciation prior
(rate = 20), beta-distributed extinction fraction (extinction rate / speciation rate) and fossilization
priors (shape and rate = 1), ‘samplestrat’ parameter set to ‘fossiltip’ to indicate the fossil lineages
ending in distinct tips rather than as ancestors, and sample probability of 0.25 (approximating the
proportion of extant primates in the sample, ~100 sampled / ~400 total extant recognized species,
IUCN RedList accessed February 2015). MrBayes v3.2.6 was run on the CIPRES Science
Gateway (Miller et al. 2010). Codes are available in OSM File S6.
by Liliana Davalos on April 25, 2016 from
We verified convergence of the MC3 search by: 1) plotting the time series of parameter
values sampled from each chain to assess stationarity; 2) quantifying the effective sample sizes
(ESS) for all model parameters, representing the number of independent estimates of the
parameter values drawn from the posterior, with ESS values >200 being ideal (quantified in
Tracer v1.6, Rambaut et al. 2014); 3) verifying the average standard deviation of split
frequencies (ASDSF) were <0.01 and potential scale reduction factor (PSRF) values were stable
around 1.00; and 4) examining the split frequencies among chains and generations using the
utilities in the online application Are We There Yet (Nylander et al. 2008). For all parameters,
independent runs had exhibited mixing and stationarity with ASDSF < 0.01 and PSRF ~1 by ~30
million generations. ESS values combined from the two runs were > 200 for most parameters
and the split frequencies of tree comparisons suggested trees converged between runs. We
discarded the first 50% of generations as burn-in and summarized the posterior distribution of
topologies as the mean clade credibility (MCC) tree (i.e., contype=allcompat command in
MrBayes v3.2.6).
Table 1. Fossil taxa included in phylogenetic analysis and age-range used for divergence-time
estimation, in millions of years ago (Ma).
Genus Species Min age (Ma) Max age (Ma) Reference
Adapis parisiensis 33.9
Aegyptopithecus zeuxis 28.1
Altiatlasius koulchii 56
Anchomomys frontanyensis 37.2
Archaeoindris fontoynonti 0.002149
Archaeolemur edwardsi 0.001
by Liliana Davalos on April 25, 2016 from
Archaeolemur majori 0.0014
Babakotia radafolia 0.00484
Branisella boliviana 26.4
Cantius abditus 50.3
Carpolestes simpsoni 55.8
Daubentonia robustus 0.001
Djebelemur martinezi 41.3
Donrussellia provincialis 48.6
Hadropithecus stenognathus 0.0016
Karanisia clarki 33.9
Komba sp. 20
Leptadapis magnus 33.9
Megaladapis edwardsi 0.001
Megaladapis grandidieri 0.001
Megaladapis madagascariensis 0.00276
Mesopropithecus pithecoides 0.0014
Mesopropithecus dolichobrachion 0.0014
Nycticeboides simpsoni 5.3
Pachylemur jullyi 0.0117
ingens 0.001
maximus 0.00216
Plesiadapis tricuspidens 56
Plesiopithecus teras 28.1
Pronycticebus gaudryi 38
Purgatorius unio 63.3
Saharagalago misrensis 33.9
Teilhardina americana 50.3
Wadilemur elegans 28.1
1: Behrensmeyer and Turner accessed 2015, 2: Hartwig 2002, 3: Godfrey et al. 2010, 4: Seiffert
et al. 2003
Topology comparison
We investigated the congruence of phylogenetic inferences derived from the total evidence (TE)
dataset and the separate analyses of the mtDNA, nDNA and morphological datasets, as well as
topologies inferred by maximum likelihood and parsimony (see OSM for details). First, we
found the best partitioning schemes for each molecular dataset separately using PartitionFinder
(OSM Table S3-S5). Unconstrained phylogenies were inferred for each dataset and the total
by Liliana Davalos on April 25, 2016 from
evidence dataset using two exponentially distributed priors on branch lengths, such that internal
branches had a prior of 0.01 (shape parameter=100) and external branches had a prior of 0.1
(shape parameter=10), since a single prior on branch lengths may be inappropriate when internal
branches are shorter than external branches (Yang & Rannala 2005). We then used Bayesian
concordance analysis to infer the tree that maximized the relationships in common among trees
inferred from separate loci (BUCKy, Larget et al. 2010). The primary concordance tree (PCT)
consisted of a reduced set of 36 extant taxa which had data in all three data types. Trees from
each dataset and the total evidence dataset were then pruned to this 36 taxon set.
We compared the topological similarity of 1000 trees from the posterior distribution of
trees from analyses of each dataset separately and the TE dataset to the PCT by computing the
Robinson-Foulds tree distance (Robinson & Foulds 1981) using the multiRF function in the R
package phytools (v0.4-45, Revell 2012). For null distributions, we also generated two sets of
1000 random trees, first simply by randomly shuffling the tip labels on the PCT 1000 times
(using the phyloshuffle function in the R package phylotools, Zhang, Mi & Pei 2010), and also by
simulating 1000 trees with 36 tips under a birth-death model (speciation=0.15, extinction=0.05
similar to the results found from the FBD analysis described below, using the diversitree
package in R, FitzJohn 2012), applying the tip labels from the PCT and randomly shuffling the
tip labels again. This allowed us to investigate the differences among trees inferred from
different datasets. We used this same methodology to compare the trees inferred from the total
evidence dataset with all taxa using Bayesian, maximum likelihood, and parsimony techniques
(OSM Methods).
To test for significant differences in the fit of the data to the different tree topology
models, the marginal likelihood of the total evidence dataset was compared under the pruned
by Liliana Davalos on April 25, 2016 from
topology inferences from each separate dataset, the PCT, and the total evidence tree using
stepping stone analyses. Topologies in the stepping stone analyses were fixed by specifying node
constraints for the nodes from each dataset using the createMrBayesConstraints function in the
R package paleotree (Bapst 2012). Lastly, our analyses suggested two especially surprising
results: (1) with the time-calibrated total evidence analysis and the reduced morphological
dataset (but not the full dataset), a sister relationship between the African fossil primate
Plesiopithecus tricuspidens and the extant lemuriform Daubentonia madagascariensis, and (2)
the extinct lemur genus Megaladapis was inferred to be sister to all lemuriforms after the most
basal split of Daubentonia from other lineages, rather than inferred to be sister to Lemuridae, as
was the case with ancient DNA (Orlando et al. 2008, Kistler et al. 2015). We used stepping stone
analyses to estimate the marginal likelihood of the total evidence data with four different
topology constraints: (1) the FBD tree inferred from the total evidence analysis with the reduced
morphological dataset (with the Plesiopithecus + Daubentonia node), (2) the tree inferred from
the total evidence analysis with the reduced morphological dataset but with Megaladapis
constrained to be sister to Lemuridae, (3) enforcing a negative constraint on the Daubentonia +
Plesiopithecus relationship, such that the probability of a tree containing this relationship was 0,
and 4) the tree inferred from the total evidence analysis with the reduced morphological dataset
and constraints based on the PCT. Further details are available in the OSM.
Morphological character evolution
To further investigate the surprising sister relationship between Plesiopithecus tricuspidens
and Daubentonia madagascariensis, we mapped synapomorphies on the most parsimonious trees
that included the Plesiopithecus + Daubentonia sister relationship. To further validate the
inferences of synapomorphies from the parsimony analysis, we found the posterior probabilities
by Liliana Davalos on April 25, 2016 from
of character state estimates at the Plesiopithecus + Daubentonia node for the characters found to
be synapomorphies. We used MrBayes to estimate the ancestral states of each morphological
character at the node by constraining that node, using the report command for the morphological
partition and setting ancstates to ‘yes’. We then ran two MCMC searches for 10 million
generations, sampling every 1000 generations, discarded the first 50% as burnin, and combined
the post-burnin generations.
Additional methods and results, as well as input data matrices and code to run analyses,
and output trees are available in the OSM files archived in the Dryad Digital Repository
(, trees and data matrices are archived
in TreeBASE (, ref. # 17704), and the morphological data are available in
MorphoBank (, project P2167).
Clock Model Comparison
The relaxed-clock model with the highest marginal likelihood was the TK02 model, with the
Bayes factors in support of the TK02 model >2X1023, indicating there was strong evidence for
the TK02 model being a better fit to the data than other models (Table 2). For all model
comparisons, the differences between models were always greater than the differences between
runs within analyses (1-10 log likelihood units difference between runs within analyses
compared to >50 log likelihood units difference between model comparisons). We therefore
chose the TK02 model for divergence time analyses.
Parameter rate estimates
by Liliana Davalos on April 25, 2016 from
Posterior estimates of model parameters, such as substitution rates for molecular and
morphological partitions and base frequencies, were consistent across analyses. The 95% HPD of
substitution rates for the molecular partitions were 0.35-0.38 substitutions/site/Ma for the slow-
evolving partition (nuclear genes, 1st and 2nd codon positions of cytochrome B, 2nd and 3rd codon
positions of NADH dehydrogenase-4), and 3.99-4.24 sub./site/Ma for the fast-evolving partition
(3rd codon position of cytochrome B, 1st codon position of NADH dehydrogenase-4). The
morphological partition substitution rate was intermediate, at 2.22-2.63 sub./character/Ma.
The mean TK02 variance parameter of 0.2 (95% HPD 0.08-0.38) and the clock rate parameter of
1.2X10-2 (8X10-3-1.2X10-2) indicated low rates of change across the tree. The net speciation rate
estimate from the FBD analysis was 0.06 species/Ma (0.03-0.11), the relative extinction fraction
(extinction / speciation) was 0.75 (0.40-0.96), and the relative fossilization rate was 0.09 (0.004-
Phylogenetic Inferences
The data under the FBD model had a higher mean marginal likelihood (-89950.87) than
under the TD model (-89984.41), and the difference between models (33.54 log likelihood units)
gives a Bayes factor of 3.68X1014, suggesting the FBD is a better fit to the data than the TD
(following Kass and Raftery 1995). The MCC trees generated from the TD and FBD analyses
differed in the topology inferred for some fossil taxa, and those nodes had low posterior
probabilities (Table 3, Fig. 3, OSM Fig. S1-4, File S7-9, File S16-19, TreeBASE submission #
17704). Especially significant are the placements of the fossil taxa Plesiopithecus teras and
Djebelemur martinezi. Both taxa were suggested to be stem strepsirrhines in previous studies
(e.g., Seiffert et al. 2003) but here we inferred Plesiopithecus to be sister to Tarsius in the FBD
analyses with the full morphological dataset (Fig. 3), and sister to Daubentonia in the FBD
by Liliana Davalos on April 25, 2016 from
analyses with the reduced dataset and in all TD analyses (OSM Fig. S2-4). Djebelemur was a
stem strepsirrhine in the FBD analysis with the full morphological dataset (Fig. 3), while it was a
stem lorisiform with the reduced morphological dataset and in the TD analyses (OSM Fig. S2-4).
Table 2. Marginal likelihood of each clock model for divergence-time estimation, calculated as
the mean of the summed marginal likelihoods across 50 steps of a stepping-stone analysis. Each
model was compared to the model with the lowest marginal likelihood (Thorne-Kishino 2002,
TK02) with Bayes factors (TK02 likelihood versus alternate model likelihood). The TK02 model
had the highest marginal likelihood, exceeding the next-best model (CPP) by ~54 log likelihood
units and Bayes factor ~2X1023, indicating strong support for the TK02 model over other
Model Marginal likelihood (ln) Bayes factor (TK02/alternate model)
a Thorne-Kishino 2002, b Compound Poisson Process, c Inverse Gamma Rate, d strict molecular
Topology comparisons
We investigated the congruence of tree topologies inferred from the mtDNA, nDNA and
morphological datasets separately (OSM Figs S6-S8, Files S10-S12) using a Bayesian
concordance analysis (BUCKy). Nodes in the 36-taxon primary concordance tree (PCT, OSM
Fig S9, File S13) were supported by concordance factors (CFs) between 0.292 and 1 (median =
0.669), which can be interpreted as the mean proportion of data types for which the same nodes
were inferred. CFs in the PCT suggest that 35% of nodes were congruent among all data types,
while 56% of nodes were congruent among two data types, and 0.09% of nodes were found in
only one data type, on average.
by Liliana Davalos on April 25, 2016 from
We compared the PCT to the posterior distribution of trees inferred from analysis of the
total evidence dataset (TE), mtDNA, nDNA, and morphological data using Robinson-Foulds
distances (OSM Figure S10, Table S6). Trees inferred from morphology alone had the greatest
distance from the PCT (mean distance=40.75, SE=0.08), indicating the morphological trees were
least congruent with the PCT. Trees inferred from the TE had the lowest distance from the PCT
(mean=6.04, SE=0.05), followed closely by trees from the mtDNA (mean=7.56, SE=0.04), and
the trees inferred from the nDNA had intermediate distance values (mean=20.89, SE=0.04, OSM
Table S6). All trees inferred from data were closer to the PCT than random trees (mean=65-68,
OSM Figure S10, Table S6).
