[Taxacom] Hauenschild et al 2018

[John Grehan]

OK, so I now have the Hauenschild et al 2018 paper.

(a) Age of taxa

Rich fossil record should constitute ideal model groups, because they are
more likely to uncover accurate age estimates”

But if fossils are minimum ages then they age estimates are still going to
be minimums.

"We used macrofossils to calibrate a maximum of seven internal nodes"

But fossils cannot do that. They can only provide minimums for the group as
a whole and any sub clade within.

Last, we used the Albian‐Aptian boundary (113 Ma) as hard bound maximum age
for our temporal reconstruction (truncated), and the Cenomanian‐ Albian
boundary (approximately 100 Ma), where the fossil record still includes
fossils with affinities to buckthorns, as soft maximum constraint
(log‐normal prior to the stem, truncated to 113 Ma)

They can set those limits, but they are not empirically based.

Results - The temporal origin of Rhamnaceae was recovered between 91.4 and
102.6 Ma (95% highest posterior density, HPD, mean: 101.0 Ma), and of the
ziziphoid clade between 74.6 and 93.2 Ma (HPD, mean: 89.8 Ma).

The minimums provided by fossils are transmogrified into actual or maximal
ages. None of this precludes earlier origins.

(b) Vicariance and chance dispersal

‘BioGeo-BEARS’ was used for all biogeographic reconstructions.

To this I would state the following excerpt from a forthcoming article
(this has been posted earlier also):

"Modern ‘ancestral area’ algorithms, such as DIVA (Ronquist, 1997), DEC
implemented in LAGRANGE (Ree & Smith, 2008), and DEC+j in BioGeoBears
(Matzke, 2013) all calculate the centre of origin at the location of a
paraphyletic basal grade. This can be seen in the phylogenies of countless
individual studies; differentiation in a clade with an area phylogeny: A (A
(A, B))) is always attributed to a centre of origin in A, followed by
dispersal to B."

However, while paraphyletic basal grades are indeed often restricted to
particular areas, such as A in Figure 1, these are not necessarily centres
of origin. It is just as likely that they are centres of differentiation in
ancestral metapopulations that were already widespread in A and another
area B. This process (Fig. 1) explains the observed phylogeny without
invoking a centre of origin in A or B, or dispersal between A and B.

In other words, the vicariance and dispersal findings by Hauenschild et al
are the product of an artificial model that cannot distinguish cases of
ancestral vicariance from dispersal for a paraphyletic basal grade. And
layered on top of that is the continued misrepresentation of minimum ages
as actual or maximum (not sure what to make of a 'soft maximum' – if it's
'soft' then it's not a maximum at all). So, yes I would ignore the result
as of no empirical standing.

John Grehan