302 
J.A. Long, G.C. Young 
Figure 11 Cladogram summarising relationships 
between taxa discussed in the text, including 
the new forms Portalodus, Aztecodus, and 
Anareodus described here. Synapomorphies 
(as discussed in the text) are: 1, phoebodont 
tooth crown (having three main cusps, 
central cusp slightly smaller). 2, single 
button on lingual torus. 3, base with labial 
extension. 4, lingual torus lost. 5, crenulated 
mesial cutting ridge. 6, accessory lateral 
cusps. 7, diplodont tooth crown (central 
cusps much reduced or absent, two main 
lateral cusps). 8, diplodont tooth crown 
(inferred parrallelism). 9, spine detached 
from dorsal fin, with pectoral or occipital 
attachment. 10, squamation lost. 
that a ctenacanth-like spine is primitive for 
xenacanths is not inconsistent with Zangerl's 
suggestion that their dentition is derived from a 
cladodont tooth type (see below). We therefore 
conclude that objections to xenacanth affinity for 
Antarctilamna based on its 'ctenacanth-like' spines 
are concerned with symplesiomorphy, and have no 
foundation. 
Teeth 
Teeth presumably originated as modified dermal 
denticles, and primitively can be assumed to have 
resembled scales in both size and morphology (e.g., 
Williams 1985: 141). However, faced with the 
morphological range of known Devonian shark 
teeth (from diplodont to cladodont), either one 
condition must be interpreted as primitive and the 
others derived, or special arguments may be 
invoked to justify a less parsimonious hypothesis 
of an unknown primitive morphological type. 
Zangerl (1981: 7) proposed that the simplest 
cladodont tooth form was a single elongate crown 
and a small base, which could then be 'enhanced', 
first by expansion of the base and addition of 
cusps, and then by modifications in cusp size. Thus 
he regarded xenacanth teeth as of 'modified 
cladodont design' (1981: 63), thereby implying that 
the cladodont condition was primitive. Lund (1985: 
15) agreed that a single simple cone was the 
primitive tooth crown condition, using outgroup 
comparison with osteichthyans, but noted that the 
'simplest condition known among mandibular 
teeth of chondrichthyans was the coronodont state: 
a distal-proximal series of subequal cusps fused 
into a 'multicuspid unit'. Lund suggested that this 
is plesiomorphic for elasmobranchs, and that 
protacrodont (low, subequal cusps), diplodont, and 
cladodont types are alternate derived conditions of 
the tooth crown developed on a 'synapomorphic' 
base. These alternative hypotheses of Zangerl and 
Lund were both apparently accepted by Williams 
(1985), who noted a tendency to fusion in typical 
anacanth branchial denticles, which could then be 
modified into a typical cladodont dentition by 
enhancement of the central cusp, and development 
of a lingual torus. Williams regarded these two 
features as advanced for all anacantlious sharks 
except Cobelodus, which he suggested exhibited the 
primitive condition for anacanthous sharks (where 
most teeth are small simple cones, with poorly 
developed bases, resembling the small Petrodus), 
like scales on the head of Stethacanthus and other 
form. However this interpretation has the 
unparsimonious consequence of requiring the 
evolution of multicuspid teeth in anacanth sharks 
independently of that in other elasmobranch 
groups. 
Support for Lund's hypothesis is provided by the 
multicuspid branchial denticles observed in many 
forms, including Antarctilamna (Young, 1982: plate 
87, figures 9, 10). However to interpret diplodont, 
cladodont and other types as alternate derived 
conditions of the tooth crown requires the same 
interpretation for the 'phoebodontid' tooth type as 
well, or any other combination. For heuristic 
reasons therefore, we provisionally follow 
Zangerl's interpretation that the cladodont tooth 
type (seen in a diversity of Palaeozoic sharks) is 
plesiomorphic relative to the diplodont type. We 
note, however, that the known fossil record 
(Leonodus of Mader 1986) suggests the opposite. 
The differences in tooth morphology and histology 
within the Family Xenacanthidae (e.g., Johnson 
1980: 930; Zangerl 1981: figure 69; Hampe 1991) 
may be seen as variations on the diplodont theme. 
However the fact that the crown in xenacanthids 
always comprises three cusps with an orthodentine 
histology, which emerge separately from the base, 
may be derived features characterising the family 
(Hampe 1991), by which they are distinguished 
from more primitive tooth types of Antarctilamna 
and phoebodontids (but histology is not yet known 
in these). 
Under this interpretation, the phoebodont tooth 
type would represent an intermediate stage of cusp 
reduction. Phoebodontid teeth are characterised by 