Topology tests were conducted by comparing the marginal likelihood of the full total
evidence dataset with the 36 taxa under topology constraints to match the TE, PCT, mtDNA,
nDNA and morphology trees. The marginal likelihood of the data was higher with nodes
constrained to match the PCT tree than under topologies inferred from the TE and trees from
each dataset separately (Bayes factors=6.66 for the PCT tree compared to TE and >1019
compared to alternate topologies, OSM Table S8a). This result suggests that the PCT is a better
fit to the data than trees inferred under each dataset separately or all data combined for this
subsample of 36 taxa. When comparing the data for the full taxon set, however, we found that
enforcing nodes inferred from the FBD analysis with reduced morphological characters is a
better fit than alternatives (Bayes factors > 1X1090 for the FBD tree compared to alternate
topologies, OSM Table S8b). These topology tests lend strong support for the relationships
inferred from the concatenated dataset over alternatives including constraints based on the
primary concordance among data types.
Divergence time inferences
by Liliana Davalos on April 25, 2016 from
The divergence times estimated from the TD analyses were older than those estimated
from FBD (Fig. 4, Table 4), and the TD dates were older than previously inferred using node
dating (Table 5). The results of the FBD analysis with wide and fixed date priors were
comparable, with a mean difference in the median estimates of 0.53 Ma, and the 95% HPD range
was 1 Ma wider on average with fixed dates compared to HPDs estimated using age distributions
(OSM Fig. S5).
by Liliana Davalos on April 25, 2016 from
Table 3. Summary of the phylogenetic placement of taxa in this study compared to previous hypothesized topologies. Results from
different analytical techniques are as follows: TD: tip-dating method, FBD 1: fossilized birth-death process with the full 421 character
morphological dataset, FBD 2: FBD analysis with the reduced 369 character dataset.
Taxon Previous placement Ref TD placement FBD placement 1 FBD placement 2
Stem primates,
Outside crown
primates 1
Sister to Crown
Sister to Crown
Sister to Crown
Early primates:
Donrusselia Adapiformes 2 Sister to Adapiformes Sister to Adapiformes Sister to Adapiformes
Teilhardina Omomyiformes,
Sister to Tarsiidae 3
Sister to Crown
Sister to Crown
Sister to Crown
Altiatlasius Problematic,
Sister to
Anthropoidea 2,4 Sister to Adapiformes
Sister to Crown
Primates Sister to Adapiformes
Adapiformes Sister to
Sister to Haplorhini 5
Sister to Crown
Sister to Crown
Sister to Crown
Djebelemur Sister to Strepsirrhini
(Stem Strepsirrhini),
sister to Lorisiformes
Strepsirrhini) 6-8
Sister to Lorisiformes
(Crown Strepsirrhini) Sister to Strepsirrhini
Sister to Lorisiformes
(Crown Strepsirrhini)
by Liliana Davalos on April 25, 2016 from
Plesiopithecus Sister to
Lorisiformes, Stem
unresolved 8-11
Sister to Daubentonia
(Lemuriformes) Sister to Tarsius
Sister to Daubentonia
Wadilemur Crown Lorisiformes,
stem galagid 12 Sister to Galagidae Sister to Galagidae Sister to Galagidae
Komba Crown Lorisiformes,
stem galagid 12
Sister to Euoticus
(crown Galagidae)
Sister to Euoticus
(crown Galagidae)
Sister to Euoticus
(crown Galagidae)
Nycticeboides Sister to Nycticebus
(crown Lorisidae) 12
Sister to Nycticebus
(crown Lorisidae)
Sister to Nycticebus
(crown Lorisidae)
Sister to Nycticebus
(crown Lorisidae)
Saharagalago Crown Galagidae 11 Sister to Lorisiformes Sister to Lorisiformes Sister to Lorisiformes
Karanisia Crown Lorisidae 11 Sister to Lorisiformes Sister to Lorisiformes Sister to Lorisiformes
Lorisidae / Galagidae Paraphyletic,
14 Monophyletic Monophyletic Monophyletic
Daubentonia robustus
Sister to
madagascariensis 15
Sister to D.
Sister to D.
Sister to D.
Megaladapis Sister to
Sister to Lemuridae
Sister to all lemurs
sans Daubentonia
Sister to all lemurs
sans Daubentonia
Sister to all lemurs
sans Daubentonia
Archaeolemuridae Sister to
+ Indriidae
Sister to
Palaeopropithecidae +
Sister to
Palaeopropithecidae +
Sister to
Palaeopropithecidae +
Sister to Indriidae
Indriidae paraphyletic
with Indri sister to
Indriidae paraphyletic
with Indri sister to
Indriidae paraphyletic
with Indri sister to
Pachylemur Sister to Varecia,
Sister to Varecia,
Sister to Varecia,
Sister to Varecia,
Hapalemur simus Hapalemur
paraphyletic, H.
simus sister to Lemur 20
monophyletic, H.
simus sister to other
monophyletic, H.
simus sister to other
monophyletic, H.
simus sister to other
by Liliana Davalos on April 25, 2016 from
catta Hapalemur Hapalemur Hapalemur
Phaner Cheirogaleidae,
Sister to
Sister to
Sister to
1: Bloch et al. 2007; 2: Hartwig 2002; 3: Beard 2008; 4: Ni et al. 2013; 5: Gebo 2002; 6: Marivaux et al. 2013; 7: Seiffert 2012; 8:
Pattinson et al. 2015; 9: Godinot 2005; 10: Simons & Rasmussen 1994; 11: Seiffert et al. 2003; 12: Seiffert et al. 2005; 13: Yoder et
al. 2001; 14: Masters et al. 2005; 15: Godfrey et al. 2010; 16: Tattersall & Schwartz 1974; 17: Kistler et al. 2015; 18: Karanth et al.
2005; 19: Jungers et al. 1991; 20: Pastorini 2000; 21: Tattersall & Schwartz 1991; 22: Horvath et al. 2008; 23: Springer et al. 2012
by Liliana Davalos on April 25, 2016 from
Table 4. Comparison of the differences in age estimates for 21 nodes in the phylogenies among
dating techniques and datasets. “Full” refers to the complete 421 character morphological data
matrix, and “reduced” refers to the subset of 369 characters that were found to be independent
based on the Gower dissimilarity of state changes among species. Mean difference of age
estimates in millions of years ago.
Comparison Mean
FBD full vs FBD reduced -2.55
FBD full vs TD full -12.10
FBD reduced vs TD reduced -9.48
TD full vs TD reduced 0.07
Morphological character evolution
We investigated the evolution of morphological characters, focusing on the node that
supports the relationship between Daubentonia and Plesiopithecus to assess what characters
support this node in the reduced morphological dataset. The node was present in 36% (20 / 55) of
most parsimonious trees (OSM Fig S12, File S15). Across the 20 trees, we found that the
synapomorphies linking these taxa included simplification of the lower molar structures (OSM
Table S9). This included lower first and second molar cristid obliqua that terminate at the base of
the trigonid, lower second molar trigonids and talonids of approximately equal height, weak or
rounded cristid obliqua, lower third molars slightly shorter than second molars, and no
hypoconulid on the lower third molar (see OSM Table S9). We found that these character states
had high posterior probabilities (>0.80) for ancestral estimates at the Plesiopithecus +
Daubentonia node (OSM Table S9), indicating that the parsimony-based synapomorphies are
supported in a probabilistic modelling framework. Characters supporting Plesiopithecus as a
strepsirrhine include mesiodistally compressed lower canines, high crowned and procumbent
by Liliana Davalos on April 25, 2016 from
lower canines, lower third premolar mesial roots lateral to the distal roots, two lower third
premolar roots, and no cristid obliqua on the lower fourth premolar (OSM Table S9). It is
important to note that Plesiopithecus lacks a toothcomb (procumbent, laterally compressed
anterior lower dentition used primarily for grooming fur), which is one synapomorphy of
This study sought to compare two new methods of inferring phylogeny and divergence
times with living and extinct taxa, and to infer the phylogeny and divergence times of primates,
focusing on lemurs and their close relatives, lorisiforms. The results revealed striking differences
in the divergence time estimates and substantially different statistical support for the two models.
The new phylogenies we inferred for lemurs are the most taxonomically complete to date,
including representatives of every genus of extant and extinct lemurs as well as inferring the
positions of these taxa with strong support. We inferred the split between Haplorhini and
Strepsirrhini to be post-Cretaceous, albeit with highest probability density including up to 70Ma,
instead of the pre-Cretaceous-Paleogene dates usually inferred from molecular data. Further, we
inferred divergence times for lemurs that are more recent than previously estimated from
molecular data only. The divergence of the families was concentrated around the Eocene-
Oligocene boundary, a geological time period associated with major faunal turnover in many
primate clades (Seiffert 2007). These results have implications for the drivers of diversification
in primates, especially extant and extinct lemurs.
Differences in divergence times between TD and FBD
The divergence dates for lemurs using TD were ~9-12 My older than those inferred from
FBD, on average. While previous node-calibrated estimates of lemur origins suggested the most
by Liliana Davalos on April 25, 2016 from
recent common ancestors (MRCAs) of all lemurs occurred 41-75 Ma, we estimated the MRCA
of lemuriforms to exist 58-72 Ma using TD, and 40-56 Ma from FBD (Table 5). Node-calibrated
molecular phylogenies dated the subsequent divergence of lemur families to ~30-40 Ma, and the
relationships among families were unresolved (Yoder and Yang 2004; Horvath et al. 2008;
Chatterjee et al. 2009; Perelman et al. 2011). Our TD analyses suggested the divergences among
families occurred ~50-62 Ma, while the FBD analysis inferred divergence times ~34-49MA
(Table 5). Our comparison of the marginal likelihoods for these models suggested that the FBD
was a better fit to our data than the TD (Bayes factor > 104). The non-overlapping divergence
time estimates suggest there are some major differences in the way that time is modelled by these
Recent studies using the TD method have also recovered earlier dates than those inferred
using node dating (Ronquist et al. 2012a; Slater et al. 2012; Wood et al. 2012; Slater and Harmon
2013; Beck and Lee 2014). One possible explanation is that the TD method may overestimate
divergence times because the Markov-k model of discrete morphological evolution may
underestimate the morphological rate of change (Beck and Lee 2014). We found that the
morphological substitution rate was intermediate between the slowest and fastest evolving
molecular partitions. If there is character state saturation in the morphological data, the same
character states will appear in different species through homoplasy, leading to underestimated
rates of morphological substitution (Wagner 2000), and in turn to overestimated dates.
