Turinia pagei (Powrie): a new reconstruction of the soft organs of the cephalothorax

12 Pages·1998·3.4 MB·

by  L I Novitskaya

TURINIA PAGE!(POWRIE): A NEW RECONSTRUCTION OF THE SOFT ORGANS
OF THE CEPHALOTHORAX
L.I.NOVITSKAYA AND S. TURNER
Novitskaya, L.I. &Turner, S. 1998 0629: Turiniapagei(Powrie): A new reconstruction of
the softorgans ofthe cephalothorax. hiemoirs ofthe QueenslandMuseum 42(2)- 533-544
Brisbane. ISSN 0079-8835.
The possible structure and mutual arrangementofsomeofthe innersoft organs on'urinia
pagei (Powrie) including the brain, pineal organ, olfactory sacs, semicircular canals, and
branchialsacsare reconstructed. 'Fhedetailsofmorphologyofdepartmentsofthebrainand
organisation ofthe branchial system are analysed. We interpret T. pagei as having paired
nasal sacs and a telencephalon possessing well-developed olfactory tracts. Traces ofthe
labyrintharepresentatthelevelofthesecondandthirdgillpouches. Inboththesecharacters
T. pagei was similarto heterostracans and early gnathostomes (placoderms, chondrichthy-
ans). On the basis ofa suite ofcharacters including new data on the inner organisation of
TuriniaandLanarkia.thelodonts representamonophyleticgroupphylogenetically nearest
to, and thus should be considered as the sister group of the gnathostomes. Silurian,
Devonian, agnathans, Thelodonti, brain, nasalsacs, stomach.
Lanssa 1. Novitskaya. PaleontologicalInstitute ofthe RussianAcademyofSciences, Prof-
soyuznaya123,Moscow117647,Russia;SusanTurner, QueenslandMuseum. POBox3300,
SouthBrisbane 4101, Australia; 2^March 1996.
Thelodonts are extinct agnathan fishes (Late thelodonts are most like chondrichthyans, gener-
Ordovician-Late Devonian [early Frasnian]) ally called ‘sharks’ in the older literature. This
wschailcehs pwoistshesasabnaseexoosfkealceetlolnuloafrdbisocnreeltiekdeenttiisnsaule phoyrptoetdhesbiysNwoavsitresiktaeyraate(d1b9y83S,tept.so1n48()1,93a1n)d, suepx--
(aspidin) which is capable of growth, and the pandedon by Turner(1985, 1991). Mostofthese
production of simple to complex anchoring de- studieshavebeendoneonawealthofoldandnew
vices. The squamation is subdivided into several specimens of Loganellia scotica (Traquair),
specialised scale areas. These characters were Shieliataiti (Stetson) andLanarkiaspecies from
used byTumer(1991)todefinetheThelodontias the classic Silurian Lagerstalte in the Southern
a monophyletic group. Thelodonts are known Uplands, Scotland, which have now been newly
ambauinndlayntflryomin imsaolnayteddepsocsailtess awrhoiucnhd tahree wfoorulnd.d described by Marss & Ritchie (1998), who also
Taxa known from articulated specimens are few find evidence for gnAathostome characters in the
s(aelrtvheoduognhltyhienneuxmcebpetrioonfalspeencviimreonnsmeinstlsarfger)o,mptrhee- CSialnuardiaa,ntthheelosdoo-nctasl.led s‘ufiotrek-otfaifloesds’iltshfelroodmonAtrsc,tiics
nSiial,urainadn offroSmcotthleanEda,rlNyorDweavyo,niCaannoafdaBraintadinEsatnod- pinrotvhiedliondgonntewmoi&rnspihgohtlsoginyto(tWhielrsaonnge&ofCvaalrdiwaetlilo,n
Canada(Turner, 1976, 1982; Wilson&Caldwell, 1993; Caldwell Wilson, 1995)butthesefossils
1993; Caldwell & Wilson, 1995). Recent studies have not yet been fully described and will not be
on the general morphology, squamation, and in- considered here in detail.
