GEOL. CROAT. 51/2 105 . 134 23 Figs. ZAGREB 1998 Remains of Sauropoda (Reptilia, Saurischia) in the Lower Cretaceous (Upper Hauterivian/Lower Barremian) Limestones of SW Istria (Croatia) Fabio Marco DALLA VECCHIA Key words: Dinosauria, Sauropoda, Titanosauriformes, Diplodocimorpha, Upper Hauterivian- Lower Bar­ remian, Cretaceous, Istria (Croatia). Abstract Remains belonging 10 sauropod dinosaurs have recently been dis­ covered in Upper Hauterivian/Lower Barremian (Lower Cretaceous) limestones of SW Istria (Croatia). The material consists of a complete cervical vertebra, a nearly complete cervical centrum, fragmcnls of poss ibl e cervical ribs, three partial dorsa! and five more or less in­ complete caudal vertebrae, parts of caudal neural spines, a chevron, the distal pari of a femur, the proximal portion of a tibia and other fragments of bones. The bones were collected randomly from the sea bottom, therefore despite the fact that they come from the same out­ crop, the same level and probably the same bed, they cannot be assigned with certainty to the same taxon. Their vastly different sizes indicate the presence of several individuals while different morpholo­ gies suggest the probable presence of more than one taxon. The com­ plete cervical and the anterior 10 mid-caudal vertebrae present a more strict affinity with Brachiosauridae, a proximal cervical centrum rese­ mbles those of "Cholldros(eosourtls", and a caudal neural spine is si­ milar to those of the camarasaurids. The dorsal vertebrae have pecu­ liar features (a very tall neural arch, well developed laminar complex, etc.) and characters suggesting their assignation to basal Titanosauri­ formes and, possibly, to Diplodocimorpha. A posterior dorsal verte­ bra testifies the presence of a new Diplodocimorph similar to Reh­ bachisollrlls but more primitive. 1. INTRODUCTION Dinosaur tracks of Cretaceous age are relatively common in the lstrian peninsula (NW Croatia), and have recently been reviewed by DALLA YECCHIA et al. (1993), DALLA YECCHIA & TARLAO (1995) and DALLA YECCHIA (1997a, b. c). They are preserved in Upper Barremian, Upper Albian and Upper Ceno­ manian limestones at several sites along the coast of the peninsula, and on the islands (Main Brijuni/Brioni island, Fenoliga isle[, e[c.). The first discovery of dinosaur bones, on the Adriat­ ic sea bottom at the Kolone locality near Bale/Valle vil­ lage, south o/" Rovinj, was reported by BOSCAROLLl e[ al. (1993), and was fo llowcd by preliminary notes concern ing the fossils and their stratigraphic and pala- Museo Pa[eontologico Cittadino, Via Valentinis 134, 1-34074 Mon­ fa1eone (Gorizia), Italy. eoenvironmental context (DALLA YECCHIA, 1994a. b; DALLA YECCHIA et a!., 1993; TUNIS et a!., 1994; KOZARIC et aI., 1996). Some specimens were identi · fied as sauropod bones (DALLA YECCHIA, 1994b. 1997a. b, c; DALLA YECCHIA & TARLAO. 1995) but never described in detail. This paper presents a deta­ iled description of the identifiable and attributable remains of the sample. All of the material comes from the same outcrop, from the same stratigraphic level and, probably from the same bed (see below). However, most were collected randomly as scattered fragments on the beach and on the sea bottom, where the fossi liferous layer crops out. No systematic attempt had been made to collect the numerous bones still embedded in the rock. Since the fossiliferous bed, which seems to be particularly laden with bony remains (D. BOSCAROLLI, pers. comm.). lies below sea level, its excavation is difficult. Howev­ er, this outcrop promises to be one of the most rich and interesting of present Europe. 2. DEPOSITIONAL ENVIRONMENT AND AGE The age and depositional environment of this depo­ sit are discussed in BOSCAROLLI e[ al. ( 1993), TU· NIS et a!. (1994) and DIN! e[ a!. (1998). The fossilifcr· ous outcrop is characterized by the presence of beds of oncolitic rudstone with bone debris, and thinly laminat­ ed limestones with plant, shrimp and fish remains. So­ me bones (for example, the complete cerv ical vertebra WN·YI, the small clorsal MPCM·YI , [he mid·caudal Nos IG -I and other fragments) were preserved wholly in a yellow, thinly laminated limestone or with a side of the fossil in this laminated limestone and the other side in a grey, hard oncolitic rudstone. This seems to indi­ catc that the preservation of comp lete bones is due to