In conjunction with the potential issues related to estimating rates of morphological
evolution, the branching process prior in the TD analysis assumes a uniform prior distribution on
branching events, in contrast to the birth-death prior in the FBD (Zhang et al. 2015). The choice
of the branching prior in divergence time estimation is not trivial, and a birth-death prior is more
by Liliana Davalos on April 25, 2016 from
appropriate than a pure-birth prior when extinction is non-zero (Condamine et al. 2015). If the
TD method is biased by an underestimated morphological substitution rate and the probability of
branching is assumed to be equal through time, then the TD methodology may be prone to
pushing nodes deep into the past because the posterior probabilities on node ages are
inadequately constrained and sensitive to the time prior. As pointed out by one reviewer of this
article, this may result from too much flexibility in the model with the uniform branching prior
because node ages are not constrained as they are in node calibration. In contrast, the FBD
analysis assumes a birth-death prior for the branching process, such that the probability of a
branching event is conditional on the parameterization of the inferred speciation and extinction
rates of the tree. This constraint on the branching process means that nodes cannot be pushed too
deep in the phylogeny if extinction rates are inferred to be greater than zero because a phylogeny
with deep nodes and long external branches indicates low extinction (Pybus and Harvey 2000).
Lastly, the uncertainty in placing fossils on the tree given their patchy and sometimes
uninformative character data compounds the uncertainty in branching times and substitution
rates. For example, the positions of some fossils breaking up long branches of the tree may draw
nodes deeper into the past than if no fossils were considered. The placement of Megaladapis as
sister to all non-aye-aye lemurs certainly changes our interpretations of crown ages. These
methodological considerations may explain the earlier divergence times inferred using TD
compared to FBD and previous node-dating techniques.
We argue that the benefits conferred by the ability to place important extinct taxa (e.g.,
Djebelemur, Saharagalago, Karanisia, Wadilemur, Komba, extinct lemurs) outweigh the
disadvantages of the artifacts that drive the differences between TD and FBD, especially in
comparison to node-based divergence times from extant-only datasets. Previous molecular
by Liliana Davalos on April 25, 2016 from
analyses could not include calibration information for stem taxa like plesiadapiforms, adapiforms
or Djebelemur, despite the importance of these fossil taxa in the evolution of primates. Further,
the lack of fossils limited node-calibrated molecular analyses of lemurs. The divergence times of
extinct and extant lemurs were recently inferred from mitochondrial genomes and the results
were similar to those we report, with the exception of the position and divergence time of
Megaladapis (Kistler et al. 2015) as discussed below. Mitochondrial genomes are known to
evolve faster than nuclear genomes, leading to saturation and bias in divergence times towards
the calibration points; divergences that are older than calibration points are underestimated and
younger divergences are overestimated (Arbogast et al. 2002; Zheng et al. 2011). The four node
calibrations used previously (Kistler et al. 2015) were based on fairly recent divergences in the
Haplorhini (human-chimp ~ 5-8 Ma, baboons ~ 1-3.5 Ma), and two older calibrations (apes-Old
World Monkeys ~ 21-34 Ma, Lorisiformes ~ 37 Ma). If divergence times previously derived
from mitochondrial DNA sequences were biased towards calibration points, then the inferred
divergences of lemurs should be close to those calibrations, which they are (Table 4). We
included the mitochondrial sequences from previous studies, and combined them with
morphological data that are most likely coded by multiple nuclear loci. In addition, with our
calibrations based on 33 fossils actually in the tree, spread across the chronology of early to
recent primate diversification, it is expected that lineage divergence times should be earlier and
spread more evenly through time than observed in the mitochondrial node-dating divergence
time estimates, which is the case in our results.
Joint Inference of Phylogeny and Divergence Times: Ancient Primate Fossils
Support for the phylogenetic placement of possible stem and early crown primate fossil
taxa was weak, and this was expected given that many taxa had high proportions of missing data,
by Liliana Davalos on April 25, 2016 from
few synapomorphic characters linking them with extant lineages, and autapomorphic character
states that make their derived morphology difficult to place (e.g., Altiatlasius, Teilhardina, some
adapiforms). Including them in these analyses allowed us to place the taxa in the tree with
empirical data and the temporal occurrence information of the fossils calibrated the speciation
and extinction rates of the tree. This is the first study that could use the temporal information of
possible stem primates such as plesiadapiforms, particularly the oldest known fossils, in
estimates of divergence times. While previous studies have calibrated the root of crown primates
based on the oldest known crown fossils (e.g., Wilkinson et al. 2011), placing those fossils on the
tree and inferring their stem ages is unique to this study. Some inferred relationships were
unexpected and most likely due to the paucity of fossils in this sample compared to previous
studies focused on fossils (e.g., Seiffert et al. 2010). For example, some fossil clades which are
considered to be crown clades sister to Haplorhini (Omomyiformes such as Altiatlasius and
Teilhardina, Hartwig 2002; Beard 2008, or Altiatlasius sister to Anthropoidea, Ni et al. 2013)
and Strepsirrhini (Adapiformes sister to Strepsirrhini, Seiffert et al. 2009) were inferred to be
outside crown primates (Table 3). The underrepresentation of omomyiform and adapiform
species in the present sample precludes conclusions regarding relationships for those taxa. Other
fossils were well supported, firmly anchoring the topology and divergence times for catarrhines
(Aegyptopithecus) and platyrrhines (Branisella). The plesiadapiforms, which are accepted to be
outside crown primates (Bloch et al. 2007), were well supported as sister to the other lineages.
The occurrence of Plesiadapiformes and early primate taxa in the Paleocene/Eocene and their
placement at the base of the tree calibrates the total tree depth to ~75 Ma, and key divergences
among crown primate lineages occurred after the Cretaceous-Paleogene boundary.
by Liliana Davalos on April 25, 2016 from
The divergence times we estimated with the FBD model for the deepest nodes are
generally more recent than previously suggested using molecular data for only extant taxa. Total
evidence dating techniques do not assume that fossil species represent minimum ages for the
MRCAs of living taxa, as node dating does. Rather, the extinct taxa share a common ancestor
with sister lineages sometime before their appearance in the fossil record (Ronquist et al. 2012a).
Fossil taxa represent a minimum bound for a node, but the maximum bound may be much earlier
than allowed by most hard prior distributions used to date. Our results suggest the divergence of
crown Haplorhini and Strepsirrhini ~54-70 Ma, with a rapid subsequent divergence among
lineages during the Paleocene and Eocene. These dates are more concordant with the fossil
record than the deep Cretaceous estimates found by some molecular studies (Horvath et al. 2008;
Wilkinson et al. 2010; Perelman et al. 2011).
by Liliana Davalos on April 25, 2016 from
Table 5. Comparison of divergence time estimates at key nodes in the phylogeny, in millions of years ago. The results of this study
using the fossilized birth-death process and combined data are compared to those published previously using node dating and
molecular data.
Node This study Kistler et
al. 2015
Yoder and
Yang 2004
Horvath et
al. 2008
et al. 2011
Springer et
al. 2012
et al. 2009
64 (48,70) 68 (60,76) 85* (77,90) - 87 (76, 99) 68 (63,71) 67 (64,73)
Crown Strepsirrhini 61 (56,67) 59 (52,66) 69 (61,75) 75 (67,84) 69 (59,77) 54 (53,55) 52 (48,56)
Lorises + Galagos 38 (32,39) 38* (37,41) 39* (38,42) 39* (37,42) 40* (35,46) 35* (31,37) 38 (37,39)
Lemuriformes 55 (49,61) 50 (42,57) 62 (58,73) 66 (55,75) 59 (39,77) 50 (49,51) 46 (41,51)
Lemurs (sans
42 (34,50) 31 (27,35) 42 (35,50) 39 (33,46) 39 (26,50) 32 (27,37) 32 (29,34)
Archaeolemuridae 28 (21,35) 24 (20,28) - - - - -
Palaeopropithecidae 23 (17,29) 21 (17,24) - - - - -
Indriidae 23 (17,28) 17 (14,20) 39 36 17 (10,26) 18 (12,26) 21 (17,25)
Lemuridae 26 (19,33) 19 (16,22) 32 (26,39) 23 (19,29) 26 (16,37) 21 (15,26) 21 (18,25)
Lepilemuridae 16 (12,22) 12 (9,15) 37-38 32 (26,38) 12 (6,17) 9 (6,13) 16 (13,19)
Cheirogaleidae 31 (24,39) 25 (21,30) 29 (23,36) 23 (19,28) 25 (15,35) 22 (17,27) 24 (20,27)
* Node used as calibration point in previous studies.
by Liliana Davalos on April 25, 2016 from
The origin of strepsirrhines is still poorly understood. The djebelemurid clade of northern
Africa is the oldest stem strepsirrhine in this analysis at 45-49 Ma (Seiffert 2012; Marivaux et al.
2013). Djebelemurid fossils have not been informative for studies using node dating because
they cannot be assigned to any node in extant-only phylogenies. In the FBD analysis with the full
morphological dataset, Djebelemur was found to be sister to all other strepsirrhines (a stem
strepsirrhine), while analyses with the reduced morphological dataset suggested Djebelemur was
sister to Lorisiformes, and thus a crown strepsirrhine (similar to Pattinson et al. 2015).
Djebelemur lacks a toothcomb, the synapomorphy that unites crown strepsirrhines (Seiffert
2012). Thus, either the toothcomb evolved once after the divergence of Djebelemur from the
crown strepsirrhine lineage as suggested by the FBD analysis with the full dataset, or it evolved
with crown strepsirrhines and was independently lost in Djebelemur. While the toothcomb is a
defining crown strepsirrhine synapomorphy, it has been lost, presumably independently, in
several lemur lineages including Daubentonia and some extinct giant lemurs including
Archaeolemur and Hadropithecus. The hypothesis that Djebelemur was a stem strepsirrhine is
most likely better supported, given the results of previous studies with greater sampling of stem
strepsirrhine fossils (e.g., Seiffert et al. 2003).
One surprising result in this study was the placement of Plesiopithecus as sister to
Daubentonia in the FBD analysis with the reduced morphological dataset and the TD analyses
with both full and reduced datasets. Plesiopithecus is an African Eocene fossil with a unique
suite of derived and plesiomorphic characters that has made inferring its phylogenetic position
relative to other primates difficult. Plesiopithecus has been hypothesized to be an early
anthropoid (Simons 1992), a lorisiform (Rasmussen & Nekaris 1998), or, more commonly, a
stem strepsirrhine (Simons & Rasmussen 1994; Seiffert et al. 2003, reviewed in Seiffert 2012).
by Liliana Davalos on April 25, 2016 from
Recent total evidence analyses using partitioned Bayesian analysis with extensive fossil
sampling found Plesiopithecus to be unresolved (Pattinson et al. 2015). The
Daubentonia+Plesiopithecus relationship has been suggested before (Godinot 2005), and many
of our analyses support this sister relationship. The synapomorphic characters linking
Plesiopithecus and Daubentonia include simplification of molar features compared to other taxa
in the analyses. Simplifications of the molars may be convergent adaptations to food items that
are structurally defended but processed with the anterior teeth. The chisel-like incisors of
Daubentonia are used to bore holes in tree bark and seeds, but the food items obtained require
little physical processing by the posterior teeth (insect larvae, soft inner flesh of seeds, Sterling
1994). A similar adaptive function may explain the enlarged canines of Plesiopithecus. The lack
of other, non-functionally related synapomorphies indicates that the inferred relationship
between Plesiopithecus and Daubentonia may be the result of convergent evolution, rather than
common descent. The biogeographic implications of this result are of great importance to
understanding lemuriform origins. The presence of a lemuriform primate in Africa after the split
of the most basal lineages would suggest either a single origin of lemurs and a back-dispersal of
Plesiopithecus to Africa, or two independent dispersals to Madagascar. Further, if Plesiopithecus
is accepted to be a lemuriform, it is the first and only true fossil lemuriform. This result is
controversial in view of the strongly supported single-origins hypothesis from molecular and
total evidence analyses (e.g., Yoder et al. 1996). Further validation is necessary to confirm the
placement of this taxon by including more complete character sampling (both taxa were sampled
for ~60% of characters, mostly dental) and more stem strepsirrhines such as anchomomyine taxa.