ternal organisation ofthe thelodonts are provid- In late 1994, the authors hadtheopportunityto
ing new information on the nature ofthese fossil review togetherthelodont collections held at the
fishes, leading to diverse interpretations of the Queensland Museum and the data from new
structuresandconsequently,differentideasabout specimens. Amongthe studied materials was the
their relationships, some regarding them as mo- castoftheholotypeofTuriniapagei(Powrie)the
nophyletic (e.g,, Turner, 1991; Turner& van der originalofwhichexhibitsclearlythetracesofsoft
Brugghen, 1993; van der Brugghen, 1993) and innerorganson itssurface(seee.g..Turner 1982,
others as paraphyletic (van der Brugghen & Jan- pi. 97). In addition, although not pointed out
vier, 1993; Wilson & Caldwell, 1993; Caldwell previously, the rounded clay massjust ventral to
8c Wilson, 1995). Closerscrutiny ofthe squama- the cephalothorax is interpreted here as remains
tion and internal morphology seems to usto sup- ofstomach contents in this thelodont. The pres-
port the impressions ofmany early workers that ence of a true stomach in thelodonts has only
534 MEMOIRS OFTHE QUEENSLAND MUSEUM
FIG. 1. Comparison oftwo former different interpretations ofthe holotype ofTuriniapagei: A, Powrie, 1870,
as ‘ventralview’;B,Westoll 1945,asa‘larvalcephalaspid’ withhisdrawingoftheoutlineofthecephalothorax
‘afterTraquair, 1899a’;comparedwithC,acephalaspidbasedonStensio(1927)figuresofKiaeraspis,showing
conformation ofroofofbranchial cavity etc.’; Westoll figured ‘br r=branchial fossa, ‘i br r’-interbranchial
ridge, oes=oesophagous, pect=pectoral fin, tr=opening forIruncus arteriosus.
recently been verified by Wilson & Caldwell preservedasanatural mould inthree-dimensions
(1993) and van der Brugghen (1994). with friable remnants and impressions ofscales
Turinia pagei was discovered and first de- present. Therangeofscale variation can becom-
scribed by Powrie (1870, Fig. la); the specimen paredwiththatofotherarticulatedbutincomplete
is preserved in fine-grained Lower Devonian specimensofT. pagei(e.g.,Traquair, 1899; 0rvig,
sandstone from Turin Hill, Forfar, Scotland. We 1969; Turner in Allen et al., 1968) and with
concentrate on this example because it is one of isolated scales from numerous marine and mar-
thefew completespecimensofaDevonianthelo- ginal or non-marine localities from around the
dont and because, unlike other complete thelo- formerOldRed Sandstonecontinent(e.g..Gross,
donts mentioned above, this specimen is 1967; Turner, 1973; Karatajute-Talimaa, 1978).
SOFT ORGAN RECONSTRUCTION OF WRINIA PAGEI 535
TABLE 1. Differences ofinterpretation ofthe holotypeofTuriniapageiin earlierliteraturecomparedwiththe
presentstudy. Abreviations: ant.=anterior; cart.=cartilaginous; d=dorsal; interbr.=interbranchial; longit.=long-
itudinal; oesoph.=oesophagial; pters.=pteraspid; v= ventral.
Author Identity View Mouth Eyes Nsaascasl Lornigdigte. Brain Otic Br-aianlch Brainalch-Brainalch- Gut Paepcpteonrda-l Tail Other
sacs arches openings age
no
Po1w8r7i0e faomwinly V ant. 7 7 brainalch- - - - ‘ex7p/o8sed’ - - fins ‘beentliorwely vaennatlraolr,
ray-lilce skeleton dorsal
fins
early
La1nk8e7s0ter sharlo V? - - - - - - - - - -
ray
IhelodiLs hetero- no
Tra1q8u9a9ir Hsettraecrio- d 7 7 - - - sk8eclaertto,n - - laftienr-al bcreorckaeln vaennatlraolr,
shark folds dorsal
off
origin fins
Ke19m0n3a Ihelochis d - - - - - - ‘poOc8hRes’ 8arcs? none fins bsrkeailnaelctha-l
St1e9n2s7io Hesttrearcoi- d ptcefr.s. ?pstmcefar.ls.l ptcefr.s. ocrceigpiitoanl en‘sdiooulcimrd’an sphtnceoofr.tws.n dmese7ras+mcosa-l uirnnidtediedvrgbierds.- nosign ofeosroapmhe.n cenrtoatin ccah(rpbotetnhoridinilnr-ae)agle
aortic roof
We1s9t4o5ll cleaapsrhpvaialdl- V none gKricoafeo.rv-e ppoacikreetds irntiedrgbers. brainalch-
aspis cliamber
visceral oesoph.
