Taxacom: vicariance, dispersal, ancestral areas etc

[John Grehan]

I suspect not many have looked at the supplementary files inHeads' (2025)
biogeography of oranges article. But to be of assistance to those for whom
the content is relevant to their studies I have appended below the text
concerning some important concepts and how they affect
biogeographic analysis. I challenge anyone to demonstrate a critical flaw.
This had never been done in more than half a century, so I don't expect it
will happen now.

Cheers, John

*Appendix S1. Methods in biogeography*



*Ancestral distribution*

Consider another hypothetical example of vicariance, in which an African
clade and its Asian sister originated from an ancestor widespread through
Africa and Asia. The origin of the wide range *in this ancestor* would
require establishing the range of *its* sister. The Africa-Asia ancestor
may have developed its wide range either through range expansion by
dispersal, or through vicariance with a sister in, for example, America.



*Vicariance and dispersal explanations of allopatry *

There have been two main explanations for allopatry, the first being
*dispersal* in the sense of Darwin (1859): ?dispersal by occasional means?.
This mode of allopatric speciation is also termed ?long-distance dispersal?
(although the distance covered may be short), ?peripatric speciation?,
?founder-event speciation?, ?cladogenetic dispersal?, ?jump dispersal? and
?lineage dispersal? (Hackel & Sanmartin, 2021). In this process, an
individual, or a small number of individuals, crosses what is otherwise a
barrier for the species. Chance dispersal events are not related to
environmental or biological factors (including group?s usual means of
dispersal), and the effects of the process are determined by its chance,
one-off nature.

   This type of ?chance? or ?long-distance? dispersal is accepted as a mode
of speciation and an explanation of allopatry in many biogeographic
studies. However, it has been criticised for not explaining one
observation: similar patterns of allopatry are repeated in large numbers of
groups, including groups that have very different ecology and means of
dispersal.

   In the second explanation for allopatry, *vicariance*, a widespread
ancestor breaks down by *in-*situ, allopatric differentiation into
descendant groups. Vicariance has been discussed in many studies since the
key paper by Croizat et al. (1974); the term was cited 7 times in 1973, but
2060 times in 2023 (GoogleScholar).

   The process of vicariance can have considerable explanatory power, as it
accounts for many distributions with a single event. The observed
repetition of distribution patterns in different clades suggests that whole
communities have been affected in similar ways by a common cause. Such
causes could include major topographic and environmental changes, such as
the opening of a tectonic rift or a phase of uplift. Repeated patterns in a
community would then be explained by a single event of vicariance, rather
than many, separate events of chance dispersal in each group. In a
vicariance approach, the focus is on large-scale environmental changes ?
geological, climatic, and anthropogenic ? that may explain the distribution
patterns.

   Allopatry is most widespread at lower levels in the taxonomic hierarchy.
Sister species are often allopatric, sister families less so, ad in the
oldest groups, such as orders and classes, there is often no allopatry ?
overlap is global. This observation is consistent with the idea that clades
originate in allopatry by vicariance, and that some clades then undergo
secondary range expansion leading to clade overlap.



*Normal, observed dispersal vs chance, long-distance dispersal*

In vicariance theory, range expansion occurs through normal, observed
dispersal. It is important to distinguish this process from chance,
long-distance dispersal (LDD) (Heads, 2014; Hackel & Sanmart?n, 2021).
Differences between the two include the following:



?(i) Normal dispersal is observed every day, whereas LDD is a theoretical
process; (ii) normal dispersal is not a mode of speciation (a weed
dispersing into a garden does not speciate there), whereas LDD is a mode of
speciation; (iii) normal dispersal events are mediated by a group?s
standard, observed means of dispersal, whereas LDD events are mediated by
means of dispersal that are ?non-standard? (Higgins, Nathan & Cain, 2003)
and rather enigmatic; and (iv) slowing or cessation of normal dispersal
(potentially leading to speciation) is caused by environmental or
biological change, whereas cessation of LDD (potentially leading to
speciation) is caused by chance ? the unique, idiosyncratic nature of the
LDD event. Nearly all individual organisms have moved to their present
locality, using their normal means of dispersal, but clades can evolve *in
situ*? (Heads & Saldivia, 2024).



*Vicariance, dispersal and barriers*

In the chance dispersal paradigm, clade distribution is established by
members of a clade migrating across a *pre-existing* barrier, such as a
mountain range, and this leads to isolation and allopatric speciation.
Instead, in the vicariance model, the differentiation of a group is
triggered by the *formation* of a barrier, such as the uplift of a mountain
range or the opening of an ocean basin. Here the allopatric speciation is
caused not by dispersal, but by the cessation, or at least slowing, of
dispersal across the area that becomes a new barrier.

