Testing Relationships Between Biotic Components - Evolutionary Biogeography - page 171

Figure 5.2 Possible resolved area cladograms obtained for a taxon-area clado-
gram with a widespread taxon, as monophyletic (assumption 0), monophyletic and
paraphyletic (assumption 1), and monophyletic, paraphyletic, and polyphyletic
groups of areas (assumption 2). AF, Africa; AUS, Australia; NA, North America; SA,
South America.
Van Soest (1996) and van Soest and Hajdu (1997) proposed an altern-
ative treatment for widespread taxa called “no-assumption coding.” These
authors reasoned that because taxa have different means of dispersal and
are affected differentially by the environment, when comparing different tax-
on-area cladograms, usually there will be areas of greater and smaller size
partially superposed. If the smaller areas are chosen as units of the ana-
lysis, then the greater ones will generate widespread taxa, and later ma-
nipulation with assumptions 0, 1, and 2 will imply the assumption of some
of the processes previously indicated. Additionally, current biogeographic
methods allow areas to occur on a single position in the general area clado-
gram, whereas the history of the biota may indicate various positions. To
remedy these limitations, van Soest (1996) proposed a different coding of
widespread taxa. How does it work? If two or more taxon-area cladograms
have taxa widespread in the same areas, this combination can be treated as
a single area ( figs. 5.3a - 5.3d ). The comparison between the different area
cladograms will determine whether the large area is united in a general area
cladogram with its constituent areas (demonstrating that in fact it is a single
area)orit constitutes aparaphyletic orpolyphyletic group(indicating that the
individual areas are independent). In the example, both general area clado-
grams obtained ( figs. 5.3e and 5.3f ) show that area E has two different re-
lationships: It is the sister area of A ( fig. 5.3e ), and the sister area of D ( fig.
5.3f ) .
Sanmartín and Ronquist (2002) proposed another treatment for wide-
spread taxa, within the approach of the event-based methods, because as-
sumptions 0, 1, and 2 do not specify the costs these methods entail. For
these authors, the problem of a widespread taxon is solved by treating it as
an unsolved polytomy, consisting of a unit for each of the involved areas.
In order to optimize the ancestral distribution of the widespread taxon, three
options are applied ( fig. 5.4 ). These options allow different sets of possible
ancestral distributions, each with different associated costs. The option “re-
cent” assumes that the wide distribution is recent, due to dispersal. The op-
tion “ancient” assumes that the wide distribution is ancestral, due to vicari-
ance and extinction. The option “free” assumes that the wide distribution