S1t19e9n56s84i/o Hesttrearcoi- d eccrhuarne- eoncrcdieiogupciimrtoaanlnlabyrinth ‘fos8ses’ iasrn1trih-ecdofrhgboweer?sss, 8braeatixnartlicraha-- bfCrooanrfindaaolcurmihoti- mahaunyrldocaiirhdb-
9 vessels
PTahpiesr tbdhaoesnlaotl- d tesrumbi-al 7 paired bsraeien epnrwdeiostcehanstt bpetatriwareceeedn 7(8?) nlaornrgo,w sepa7rate sptrceMsneancht fins unhceyrptoa-in pobenrodirncaehl-
ogsntaothm-e pineal br.2/3 cercal implied
The traces ofthe endoskeleton and various in- new reconstruction (Fig. 2a). The new data from
ner organs are either rarely preserved in thelo- theholotypeof71pageiconcernmainlythebrain,
donts or, because of the nature of their nasalsacsandbranchialsystem.Thisinformation
preservation, are difficult to interpret. Despite diminishes the gaps in our knowledge of the
being known for over 160 years, the group re- internal organisation ofthe thelodonts. We thus
mains enigmatic and was until recently poorly regardthelatterasbeingsignificantforthefurther
characterised. Turner(1991)proposedthemono- study of the problem of phylogenetic relation-
phylyoftheThelodontibasedonfeaturesofgross ships of the ancient agnathan and gnathostome
and scale morphology (see also Forey, 1984). vertebrates.
Ideas about the inner organisation ofthelodonts The first turiniid thelodont was found in the
have been mostly reduced to creation of very classic Old Red Sandstone area ofthe northern
generalised schemes which had as their founda- hemisphere in depositswhich until recentlywere
tion information from only a few specimens. thought to be non-marine sediments. The type
Various interpretations of the preserved struc- specimen, ‘^Cephalopterus’' pagei, was discov-
turesoftheholotypeandonlycompletespecimen ered by Powrie in 1870 intheLowerDevonian of
ofT. pagei have been given (e.g., Powrie, 1870; Turin Hill, eastern Scotland. That year both
Traquair, 1899; 1906; Westoll, 1945; Stensio, Powrie and Lankester considered that the holo-
1927, 1964;Fig. 1,Table 1). Thepresentstudyof typewas preserved in ventral view; Powrie inter-
the cast ofthe holotype T pagei concludes that pretedsevenoreight pairs of‘exposed’ branchial
the arrangement ofridges and hollows shown on archesandthought it ‘strangelyalliedtothemod-
the cephalothorax can be interpreted as various em Rays’ (see Powrie’s restoration in Fig. la);
endoskeletal features and soft organs allowing a Lankester (1870) noted the resemblance to the
536 MEMOIRS OFTHE QUEENSLAND MUSEUM
ventral surface ofa cephalaspid but thought that Aswill beapparentinourdescriptionofthesoft
thefossilwasanearlyrepresentativeofthesharks parts which follows we now interpretthe cepha-
and rays. Traquair (’899) was the first to seri- lothorax of the holotype as being preserved in
ously considerthe significance ofT. pagei in the dorsal view because ofthe imprints ofthe olfac-
light of new Silurian thelodonts he was also tory tracts and aspects ofthe brain.
studying. He had renamed the holotype T.