deposition in a relatively protected, low energy environ­ ment (laminated limestones) and to the probably rapid covering by the rudstone which testifies a high energy environment. However this shou ld be confirmed by a detailed sedimentological and taphol1omical study of the fossiliferous beds which is beyond the scope of this paper. The stratigraphic sequence of the OLilcrOP at Bale is composed of subt idal, intertidal and lacustrine limesto­ nes, The lower section is of Late Hauterivian/Early 106 Barremian age due to the presence of the foraminifer Campanellll/a capllensis DE CASTRO; the upper scc­ tion is probably Lower Barremian (BOSCAROLLI ct aI., 1993; TUNIS et aI., 1994). Therefore the bones des­ cribed here arc the oldest record of dinosaurs curremly known in Istri a (DALLA VECCHIA & TARLAO, 1995) if possible sauropod tracks in the Berriasian of Fantazija Quarry (LOCKLEY et aI. , 1994) arc exclud­ ed. The bones arc approximately coeval with the thero­ pod pes and sauropod manus prints found in a quarried limestone block from the Cansiglio Plateau (Northeast­ crn Italy, Pordenone) described by DALLA VECCHlA & VENTUR INI ( 1995) . The sauropods of Bale are a rare case of dinosaur bone remains that are dated in cor­ relation w ith the marine biochronology. 3. DESCRIPTION AND COMPARISON The sample consists of more than 200 specimens, most of whi ch are just bone fragments. The collected material was exposed to recent marine and shore weath­ eri ng and was encrusted (and som etimes pierced) by living mar ine ani mals and algae. It was c leaned and prepared by the preparators of the Gmppo SpeJcologico Monfalconese A.D.F. at the Museo Paleontologico Cit ­ tadino of Monfalcone (Gorizia). Most of the bones are crushed, sometimes strongly, but olherwise the state of preservation of the bones still imbedded in the rocky matri x is very good. Onl y bones attributable to sauro­ pods are describcd hcre; several specimens are too frag­ mented to ident ify the skeletal element to which they belonged, other fragments belong obviously to long and large bones which remain indetenninate. The specimens wi ll finally be stored or exhibited in a Museum dedicated to the local dinosaur rema ins in the vi llage of Bale (Istria, Croatia). At present they are without the defini tive number of this Museum therefore J will report here the numbers used during preparation (MPCM-V = Museo Paleontologico Cittadino di Mon­ falconc - Valle). The specimens that were nOl numbered during preparation and were at the Mu seum in Ba le during the final version of this paper are identified with the abbrev iat ion WN-V. Two bones are presently in the collection of the Institute of Geology, Zagreb (Nos lG- I and Nos IG-2). Because of the way the specimens were collected it is im possible to be immediately sure that they belong to the same taxon, or to a precise number of different taxa. In some occurrences, for example in some levels of the Morrison Formati on (Upper Jurassic) of North Ameri ­ ca, six sauropod genera were found together (see CUR­ TICE & WILHITE, 1996). Therefore each bone is com­ pared with the corresponding bones of described sauro­ pod taxa, in o rder to determine affin iti es particul arl y with the better known forms (Brachiosaurus RIGGS , Haplocanfhosaurus HATCHER, Camarasaurus COPE, Dip/odocll s MARSH , and Apatosallrus MARSH ) and the Neocomian- Barremian ones. Geologia Croatica 51n Acronyms: CD = caudal rib , CDL= centrodiapo­ physial lamina, CPR = centroprezygapophys ial lamina, CO=eondyle, COPR=condyloprezygapophysial lami­ na, DP=diapophysis, Has=art icular surface for the hy­ posphene, HL= horizon tal lamina (= diapo-prezygapo­ physial lamina , diapo-postzygapophys ial lamina), I-IPN = hyposphene, HYP = hypantrum , IDL = infradi­ apophysial lamina, IHPNL = infrahypos phellal lamina, IPDL = infrapostdiapophysial lam ina, lPDRL = infrapre­ diapophysial lamina, IPPLa = infraparapophysial anteri ­ or lamina, IPPLb = infraparapophys ial posterior lamina, JPRL = infraprezygapophysial lamina, IPZL= infrapost­ zygapophysial lamina, ISPZL = inner suprapos tzygapo­ physial lamina, ITPZ= illtrapostzygapophysial lamina, L1PPL = lateral in fraparapophysial lam ina; L1PZL = lat­ eral infrapostzygapophysiaJ lamina, NA = neural arch, NC= neural canal, NS = neural spine, OSPZL= outer suprapostzygapophys ial lamina, PL = pleurocoel, PP = parapophysis, PR = prezygapophysis, PRL= prezygapo­ physiallam ina, PRSPL =prespinal lam ina, PSPL = post­ spinal lamina, PZ= postzygapophysis, SCL = "core" la­ mi na of the neural spine, SDL=suprad iapophysiallam­ ina, SIPRL=subinfraprezygapophysial lamina, SPRZ= supraprezygapophysial lamina, SPZL = suprapostzyga­ pophysial lamina. AXIAL SKELETAL ELEMENTS Cervical vertebrae WN-Vl (Figs. I & 2; photographs in DALLA VECCHIA, 1994b, fi g. 3, and DALLA VECCHIA, 1997c, fi g. 2): two cervical vertebrae were preserved nearly in anatomical articulat ion. The posterior of the two is entire while only a posterior fragment (with the eoty le and the left postzygapophysis) of the other is preserved. The following description concerns the com­ plete vertebra (Fi gs. I & 2). This is crushed latera lly, the neural arch is bent to the Icft s ide and the left diapophysis with the corresponding ho ri zon tal lamina are crushed and bent to the centrum. The right s ide has been exposed to eros ion on the sea bOil om in recent times and was strongly weathered. The centrum is opisthocoe lous, very e longate and tubular. Its length is 350 mm , its height at the posterior end is 50 mm , the length/height ratio is therefore 7; maximum depth of the neural arch as preserved is 100 mm. The centrum has a cavernous, cancellate structure wi th thin external walls and re lative ly smail, irregular internal cavities bordered by bone septa and ridges; the septa and wall s are compa rat ive ly thicker than in the posterior cervical vertebra MPCM-V2 described below. The inner cavities are, at least in the anterior part of the centrum , antero-posteriorly e longated. A small pleuro­ coel, al so antero-posteri orl y e longated, is identifiable on the middle ventral part of the lateral side of the cen­ trum (Fig. 2). "Pleurocoels" are also present anteriorly on the centrum, just above o r at the base of the para­ pophys is. They ac tuaIJy are deep depressions and do not appear to communicate with the inner part of the Dalla Vecchia: Remains of Sauropoda (Reptilia, Saurischia) in the Lower Cretaceous .. NS COPR CPR CDL DP IPDL...: .. >,'. PZ ~ HL " ..1~/ PR"i<{t;'i,;.:~~ ... '~~ ... ·.:!.,.~tti"-.:.L.'~~;~c PR~',h .. :· ~, - PP NS IPZL lii!1T~M~\- NC DP PP 107 Fig. J Cervical vertebra WN-VI . Views: A) left laleral. B) dor­ sal, C) ventral, D) posterior. E) Recon­ structi on in pos teri­ or view. Acronyms: ASPZ= additional su prapostzygapo­ phys ial lamina. 108 Geologia Crorllic(l 51n ~ ... .. ...... ··· ·· <:'?:' )"~;'('~"6~~;.,~;c,';~;t'i ···I~jjKLii::.;-·:·· · . centrum (they do not pierce the external wall) and are probably the Aussenkaverne reported in the ccrvicals of Brachiosaurlls branca; JANENSCH by JA NENSCH ( 1947, fig. I) . Small , rounded or oval (4-10 mm) shal­ low depressions arc present in the anterior part of the centrum , and are most prevalent on the latera l side of the neural arc h. Ventral excavat ions arc not present. The " pJcurococls" and the depressions on the centrum are fi gured in Fig. 2. The cotylc (the posterior articular cav ity on the cen­ trum) is deep and ova l (main vertical diame ter;:::: 50 mm) but the shape is probably biased by compression, and the condyle is well developed, ball-like and small (maximum diameter= 35 mm). The neural arch occupies nearly the entire dorsal surface of the centrum. The neural spine is low and not bifid. The parapophysis lacks the distal part and is ori­ ented in the vertical plane; despite any compressional e ffects it is unlike ly that this was originally oriented in the hori zontal plane. The diapophys is is triangular in dorso-Iate ral view and tapers at the di stal tip , which is broken and shows a ci rcular outline of the section and a hollow insidc. The tip of the diapophysis and the poste­ rior margin of the anterior horizontal lamina are rough. From the diapophys is a well dcveloped wing-like hori ­ zontal lamina is direc ted anteriorly to the prezygapoph­ ys is and posteriorly to the poslzygapophysis. Where the posterior horizontal lamina begins a strong infrapostdi­ apophysial lamina is al so obvious (Fig. I C); this ends as a lamina at the dorsal-posterior third of the centrum and continues as a ridge with the rel ief tapering caudal­ ly, ending before reaching the caudal margin of the cen­ trum. Where the infraposld iapophysiaJ lamina attaches to the diapophysis, the posterior side of the latter beco­ mes