If a fossil record is well-sampled, first principles predict that there is then a reasonable expectation that some observed morpho-species might be ancestral to other taxa. Modern tip-dating Bayesian phylogenetics employ explicit models of sampling in the fossil record, thus providing a quantitative framework for estimating sampled ancestors. Furthermore, the processes by which new morphologically-distinguishable chrono-species arise can be distinguished by comparing estimated divergence times against the stratigraphic ranges of their inferred ancestors. Ancestor-descendant relationships are often categorized into three possible modes, with morphotaxa originating (1) via ‘anagenetic’ changes independent of branching, (2) by symmetrically ‘bifurcating’ into two new morphotaxa, or (3) by morphologically distinct descendants ‘budding’ off their morphologically-persistent ancestor. Specific cases have been attributed to each pattern, but there is little consensus on which mode is most common, or whether these dynamics vary across and within major fossil-rich groups. Our assumptions about the mode of ancestor-descendant relationships have broad impacts on our interpretation of evolutionary history, from whether we expect morphological change to correlate with speciation, or whether taxonomic turnover in the fossil record might be artificially inflated by the pseudo-extinction implicit to anagenesis and bifurcation. Here, we assemble morphological matrices and stratigraphic range data from more than thirty studies of various invertebrate groups with well-sampled fossil records, predominantly Paleozoic groups, such as brachiopods, graptolites, trilobites, crinoids and other echinoderms. These datasets differ not only in the particular clades assessed, but also the taxonomic resolution and sampling schema. Using tip-dating Bayesian analyses, we address the extent to which sampled ancestors are inferred to occur in these fossil records, as well as the degree to which these records provide support for particular modes of morphological differentiation among taxa. Authors: BAPST, David W., Geology & Geophysics, Texas A & M University, 3115 TAMU, College Station, TX 77843-3115, WRIGHT, David F., Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, DC, COLE, Selina R., Paleontology and HOPKINS, Melanie J., Division of Paleontology (Invertebrates), American Museum of Natural History, Central Park West at 79th Street, New York, NY, MELCHIN, Michael, Dept. of Earth Sciences, St. Francis Xavier Univ, Antigonish, Canada, CONGREVE, Curtis R., Geosciences, The Pennsylvania State University, University Park, PA 16802, LAMSDELL, James C., Geology & Geog/, West Virginia University, 98 Beechurst Avenue, Brooks Hall, Morgantown, WV, MITCHELL, Charles E., Geology, University at Buffalo, SUNY, 126 Cooke Hall, Buffalo, NY, and CARLSON, Sandra J., Earth and Planetary Sciences, University of California, Davis, One Shields Ave, Davis, CA

1. On the Origin of Morphospecies
Using Bayesian Phylogenetics to Quantify
the Mode of Ancestor Descendant
Relationships in Paleozoic Invertebrates
David W Bapst,
Selina Cole, David Wright, Melanie Hopkins,
Michael Melchin, Curtis Congreve, James Lamsdell,
Charles E. Mitchell, Sandra Carlson
Feel free to
tweet this talk!
@dwbapst

2. Morphotaxa in
the Fossil Record
• Often, we find specimens
with similar morphology,
but from different localities
and stratigraphic height
• We use those features to
define morphotaxa that
persist over geologic time
Dicellograptus

3. ForteyandCooper1986
The Question of Ancestors
in the Fossil Record
Do we read the fossil record too literally?

4. A New Era of Ancestors on Trees
• We expect to sample
ancestors (especially
indirect ancestors) from
first principles
Foote, 1996

5. Beast2
(PP)
MrBayes
(PP)
cal3
(prop)
Bapst,Wright,Matzke&Lloyd,2016.Biol.Lett.
A New Era of Ancestors on Trees
• We expect to sample
ancestors (especially
indirect ancestors) from
first principles
• Bayesian tip-dating and
other methods let us
quantify support for
ancestor-descendant
relationships
• Different methods seem to
assign similar support

6. AsaphidtrilobiteimagesfromGon,2009
‘Budding’
CladogenesisBut what sort of
ancestor-descendant
relationships?
Anagenesis

