Lower Devonian acanthodians from New South Wales 
or to a closely related genus. This hypothesis is 
supported by the composition of residues from 
Late Silurian beds of Cornwallis Island, Arctic 
Canada (Sample Tc415[4+5], Tom Uyeno 
collection), in which the only acanthodian remains 
are large scales of Poracatithodes cf P. porosus co¬ 
occurring with jaw fragments of Nostolepis-type 
seiisu Gross (personal observation). The occurrence 
of this type of jaw bone in the samples from 
Trundle beds, which have no Poracatithodes scales, 
but which do have scales of "Nostolepis" 
guangxiensis (Wang), resembling those called 
Comphonchm cf G. hoppei by Vieth (1980), and 
scales of a new ischnacanthid (Burrow in press), 
suggest that Poracatithodes and one of these species 
are closely related. 
Histological structure of the jaw bones is not 
readily of use in determining their affinities; Gross 
(1971b:27) recognized similar hard tissues in both 
Gomphonchus and " Nostolepis " tooth cusps, and 
wide variation in the development of vascular 
canals and denteons in "Nostolepis" jaw bones. Also 
0rvig (1973:125) noted that the side cusps in both 
Gomphonchus and "Nostolepis" are formed of a 
similar dentinous tissue. 
Jaw bone fragments similar to those described 
here have also been recorded from the Lower 
Devonian Obere Carazo Formation, Profil Arauz, 
of Palencia, Spain (figured as "Ischnacanthidae 
indet." by Mader 1986, plate 3, figures 15,17a-c). 
Tooth whorls 
By comparison with the lower jaw symphysial 
tooth whorls of articulated ischnacanthids, the 
symmetrical tooth whorls (e.g.. Figure 4A) were 
probably positioned in the lower jaw symphysis. 
These tooth whorls are similar to those attributed 
to Gomphonchus sandelensis (Gross 1967), but most 
of them differ in having more side cusps (up to 
eight per tooth row). Similar tooth whorls have 
also been described from the Lower Devonian 
Nakaoling Formation ("Siegenian") and Yukiang 
Formation (early Emsian) in the Liujing area of 
Hengxian County, China, assigned to Gomphonchus 
liujingensis Wang 1992, a species described on the 
basis of scales and a tooth whorl. The cusps of the 
paratype tooth whorl figured by Wang (plate 3, 
figure 2a-c), have a circular parabasal section 
rather than the sub-triangular section of the 
Trundle beds symmetrical tooth whorl 
MMMC02249 (Figure 4A), and are conspecific with 
scales that have a pore canal system in the crown. 
As mentioned above, the robustness, size, and 
relative size and number of tooth cusps in the 
Trundle beds specimens is quite variable. 
Histological structure of these tooth whorls (e.g.. 
Figure 6C) appears identical to that of "Gomphodus" 
tooth whorls as illustrated by Gross (1957, plate 3, 
figures F-H). While some of the Trundle beds tooth 
339 
whorls have cusps with a sub-triangular parabasal 
section, others, which lack the distinctive ridges, 
have a circular parabasal section. The triangular 
shape of the bone base of the tooth whorls indicates 
they are ischnacanthid symphysial whorls, as 
climatiid tooth whorls have parallel sides (e.g.. 
Figure 7B,C). However, the variation in the 
Trundle beds tooth whorls indicates they could 
have derived from more than one ischnacanthid 
species. 
The multiple tooth whorl (MMMC02250, Figure 
4D,E) appears comparable to the spiky 'hedgehog'- 
like tooth whorls in the mouth cavity of 
Gomphonchus (Gross 1957, plate 3, figures 1,4-6). 
The tooth whorls described by Gross have strongly 
spiralled bases, but the Trundle beds tooth whorl 
is incomplete. In other features - three tooth rows, 
smooth recurved teeth and gradual increase in 
tooth size anterior to posterior - they are similar. 
In all ischnacanthid genera described from 
articulated specimens the symphysial tooth whorl 
(where present) is positioned in the lower jaw. 
Several types of isolated jaw bones have been 
figured (e.g., "Gomphodus" in Gross 1957, and 
Taemasacanthus in Long 1986, fig. 4A,D) with a 
cusp-free anterior segment. Perhaps some genera 
had parasymphysial tooth whorls/tooth whorl 
complexes on, but not ankylosed to, the jaw bone 
or cartilage in a comparable position to the tooth 
plates of dipnoans, or the tooth whorls of other 
higher osteichthyans (e.g., onychodontids). In its 
shape and robustness the asymmetrical tooth 
whorl described here certainly bears comparison 
with such tooth plates. These similarities support 
the possibility that the element had a 
parasymphysial position; alternatively, it may have 
been a pharyngeal element. 
Denticulated plates plus tooth whorl 
These elements (Figures 4F, 6D,E) do not match 
any previously described dental elements. Mader 
(1986, plate 3, figure 6a,b) figured a tooth whorl 
from the Lower Devonian Obere Carazo Formation 
of Spain which appears morphologically 
intermediate between this tooth whorl type and the 
fragile 'Borstenplattchen' (i.e., bristle plates) which 
Gross (1971b) suggested may have been supported 
by the branchial arches. Mader made no 
suggestions regarding their original anatomical 
position. The asymmetry of the Trundle beds 
element type indicates that it was paired, and its 
worn tubercles suggest that it opposed another 
(possibly identical) dental element. It might 
originally have been positioned in the pharynx; 
however, if this element, and the multiple tooth 
whorl (MMMC02250) which has similar cusps, 
came from a different species to the rest of the 
described elements, there are other possibilities. 
The shape of the base, the denticles on the narrow 