There were discrepancies among topologies inferred with the full and reduced
morphological matrices. The morphological characters were variable in all analyses and included
by Liliana Davalos on April 25, 2016 from
autapomorphic characters. In the reduced dataset, characters were culled based on an objective
approach that tracks co-distributed state changes among characters. Using the Gower
dissimilarity of character state changes among taxa, characters were omitted if, for example,
character Y always changed from state 0 to state 1 when character X changed from 0 to 1 in all
taxa. This situation supports non-independence of characters X and Y, and we removed one
character of each pair. If taxa change positions in analyses of the full and reduced matrices,
support for those taxa in the analysis of the full matrix may be biased by effectively up-
weighting correlated characters. This had the biggest effect on fossils, and one explanation is that
these taxa had few synapomorphic characters to link them with strong support to other lineages
so removing correlated characters left too few characters to secure their positions. Also, the
characters were taken from previous studies that sought to identify those characters which most
strongly distinguished major clades, such as Haplorhini / Strepsirrhini, such that many of these
characters have only a single transition. Some examples include: (1) allantois development is
rudimentary in all haplorhines while all strepsirrhines have large, vesicular structures; (2)
primordial amniotic cavity present in haplorhines and absent in strepsirrhines; (3) retinal fovea
found in haplorhines and not strepsirrhines. Lastly, this analysis is dataset-dependent; there are
taxa not included in the present matrix that may break up the 1:1 state change pattern and future
analyses with greater taxonomic sampling of intermediate fossil forms will likely change which
characters are considered correlated.
Joint Inference of Phylogeny and Divergence Times: Strepsirrhini and Lemuriformes
Total evidence analyses can test the assumption that the placement of fossil taxa in the
phylogeny corresponds to nodes linking extant taxa. An especially important example of fossils
representing minimum age bounds in this dataset concerns the MRCA of Lorisiformes, which
by Liliana Davalos on April 25, 2016 from
was previously calibrated to approximately 37 Ma based on Saharagalago and Karanisia
(Seiffert et al. 2003; Pozzi et al. 2014a). In our analyses, the relationships of these fossils and
their MRCAs with crown sister lineages were inferred jointly, and the results showed these
fossils shared a common ancestor with crown lorisiforms ~35-56 Ma. We inferred the MRCA of
crown Lorisiformes ~31-39 Ma, concordant with estimates from node calibration. By including
these fossils in this study, their placement in the tree was inferred empirically and the divergence
times for lorisiforms was estimated from the data, rather than calibrated a priori. Before the
discovery of Saharagalago and Karanisia, the MRCA of Lorisiformes was inferred to exist ~40
Ma based on calibrations from non-strepsirrhine primates (Yang & Yoder 2003), further
validating the Eocene origins of the clade.
This study is the first to infer the position of nearly all subfossil lemurs and their
divergence times jointly from empirical analysis of combined data. The strepsirrhine phylogenies
were generally well supported, especially at key nodes within Lemuriformes which have been
contentious until now. Many extinct species were placed with moderate to strong support,
corroborating inferences from both morphological affinities (e.g., Jungers et al. 1991; Jungers et
al. 1997) and ancient DNA (Karanth et al. 2005; Orlando et al. 2008), especially
Archaeolemuridae, Palaeopropithecidae, Pachylemur, Daubentonia robustus. One exception was
the paraphyly of Indriidae; while Indriidae was previously hypothesized to be sister to
Palaeopropithecidae, here we inferred that Indri was sister to a clade consisting of
Palaeopropithecidae and Propithecus + Avahi, but with low posterior probability. The
relationships among indriid genera have always been contentious, and new data are needed to
resolve this issue. For example, there are no nuclear loci available for Indri, and the recent
by Liliana Davalos on April 25, 2016 from
recovery of ancient nuclear DNA from Megaladapis holds promise for acquiring those data for
other subfossil taxa as well (Perry et al. 2015).
Another unique finding in this study was the placement of Megaladapis as sister to all
lemurs other than Daubentonia, a hypothesis that conflicts with its morphological similarities to
Lepilemuridae (e.g. Tattersall and Schwartz 1974), and the sister relationship to Lemuridae
found with ancient mitochondrial DNA. This result was surprising, given that we included the
published molecular data that had recovered the Megaladapis+Lemuridae relationship (Kistler et
al. 2015). These differences may be related to different data partitioning schemes between
studies; previous studies had applied a single molecular model to the entire mitochondrial
genome or partitioned the genome by codon position (Kistler et al. 2015), instead of partitions
based on the best subset of substitution rate categories as in this study. The specification of
molecular models in phylogenetic inference is an important yet often overlooked issue, and
misspecification of the molecular partitions and models can lead to poor inferences (Brown and
Lemmon 2007, Lanfear et al. 2012). In this study, the best partitioning scheme of the multi-gene
alignment included a fast-evolving partition (cytochrome B third codon position and NADH
dehydrogenase 4 first codon position), and a slow-evolving partition (all other loci together).
With this partitioning scheme, the position of Megaladapis we inferred was more strongly
supported by the data than the alternative Megaladapis+Lemuridae relationship (OSM Table S8).
These differences in molecular evolution and partitioning scheme between previous studies and
this study may account for the discrepancies in fossil placement observed.
Ours are among the most complete phylogenetic inferences for lemurs to date. Accurate
and complete dated phylogenies are necessary for testing hypotheses about lineage and character
evolution (Felsenstein 1985; Nunn 2011). Our time-tree inferences have important implications
by Liliana Davalos on April 25, 2016 from
for the diversification dynamics in this biologically diverse and endangered primate group. For
example, the tree shape and balance is indicative of the tempo of diversification and possible
shifts in diversification rate through time (Pybus and Harvey 2000; Rabosky 2014). Including
fossil species in phylogeny-based inferences of lineage diversification rates is at the forefront of
macroevolution (Pyron and Burbrink 2012; Silvestro et al. 2014). With increasing availability of
molecular and morphological data, paleontological databases, and innovative models of
divergence times and character evolution, researchers in phylogenetic systematics and
macroevolution are primed to clarify the structure and the ages of the tree of life.
Supplementary material, data files and/or online-only appendices, can be found in the
Dryad Digital Repository: 10.5061/dryad.10.5061, MorphoBank project P2167,
and TreeBASE submission # 17704.
This work was supported by the Museum of Comparative Zoology (Ernst Mayr Travel
Grant in Animal Systematics), National Science Foundation (Graduate Research Fellowship),
Stony Brook University (Turner Fellowship and AGEP T-FRAME Scholarship) (J.P.H.) and
National Science Foundation (DEB-0949759, DEB-1442142) (L.M.D).
We thank the faculty and staff of the following institutions housing specimens used in
this study: American Museum of Natural History Department of Mammalogy, Museum of
Comparative Zoology, Harvard, The Duke University Division of Fossil Primates, and the Stony
Brook University Anatomical Museum. We thank L. Kistler and P.J. Perry for early access to
annotated alignments of mitochondrial genomes for the subfossil lemurs. Thanks to S Nash for
providing his wonderful illustrations of primates to bring the extinct lemurs back to life. For
by Liliana Davalos on April 25, 2016 from
training in phylogenetic systematics, comparative methods and statistical analyses we thank: the
AnthroTree workshop held by C. Nunn and supported by the NSF (BCS-0923791) and the
National Evolutionary Synthesis Center (NSF grant EF-0905606); we thank the UC Davis
Bodega Bay Applied Phylogenetics Workshop leaders, especially P. Wainwright, L. Mahler, S.
Price, and B. Moore. We thank D. Rojas and members of the Dávalos lab, J. Smaers, E. Seiffert,
W. Jungers, and P.C. Wright for valuable insights, discussions, and revisions to early versions of
the manuscript. We thank F. Anderson, T. Near, A. Yoder, and one anonymous reviewer for
valuable feedback, insightful comments, and helpful suggestions on the first draft that greatly
improved this manuscript.
Andriaholinirina N., Fausser J.L., Roos C., Zinner D., Thalmann U., Rabarivola C.,
Ravoarimanana I., Ganzhorn J.U., Meier B., Hilgartner R. 2006. Molecular phylogeny and
taxonomic revision of the sportive lemurs (Lepilemur, Primates). BMC Evol. Biol. 6:17.
Andriantompohavana R., Runhua L., Zaonarivelo J.R., Engberg S.E., Nalanirina G., McGuire
S.M., Shore G.D., Andrianasolo J., Herrington K., Brenneman R.A., Louis E.E. 2007. Molecular
phylogeny and taxonomic revision of the woolly lemurs, Genus Avahi (Primates: Lemuriformes).
Special Publictions of the Museum of Texas Tech University 51:1-64.
Arbogast B.S., Edwards S.V., Wakeley J., Beerli P., Slowinski J.B. 2002. Estimating divergence
times from molecular data on phylogenetic and population genetic timescales. Annu. Rev. Ecol.
Syst. 33:707-740.
Arcila D., Pyron R.A., Tyler J.C., Ortí G., Betancur-R R. 2015. An evaluation of fossil tip-dating
versus node-age calibrations in tetraodontiform fishes (Teleostei: Percomorphaceae). Mol.
Phylogen. Evol. 82:131-145.
by Liliana Davalos on April 25, 2016 from
Bapst D.W. 2012. paleotree: an R package for paleontological and phylogenetic analyses of
evolution. Methods in Ecology and Evolution 3:803-807.
Beard K.C. 1990. Gliding behaviour and palaeoecology of the alleged primate family
Paromomyidae (Mammalia, Dermoptera). Nature 345:340-341.
Beard K.C. 2008. The oldest North American primate and mammalian biogeography during the
Paleocene–Eocene Thermal Maximum. Proc. Natl. Acad. Sci. U.S.A. 105:3815-3818.
Beck R.M., Lee M.S. 2014. Ancient dates or accelerated rates? Morphological clocks and the
antiquity of placental mammals. Proc. Royl. Soc. B: Biol. Sci. 281:20141278.
Behrensmeyer A.K., Turner A. accessed 2015. Taxonomic occurrences of Primates recorded in
the Paleobiology Database. Fossilworks.
Bloch J.I., Silcox M.T., Boyer D.M., Sargis E.J. 2007. New Paleocene skeletons and the
relationship of plesiadapiforms to crown-clade primates. Proc. Natl. Acad. Sci. U.S.A. 104:1159-
Cartmill M. 1975. Strepsirhine basicranial structures and the affinities of the Cheirogaleidae. In:
Luckett W. P. and Szalay F. S. editors. Phylogeny of the Primates: A Multidisciplinary
Approach. US, Springer, p. 313-354.
Cartmill M. 1978. The orbital mosaic in prosimians and the use of variable traits in systematics.
Folia Primatol. 30:89-114.
Catlett K.K., Schwartz G.T., Godfrey L.R., Jungers W.L. 2010. “Life history space”: a
multivariate analysis of life history variation in extant and extinct Malagasy lemurs. Am. J. Phys.
Anthropol. 142:391-404.
Chatterjee H., Ho S., Barnes I., Groves C. 2009. Estimating the phylogeny and divergence times
of primates using a supermatrix approach. BMC Evolutionary Biology 9:259.
by Liliana Davalos on April 25, 2016 from
Condamine F.L., Nagalingum N.S., Marshall C.R., Morlon H. 2015. Origin and diversification
of living cycads: a cautionary tale on the impact of the branching process prior in Bayesian
molecular dating. BMC Evol. Biol. 15:65.
Craul M., Zimmermann E., Rasoloharijaona S., Randrianambinina B., Radespiel U. 2007.
Unexpected species diversity of Malagasy primates (Lepilemur spp.) in the same biogeographical
zone: a morphological and molecular approach with the description of two new species. BMC
Evol. Biol. 7:83.
dos Reis M., Donoghue P.C., Yang Z. 2014a. Neither phylogenomic nor palaeontological data
support a Palaeogene origin of placental mammals. Biol. Lett. 10:20131003.
dos Reis M., Zhu T., Yang Z. 2014b. The Impact of the rate prior on Bayesian estimation of
divergence times with multiple loci. Syst. Biol. 63:555-565.
Drummond A.J., Ho S.Y., Phillips M.J., Rambaut A. 2006. Relaxed phylogenetics and dating
with confidence. PLoS Biol. 4:e88.