(Turner, 1976) and, in 1899, gave a full descrip- DESCRIPTION
tion ofthe specimen, reverting to the type genus
Thelodus. Henoted the absenceofjaws andteeth Thetypespecimenof7.pageiisthenaturalcast
and deduced that the arrangement ofridges and ofan intactanimal notdeformed in anyway. The
grooves was simply an indication ofthe presence counterpart apparently was not found. Table 2
ofacartilaginous branchial skeleton. Contraryto displaysmeasurementsof7pagei. Thetotal length
Powriehedidnotbelievethatthebranchialarches was taken from the middle ofthe anteriorborder
were exposed but, because he believed that the ofthe head to the (broken) end ofthe lower lobe
thelodonttail washeterocercal. he interpretedthe ofthecaudalfin; themaximumwidth isacrosson
fossil aspreserved indorsalviewexhibitingeight the level of postero-lateral points of the lateral
brim. The caudal peduncle was taken across the
pairs ofcartilaginous arches, the last pairtending
backwards in a reversed V-shape. Kemna(1903) narrowestpoint in frontofthe tail. As preserved,
went on to study T. pagei by comparison with theanteriorpartofthebody isrelatively flat. The
otherthelodonts (‘coelolepids’) known atthe be- well-developed lateral brim (ofthe pectoral fin)
ginning of the 20th century. He surveyed the startsjust posteriorto aslight bulge intheoutline
ofthe specimen, interpreted here as the possible
histor>' of T. pagei and noted the differences of position of the orbit, shown in Fig. 2a wnth a
opinion on the detenuination ofthe visible side
dashed line. Alternatively, the eyes might be
of the holotype. Kemna, following Traquaifs
placed directly on the antero-lateral borderofthe
(1899) opinion on the caudal fin, interpreted itas cephalothorax in amore lateral positionandthus,
the dorsal side. While giving arguments pro and
tcuornetsraasthveisicnetrearlproertabtriaonnchoifalthaercbhreasn,chhieallesfttrtuhce- aTosrhe‘Tfripane’qcutiaonirtarhle(e1lx8itt9ee9rn)astiunrooent;e(dhv,earrteihoweuyeslauyrseceantlohlteedtpreterhseme'rTfvmlea’dp).’
question open for further discussion. Sub-
thusstarts alittle in frontofthe firstpairofraised
sequently, therehavebeen several interpretations ridges orbranchial structures,which we interpret
ofthe morphology ofT. pagei and its systematic as branchialsacs (see below), becominglarger in
position compared to other thelodonts and in the a caudal direction. This lateral rim ofthe fin is
overall interrelationships of agnathans (e.g., flatter than theraised centra! portion ofthe large
Westoll, 1945; Stensib, 1927, 1964; Karatajute- cephalothorax. The cephalothorax in thelodonts
Talimaa, 1978; Turner, 1982, 1991; Janvier, usually comprises around one-quarter to one-
1996). Information on the soft organs, however, third oftotal body length (Turner, 1991); at least
is almost absent from the preceding literature. A the latter in the case of 7 pagei. The external
briefinterpretation ofsomeofthem wasgiven by borderofthelateral brim isrounded. Thepostero-
Stensib(1958, 1964). Stensibconsidered T.pagei lateral lobes ofthe fin have a rounded-triangular
as separate from other thelodonts; he placed it form. Behind the widest part ofthe lateral brim
within the Heterostraci based on the configura- and cephalothorax the body tapers suddenly to a
tion of the branchiae and the possession of an quite narrow peduncle. On the type specimen in
endocranium. Westoll (1945), how'ever, pre- between the lateral brim (distal pectoral lobes of
ferred to return to the similarity to cephalaspids, the fin) at the posterior part ofthe cephalothorax
notablyKiaeraspis(Fig. lb, c). However, in con- can be seen a large rounded mass preserved as a
sidering the Turiniidae, Stensib gave his main darker clay-like sediment. The configuration of
attention to the exoskeleton, which is not the themedian fins is notclearbutthere isaprobable
subjectofthispaper. Onlyontheplate illustrating anal fin preserved on the right hand side ofthe
the holotype did he interpret the morphology of specimen. The upper lobe ofthe tail fin is short,
the branchial apparatus (e.g., Stensib 1964, fig. the lower lobe is well formed and elongated but
89) where he distinguished at least eight pairs of not complete as the rock is broken (or trimmed)
branchial sacs. His hypothesis will be considered atthis point. These lobesare apparently rounded.
furtherbelow inconnectionwithanalysisofques- They are connected by a series of undulations
tions remaining in the reconstruction ofthe inner which might represent a tail web. The tail fin is
organs ofT. pagei. apparently hypocercal.
