Geography Reference
In-Depth Information
Figure 5.10
Secondary Brooks parsimony analysis. (a) Data matrix with Africa rep-
resented three times (AF', AF'', and AF'''); (b) general area cladogram obtained. AF',
AF'', and AF''', Africa; AUS, Australia; NA, North America; SA, South America.
An alternative implementation of BPA was proposed by Kluge (1988). It
differs in three aspects. It considers missing areas to be uninformative, cod-
ing them with “0.” It considers widespread taxa, caused either by dispersal
orbynothavingrespondedtovicariance,tobeirrelevantandthereforedoes
not take them into account. Because for redundant distributions it is impos-
sible to determine which distribution is irrelevant (by being due to dispersal)
and which one is not, Kluge (1988) suggested that they be eliminated one
at a time, with the resulting columns weighted in proportion to their number
(e.g., if there are two redundant distributions, each of the columns will weigh
0.5, and if there are three, 0.33).
Lieberman (1997, 2000, 2003a, 2004) proposed another modification
of BPA, named modified BPA, intended to interpret geodispersal within a
cladistic biogeographic framework. The biogeographic analysis is divided in-
to two separate analyses: one to retrieve congruent episodes of vicariance
and another to retrieve congruent episodes of geodispersal (Lieberman
2004). The vicariance analysis produces a cladogram that makes predic-
tions about the relative sequence of vicariance events that fragmented bi-
otas. The geodispersal analysis produces a cladogram that provides inform-
ation about the relative sequence of vicariance events that joined the biotas.
Both cladograms provide complementary information about the biotic his-
tory and can be placed in a geological framework. The procedure implies
optimizing the ancestral states in the area cladograms in order to estimate
whether distributions implied expansions or contractions of the ancestral
areas and building the vicariance and geodispersal data matrices, following
the same procedure as BPA. The best-supported patterns of vicariance and
geodispersal emerge from the parsimony analysis of these matrices. If both
cladograms are similar, the same geological processes may have produced
vicariance and geodispersal at different times (Lieberman 2004) (e.g., cyc-
licalsealevelriseandfall).Ifthecladogramsareverydifferent,theymayim-
ply that vicariance and geodispersal have been caused by noncyclical pro-
cesses (e.g., continental collisions).
BPA has received some criticism (Carpenter 1992; Cracraft 1988; Ebach
and Edgecombe 2001; Ebach and Humphries 2002; Ebach et al. 2003; Eng-
hoff 2000; Miranda Esquivel et al. 2003; Nelson and Ladiges 1991c; Page
1993a; Parenti 2007; Platnick 1988; Ronquist and Nylin 1990; Siddall 2005;