Eernisse D.J., Kluge A.G. 1993. Taxonomic congruence versus total evidence, and amniote
phylogeny inferred from fossils, molecules, and morphology. Mol. Biol. Evol. 10.
Fabre P.H., Rodrigues A., Douzery E.J.P. 2009. Patterns of macroevolution among Primates
inferred from a supermatrix of mitochondrial and nuclear DNA. Mol. Phylogen. Evol. 53:808-
Felsenstein J. 1985. Phylogenies and the comparative method. Am. Nat. 125:1-15.
Felsenstein J. 2004. Inferring phylogenies. Sunderland, Massachusetts, Sinauer Associates.
FitzJohn R.G. 2012. Diversitree: comparative phylogenetic analyses of diversification in R.
Methods Ecol. Evol. 3:1084-1092.
by Liliana Davalos on April 25, 2016 from
Foote M. 2000. Origination and extinction components of taxonomic diversity: general
problems. Paleobiology 26:74-102.
Gebo D.L. 2002. Adapiformes: Phylogeny and adaptation. In: Hartwig WC editor. The Primate
Fossil Record. UK, Cambridge University Press, p. 21-43.
Godfrey L.R., Jungers W.L., Burney D.A. 2010. Subfossil lemurs of Madagascar. In: Werdelin
L, Sanders W editors. Cenozoic Mammals of Africa. Berkeley, The University of California
Godinot M. 2005. Lemuriform origins as viewed from the fossil record. Folia Primatol. 77:446-
Grimm G.W., Kapli P., Bomfleur B., McLoughlin S., Renner S.S. 2014. Using more than the
oldest fossils: Dating Osmundaceae with three Bayesian clock approaches. Syst. Biol. 64:396-
Groeneveld L.F., Blanco M.B., Raharison J.L., Rahalinarivo V., Rasoloarison R.M., Kappeler
P.M., Godfrey L.R., Irwin M.T. 2010. MtDNA and nDNA corroborate existence of sympatric
dwarf lemur species at Tsinjoarivo, eastern Madagascar. Mol. Phylogen. Evol. 55:833-845.
Groves C.P., Eaglen R.H. 1988. Systematics of the Lemuridae (primates, Strepsirhini). J. Hum.
Evol. 17:513-538.
Heath T.A., Huelsenbeck J.P., Stadler T. 2014. The fossilized birth-death process for coherent
calibration of divergence-time estimates. Proc. Natl. Acad. Sci. U.S.A.:E2957-E2966.
Horvath J.E., Weisrock D.W., Embry S.L., Fiorentino I., Balhoff J.P., Kappeler P., Wray G.A.,
Willard H.F., Yoder A.D. 2008. Development and application of a phylogenomic toolkit:
Resolving the evolutionary history of Madagascar’s lemurs. Genome Res. 18:489-499.
by Liliana Davalos on April 25, 2016 from
IUCN 2015. The IUCN Red List of Threatened Species. Version 2015-3.
<>. Downloaded February 28 2015.
Johnson S.E., Lei R., Martin S.K., Irwin M.T., Louis E.E. 2008. Does Eulemur cinereiceps exist?
Preliminary evidence from genetics and ground surveys in southeastern Madagascar. Am. J.
Primatol. 70:372-385.
Jungers W.L., Godfrey L.R., Simons E.L., Chatrath P.S. 1997. Phalangeal curvature and
positional behavior in extinct sloth lemurs (Primates, Palaeopropithecidae). Proc. Natl. Acad.
Sci. U.S.A. 94:11998.
Jungers W.L., Godfrey L.R., Simons E.L., Chatrath P.S., Rakotosamimanana B. 1991.
Phylogenetic and functional affinities of Babakotia (Primates), a fossil lemur from northern
Madagascar. Proc. Natl. Acad. Sci. U.S.A. 88:9082.
Karanth K.P., Delefosse T., Rakotosamimanana B., Parsons T.J., Yoder A.D. 2005. Ancient
DNA from giant extinct lemurs confirms single origin of Malagasy primates. Proc. Natl. Acad.
Sci. U.S.A. 102:5090-5095.
Kass R.E., Raftery A.E. 1995. Bayes factors. J. Amer. Stat. Assoc. 90:773-795.
Katoh K., Kuma K.-i., Toh H., Miyata T. 2005. MAFFT version 5: improvement in accuracy of
multiple sequence alignment. Nucleic Acids Res. 33:511-518.
Kearse M., Moir R., Wilson A., Stones-Havas S., Cheung M., Sturrock S., Buxton S., Cooper A.,
Markowitz S., Duran C. 2012. Geneious Basic: an integrated and extendable desktop software
platform for the organization and analysis of sequence data. Bioinformatics 28:1647-1649.
Kistler L., Ratan A., Godfrey L.R., Crowley B.E., Hughes C.E., Lei R., Cui Y., Wood M.L.,
Muldoon K.M., Andriamialison H. 2015. Comparative and population mitogenomic analyses of
Madagascar's extinct, giant ‘subfossil’ lemurs. J. Hum. Evol. 79:45-54.
by Liliana Davalos on April 25, 2016 from
Ksepka D.T., Parham J.F., Allman J.F., Benton M.J., Carrano M.T., Cranston K.A., Donoghue
P.C., Head J.J., Hermsen E.J., Irmis R.B. 2015. The Fossil Calibration Database, a new resource
for divergence dating. Syst. Biol.: 10.1093/sysbio/syv025.
Lambert S.M., Reeder T.W., Wiens J.J. 2015. When do species-tree and concatenated estimates
disagree? An empirical analysis with higher-level scincid lizard phylogeny. Mol. Phylogen.
Evol. 82:146-155.
Lanfear R., Calcott B., Ho S.Y.W., Guindon S. 2012. PartitionFinder: combined selection of
partitioning schemes and substitution models for phylogenetic analyses. Mol. Biol. Evol.
Larget B.R., Kotha S.K., Dewey C.N., Ané C. 2010. BUCKy: gene tree/species tree
reconciliation with Bayesian concordance analysis. Bioinformatics 26:2910-2911.
Lei R., Engberg S.E., Andriantompohavana R., McGuire S.M., Mittermeier R.A., Zaonarivelo
J.R., Brehheman R.A., Louis E.E. 2008. Nocturnal lemur diversity at Masoala National Park.
Special Publications of Texas Tech University 53:1-48.
Lewis P.O. 2001. A likelihood approach to estimating phylogeny from discrete morphological
character data. Syst. Biol. 50:913.
Lihoreau F., Boisserie J.-R., Manthi F.K., Ducrocq S. 2015. Hippos stem from the longest
sequence of terrestrial cetartiodactyl evolution in Africa. Nature Communications 6.
Louis E.E. 2006. Molecular and morphological analyses of the sportive lemurs (Family
Megaladapidae: Genus Lepilemur) reveals 11 previously unrecognized species. Special
Publictions of the Museum of Texas Tech University.
by Liliana Davalos on April 25, 2016 from
Louis E.E., Coles M.S., Andriantompohavana R., Sommer J.A., Engberg S.E., Zaonarivelo J.R.,
Mayor M.I., Brenneman R.A. 2006. Revision of the mouse lemurs (Microcebus) of eastern
Madagascar. Int. J. Primatol. 27:347-389.
Maechler M., Rousseeuw P., Struyf A., Hubert M., Hornik K., Studer M., Roudier P. 2015.
cluster: Cluster Analysis Basics and Extensions, R package version 2.0.1, CRAN.
Markolf M., Rakotonirina H., Fichtel C., von Grumbkow P., Brameier M., Kappeler P.M. 2013.
True lemurs... true species - species delimitation using multiple data sources in the brown lemur
complex. BMC Evol. Biol. 13:233.
Marivaux L., Ramdarshan A., Essid E.M., Marzougui W., Ammar H.K., Lebrun R., Marandat
B., Merzeraud G., Tabuce R., Vianey-Liaud M. 2013. Djebelemur, a tiny pre-tooth-combed
primate from the Eocene of Tunisia: a glimpse into the origin of crown strepsirhines. PLoS ONE
Masters J., Anthony N., De Wit M., Mitchell A. 2005. Reconstructing the evolutionary history of
the Lorisidae using morphological, molecular, and geological data. Am. J. Phys. Anthropol.
Meredith R.W., Janečka J.E., Gatesy J., Ryder O.A., Fisher C.A., Teeling E.C., Goodbla A.,
Eizirik E., Simão T.L., Stadler T. 2011. Impacts of the Cretaceous Terrestrial Revolution and
KPg extinction on mammal diversification. Science 334:521-524.
Miller M.A., Pfeiffer W., Schwartz T. 2010. Creating the CIPRESS Science Gateway for
inference of large phylogenetic trees. Proc. Gateway Computing Environments Workshop. New
Orleans, LA, GCE, p. 1-8.
Murphy W.J., Eizirik E., Johnson W.E., Zhang Y.P., Ryder O.A., O'Brien S.J. 2001. Molecular
phylogenetics and the origins of placental mammals. Nature 409:614-618.
by Liliana Davalos on April 25, 2016 from
Ni X., Gebo D.L., Dagosto M., Meng J., Tafforeau P., Flynn J.J., Beard K.C. 2013. The oldest
known primate skeleton and early haplorhine evolution. Nature 498:60-64.
Novacek M.J., Wheeler Q. 1992. Extinction and Phylogeny. NY, Columbia University Press.
Nunn C.L. 2011. The comparative approach in evolutionary anthropology and biology. Chicago,
University of Chicago Press.
Nylander J.A., Wilgenbusch J.C., Warren D.L., Swofford D.L. 2008. AWTY (are we there yet?):
a system for graphical exploration of MCMC convergence in Bayesian phylogenetics.
Bioinformatics 24:581-583.
Olivieri G., Zimmermann E., Randrianambinina B., Rasoloharijaona S., Rakotondravony D.,
Guschanski K., Radespiel U. 2007. The ever-increasing diversity in mouse lemurs: three new
species in north and northwestern Madagascar. Mol. Phylogen. Evol. 43:309-327.
Orlando L., Calvignac S., Schnebelen C., Douady C.J., Godfrey L.R., Hänni C. 2008. DNA from
extinct giant lemurs links archaeolemurids to extant indriids. BMC Evol. Biol. 8:121.
Parham J.F., Donoghue P.C., Bell C.J., Calway T.D., Head J.J., Holroyd P.A., Inoue J.G., Irmis
R.B., Joyce W.G., Ksepka D.T. 2011. Best practices for justifying fossil calibrations. Syst. Biol.
Pastorini J., Forstner M.R.J., Martin R.D. 2001a. Phylogenetic history of sifakas (Propithecus:
Lemuriformes) derived from mtDNA sequences. Am. J. Primatol. 53:1-17.
Pastorini J., Martin R.D., Ehresmann P., Zimmermann E., Forstner M.R.J. 2001b. Molecular
phylogeny of the lemur family Cheirogaleidae (Primates) based on mitochondrial DNA
sequences. Mol. Phylogen. Evol. 19:45-56.
Pastorini J., Forstner M.R., Martin R.D. 2002. Phylogenetic relationships of gentle lemurs
(Hapalemur). Evol. Anthrop. 11:150-154.
by Liliana Davalos on April 25, 2016 from
Pastorini J., Thalmann U., Martin R.D. 2003. A molecular approach to comparative
phylogeography of extant Malagasy lemurs. Proc. Natl. Acad. Sci. U.S.A. 100:5879-5884.
Pattinson D.J., Thompson R.S., Piotrowski A.K., Asher R.J. 2015. Phylogeny, paleontology, and
primates: do incomplete fossils bias the tree of life? Syst. Biol. 64:169-186.
Patzkowsky M.E., Holland S.M. 2012. Stratigraphic paleobiology: understanding the distribution
of fossil taxa in time and space. Chicago, University of Chicago Press.
Perelman P., Johnson W.E., Roos C., Seuánez H.N., Horvath J.E., Moreira M.A., Kessing B.,
Pontius J., Roelke M., Rumpler Y., Schneider M., Silva A., O'Brien S.J., Pecon-Slattery J. 2011.