SOFTORGAN RECONSTRUCTIONOF TURINIA PAGE! 537
en
FIG.2.Diagrammaticrepresentationoftheinnerorganisationofthelodonts,helerostracansandsharks.A,stylised
drawing ofmain features ofTitriniapagei (Powrie) based on the cast ofthe holotype, National Museums of
Scotland RSM 1891. 92. 133. B. Lanarkia horrida (from Turner & van der Brugghen, 1993). C, pteraspid
heterostracan ofPodolaspis type (modified fromNovitskaya, 1993). D, shark, Squalnsacanthias (simplified
fromMarinelli&Strenger, 1959).Abbreviations:bap^branchialapparatusorrowwithbranchialcompartments;
bar=possible position ofvisceral arch; bp=possible branchial pouch; dien=diencephaIon; Ic=lateral branchial
canal; med=medulla (myelencephalon); mes=mesencephalon; nc=nasal capsule; oe=oesophagus; or=orbit;
pin=pineal macula (probable position in thelodonts: A, B); sc=location ofsemicircular canals (uncertain in
thelodonts: A, B); teHelencephalon; trolf=olfactory tracts.
The traces ofsome soft organs are clearly pre- THE BRAIN. In his explanations ofinner struc-
served on the dorsal side ofthe holotype. These tures of T. pagei, Stensid (1964, p. 272, fig. 89:
are discussed below. ‘r.oc’)notedtheexistenceofathickeningformed
538 MEMOIRS OFTHEQUEENSLAND MUSEUM
TABLE2.MainmeasurementsofTuriniapagei-RSM since, as far as it is known, the pineal organ in
1891.92.133. agnathans is always associated with the di-
Cmheaarsaucrteedr Meas(umrme)ment Ratio ecansctepbhaelcoonm.esThesopmreewsheravtedlosuwrefraciensfhroowntn oofn tthhee
medialtotal length(mtl) 320 diencephalon. In all probability the telencephalon
cephalothoraxlength(cl) 120 mtl/cl=2.6^ in T. pageiwas placedinthe regionofthis depres-
maximumwidth(mw) 155 mtl/mw= sion,thatis,theuppersurfaceoftelencephalonwas
approx.2 lowerthan the upper surface ofdiencephalon. Al-
lengthofpinealorgan(po) 5 ternatively,thedepressionmightbethebackofthe
widthofpo 1-1.5 mesencephalon which is often opposite gill pouch
lengthofpinealmacula(pm) 8-9 #1 in gnathostomes (Mallatt, pers. comm.). The
widthofpm 7 posteriorlimitofdiencephalonismarkedbyaweak
widthatlevelofpo 95 mw/po= 1.6 groovebehindthepinealmacula.Thustheposterior
widthofcaudalpeduncle(cp) 30 edge ofthe diencephalon is placed approximately
l(lein)gthoflowercaudallobe 45 apmptrlo/1x1.=7.5 oofnbtrhaenclheivaellsbaectsw.een the first andthe secondpairs
mlaatexrialmburmimwi(dlbt)hof mw/lb=5 myTehleencnaetpuhraalloncasotrs tohfeitrheenmdeoscreanncieaplhaclaostns aanrde
visible on the holotype of T. pagei behind the
by the occipital region ofthe endocranium. This diencephalon althoughthe limit between theme-
thickeningcontinues forward intotheoticregion sencephalon and myelencephalon is poorly dis-
of endocranium. Our study of the cast of the tinguished. Thejunction ofmesencephalon and
holotype shows that the traces of the compart- myelencephalon was on a level approximately
ments of the brain can be distinguished in this between the second and third pairs ofbranchial
thickening. Here weproposethe following inter- sacs. Further back the cast ofmyelencephalon is
pretation ofthe brain. seen distinctly, extending back as far as the pos-
Apairofshallowbutcleargroovesisvisibleon teriorbranchialregion. Thetraces ofanteriorand
the anterior part ofthe head. They diverge from posteriorsemicircularcanalsareseenonthelevel
theregionsituatedinfrontofthepinealorgan(see of second and third branchial sacs (Fig. 3, sc.
below), continue in the direction ofthe anterior (lab.)), but the imprints of the canals are not
border ofthe head and terminate not far from it perfectlyclear. Thepossiblepositionofthelaby-
(Fig. 2a, Fig. 3, trolf). We equate these grooves rinth in the same place was also indicated by
with the tracti olfactorii of modem sharks and Stensio (1964, fig. 89, ‘lab?’) although our nu-
otherdiplorhinal animals, forexample, Lanarkia meration ofthe branchial sacs does not coincide
(Turner&vanden Brugghen, 1993;Fig.2b),and with his.
