surrounding South American sigmodontines relates to their
diversity and timing of their invasion. For instance, both
Hershkovitz (1966) and Reig (1986) suggested an old in-
vasion (early Miocene) by a sweepstakes route prior to
the Panamanian land bridge, whereas Simpson (1950) and
Flynn et al. (1985) argued for a more recent invasion (Plio-
Pleistocene) across the land bridge, thus suggesting rapid
speciation subsequent to the invasion. According to a mo-
lecular clock based on sequences of mitochondrial genes
(Engel et al. 1998), dates for the origin of South American
sigmodontines are intermediate between the late and early
hypotheses, suggesting an origin of ancestral sigmodontines
in North America (between 9 and 14.8 mya) followed by a
somewhat newer adaptive radiation in South America (be-
tween 5 and 8.3 mya). In addition, the molecular phylog-
eny suggests that South American sigmodontines are a
monophyletic group derived from a single ancestor.
Origin of the Malagasy Nesomyinae
Although Madagascar and Africa were part of Gondwana,
they separated approximately 150 –165 mya, with Mada-
gascar occupying its current position since the Cretaceous
(Rabinowitz et al. 1983). Large portions of the endemic
mammalian fauna in Madagascar represent invasions by
mainland sources derived from Africa and /or Asia. Al-
though all rodents in Madagascar are endemic and placed
in the subfamily Nesomyinae, two major controversies con-
cerning their origin and evolution persist. First, monophyly
of the Nesomyinae is disputed. Carleton and Musser (1984)
placed all genera and species from Madagascar in the same
subfamily but indicated that the primary basis for this clas-
sification was the unique distribution of nesomyines. They
also noted the high level of morphological differences sepa-
rating Malagasy genera. At least two earlier classifications
also recognized Malagasy rodents as a monophyletic group
(Simpson 1945; Lavocat 1978). Second, the traditional view
of nesomyine biogeography advocates a single invasion from
African cricetodontid rodents followed by an adaptive ra-
diation (Simpson 1952; Lavocat and Parent 1985). A con-
trasting viewpoint is that nesomyine rodents are the prod-
uct of multiple invasions, possibly from different colonizing
sources (Jansa et al. 1999). Recent studies of Malagasy ro-
dents have attempted to address these controversial issues
with the use of molecular phylogenetics. Two studies ex-
amined a small number of nesomyines in comparison to
other murid subfamilies and concluded that the subfamily
was monophyletic and related to either the African subfam-
ilies Mystromyinae or Cricetomyinae (Dubois et al. 1996;
Michaux and Catzeflis 2000). Jansa et al. (1999) examined
15 of the 21 species of nesomyines, and produced a phylo-
genetic tree based on nucleotide sequences from the cyto-
chrome bgene. Their resultant phylogeny revealed a para-
phyletic Nesomyinae, with some Malagasy taxa grouping
closer to the subfamilies Rhizomyinae and Dendromurinae.
Furthermore, an overlay of the cladogram on a map of Indo-
Africa suggested that rather than Africa, Malagasy rodents
are the result of a single invasion from Asia followed by an
adaptive radiation in Madagascar and a secondary invasion
of Africa by Malagasy lineages.
Biogeography of squirrels
Based on the fossil record, the family Sciuridae originated
in North America in the Late Eocene (35 mya) and had a
Holarctic distribution by the Early Oligocene (Thorington
et al. 1997); ancestral stocks entered Africa, parts of south-
east Asia, and South America more recently. Today, the
family is the second most diverse group of rodents (table 2.1
and fig. 2.1) and is distributed worldwide, except for Mad-
agascar, Australia, and the southern regions of South Amer-
ica (Hoffmann et al. 1993). As stated by Mercer and Roth
(2003, p. 1570), “The geographic coherence of modern
clades of squirrels suggests that dispersal over large ex-
panses of water or desert is uncommon and that squirrel
distributions may therefore be important indicators of geo-
logical and environmental history.” Therefore, based on
their distribution and fossil record, the family Sciuridae
provides an ideal model for reconstructing patterns of con-
tinental interchange in response to vicariant events initi-
ated during changes in climate and the formation of land
bridges. Two recent molecular phylogenetic studies (Mer-
cer and Roth 2003; Steppan et al. 2004) have established
relationships among the major lineages of squirrels oc-
curring worldwide (fig. 2.3), and estimates derived from a
molecular clock show congruence between the overall phy-
logenetic pattern and the establishment of overland disper-
sal pathways linking various continents (Mercer and Roth
2003). These patterns include: (1) two separate invasions
of sciurids from Asia to Africa, establishing monophyletic
groups of both ground squirrels and tree squirrels, and oc-
curring during a period when a land bridge linked these two
continents; (2) single invasion of sciurids into South Amer-
ica subsequent to the formation of the Panamanian land
bridge; (3) exchange of both flying squirrels and tree squir-
rels between North America and Eurasia during periods
when Beringea provided a land bridge for dispersal; (4) col-
onization of islands during periods of lowered sea levels and
connections with the mainland, such as seen for a mono-
phyletic group of tree squirrels occupying Sundra Shelf is-
lands in Southeast Asia. The only exception to this general
pattern is the basal position of two tree squirrel lineages
that display the extreme ranges of body size, the giant squir-
rels (Ratufa) from Borneo and the pygmy tree squirrels (Sci-
Rodent Evolution, Phylogenetics, and Biogeography 13