   Following the introduction of plate tectonics and vicariance
biogeography in the 1960s-1970s, many authors have integrated results from
the two fields. Scora (1975) observed that in Rutaceae, the existence of
African aurantioids (five genera, four endemic) has ?long puzzled?
taxonomists. The question was: how, and along which route, did they get to
Africa from the main centre of diversity of aurantioids (and assumed centre
of origin) in southern Asia? Scora (1975) instead proposed a vicariance
model, with generalized aurantioids already present in both Africa and Asia
before Gondwana breakup. Other authors have also proposed vicariance in
Rutaceae (Hartley, 2000; Ladiges & Cantrill, 2007; Beattie et al., 2008;
Kubitzki et al., 2011; Heads, 2012, p. 413).



*The use of a priori areas in analysis*

Many current studies in molecular phylogenetics and biogeography map the *a
priori* areas used for the biogeographic analyses, but they do not map the
clades, the main findings of the work (e.g. Liang et al., 2024; Willink et
al., 2024; Folk et al., 2024). The details of the clade distributions are
not considered. How has this strange new development ? ?biogeography
without distributions? ? come about?

   The reliance on *a priori* areas in biogeographic analysis is determined
by the current ?ancestral area? algorithms. In order to use the algorithms,
distributions must be summarised into a very small number of areas, usually
fewer than 9; for example, Schwartz et al. (2015) analysed the
distributions of Aurantioideae genera using just six. Huang et al. (2023)
analysed distributions in *Citrus*, again using just six areas. Such
drastic simplification will obscure many critical aspects of the
distribution, including allopatry, sympatry, disjunctions, and centres of
diversity.

   The particular *a priori* areas that are selected in ?ancestral area?
studies also have serious implications for analysis. For example, if the
areas used are ?Australia? and ?China?, a group found in, say, western
Australia and western China will be coded as sympatric with one found in
eastern Australia and eastern China, even though the two groups display
perfect allopatry. (This principle is important in analysing the
distribution of *Citrus*; Sections 3.4, 3.5, 3.6).

   As discussed, basal paraphyletic grades are often restricted to a
particular region. In analysing these, the selection of the *a priori*
areas used is, again, critical. For example, consider a hypothetical group
with perfectly allopatric members and a taxon-area phylogeny: (North China
(South China (East China (West China + Eurasia)))). Here there is a
phylogeny China (China (China (China + Eurasia))) and a paraphyletic basal
grade of overlapping groups in ?China?. However, overlap is inferred only
if ?China? is accepted *a priori* as a biogeographic region. If the *a
priori* areas accepted are instead North China, South China etc., the
simple allopatry of the groups will be indicated. In this case, and there
will be no reason to assume a centre of origin in China or Eurasia. A
hypothesized derivation from a widespread China ? Eurasia ancestor would be
equally plausible.

   The present study compares and analyses distributions, but these are
examined directly, without first summarising them into *a priori* areas.
The clade distribution boundaries (based on collection localities) are
indicated in the maps, and it is these boundaries that are examined.



*?Primitive? and ?advanced?; ?basal? and ?derived?*

In a fully-resolved phylogeny, each node connects a clade and its sister.
The smaller, less diverse (species-rich) of the two is referred to as
?basal?.  However, despite the confusing term, there is no reason to assume
that a ?basal? group is any older or more primitive than its sister, that
it is ancestral to it, or that it occupies a centre of origin. For example,
a single genus, *Amborella* of New Caledonia, is sister to all other
angiosperms and so is the ?basal angiosperm?, but its sister group ? ?all
other angiosperms? ? is just as ancient. There is no reason to think that a
vicariance event should split an ancestor into two, even halves. In many
cases the division is not located at the geographic centre of the ancestor,
and while sister groups are often allopatric, the sizes of their respective
distribution areas (and their levels of diversity) are often very unequal.



*Summary on vicariance and dispersal*

Traditional (dispersal) theory and most current (program-based)
biogeography explain allopatry by chance dispersal, and overlap by normal
dispersal.  Instead, the vicariance approach followed here explains
allopatry by vicariance, overlap by normal dispersal.

   The centre of origin/dispersal model relies on idiosyncratic, chance
explanations for each individual clade. Long-distance dispersal involves
one-off events, that all occur at different times and are not related to
means of dispersal, or to any other biological or environmental factor. It
does not make any particular predictions. In contrast, vicariance predicts
that similar distributions will be shared by large numbers of taxa in a
community, and that the taxa will show a wide range in their ecology and
means of dispersal.

-- 
https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fhepialidsoftheworld.com.au%2F&data%7C02%7Ctaxacom%40lists.ku.edu%7C2b2cfb9cad534078106f08ddd1db9342%7C3c176536afe643f5b96636feabbe3c1a%7C0%7C0%7C638897463196737016%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=unR7V4nw5cKeJmo32Zb25z8bmH9zVwIOvE9V9UPX7%2Bw%3D&reserved=0 (use the 'visit archived web site'
link, then the 'Ghost Moth Research page' link.