the Palaeozoic heterostracans (Fig. 2c), by com- BRANCHIAL STRUCTURE. The casts ofthe
parisonwith theirposition relativetothepineal branchial sacs are clearly visible onthe holotype
organ and the anterior border ofthe head, and of T. pagei. The branchial sacs were large and
bytheir length. The position ofanteriorends of situatedoverall transversally.Theydiminishedin
olfactorytracts indicatesthe place ofthepaired sizeinacaudaldirection(Figs3-4).Thebranchial
nasalsacswhichweresituatedneartheanterior sacs are situated neartoone another. Seven pairs
border of the head on either side of the pre- ofthemareclearlydistinguishable. Thepresence
sumedjustventralmouthopening.Thebulbous oftheeighthmostposteriorpairisnotdefinitebut
shape ofthe nasal sacs is apparent. The telen- the possible position is indicated in Fig. 2a. The
cephalon isnotdistinguishablebutitispossible anterior pair ofbranchial sacs is the largest, ori-
toestimate its place indirectly,basedontheposi- ented transversally and obliquely; their lateral
tion ofthe pineal organ and, therefore, from the endsaredirectedforward. Theorientationofsacs
position of the diencephalon. The pineal organ changes from anterior to posterior: the lateral
was situated on a level ofthe anteriorpairofthe ends of posterior branchial sacs are directed
branchial sacs. Its trace has the appearance of obliquely and back.
elongated oval pit bordered by a low convex Unfortunately,thepreservationofmaterialstill
border visible on the surface of the cast. The does not allow usto answerthe question ofhow
dimensionsofthepitandpinealmacula(including the branchial sacs ofT. pagei opened to the out-
convexborder)are shown in Table2. Theposition side. We still cannotdeterminewhethertheyhad
ofpinealmaculaindicatesthatofthediencephalon separate branchial openings or individual canals
SOFTORGAN RECONSTRUCTIONOF TURINIA PAGEI 539
FIG.3 PhotographofthecastoftheholotypeofTuriniapageitPowrie)exhibitingthepositionofolfactorytracts
.
(trolf),brain,pinealorgan(pin),possiblesemicircularcanals(sc(lab)).Impressionsofbrainstemandbranchial
sacs can be seen. Holotype (RSM1891.92.133; Museum ofScotland, Edinburgh) from the Early Devonian
(Lochkovian) Old Red SandstoneofTurinHill, Forfar(Scotland).
540 MEMOIRS OFTHEQUEENSLAND MUSEUM
leading out to a common atrial canal. The slight skaya, 1983; Gagnier, 1995). Thearrangement is
thickeningwhich is seen on the lateral brim ofT. comparable with living early gnathostome em-
pagei was interpreted as the anterior branchial bryos with first gill arch and pouch just behind
canal by Stensio (1964, ‘conduit branchial anter- the eye, second arch (the hyoidean)Just in front
ieur’) . Such an interpretation sat well with his oftheoticcapsuleandsoon(e.g., Mallatt, 1996).
idea of T. pagei being an heterostracan. Our in- Thus, the primitive vertebrate pattern might be a
vestigations show, however, that this canal was high number ofpaired branchial structures, as in
not continuous but consisted ofshort separated galeaspids and some Ordovician forms(Gagnier,
segments (Fig. 4, icbr). The curved imprints, 1995; Janvier, 1996). A small number of eight
which were explained by Stensid (1964, fig. 89) paired branchial structures, as seen in Turinia,
as extrabranchialatria, areseentobemore lateral which might functionally have been only seven
relativetoeach branchialsac. Similarimprints or pairs; this recalls the suggested pre-gnathostome
natural casts can be seen in many heterostracans pattern (Mallatt, 1996)which isclosetothatseen
(e.g., Novitskaya, 1983). In all probability these in modem primitive chondrichthyans such as
imprints represent the traces of individual bran- Heptranchias and Chlamydoselachus.
schoimalectahneallodsonltesa,difnogrefxraommptlheeibnrPahnlcehbiaollespaicss.eleIn- Consideration ofthe morphology of 7! pagei
gam(Ritchie, 1968), andtherecentlydiscovered indicates the internal organisation ofone thelo-
& dont, although to understand some organs, prin-
fork-tailed thelodonts (Wilson Caldwell,
cipally the branchial system, we need more
1993), the branchial sacs each opened out sepa-
precise data. At the same time the new interpre-
rately.