7. How often do we infer sampled ancestors in real
fossil records? Which mode of differentiation is
most common among persistent morphotaxa?
• Implications for our understanding of speciation,
trait evolution, extent that taxonomic turnover
could be inflated by pseudo-extinction
• Apply tip-dating to morphological matrices from
the well-sampled marine invertebrate record
• More than 35 pre-existing datasets; 26 examined today
• Predominantly Paleozoic groups, such as brachiopods,
graptolites, trilobites, crinoids, other echinoderms
• Varying matrix size, taxonomic level (species vs genera),
sampling schema (e.g. clade-focused vs interval focused)

8. Sampled-Ancestor Tip-Dating with the
Fossilized Birth-Death Model in MrBayes
• Currently available tip-dating methods limited to
considering taxon units as point occurrences in time
• To consider relationships among persistent morphotaxa,
we treat first & last appearances as two separate taxon
units, with identical morphological characters but
different time of appearance
• Functions for automating this are in R package paleotree
Gavryushkina et al., 2014

9. MCCT Tree for Analysis of
Blastozoa dataset from
Sumrall and Brett 2002

10. MCCT Tree for Analysis of
Blastozoa dataset from
Sumrall and Brett 2002
Zero-Length Tip
(=Sampled Ancestor)
Paraphyletic taxon
(=Must be budding, unless
last appearance is an SA)
Of 11 original OTUs, five
are paraphyletic,
three of which must be
budding, and two might
be examples of pseudo-
extinction

11. MCCT Tree for Analysis of
Blastozoa dataset from
Sumrall and Brett 2002
Zero-Length Tip
(=Sampled Ancestor)
Paraphyletic taxon
(=Must be budding, unless
last appearance is an SA)
Of 11 original OTUs, five
are paraphyletic,
three of which must be
budding, and two might
be examples of pseudo-
extinction
But this is just a single tree pulled
from a very large posterior – need to
measure and summarize these
patterns across the posteriors

12. 50% First & last appearances are separate
OTUs, thus might expect 50% to be
placed as sampled ancestors

13. Upper limit on possible
extent of pseudo-extinction

15. The Pattern of
Ancestor-Descendant Relationships
• Treating fossil morphotaxa as more than
single OTUs sheds light on patterns of
ancestor-descendant relationships
• Ancestor-descendant relationships
considerably across different datasets
• Effects of taxonomic practices? Sampling?
Biological differences?
Thanks for listening! Questions?

18. AsaphidtrilobiteimagesfromGon,2009
‘Budding’
Cladogenesis
Modes of
Differentiation
Anagenesis

19. AsaphidtrilobiteimagesfromGon,2009
Anagenesis
Notice that budding can
look like anagenesis
(but not vice versa)
in an incomplete record
‘Budding’

20. Each pair is a
stacked barplot
Dots indicate
putative pairs
Evidence for all
a priori AD pairs,
& a few extra
cal3 finds very
little support for
anagenesis
Given biases,
perhaps entirely
budding?
Bapst & Hopkins 2017

22. The problem is, very rarely
can we read the fossil
record as literally as this

23. How do we infer the
relationships among
ancestors & their descendants,
given the incompleteness
of the fossil record?

24. Anagenesis
‘Budding’
Cladogenesis
Previous analyses find support for
budding (Wagner & Erwin, 1995; Bapst & Hopkins, 2017)
Do morphotypes mainly arise from
asymmetric cladogenetic change?

25. ‘Budding’
Cladogenesis
Anagenesis
Budding evidenced by
paraphyly of occurrences
assigned to same morphotaxon

26. ‘Budding’
Anagenesis
Notice that budding can
look like anagenesis
(but not vice versa)
in an incomplete record

27. Gaps in Densely-Sampled
Fossil Records
Maletz and Zhang, 2003; Vandenberg, 2003; C.E. Mitchell
• Closest relatives separated by a 15 to 20 million year gap in
this lineage:
• Were the intermediates living somewhere else? Open ocean?
Bergstromgraptus
Middle Darrwillian
Sinoretiograptus
Latest Katian

28. In The Age Of Ancestor Inference…
Beast2
(PP)
MrBayes
(PP)
cal3
(prop)
Bapst&Hopkins,2017,Paleobio.
(cal3 is an off-brand
tip-dating lite)
• Different methods agree on
placing ancestors [dinosaurs]
• Quantitative inferences agree with
previous putative pairs of
ancestor-descendants [trilobites]
Bapst,Wright,Matzke&Lloyd,2016.Biol.Lett.