wider and presents a shallow depression (this part has therefore a somewhat spoon-like aspect). There is a double, V -shaped inner centradiapophysial lamina (Fig. I C), with the point placed in correspondence to the nar­ rowing of the tip of the diapophysis. The two branches of the lamina end at the dorsal-lateral part of the cen­ trum; the anterior one is wider and very thin. There is a Fig. 2 Cervical vertebra WN­ V I. Pattern of pleurococls and small depressions. well developed centroprezygapophysial lamina, crushed against the anterior horizontal lamina, with a wide basal attachment on the dorsal-anterior part of the centrum, just caudal to the condyle. There is also a short condy­ loprezygapophysial lamina. The supraprezygapophysia l and suprapostzygapophys ial laminae are thin , well developed and separated (right from left) by very deep infra-supraprezygapophysial and infra-suprapostzy­ gapophysial cavit ies (Fig. I B). The supraprezygapo­ physial laminae are thinner than the supraposlzygapo­ physial laminae. The ventra-poste rior part of the medial side of the supraposlzygapophysial lamina presents a deep cavity bordercd medially by a very thin vertica l lamina (ASPZ = additional suprapostzygapophys ial lamina). There are two thin , paralle l and short verlical lami­ nae just above the bony arch surrounding the neural cavi ty. The intrapostzygapophysial lamina reaches the­ se laminae medially fanning a Y -shaped structure. The short verti cal lamina, infrapostzygapophys ial lami na and intrapostzygapophys iaJ lamina surround a large, deep cav it y (Fig. I E). Right infrapostzygapophysial lamina and intrapostzygapophysial lamina arc crushed against one another and the corresponding cavity bor­ dered by the two laminae is nearly closed. The prezygapophys is is long and project s we ll beyond the anterior tip of the centrum. T he art icular surface is drop-shaped, fac ing medio-dorsally (orienta­ tion possibly partly modified by crushing). The postzy­ gapophysis is similar to the prezygapophysis in that the articular surface faces ventra- laterally. The well devel­ oped, thin and wide (wing-like) horizontal lamina con­ nects the prezygapophysis and postzygapophys is to the diapophysis. Comparisons - elongated , " tube-like" cervical cen­ tra are present in Brachiosaurus brancai, Diplodocus and Barosallnts lellllls MARSH (MclNTOSH, 1990a, b). Cervical centra of the camarasaurids are relative ly short and wide (OSBORN & MOOK , 1921 ; MclN­ TOSH, 1990a, b). Diplodoeids and Camarasallrus have bifid neural spines in the cervi cal vertebrae (McIN- Dall:t Vecchi3: ReJn3ins of S3uropoda (Replil i3. S3urischi3) in the Lower CrClaceous ... 109 A TOSH, 1990a, b; in Diplodocus and Camarasaurus they arc bifid from cervical 3 onwards) whereas they are single in Brachiosaurus . T he distribution of the "pleurocoels" in WN-V I is similar, but not the same, as that of the anterior cervical vertebrae of Brachiosaurus brollcoi (JANENSC H, 1950, figs. 20, 23, 26, 29). In fact, the size of the posterior pleurococl of the cenlrum is much small er in the descri bed specimen. Al so the overall shape of the vertebra, the anterior elongation of the prezygapophysis, and diapophysis d ista ll y narrow­ ing in a bottleneck manner, are si milar to those of the anterior cervical s of Brachiosaurus brancoi (JANEN­ SCI-!, 1950, rigs . 20, 23, 26, 29). The internal cavi ties are probably the same as the " longi tudinal pneumatic tubes" observed in a presumed specimen of Mamen ­ chisourus YOUNG by RUS SELL & Z I-!ENG ( 1993, p. 2089) and are present also in "Chondrosleosaurus" (e.g. HULKE, 1879, p. 756) and in another cervica l cenlmm described below. M PCM-VS (Fig. 3A): this is the posterior part with the coty le o f a small centrum, w ith a basal-posterior fragment of the neural arch. The spec imen is strongly crushed; it is 11 4 mm long and its height at the coty le is 70mm. Comparisons - it is simi lar to the same region of the vertebra WN -V 1 and represents another cerv ica l verte­ bra belonging to a relatively small sauropod. MP CM -V6: th is speci men is probably the anterior portion (110 mm long) of a right in frapostdiapophysial lam ina from a rather large vertebra : the correspond ing part on the complete cervical WN-V 1 is no longer than 20mm. MPCM -V7 (Fig. 3B): an incomplete posterior part of a postzygapophys is (48 mm long, 60 mm wide) cle­ arly belonging to a vertebra much larger than the com­ plete cervical vertebra WN-V I. Fig. 3 A) MPCM -VS poslerior part of a small cervical centrum, right lateral view; B) MPCM -V7 incomplete post zygapophys is of a cerv ical vertebra, lateral view. The scale bar is in centimetres. M PC M- V2 (Figs. 4 & 5): a nearly complete cen­ trum, rather short , wide and low (leng th = 300 mm, height = 105 mm , and width = 175 mm at the posteri or end) (Fig. 4). Its low profi le is only partly due to dorso­ ventral crushing (the specimen, mainly in it s an terior part, is crushed becau se of its ex tremely cavernous internal struclllre). A small pal1 of the base of the neural arch is also preserved in the posterior part. T he neural arch, the parapophyses, and the external bony wall in the cranial hal f have all been weathered away. The posterior cotyle is rather deep and probably had an ell ipt ical shape, wider than high, with a rat io Wjh = J .67. It s ventral side projects posteriorly more than the dorsal one. There are three large pleurocoe ls on both la te ra l sides (F ig, SA). T he external margins of the anterior pleuroeoel (APL) arc weathered away, there fore what we see now is probably slightly different to the original external shape. It is the shallowest of the th ree openings and is separated [rom the posterior (PPL) and lower pleuroeoel (LPL) by thin bony laminae. T he posterior pleurocoel is the largest and deepest openi ng (about 40 mm), ell ipt ical, craniocaudally elon­ gate, ex tending along most of the caudal half of the centrum. Though deep, the posterior pleurocoels do not occupy the whole inner part of the centrum, and are not separated from each other only by a medial lamina (as in the dorsals of Camaras(JI.II"IIS or Brachiosaurus) . In faci the interior of the cent rum of MPCM-V2 is wholl y composed of smal l, honeycomb- l ike chambers. T he dorsal and ven tral rims of the pleurocoeJ are thick and lip- like. T he anterior pl eurocoel is elliptica l and also anteroposlerior ly elongated. Th is pleurocoel is deeper posterio rl y and becomes more and more shallow ven­ tro-ameriorly. It is separated from the posterior plcuro­ coel by a thin lamina. T he lower pleurocoel opens lat­ era-vent rally in the mid-anterior part of the centrum and is more developed ventrally (Fig. 4D). Its true out ­ line has probabl y been affected by weat hering on the lateral side where it poss ibly ex tended on the pa ra- 110 NA PPL 10 em A B NC c o E Geologia Croa!ica 51/2 Fig. 4 MPCM-V2, posterior cervical centrum. Views: A) right lateral, B) left lateral , C) dorsal, D) ventral, E) posterior. Acronyms: APL=antc­ rior pleurocoel, LPL =iower plcufocoel. PPL= posterior pleurocoel. Dalla Veeehia: Remains of Sauropoda (Reptilia, Saurischia) in the Lower Cretaceous .. , 111 NA , , --- --- -', pp A Fig. 5 MPCM-V2. posterior cervical centrum. A) Pattern of the pleu­ rococl s, B) cancellate structure of the condylar region (in the pic­ ture the vertebra is upside down and shows the ventral-anterior surface). Acronyms as in Fig. 4. pophyses. Its present shape is elliptical and antero-pos­ teriorly elongate. The left one is partly subdivided into cells by thin bony septa. The weathered dorsal surface of the centrum shows the cancellate structure of the inner vertebra. More or less irregularly shaped cells are separated by very thin septa; in some places (as in the left upper-mid part of the centrum) these cells are regularly divided by verti ­ cal septa into a "honeycomb" pattern. A regular "hon­ eycomb" pattern, with antero-posteriorly elongated, tubular cel ls, is evident in the weathered anterior con­ dylar part (Fig. 5B). The trace of the neural canal on the dorsal side of the centrum shows that it is very expanded at both extremities: maximum posterior (at the exit from the neuraJ arch) width is about 50 mm , maximum anterior width is 55 mm, minimum width is at midlenglh (about 25 mm); height in correspondence to the posterior exit from the neural arch is about 35 mm. The posterior opening of the neural canal is therefore elliptical, wider than high. The ventral side of the centrum is very flat and wi thout longitudinal depressions or ridges. In the ante­ rior third the external, compact wall is weathered and the inner tubular, "honeycomb", cancellate structure is exposed (Fig. 5B). In the mid-anterior part there is a median, antcroposteriorly