tation makes itpossibletocompare T.pageiwith
Asalreadynoted,thecastsofthebranchial sacs Lanarkiahortida,anotherthelodont inwhichthe
in T.pageiaresituatedneartooneanother. Ifthis soft organs have recently been reconstmcted on
state was not due to posthumous alteration such thebasisofgoodnew material (Turner&vander
as contraction, the branchial arcs (arches) appear Brugghen, 1993). Comparison shows that Tur-
to have been very long and narrow and subequal iniaandLanarkiaweresimilarinthemainpattern
in size. The first visceral arch was apparently oftheir inner organisation. The presence ofpos-
situated immediately behind the eye. In T.pagei, sible well developed long olfactory tracts and
the posterior arches were shorter than the more
paired, separate large nasal sacs ofgnathostome
mamnaettnectrhieoirsnwomenleolsd.weiAmtshtpwhrieitmehixtttirhvaeebrpsahhnaacrrhkyisna,glectaahlritsialrparagatentsgeeron-f tcityopimsemcilonenabroctthhhaartatcththeeelrosrd.eocnFotnrssotamrrueocuttrihoesntmouodfsytthoesfilTg.anbipyfarigicenaitn,ht
sharks (Mallatt, pers. comm. 1996). inL. horridawasprobablyplacedtoofarbackby
DISCUSSION Turner and van den Brugghen (1993; the brain
configuration needs to be slightly alteredto abut
On the whole, the organisation ofthe visceral the semicircular canals with the level ofthe 2nd
system in T.pageihasoverall similaritytothatof and 3rd gill pouches. However, as shown on the
heterostracans and some other agnathans (Gag- comparative scheme(Fig. 2), both thelodonts are
nier, 1995). Such common traits are observed in similarinthecharacteristicsofthecephalothorax
thepresenceofafewbranchialsacs, intheirfomi to heterostracans and basal gnathostomes. The
anddisposition, in thepresenceofindividual (for same position ofthe telencephalon in relation to
eachsac)lateralcanalsandinthesimplestmcture the upper surface of diencephalon is found in
of cartilaginous non-differentiated branchial heterostracans (Novitskaya, 1974, 1983) and
arches. Whethersimilarcharacterswere inherent lampreys (Marinelli & Stronger, 1954; Fontaine,
also in the most archaic vertebrates is still in 1958).Thepinealorganwassituatedonalevelof
debate. To a certain degree the finds ofOrdovi- the anterior pair ofthe branchial sacs, as in het-
cianvertebratesinAustralia(Arandaspis: Ritchie erostracans (TJovitskaya, 1983, fig. 64). The pos-
&
Gilbert-Tomlinson, 1977) confirm this. Some terior limit of diencephalon occupies a similar
current analyses have led to the conclusion that position in heterostracans, e.g., Poraspis pom-
cephalaspids, with their highly specialised mor- peckji(Novitskaya, 1983,pi.4,figs2,3). Because
phologyarethesister-groupofgnathostomesand the labyrinth is notclearlyseen, it is possiblethat
&
that thelodonts are paraphyletic (e.g., Forey the semicircular canals were placed somewhat
Janvier, 1994). However, we considerthatthe T. deeper in the endoskeleton in 7. pagei than in
pattern seems to show what some regard as the someotherearlyvertebrates. Inheterostracans,for
primitive gnathostome condition (e.g., Novit- example, the semicircularcanals retain very clear
SOFTORGAN RECONSTRUCTION OF TURINIA PAGEI 541
%
FIG. 4. Photograph ofthe castofthe holotype ofTuriniapagei(Powrie). Differentdirection oflightmakesthe
tracesofbranchial apparatus clearer: icbr=individualbranchial canals; s=position ofstomach contents.