A molecular phylogeny of living primates. PLoS Genet. 7:e1001342.
Perry G., Kistler L., Godfrey L.R., Crowley B.E., Muldoon K.M., Malhi R., Schuster S., Miller
W., Yoder A.D., Louis E.E. 2015. Nuclear genome sequences from the extinct subfossil lemurs
Palaeopropithecus ingens and Megaladapis edwardsi. Am. J. Phys. Anthropol. Volume 156
(S60): 251.
Pozzi L., Disotell T.R., Masters J.C. 2014a. A multilocus phylogeny reveals deep lineages within
African galagids (Primates: Galagidae). BMC Evol. Biol. 14:72.
Pozzi L., Hodgson J.A., Burrell A.S., Sterner K.N., Raaum R.L., Disotell T.R. 2014b. Primate
phylogenetic relationships and divergence dates inferred from complete mitochondrial genomes.
Mol. Phylogen. Evol. 75:165-183.
Pybus O.G., Harvey P.H. 2000. Testing macro–evolutionary models using incomplete molecular
phylogenies. Proc. Royl. Soc. London B: Biol. Sci. 267:2267-2272.
Pyron R., Burbrink F. 2012. Traitdependent diversification and the impact of palaeontological
data on evolutionary hypothesis testing in New World ratsnakes (tribe Lampropeltini). J. Evol.
Biol. 25:497-508.
by Liliana Davalos on April 25, 2016 from
Pyron R.A. 2011. Divergence time estimation using fossils as terminal taxa and the origins of
Lissamphibia. Syst. Biol. 60:466-481.
Pyron, R. A. 2015. Post-molecular systematics and the future of phylogenetics. Trends Ecol.
Evol.: doi:10.1016/j.tree.2015.1004.1016.
R Core Team. 2014. R: A Language and Environment for Statistical Computing. R Foundation
for Statistical Computing.
Rabosky D.L. 2014. Automatic detection of key innovations, rate shifts, and diversity-
dependence on phylogenetic trees. PloS One 9:e89543.
Rambaut A., Suchard M., Xie D., Drummond A. 2014. Tracer v1. 6. Available at: http://beast.
bio. ed. ac. uk/Tracer.
Rasmussen T., Nekaris K. 1998. Evolutionary history of lorisiform primates. Folia Primatol.
Rasoloarison R.M., Goodman S.M., Ganzhorn J.U. 2000. Taxonomic revision of mouse lemurs
(Microcebus) in the western portions of Madagascar. Int. J. Primatol. 21:963-1019.
Rasoloarison R.M., Weisrock D.W., Yoder A.D., Rakotondravony D., Kappeler P.M. 2013. Two
new species of mouse lemurs (Cheirogaleidae: Microcebus) from eastern Madagascar. Int. J.
Primatol. 34:455-469.
Reeder T.W., Townsend T.M., Mulcahy D.G., Noonan B.P., Wood Jr P.L., Sites Jr J.W., Wiens
J.J. 2015. Integrated analyses resolve conflicts over squamate reptile phylogeny and reveal
unexpected placements for fossil taxa. PloS one 10:e0118199.
Revell L.J. 2012. phytools: an R package for phylogenetic comparative biology (and other
things). Methods Ecol. Evol. 3:217-223.
Robinson D., Foulds L.R. 1981. Comparison of phylogenetic trees. Math. Biosci. 53:131-147.
by Liliana Davalos on April 25, 2016 from
Ronquist F., Klopfstein S., Vilhelmsen L., Schulmeister S., Murray D.L., Rasnitsyn A.P. 2012a.
A total-evidence approach to dating with fossils, applied to the early radiation of the
Hymenoptera. Syst. Biol. 61:973-999.
Ronquist F., Teslenko M., van der Mark P., Ayres D.L., Darling A., Höhna S., Larget B., Liu L.,
Suchard M.A., Huelsenbeck J.P. 2012b. MrBayes 3.2: efficient Bayesian phylogenetic inference
and model choice across a large model space. Syst. Biol. 61:539-542.
Rose K.D. 2006. The beginning of the age of mammals. Baltimore, MD, Johns Hopkins
University Press.
Rumpler Y., Hauwy M., Fausser J.-L., Roos C., Zaramody A., Andriaholinirina N., Zinner D.
2011. Comparing chromosomal and mitochondrial phylogenies of the Indriidae (Primates,
Lemuriformes). Chrom. Res. 19:209-224.
Sansom R.S. 2015. Bias and sensitivity in the placement of fossil taxa resulting from
interpretations of missing data. Syst. Biol. 64:256-266.
Schwitzer C., Mittermeier R., Davies N., Johnson S.E., Ratsimbazafy J., Razafindramanana J.,
Louis Jr E.E., Rajaobelina S. 2013. Lemurs of Madagascar: a stragey for their conservation
2013-2016. Bristol, UK, IUCN SSC Primate Specialist Group, Bristol Conservation and Science
Foundation, Conservation International.
Seiffert E.R. 2007. Evolution and extinction of Afro-Arabian primates near the Eocene-
Oligocene boundary. Folia Primatol. 78:314-327.
Seiffert E.R. 2012. Early primate evolution in AfroArabia. Evol. Anthropol. 21:239-253.
Seiffert ER, Costeur L, and Boyer DM. 2015. Primate tarsal bones from Egerkingen,
Switzerland, attributable to the middle Eocene adapiform Caenopithecus lemuroides. PeerJ
by Liliana Davalos on April 25, 2016 from
Seiffert E.R., Perry J.M.G., Simons E.L., Boyer D.M. 2009. Convergent evolution of anthropoid-
like adaptations in Eocene adapiform primates. Nature 461:1118-1121.
Seiffert E.R., Simons E.L., Attia Y. 2003. Fossil evidence for an ancient divergence of lorises
and galagos. Nature 422:421-424.
Seiffert E.R., Simons E.L., Boyer D.M., Perry J.M.G., Ryan T.M., Sallam H.M. 2010. A fossil
primate of uncertain affinities from the earliest late Eocene of Egypt. Proc. Natl. Acad. Sci.
U.S.A. 107:9712.
Seiffert E.R., Simons E.L., Ryan T.M., Attia Y. 2005. Additional remains of Wadilemur elegans,
a primitive stem galagid from the late Eocene of Egypt. Proc. Natl. Acad. Sci. U.S.A. 102:11396.
Silvestro D., Schnitzler J., Liow L.H., Antonelli A., Salamin N. 2014. Bayesian estimation of
speciation and extinction from incomplete fossil occurrence data. Syst. Biol. 63:349-367.
Simons E.L. 1992. Diversity in the early Tertiary anthropoidean radiation in Africa. Proc. Natl.
Acad. Sci. U.S.A. 89:10743-10747.
Simons E.L., Rasmussen D.T. 1994. A remarkable cranium of Plesiopithecus teras (Primates,
Prosimii) from the Eocene of Egypt. Proc. Natl. Acad. Sci. U.S.A. 91:9946-9950.
Slater G.J. 2013. Phylogenetic evidence for a shift in the mode of mammalian body size
evolution at the CretaceousPalaeogene boundary. Meth. Ecol. Evol. 4:734-744.
Slater G.J., Harmon L.J. 2013. Unifying fossils and phylogenies for comparative analyses of
diversification and trait evolution. Meth. Ecol. Evol. 4:699-702.
Slater G.J., Harmon L.J., Alfaro M.E. 2012. Integrating fossils with molecular phylogenies
improves inference of trait evolution. Evolution 66:3931-3944.
by Liliana Davalos on April 25, 2016 from
Spaulding M., O'Leary M.A., Gatesy J. 2009. Relationships of Cetacea (Artiodactyla) among
mammals: increased taxon sampling alters interpretations of key fossils and character evolution.
PLoS ONE 4:e7062.
Springer M.S., Meredith R.W., Gatesy J., Emerling C.A., Park J., Rabosky D.L., Stadler T.,
Steiner C., Ryder O.A., Janečka J.E. 2012. Macroevolutionary dynamics and historical
biogeography of primate diversification inferred from a species supermatrix. PLoS ONE
Steiper M.E., Seiffert E.R. 2012. Evidence for a convergent slowdown in primate molecular rates
and its implications for the timing of early primate evolution. Proc. Natl. Acad. Sci. U.S.A.
Sterling E.J. 1994. Aye-Ayes: specialists on structurally defended resources. Folia Primatol.
Tattersall I., Schwartz J. 1974. Craniodental morphology and the systematics of the Malagasy
lemurs (Primates, Prosimii). Anthropol. Papers Amer. Mus. Nat. Hist. 52:1-60.
Tattersall I., Schwartz J. 1991. Phylogeny and nomenclature in the Lemur-group of Malagasy
strepsirhine primates. Anthropol. Papers Amer. Mus. Nat. Hist. 69:1-24.
Thiele D., Razafimahatratra E., Hapke A. 2013. Discrepant partitioning of genetic diversity in
mouse lemurs and dwarf lemurs–Biological reality or taxonomic bias? Mol. Phylogen. Evol.
Thiele K. 1993. The holy grail of the perfect character: the cladistic treatment of morphometry
data. Cladistics 9:275–304.
Thorne J.L., Kishino H. 2002. Divergence time and evolutionary rate estimation with multilocus
data. Syst. Biol. 51:689-702.
by Liliana Davalos on April 25, 2016 from
Wagner P.J. 2000. Exhaustion of morphologic character states among fossil taxa. Evolution
Weisrock D.W., Rasoloarison R.M., Fiorentino I., Ralison J.M., Goodman S.M., Kappeler P.M.,
Yoder A.D. 2010. Delimiting species without nuclear monophyly in Madagascar's mouse lemurs.
PLoS ONE 5:e9883.
Wiens J.J. 2000. Phylogenetic analysis of morphological data. Washington, DC.: Smithsonian
Institution Press 1:233-235.
Wiens J.J., Kuczynski C.A., Townsend T., Reeder T.W., Mulcahy D.G., Sites Jr J.W. 2010.
Combining phylogenomics and fossils in higher-level squamate reptile phylogeny: molecular
data change the placement of fossil taxa. Syst. Biol. 59:674-688.
Wiens J.J., Morrill M.C. 2011. Missing data in phylogenetic analysis: reconciling results from
simulations and empirical data. Syst. Biol. 60:719-731.
Wiens J.J., Tiu J. 2012. Highly incomplete taxa can rescue phylogenetic analyses from the
negative impacts of limited taxon sampling. PLoS One 7:e42925.
Wilkinson R.D., Steiper M.E., Soligo C., Martin R.D., Yang Z., Tavaré S. 2010. Dating primate
divergences through an integrated analysis of palaeontological and molecular data. Syst. Biol.
Wood H.M., Matzke N.J., Gillespie R.G., Griswold C.E. 2012. Treating fossils as terminal taxa
in divergence time estimation reveals ancient vicariance patterns in the palpimanoid spiders.
Syst. Biol. 62:264-284.
Xie W., Lewis P.O., Fan Y., Kuo L., Chen M.-H. 2010. Improving marginal likelihood
estimation for Bayesian phylogenetic model selection. Syst. Biol. 60:150-160.
by Liliana Davalos on April 25, 2016 from
Yang Z. 2008. Empirical evaluation of a prior for Bayesian phylogenetic inference. Phil. Trans.
Royl. Soc. B: Biol. Sci. 363:4031.
Yang Z., Rannala B. 2005. Branch-length prior influences Bayesian posterior probability of
phylogeny. Syst. Biol. 54:455-470.
Yoder A.D. 1994. Relative position of the Cheirogaleidae in strepsirhine phylogeny: a
comparison of morphological and molecular methods and results. Am. J. Phys. Anthropol.
Yoder A.D., Cartmill M., Ruvolo M., Smith K., Vilgalys R. 1996. Ancient single origin for
Malagasy primates. Proc. Natl. Acad. Sci. U.S.A. 93:5122-5126.