elongate, large (length = 45 mm) and deep (about 40 mm) hole. In the mid-anterior part of the centrum , the ventral external bony wall nar­ rows because of the lower pleurocoel; here the medial margin of the opening is rimmed by the relatively thick wall. Comparisons - the large pleurocoels are very differ­ ent [rom the titanosaurid condition, and are characteris­ tic of brachiosaurids, camarasau rids and diplodocids (mai nly Diplodocus). Brachiosaurids and Diplodocus­ like diplodocids seem to be excluded by the relative shortness and width of the centrum and its dorso-ven­ tral flatness (MciNTOSH, 1990b). The overall mor­ phology of MPCM-V2 and the size and position of the pleurocoels strongly resemble cervical 10 of Cama- rasaurlls supremus COPE (OSBORN & MOOK, 1921, figs. 7 & 32) and also the cervicals called C hondrosteo­ saUl·llS gigas by OWEN (1876, see Pis. II-V). The latter are more or less coeval with MPCM-V2, being from the Barremian Wessex Fonnation (BLOWS, 1995) of Eng­ land and have centra of the same proportions, overall outline, shape of articular surfaces, ventral aspect, very similar anterior pleurocoel and posterior pleurocoel and the particular, regular cancellate, "honeycomb" bone texture, mainly in the condylar region (see Pis. II , IV and V of OWEN, 1876 and the description by HULKE, 1879, p. 756-57). On the other hand, the cancellate tex­ ture is not present in Camarasaurus (P. UPCHURCH, pers. comm.). The size of MPCM-V2 is between that of the two specimens of C. gigas described by OWEN (1876, sec PI. V). They could differ in the presence and shape of the lower pleurocoel, but the shape and actual position of this opening in MPCM-V2 is effected by weathering which rubbed out completely the para­ pophyses, as reported above. Also variation of position of the posterior, lower and anterior pleurocoels may account for the differences. Therefore MPCM -V2 at the present state of knowledge should be referred to this taxon, whose validity, however, is doubtful (see below). C. gigas was based by OWEN (1876) on only a com­ plete and a weathered cervical centrum. HULKE (1 879) added at least another cervical centrum (n. 144, coli. H of HULKE, 1879) (actually he made a mistake in reporting the name and attributed it to C. magnus). C. magnus (= Bothriospondylus mag nils OWEN, 1875 but type of Ornithopsis hulkei of SEELEY, 1870) is repre­ sented by two dorsal centra found in the same forma­ tion and locality of the cervicals called C. gigas (OWEN, 1876, p. 7). It appears plausible that C. mag­ nus belongs to the same taxon of C. gigas as partly real­ ized by Owen himself, who referred them, "provisional ­ ly", to distinct species on the base of the incorrect state­ ment that they are both dorsal elements and that dorsals cannot be so different in the same species (p. 7). C. mag nus was based on a vertebra (BMNH 28632) which OWEN (1875) had called Bothriospondylus mag nus before recognizing its resemblance to the vertebrae of 112 c. gigas. Vertebrae of C. magnus (= Ornithopsis hulkei) have the same honeycomb-like texture in the condylar region as C. gigas (OWEN, 1875, PI. IX). BMNH 28632 was indicated as one of the two types of Ornithopsis hulkei by SEELEY (1870) but OWEN (1876) rejected the name as "misleading" because of its Greek meaning which seems to suggest it is a bird bone. HULKE (1879) considered Eucamerotus HUL­ KE and Ornithopsis to be synonyms because the type of the first (BMNH R2522, a partial neural arch), BMNH 28632 and a "mutila ted centrum , reta ining enough of the arch and superstructures" (p. 755) present an "extre­ mely large-celled cance llous ti ssue" (p. 755). L YD­ EKKER (1888) accepted the synonymy of Chol/dros­ teosaurus (both species) and Omithopsis but suggested that Pelorosaurus MANTELL should be synonimized. In his recent review of Ornithopsis BLOWS (1995) does not include in O. hulkei the cervical vertebrae of C. gigas but considers it, as SEELEY ( 1870) did, as a lectotype of O. hulkei one of the two dorsal centra (the syntype of SEELEY, 1870, BMNH28632 - see L YD­ EKKER, 1888) called C. magI/us by OWEN (1876). C. gigas was recently considered valid and tentatively assigned to the Camarasauridae by McINTOSH (I990b, p. 387) because of its close resemblance with cerv ical 10 of Camarasaurus. P. Upchurch, who is rev iewing the British sauropod material, considers C. gigas a 110m en