ANTARCTIC CRETACEOUS BIODIVERSITY CHANGE 143
Fig. 2. AIbian reconstruction of Gondwana courtesy
of R. A. Livermore, British Antarctic Survey. Polar
view illustrating the importance of the West Antarctic
microcontinents (A, Antarctic Peninsula; B, Thurston
Island; C, Marie Byrd Land; D, Ellsworth-Whitmore
Mountains; E, Haag Nunataks) in maintaining
connections between east and west Gondwana. Note
also the potential importance of the Kergeulen
Plateau (F) and the Mozambique Rise (G) in the
early stages of rifting.
Fig. 3. Alternative scenarios for the expansion of
angiosperms across Gondwana plotted on an Aptian
reconstruction provided by R. A. Livermore, British
Antarctic Survey.
opening from east to west (Veevers et al 1991;
Royer & Coffin 1992). Although seafloor
spreading was well established by Aptian times
(c. 120 Ma BP), microcontinental fragments such
as the Kerguelen Plateau (continental crust)
(Fig. 2) probably maintained terrestrial connec-
tions for a considerable period of time. Recent
work has identified terrestrial deposits of
Turonian age (c. 90 Ma) on the Kerguelen
Plateau (Shipboard Scientific Party 2000),
implying that the microcontinent was above sea
level at this time.
The reconstruction of Africa/Antarctica and
the subsequent early stages of break-up are also
problematic due to a paucity of magnetic
anomaly data (Marks & Tikku 2001). Recon-
structions such as those of Lawver et al (1997),
Muller et al (1998) and Roeser et al (1996)
differ fundamentally from those of Segoufin &
Patriat (1980), Livermore & Hunter (1996) and
Marks & Tikku (2001), and consequently have
markedly different implications for the timing of
terrestrial separation. Sea floor spreading had
started by 155 Ma BP in, for example, the
Mozambique Basin (Segoufin & Patriat 1980),
and off the Dronning Maud Land Coast (Roeser
et al 1996). However, much of the initial
African/Antarctic motion had a large transform
component that probably allowed landmasses to
be juxtaposed (e.g. Mozambique Ridge), and
recent reconstructions suggest terrestrial con-
nections were present at least until 120 Ma BP
and perhaps later (Reeves & de Wit 2000). The
recent discovery of Late Cretaceous dinosaurs in
Madagascar (Sampson et al 1998), and their
occurrence in South America and India, also
suggests that these landmasses were still con-
nected (via Antarctica) in the Late Cretaceous
(Sampson et al 1998; Krause et al 1999; Hay et
al 1999). The uncertainties in the timing of
terrestrial separation between the component
continents of Gondwana remain a major
constraint in explaining present-day disjunct
distribution patterns.
Despite the present uncertainty, conventional
reconstructions of Gondwana break-up imply
separation of Africa/South America from east
Gondwana by at least 155 Ma BP (early Late
Jurassic times), and therefore substantially
before angiosperm radiation and diversification.
The only connection between east (Africa,
South America) and west (India, Australia,