Yoder A.D., Irwin J.A. 1999. Phylogeny of the Lemuridae: effects of character and taxon
sampling on resolution of species relationships within Eulemur. Cladistics 15:351-361.
Yoder A.D., Irwin J.A., Payseur B.A. 2001. Failure of the ILD to determine data combinability
for slow loris phylogeny. Syst. Biol. 50:408-424.
Yoder A., Yang Z. 2000. Estimation of Primate Speciation Dates Using Local Molecular Clocks.
Mol. Biol. Evol. 17:1081 - 1090.
Yoder A.D., Yang Z. 2004. Divergence dates for Malagasy lemurs estimated from multiple gene
loci: geological and evolutionary context. Mol. Ecol. 13:757-773.
Zaramody A., Fausser J.-L., Roos C., Zinner D., Andriaholinirina N., Rabarivola C., Norscia I.,
Tattersall I., Rumpler Y. 2006. Molecular phylogeny and taxonomic revision of the eastern
woolly lemurs (Avahi laniger). Primate Reports 74:9-23.
Zhang J., Mi X., Pei N. 2010. Phylotools: Phylogenetic tools for ecologists. R package version
0.0 7:201019.
by Liliana Davalos on April 25, 2016 from
Zhang C., Stadler T., Klopfstein S., Heath T.A., Ronquist F. 2015. Total-Evidence Dating under
the Fossilized Birth-Death Process. Syst. Biol. doi:10.1093/sysbio/syv080.
Zheng Y., Peng R., Kuro-o M., Zeng X. 2011. Exploring patterns and extent of bias in estimating
divergence time from mitochondrial DNA sequence data in a particular lineage: a case study of
salamanders (Order Caudata). Mol. Biol. Evol. 28:2521-2535.
Figure legends
Figure 1. Simplified phylogeny of study taxa illustrating the systematics referred to in the text
and the relationships of fossil taxa (indicated with crosses). Taxa for which morphological data
were available are depicted with crossbones, and taxa for which molecular data were available
are depicted with a double helix.
Figure 2. Schematic of the study workflow illustrating data used, data processing procedures,
and analytical techniques.
Figure 3. Time-calibrated maximum clade credibility phylogeny inferred from a total evidence
dataset (421 morphological, 5767 protein-coding molecular characters) using the fossilized birth-
death process model. Node supports are illustrated with color coding. The time scale is in
millions of years ago. The family names are given with illustrations of representative taxa.
Representatives of the extinct subfossils are shown for each family. Illustrations of extant taxa by
S. Nash in Schwitzer et al. 2013, extinct subfossils are by S. Nash in Mittermeier et al. 2010.
Figure 4. Comparison of divergence-time estimates from two techniques used in this study, the
Tip-Dating (TD) and the Fossilized Birth-Death Process (FBD) methods, and the full and
reduced morphological data matrix for each technique. Circles indicate the median age estimate
and bars encompass the 95% highest probability distribution (HPD), in millions of years ago.
by Liliana Davalos on April 25, 2016 from
Nodes are referred to by taxonomic names as in Figures 1 and 3. The two techniques differ in the
divergence time estimates for key nodes in the phylogeny, with the TD method estimating ages
that are ~9-12 million years older than the FBD method, on average.
by Liliana Davalos on April 25, 2016 from
Stem strepsirrhines
by Liliana Davalos on April 25, 2016 from
by Liliana Davalos on April 25, 2016 from
Cretaceous Plio.
60 40
20 0Ma
Pachylemur jullyi
Archaeolemur majori
by Liliana Davalos on April 25, 2016 from
African/Asian Lorisidae
0 20 40 60 80
Age (Ma)
Dating technique
FBD reduced
FBD full
TD reduced
TD full
Comparisons of median node age estimates + 95% HPD
between dating techniques
by Liliana Davalos on April 25, 2016 from
... A single representative of each species was included for all except for P. diadema, where two individuals were included to span the intraspecific variation observed in our initial phylogenetic analyses. A single out-group sequence was included, RANOL6, A. peyrierasi, to root the tree and add a calibration point based on three previous estimates [66][67][68]. From Herrera and Davalos [67], the root height of the Avahi-Propithecus split was estimated as having occurred 20.64-19.34 million years ago, as such the average of 19.99 MYA was used as a prior calibration point using a normal distribution with a 5% at 18.26 MYA and a 95% at 21.54 MYA. A separate analysis was performed using the splits estimated in Kistler et al. [68], as there are significant age discrepancies between the two datasets. ...
... A single out-group sequence was included, RANOL6, A. peyrierasi, to root the tree and add a calibration point based on three previous estimates [66][67][68]. From Herrera and Davalos [67], the root height of the Avahi-Propithecus split was estimated as having occurred 20.64-19.34 million years ago, as such the average of 19.99 MYA was used as a prior calibration point using a normal distribution with a 5% at 18.26 MYA and a 95% at 21.54 MYA. A separate analysis was performed using the splits estimated in Kistler et al. [68], as there are significant age discrepancies between the two datasets. ...
... Topologies confirmed previous studies including [67] with Indri branching first, followed by Avahi, and Propithecus was monophyletic (trees were rooted on L. catta; Figure 1). Within Propithecus, the divergent east-west split was confirmed (posterior probability = 1.0), and the species composition was as predicted with P. edwardsi (pp = 1.0, clade numbers have been added to all figures and are consistently referenced hereafter, clade 1), P. candidus (pp = 1.0, clade 2), P. perrieri (pp = 1.0, clade 3), and P. diadema (clade 4 and 5, pp 1.0 for both) forming the eastern group, and P. tattersalli (pp = 1.0, clade 6), P. coquereli (pp = 1.0, clade 7), P. coronatus (pp = 1.0, clade 8), P. deckenii (pp = 1.0, clade 9), and P. verreauxi (pp = 1.0, clade 10) forming the western group. ...
Full-text available
The most comprehensive phylogenomic reconstruction to date was generated on all nominal taxa within the lemur genus Propithecus. Over 200 wild-caught individuals were included in this study to evaluate the intra and interspecific relationships across this genus. Ultraconserved Elements (UCEs) resulted in well-supported phylogenomic trees. Complete mitochondrial genomes (CMGs) largely agreed with the UCEs, except where a mitochondrial introgression was detected between one clade of the diademed sifaka (Propithecus diadema) and the Milne-Edwards sifaka (P. edwardsi). Additionally, the crowned (P. coronatus) and Von der Decken’s (P. deckeni) sifakas belonged to a single admixed lineage from UCEs. Further sampling across these two species is warranted to determine if our sampling represents a hybrid zone. P. diadema recovered two well-supported clades, which were dated and estimated as being ancient as the split between the Perrier’s (P. perrierii) and silky (P. candidus) sifakas. The reconstructed demographic history of the two clades also varied over time. We then modeled the modern ecological niches of the two cryptic P. diadema clades and found that they were significantly diverged (p < 0.01). These ecological differences result in a very limited zone of geographic overlap for the P. diadema clades (<60 km2). Niche models also revealed that the Onive River acts as a potential barrier to dispersal between P. diadema and P. edwardsi. Further taxonomic work is required on P. diadema to determine if its taxonomic status should be revised. This first genomic evaluation of the genus resolved the relationships between the taxa and the recovered cryptic diversity within one species.
... Stemming from an inferred frugivorous common ancestor, primates of the suborder Strepsirrhini underwent one of the most impressive adaptive radiations among living primates [26], coupled with significant ecomorphological diversification [25]. The primate suborder Strepsirrhini diverged from haplorrhines approximately 60Ma [26]. ...
... Stemming from an inferred frugivorous common ancestor, primates of the suborder Strepsirrhini underwent one of the most impressive adaptive radiations among living primates [26], coupled with significant ecomorphological diversification [25]. The primate suborder Strepsirrhini diverged from haplorrhines approximately 60Ma [26]. Comprising more than 120 living species, they exhibit intermediate relative brain size between non-primate mammals and more derived anthropoids [8,9], and a variety of dietary and foraging specialisations that are reflected in their dental morphology [25][26][27]. ...
... The primate suborder Strepsirrhini diverged from haplorrhines approximately 60Ma [26]. Comprising more than 120 living species, they exhibit intermediate relative brain size between non-primate mammals and more derived anthropoids [8,9], and a variety of dietary and foraging specialisations that are reflected in their dental morphology [25][26][27]. Strepsirrhines are thought to have diversified in continental Africa, experiencing an early Oligocene partial extinction event, followed by colonisation of Madagascar, perhaps in two independent events [28,29]. Once in Madagascar, strepsirrhines diversified to fill a range of dietary niches (including folivory, frugivory, gummivory and insectivory) and specialise for different activity periods [25]. ...
Full-text available
The evolution of the remarkably complex primate brain has been a topic of great interest for decades. Multiple factors have been proposed to explain the comparatively larger primate brain (relative to body mass), with recent studies indicating diet has the greatest explanatory power. Dietary specialisations also correlate with dental adaptations, providing a potential evolutionary link between brain and dental morphological evolution. However, unambiguous evidence of association between brain and dental phenotypes in primates remains elusive. Here we investigate the effect of diet on variation in primate brain and dental morphology and test whether the two anatomical systems coevolved. We focused on the primate suborder Strepsirrhini, a living primate group that occupies a very wide range of dietary niches. By making use of both geometric morphometrics and dental topographic analysis, we extend the study of brain-dental ecomorphological evolution beyond measures of size. After controlling for allometry and evolutionary relatedness, differences in brain and dental morphology were found between dietary groups, and brain and dental morphologies were found to covary. Historical trajectories of morphological diversification revealed a strong integration in the rates of brain and dental evolution and similarities in their modes of evolution. Combined, our results reveal an interplay between brain and dental ecomorphological adaptations throughout strepsirrhine evolution that can be linked to diet.
... Our study aims to help resolve the relative importance of host evolutionary history and ecology in lemurs, a sub-order of primates endemic to the island of Madagascar. The lemur ancestor rafted from continental Africa to Madagascar 50-70 million years ago [23][24][25], where they experienced an adaptive radiation and diversified into over 100 species [25,26]. Evidence suggests lemurs evolved in an ancestral climate niche characterized by high rainfall and mild seasonality [27]. ...
... Our study aims to help resolve the relative importance of host evolutionary history and ecology in lemurs, a sub-order of primates endemic to the island of Madagascar. The lemur ancestor rafted from continental Africa to Madagascar 50-70 million years ago [23][24][25], where they experienced an adaptive radiation and diversified into over 100 species [25,26]. Evidence suggests lemurs evolved in an ancestral climate niche characterized by high rainfall and mild seasonality [27]. ...
... Fig. 1 The phylogenetic and ecological diversity of lemur species sampled in this study. A Lemur phylogeny adapted from Herrera and Dávalos [25]. Bolded tip labels denote species sampled in our study. ...
Full-text available
Mammals harbor diverse gut microbiomes (GMs) that perform critical functions for host health and fitness. Identifying factors associated with GM variation can help illuminate the role of microbial symbionts in mediating host ecological interactions and evolutionary processes, including diversification and adaptation. Many mammals demonstrate phylosymbiosis—a pattern in which more closely-related species harbor more similar GMs—while others show overwhelming influences of diet and habitat. Here, we generated 16S rRNA sequence data from fecal samples of 15 species of wild lemurs across southern Madagascar to (1) test a hypothesis of phylosymbiosis, and (2) test trait correlations between dietary guild, habitat, and GM diversity. Our results provide strong evidence of phylosymbiosis, though some closely-related species with substantial ecological niche overlap exhibited greater GM similarity than expected under Brownian motion. Phylogenetic regressions also showed a significant correlation between dietary guild and UniFrac diversity, but not Bray-Curtis or Jaccard. This discrepancy between beta diversity metrics suggests that older microbial clades have stronger associations with diet than younger clades, as UniFrac weights older clades more heavily. We conclude that GM diversity is predominantly shaped by host phylogeny, and that microbes associated with diet were likely acquired before evolutionary radiations within the lemur families examined.