dubium and probably a member of the Titano­ sauriformes (Brachiosaurus + Titanosauria) (P. UPCH­ URCH, pefS. comm.). MPCM- V4 (Fig. 6): thi s is a part of a larger bone and has a peculiar and somewhat puzzling shape and structure. Its identification is rather difficuil but it appears to be part or a laminar system of the neural arch of a cervical or dorsal vertebra of a sauropod. In fact, even if it is strongly clUshed and its original shape was deformed by compression, the specimen is a ialt iceM work of bone bridges and laminae, with pleurocoel-like openi ngs. This frame, linked to the extreme lightening of the skeletal elements, is typical of sauropod cervical and dorsal vertebrae. I chose an arbitrary orientation (see the caption of Fig. 6) of the specimen to describe it and I identify the main lamina-like structures which compose it with the acronyms S I, S2, etc. That which follows should be considered a tentative description. S I is the upper, longitudinal, lamina-like structure which is strongly cru shed but not very derorm ed. Its outline in the anterior part, where it is possibly fomled by two thinner lami nae now strongly clUs hed one aga inst the other, is rec tangular in s ide view. At the anterior third, S I becomes rib-like (ar), the profile becomes inclined , the upper part enlarges and its dorsal margin is sharply acute. The back side of S I is nearly vertical; here the stlUcture is divided inLO a ribMlike arch (bs) which reaches S2 and a thin, central and inner lamM ina (i/). S2 was originally nearly perpendicular to SI but now is parallel to it because of crushing and defor- Geologia Croalica 51/2 mation. S2 resembles a thick lamina, thicker posteriorly than anteriorly, and hollow inside. A large, oval pleuro­ coel opens in its present s ide surface, revealing the hol­ low inside. S3 is the left side structure corresponding to S2 but was mostly weathered away. If so, SI, S2 and S3 originally formed an upside down Y or T-shaped structure. S4 is a lamina crushed against SI, with a strong, rib-like upper margin (urn) and the remaining part (IS) whieh is a thin sheet of bone crushed against il. The rib-like upper margin um begins in the dorsal mar­ gin of S I where the latter becomes rib-li ke; therefore the two rib-like structures form (and formed before crushing) a V -shaped st ru cture with the acute part pointing forward and downward. S4, now parallel to S 1 was originally diverging from the latter posteriorly. Comparisons M the shape of the structure S 1 resem­ bles one of a low neural spine of a cervical vertebra. However, the re lation with the structure S4 cannot be recognized exactly in any described vertebra of Hap/o­ cantllosaurus, Diplodocus, Apatosaurus and Cama­ rasaurus (pers. obs.). The shape of the specimen mostly resembles one of the bifid cervical neural spine of Apatosaurus (GILMORE, 1936). As reported above, in my opinion it is part of a complex laminar structure of a neural arch of a rather large and bizarre cervical or dor­ sal sau ropod vertebra, but nothing more can be sa id unti l a more complete part of this structure is found. Cervical ribs? (Fig. 7) Seven rod-like bone fragments (among them is MPCM- V8, the others now on exhibit at BaleNalle are without number), with oval or elliptical cross sections (about 10.5 x j 4.5 mm in the longest fragm ent, the low­ er one in Fig. 7,195 mm long). They have the same transverse sizes for all their lengths and are straight or slightly curved; they are not hollow and their surfaces present thin longitudinal striae. One surface is usually flatter (less convex) than the other. Comparisons - They cou ld be segments of the distal part of the shaft of long cervical ribs. They resemble the very characteristic, elongate rod-like shafts of the cervi­ cal ribs of Camarasaurus, Brachiosaurus, Euheloplls ROMER and Mamenchisaurus. On the other hand , MciNTOSH (1990a) indicates short cervical ribs as a diagnostic feature of Diplodocidae. The identification as part of rib shafts is supported by the presence in the outcrop of si milar bone fragments more than 500 mm long and perhaps as long as 1200 mm (D. BOSCA­ ROLLI and F. BACCHIA, pers. comm.). If the identifi­ cation is correct, their large size excludes them from belonging to a vertebra of the same size as the complete cervical vertebra WN-V I ; they belong to a much larger vertebra of a large individual. Despite