542 MEMOIRS OFTHEQUEENSLAND MUSEUM
traces which can be seen in cyathaspids such as known in Phlebolepis elegans and Loganellia
P. pompeckji (Brotzen), P. cylindrica Kiaer, scotica.
Komalaspis della nitida (Kiaer) (e.g., Stensid, In T. pagei the morphology ofthe divisions of
1964,figs81A,B-82A,B;Novitskaya, 1983,pi.IV, the brain (diencephalon, mesencephalon,
figs2, 3). myeiencephalon) is simple and similarto that in
The comparison ofthelodonts and heterostra- such groups ofvertebrates as lampreys, heteros-
cans shows that the visceral arch situated imme- tracans, and arthrodires (e.g., Novitskaya, 1993).
diately behind the eye in T.pageicorrespondsto Inthe morphology and disposition ofthesebrain
themandibulararchofheterostracans.Thisstruc- compartments, T. pagei is generally similar to
ture was homologised to that of lower gnathos- Lanarkia. TTie brain is homologous with that in
tomes (using a shark embryo) by Novitskaya heterostracans and in lower gnathostomes (e.g.,
(1983). The second arch is compared to the chondrichthyans). The pineal organ was pre-
hyoidean in sharks, the following to the first served butthere was no nasohypophysial organ.
branchial arch and so on (Novitskaya 1983, fig. The paired semi-circular canals are placed in the
71, p.133). Thesimilarity ofthelodonts, heteros- region ofthe branchiae 2 and 3. Othercharacters
tracansandbasal gnathostomes inthedisposition such as the presence ofspecialised denticles lin-
ofthe named visceral arches and the large nasal ing the mouth and pharyngial region, morpho-
sacs, leads us to considerthevisceral skeleton as logical style and complexity of the external
homologous in allthree groups. squamation, presence ofastomach(seee.g.. Fig.
The presence of these characters in the mor- 4, s), indicateclosesimilaritybetweenthelodonts
phology ofthe named groups and the evidence and chondrichthyans.
from die holotype ofT pageitestify that a com- In agreement with the above-mentioned simi-
mon nasohypophysial duct of cephalaspid type larities between thelodonts and lower gnathos-
did not form in theirontogenesis (contra van der tomes (in morphology of olfactory organ,
Brugghen & Janvier, 1993). Furthermore, the telencephalon, in apparent absence ofa common
type ofontogenesis (development) in thelodonts nasohypophysial organ) we believe that thelo-
was fundamentally similar to that ofheterostra- donts shared a similar type of ontogenesis (see
cans and basal gnathostomes. The method ofre- Novitskaya & Karatajute-Talimaa, 1989). T^is,
constructing the developmental type of an andthepresenceofexterna!removalopeningsfor
agnathan on the basis oftheir morphology was each pair of branchial sacs, seen in some thelo-
earlier demonstrated by Novitskaya & Karata- donts (but not here in the holotype ofT pagei),
jute-Talimaa (1989) and Novitskaya (1993). provides the basis for considering them as the
SUMMARY group most similar to basal gnathostomes in the
main traits of organisation. Among Palaeozoic
The organisation ofthe visceral skeleton, the agnathans the thelodonts are thus considered to
numberofbranchialsacs(atleastsevenoratmost bethegroupphylogeneticallynearesttognathos-
eight pairs), and the position ofthe anterior pair tomes(chondrichthyanfishes).Thisideahasbeen
relativeto thepinealorgan, aresimilarin Turinia considered previously by both authors on the
andLanarkia. The characters noted here, as with basis of other material and other characters
some others concerning the internal morphology (Novitskaya, 1983; Turner, 1991). A recent
and the exoskeleton (Novitskaya, 1983; Turner, cladistic analysis ofearlyvertebrates arrivedat a
1991), are similar also in these thelodonts and in similar conclusion that places thelodonts as the
heterostracans. Atthe same time the presence of sistergroup ofgnathostomes (Gagnier, 1995). The
external openings ofbranchial sacs ventraltothe new data on the internal organisationofT. pagei
pectoral fin-lobes, now known in some thelo- presents evidence supportingthis hypothesis.
donts, separates them from the specialised het-
ACKNOWLEDGEMENTS
erostracans which possessed a common removal
branchial canalwith one common external open- The display of the dinosaur collection from
ing oratrium. PaleontologicalInstituteoftheRussianAcademy
The mouth in T. pagei was apparently anterior ofSciencesinAustralia(organisersofexhibition:
andjustventral, flankedbypairednasalcapsules. Alexej Rozanov, Moscow and Patricia Vickers-
Noevidenceforanasohypophysialorganisseen. Rich,Melbourne)allowedRussianspecialiststhe
The branchial row which began just behind the opportunitytostudyfossil materials inAustralia.
eyes, was relatively foreshortened with openings The mission ofone ofus (LIN) in Brisbane, and
in all probability ventral to the pectoral fins as jointworkattheQueenslandMuseum,wascarried