... 6.3). These analyses of the Miocene fossils provide the first data sets for applying tip dating methods to the phyllostomid phylogeny (e.g., Herrera and Dávalos 2016). Until now, all dating analyses have relied on node dating methods (table 6.1), constraining particular nodes based on taxonomy or presumed phylogenetic placement, but without including fossil taxa as tips or fitting models of evolution to morphological characters (Heath et al. 2014;Ronquist et al. 2012). ...
... 6.3), support for those relationships is low, and future analyses should benefit from including several Oligocene close relatives of phyllostomids Morgan and Czaplewski 2012). Relaxed clock analyses accounting for the instability of the nodes offer a potential route forward (e.g., Herrera and Dávalos 2016) but have yet to be implemented, despite the publication of several new data sets (table 6.1). ...
... Phylogenetic analyses of morphological characters, however, assume the independent evolution of characters (O'Keefe and Wagner 2001). Although Dávalos et al. (2014) measured the expected excess of similarity, they did not exclude characters to assess their effects on phylogeny (e.g., Herrera and Dávalos 2016) or model the evolution of associated traits to determine the contribution of developmental constraint relative to natural selection in phyllostomid evolution (e.g., Bartoszek et al. 2012). The two latter approaches are promising avenues for disentangling the contributions of adaptation and constraint to evolution and uncovering the role of natural selection in trait diversity. ...
... Madagascar was also home to several larger bodied, extinct, "subfossil" lineages that were also likely characterized by folivory, including the "sloth lemurs" (family: Paleopropithecidae) and the koala lemurs (genus: Megaladapis) (Yoder, 1999;Fleagle, 2013;Kistler et al., 2015;Marciniak et al., 2021). It is difficult to estimate the number of times that folivory evolved independently in lemurs with confidence, especially given challenges to reconstructing the phylogenetic relationships among lemur families (Horvath et al., 2008;Perelman et al., 2011;McLain et al., 2012;Springer et al., 2012;Herrera and Dávalos, 2016), likely due to a series of ancient rapid divergences resulting in incomplete lineage sorting (Horvath et al., 2008;Marciniak et al., 2021). However, the ancestral lemurid is thought to be a generalist, with folivory in bamboo lemurs representing convergence with other lemur lineages (Ballhorn et al., 2016;Fulwood et al., 2021). ...
Full-text available
The lemurs of Madagascar include numerous species characterized by folivory across several families. Many extant lemuriform folivores exist in sympatry in Madagascar’s remaining forests. These species avoid feeding competition by adopting different dietary strategies within folivory, reflected in behavioral, morphological, and microbiota diversity across species. These conditions make lemurs an ideal study system for understanding adaptation to leaf-eating. Most folivorous lemurs are also highly endangered. The significance of folivory for conservation outlook is complex. Though generalist folivores may be relatively well equipped to survive habitat disturbance, specialist folivores occupying narrow dietary niches may be less resilient. Characterizing the genetic bases of adaptation to folivory across species and lineages can provide insights into their differential physiology and potential to resist habitat change. We recently reported accelerated genetic change in RNASE1 , a gene encoding an enzyme (RNase 1) involved in molecular adaptation in mammalian folivores, including various monkeys and sifakas (genus Propithecus ; family Indriidae). Here, we sought to assess whether other lemurs, including phylogenetically and ecologically diverse folivores, might show parallel adaptive change in RNASE1 that could underlie a capacity for efficient folivory. We characterized RNASE1 in 21 lemur species representing all five families and members of the three extant folivorous lineages: (1) bamboo lemurs (family Lemuridae), (2) sportive lemurs (family Lepilemuridae), and (3) indriids (family Indriidae). We found pervasive sequence change in RNASE1 across all indriids, a d N /d S value > 3 in this clade, and evidence for shared change in isoelectric point, indicating altered enzymatic function. Sportive and bamboo lemurs, in contrast, showed more modest sequence change. The greater change in indriids may reflect a shared strategy emphasizing complex gut morphology and microbiota to facilitate folivory. This case study illustrates how genetic analysis may reveal differences in functional traits that could influence species’ ecology and, in turn, their resilience to habitat change. Moreover, our results support the body of work demonstrating that not all primate folivores are built the same and reiterate the need to avoid generalizations about dietary guild in considering conservation outlook, particularly in lemurs where such diversity in folivory has probably led to extensive specialization via niche partitioning.
Full-text available
Strepsirrhine vocalisations are extraordinarily diverse and cross-species comparisons are needed to explore how this variability evolved. We contributed to the investigation of primate acoustic diversity by comparing the vocal repertoire of two sym-patric lemur species, Propithecus diadema and Indri indri. These diurnal species belong to the same taxonomic family and have similar activity patterns but different social structures. These features make them excellent candidates for an investigation of the phylogenetic, environmental, and social influence on primate vocal behavior. We recorded 3 P. diadema groups in 2014 and 2016. From 1,872 recordings we selected and assigned 3814 calls to 9 a priori call types, on the basis of their acoustic structure. We implemented a reproducible technique performing an acoustic feature extraction relying on frequency bins, t-SNE data reduction, and a hard-clustering analysis. We first quantified the vocal repertoire of P. diadema, finding consistent results for the 9 putatively identified call types. When comparing this repertoire with a previously published repertoire of I. indri, we found highly species-specific repertoires, with only 2% of the calls misclassified by species identity. The loud calls of the two species were very distinct, while the low-frequency calls were more similar. Our results pinpoint the role of phylogenetic history, social and environmental features on the evolution of communicative systems and contribute to a deeper understanding of the evolutionary roots of primate vocal differentiation. We conclude by arguing that standardized and reproducible techniques, like the one we employed, allow robust comparisons and should be prioritized in the future.
Hair (i.e., pelage/fur) is a salient feature of primate (including human) diversity and evolution—serving functions tied to thermoregulation, protection, camouflage, and signaling—but wild primate pelage evolution remains relatively understudied. Specifically, assessing multiple hypotheses across distinct phylogenetic scales is essential but is rarely conducted. We examine whole body hair color and density variation across Indriidae (Avahi, Indri, Propithecus)—a lineage that, like humans, exhibits vertical posture (i.e., their whole bodies are vertical to the sun). Our analyses consider multiple phylogenetic scales (family‐level, genus‐level) and hypotheses (e.g., Gloger's rule, the body cooling hypotheses). We obtain hair color and density from museum and/or wild animals, opsin genotypes from wild animals, and climate data from WorldClim. To analyze our data, we use phylogenetic generalized linear mixed models (PGLMM) using Markov chain Monte Carlo algorithms. Our results show that across the Indriidae family, darker hair is typical in wetter regions. However, within Propithecus, dark black hair is common in colder forest regions. Results also show pelage redness increases in populations exhibiting enhanced color vision. Lastly, we find follicle density on the crown and limbs increases in dry and open environments. This study highlights how different selective pressures across distinct phylogenetic scales have likely acted on primate hair evolution. Specifically, our data across Propithecus may implicate thermoregulation and is the first empirical evidence of Bogert's rule in mammals. Our study also provides rare empirical evidence supporting an early hypothesis on hominin hair evolution. Adherence to Bogert's rule across sifaka lemurs. Darker coat colors are more common where it is colder–a potential implication for thermoregulation.
In recent years, multiple technological and methodological advances have increased our ability to estimate phylogenies, leading to more accurate dating of the primate tree of life. Here we provide an overview of the limitations and potentials of some of these advancements and discuss how dated phylogenies provide the crucial temporal scale required to understand primate evolution. First, we review new methods, such as the total‐evidence dating approach, that promise a better integration between the fossil record and molecular data. We then explore how the ever‐increasing availability of genomic‐level data for more primate species can impact our ability to accurately estimate timetrees. Finally, we discuss more recent applications of mutation rates to date divergence times. We highlight example studies that have applied these approaches to estimate divergence dates within primates. Our goal is to provide a critical overview of these new developments and explore the promises and challenges of their application in evolutionary anthropology.
Full-text available
The gray mouse lemur (Microcebus murinus), one of the world’s smallest primates, is thought to share a similar ecological niche and many anatomical traits with early euprimates. As a result, it has been considered a suitable model system for early primate physiology and behavior. Moreover, recent studies have demonstrated that mouse lemurs have comparable cognitive abilities and cortical functional organization as haplorhines. Finally, the small brain size of mouse lemurs provides us with actual lower limits for miniaturization of functional brain circuits within the primate clade. Considering its phylogenetic position and early primate-like traits, the mouse lemurs are a perfect model species to study the early evolution of primate brains.
Full-text available
Estimation of divergence times is usually done using either the fossil record or sequence data from modern species. We provide an integrated analysis of palaeontological and molecular data to give estimates of primate divergence times that utilize both sources of information. The number of preserved primate species discovered in the fossil record, along with their geological age distribution, is combined with the number of extant primate species to provide initial estimates of the primate and anthropoid divergence times. This is done by using a stochastic forwards-modeling approach where speciation and fossil preservation and discovery are simulated forward in time. We use the posterior distribution from the fossil analysis as a prior distribution on node ages in a molecular analysis. Sequence data from two genomic regions (CFTR on human chromosome 7 and the CYP7A1 region on chromosome 8) from 15 primate species are used with the birth–death model implemented in mcmctree in PAML to infer the posterior distribution of the ages of 14 nodes in the primate tree. We find that these age estimates are older than previously reported dates for all but one of these nodes. To perform the inference, a new approximate Bayesian computation (ABC) algorithm is introduced, where the structure of the model can be exploited in an ABC-within-Gibbs algorithm to provide a more efficient analysis.
Mathematical modeling of cladogenesis and fossil preservation is used to explore the expected behavior of commonly used measures of taxonomic diversity and taxonomic rates with respect to interval length, quality of preservation, position of interval in a stratigraphic succession, and taxonomic rates themselves. Particular attention is focused on the independent estimation of origination and extinction rates. Modeling supports intuitive and empirical arguments that single-interval taxa, being especially sensitive to variation in preservation and interval length, produce many undesirable distortions of the fossil record. It may generally be preferable to base diversity and rate measures on estimated numbers of taxa extant at single points in time rather than to adjust conventional interval-based measures by discarding single-interval taxa. A combination of modeling and empirical analysis of fossil genera supports two major trends in marine animal evolution. (1) The Phanerozoic decline in taxonomic rates is unlikely to be an artifact of secular improvement in the quality of the fossil record, a point that has been argued before on different grounds. (2) The post-Paleozoic rise in diversity may be exaggerated by the essentially complete knowledge of the living fauna, but this bias is not the principal cause of the pattern. The pattern may partly reflect a secular increase in preservation nevertheless. Apparent temporal variation in taxonomic rates can be produced artificially by variation in preservation rate. Some empirical arguments suggest, however, that much of the short-term variation in taxonomic rates observed in the fossil record is real. (1) For marine animals as a whole, the quality of the fossil record of a higher taxon is not a good predictor of its apparent variability in taxonomic rates. (2) For a sample data set covering a cross-section of higher taxa in the Ordovician, most of the apparent variation in origination and extinction rates is not statistically attributable to independently measured variation in preservation rates. (3) Previous work has shown that standardized sampling to remove effects of variable preservation and sampling yields abundant temporal variation in estimated taxonomic rates. While modeling suggests which rate measures are likely to be most accurate in principle, the question of how best to capture true variation in taxonomic rates remains open.
Phylogenies reconstructed from gene sequences can be used to investigate the tempo and mode of species diversification. Here we develop and use new statistical methods to infer past patterns of speciation and extinction from molecular phylogenies. Specifically, we test the null hypothesis that per-lineage speciation and extinction rates have remained constant through time. Rejection of this hypothesis may provide evidence for evolutionary events such as adaptive radiations or key adaptations. In contrast to previous approaches, our methods are robust to incomplete taxon sampling and are conservative with respect to extinction. Using simulation we investigate, first, the adverse effects of failing to take incomplete sampling into account and, second, the power and reliability of our tests. When applied to published phylogenies our tests suggest that, in some cases, speciation rates have decreased through time.