this, the frag­ ments are very similar to the ossified tendons which strengthen the tail and the back of iguanodontids. How­ ever these ske letal elements are usually smaller and flatter. If the described specimens are ossified tendons there are two possibilities: 1) also iguanodontid s are Dalla Vecchia: Remains of Sauropoda (Reptilia, Saurischia) in the Lower Cretaceous ... II] S1 S2 S1 2cm represented in the fossilifero us outcrop but we do not have other evidence of this, 2) there arc sauropods with ossifi ed tendons in the neck (calcified muscles are pre­ served in Camarasallrlls - D. CHURE, pers. comm. III MclNTOSH, I 990b) or in the tail. Dorsal vertebrae MPCM-Vl (Figs. 8 & 9): the specimen consists of a part ly preserved, small centrum and corresponding neural arch without most of the spine. The neural arch is strongly crushed antero-pos teriorly, due to the action of lithostatic pressure on the ex tremely ho ll ow struc­ ture. The whole vertebra is in fact just a very complex latticework of thin bone laminae. After complete prepa­ ration it seems clear that in previous papers I wrongly identifi ed the posterior part as the anterior (see DALLA VECCHIA , 1994b, fi g. 4b). The an terior side of the verteb ra was strongly weathered and to identify the ® S1 S2 Fig.6 MPCM -V4, possible frag­ ment of the laminar complex of a large cervica l or dorsal vertebra. A-B) "side" views, C) "fronC view. Acronyms: a = "anlerior". p = "posterior", 5 I = lamina I . 52 = lamina 2, 53 = lamina 3, 54 = lamina 4. For the other abbreviations see lext. original st ructures is not possible or at best very diffi ­ cult. The centru m has a kidney-like outline in pos terior view, wider than high (about 80 x 55 mm) and the arti ­ cular face is concave, It is a cy linder, therefore it has convex lateral sides. On both la teral s ides there is ev idence of a very large pleurocoel, probably with an ovate outline (but see below) occupyi ng most of the lateral side of the centrum (Figs. 8D & 9C). The pleurocoel is not a hole piercing the lateral wall of the centrum and in connec­ tion with its hollow inner part (Fig. 8B). It is just a deep depression rimmed ventrally (lip-like ventral marg in), opening dorsally and partially Uust in the vent ral part) divided internally by a small, rib-like vertica l septum (as in the dorsals of Ellcamerofu5 sensu BLOWS, 1995 and many other sauropods). The inner st ructure of the centrum is ex tremely cavernous, cancellate, wi th thin ve rti cal scpta (ma inly antero-poste riorly oriented but some smaller ones arc transversally oriented) and rein­ forc ing strut s (Fig. 8B). The inner septa are as thi n as less than I mm, whcreas the external wall is thicker. The neural canal is oval, large, about 29.5 mm wide and 21 tlll1l high in posterior view, 25 and 16 mm respect ive ly in anterior view. The neural arch is comparalively very lall: the ratio 11/h (H = distance between the base of the postzygapo­ physis and the top o f Ihe centrum , and h =ccntrunl hcight ) is 1.83; the preserved part of the neural arch is 2.8 timcs the height of the cenrrum . The neural arch is formed by very thin , long laminae which arc difficult to ident ify because of the strong crushing; thcir original pos ition is tcntal ively recons tructed in Figs. 9C and lOll. The ri ght pos t zygapophysis is nearly completely preserved. It is connected to the pedicels by a strong, straight, venieallamina (infrapostzygapophysial lami ­ na). There is no suture between the basal pediccls and the upper part of the arch. Between the two infrapostzy­ gapophysia! laminae the arch is depressed. Perhaps the two infrapos rzygapophysiallaminae were connec ted by a thin , horizonlal inlrapostzygapophysial lamina btll c ru shing preveilis furl her clarification. The c ross-sec­ lion of the post zygapophysis shows pneumatic cavi ties ins ide ; the part with cav ities is found dorsally with respect 10 a 5 mm thick layer of spongy bone wh ich is just above the art icular surface. The latter faces ventro­ lateral ly and sligh lly anteriorly. There is no hypo­ sphene: medially to the postzygapophyses the (ITch is deep ly depres sed and in the middle of the concavity there is th e basal part of a rib-like (probably broken) postspinallamina (Fig. SA). The ar ticular surface of the prezygapophysis was probabl y borne by a large, triangular , dorso-Iatcrally Gcu logia Cf()