TRILOBITE BIOGEOGRAPHY OF GONDWANA 63
conditions in the Portixeddu and Tuviois for-
mations, Sardinia (Leone et al. 1991, as ICaly-
menesun sp.; Hammann & Leone 1997), and
Sarrabesia teichmuelleri occurs in coarse-
grained sandstones of the Punta Serpeddi
Formation, Sardinia, which are also indicative of
shelf facies (Hammann & Leone 1997). N. long-
inasuta occurs alongside a varied trilobite
assemblage in shales and mudstones of the Sort
Tepe Formation in Turkey, suggestive of similar
environmental conditions (Dean & Zhou 1988).
Both Neseuretinus birmanicus and Vietnamia
douvillei are known from siltstones from the
upper Na Mo Formation in Vietnam; the for-
mation's depositional environment is poorly
understood, but both lithofacies and biofacies
comparisons suggest that it represents shelf
facies (Tong-Dzuy Thanh, pers. comm. 2001).
"Neseuretus" intermedius occurs in outer shelf
facies on the Yangtze Platform, in association
with characteristically shelf-slope taxa such as
raphiophorids and Nileus (Wang et al. 1987, as N.
xiadongensis). Reedocalymene expansa is abun-
dant in shales from the Miaopo Formation of the
Yangtze Platform, which has been interpreted as
representing a deep-water aulacogenic environ-
ment by Chen & Qiu (1986). Calymenesun also
occurs in deep-water facies (e.g. Tripp et al.
1989).
However, all of these other reedocalymenine
taxa still show relatively restricted geographic
ranges, which suggests that their incorporation
into cladistic biogeographic analysis will not
lower the resolution of the pattern of area
relationships (although see the discussion on
the possible geographic range of Neseuretinus
in Hammann & Leone 1997, p. 114). Similarly,
although most of these taxa occurred during the
Caradoc or Ashgill, a time period with different
continental configurations to the Arenig-
Llanvirn interval (Scotese & McKerrow 1990),
the consensus cladogram indicates that most of
the divergences involving these lineages are
stratigraphically constrained to have occurred
during the Early Ordovician. All of the reedoca-
lymenine taxa included in cladistic analysis are
therefore also included in cladistic bio-
geographic analysis. However, the occurrence of
Calymenella in Australia during the Early
Silurian (A. Sandford, pers. comm. 2001) is
excluded from this analysis, as this would appear
to represent a range expansion which occurred
after the time period under consideration here.
The area cladogram in Figure 4 contains four
subclades containing areas that span much of
the palaeogeographic extent of Gondwana
(nodes 1-4 in Fig. 4). This is interpreted as
indicating that several separate, independent
biogeographic events occurred during the
evolution of the Reedocalymeninae, which can
thus be compared with one another to reach an
understanding of Gondwanan area relation-
ships. Node 2 consists of an unresolved poly-
tomy in the area cladogram, and so can provide
no information on the relationships between
the four different areas (France/Spain, South
America, Avalonia and Turkey) included within
this subclade. The other three nodes are either
well-resolved (nodes 1, 4) or adequately
resolved (node 3). Nodes 1 and 4 display the
easternmost Gondwanan areas at the base of
the subclade, with progressively more western
Gondwanan areas tending to be nested progres-
sively further within the subclade. At node 1,
the far eastern Gondwanan region of South
America represents the sister group to Central
Asian and eastern Mediterranean regions
and South China; at node 4, the western
Gondwanan regions of Avalonia, France, Spain
and North Africa are more closely related to
one another than any is to the more easterly
situated SE Asian regions of Indo-China and
South China. Node 3 shows the opposite pattern
of area relationships, with South America and
South China as sister areas, together sister to
Spain, with the westernmost region of Avalonia
as the sister area to the rest of the subclade.
The ordering of successively nested areas
within each of the resolved nodes matches
previous ideas on the relative geographic
positions of Gondwanan and peri-Gondwanan
tectonic units in the Early Ordovician. These
three sets of area relationships can therefore be
interpreted as supporting the idea of a faunal
cline mirroring the climatic cline across
Gondwana during this time interval. However,
the three resolved subclades appear to represent
biogeographic shifts in opposite directions
across Gondwana. The opposing nested patterns
of area relationships thus represent non-
congruent biogeographic events, and contain
paralogous areas which would conflict with
duplications of themselves to generate ambigu-
ous data if combined as a consensus cladogram,
reducing the information available from the area
cladogram in Figure 3a. As it is imperative to
limit paralogy in area cladistics in order to detect
biogeographic congruence (Ebach 1999), com-
bining all three into a single consensus area
cladogram in an attempt to provide further
information on Early Ordovician Gondwanan
area relationships is thus to be avoided.
The subclades contained within nodes 1 and 4
appear to represent faunal shifts in the same
direction across Gondwana (Fig. 5a, b). As these
two subclades share the area of SE Asia in
