The taphonomy of dinosaurs from the Upper Jurassic of Tendaguru ...

Mitt . Mus. Nat.kd. Berl., Geowiss. Reihe 2 (1999) 25-61

19.10.1999

The Taphonomy of Dinosaurs from the Upper Jurassic of Tendaguru (Tanzania)Based on Field Sketches of the German Tendaguru Expedition (1909-1913)1

Wolf-Dieter Heinriche

With 23 figures and 2 tables

Abstract

Tendaguru is one of the most important dinosaur localities in Africa . The Tendaguru Beds have produced a diverse LateJurassic (Kimmeridgian to Tithonian) dinosaur assemblage, including sauropods (Brachiosaurus, Barosaurus, Dicraeosaurus,Janenschia), theropods (e .g ., Elaphrosaurus, Ceratosaurus, Allosaurus), and ornithischians (Kentrosaurus, Dryosaurus) . Con-trary to the well studied skeletal anatomy of the Tendaguru dinosaurs, the available taphonomic information is rather limited,and a generally accepted taphonomic model has not yet been established . Assessment of unpublished excavation sketches bythe German Tendaguru expedition (1909-1913) document bone assemblages of sauropod and ornithischian dinosaurs fromthe Middle Saurian Bed, Upper Saurian Bed, and the Transitional Sands above the Trigonia smeei Bed, and shed some lighton the taphonomy of the Tendaguru dinosaurs. Stages of disarticulation range from incomplete skeletons to solitary bones,and strongly argue for carcass decay and post-mortem transport prior to burial . The sauropod bone accumulations are domi-nated by adult individuals, and juveniles are rare or missing . The occurrence of bones in different superimposed dinosaur-bearing horizons indicates that skeletal remains were accumulated over a long time span during the Late Jurassic, and themajority of the bone accumulations are probably attritional. These accumulations are likely to have resulted from long-termbone imput due to normal mortality events caused by starvation, seasonal drought, disease, old age and weakness. The deposi-tional environment of the Middle and Upper Saurian Bed was mainly limnic to brackish in origin, while the palaeoenviron-ment of the Transitional Sands was marginal marine .

Key words : Dinosauria, taphonomy, Late Jurassic, Tendaguru, East Africa .

Zusammenfassung

Tendaguru zählt zu den bedeutendsten Dinosaurier-Lagerstätten Afrikas. Aus den Tendaguru-Schichten sind zahlreicheSkelettreste von Sauropoden (Brachiosaurus, Barosaurus, Dicraeosaurus, Janenschia), Theropoden (z.B . Elaphrosaurus, Cera-tosaurus, Allosaurus) und Ornithischiern (Kentrosaurus, Dryosaurus) geborgen worden . Sie stammen aus der späten Jura-Zeit(Kimmeridge - Tithon) . Während der Skelettbau der Tendagurusaurier gut untersucht ist, wirft die Taphonomie des Saurier-vorkommens von Tendaguru noch immer Fragen auf. Unklar ist bislang, wie die enormen Anreicherungen von Dinosaurier-knochen in den Tendaguru-Schichten zustandekamen . Unveröffentlichte Grabungsskizzen der Deutschen Tendaguru Ex-pedition (1909-1913) erweitern unsere Kenntnisse über die Taphonomie der Tendagurusaurier . In den ausgewertetenGrabungsskizzen sind Knochenansammlungen von Sauropoden und Ornithischiern aus dem Mittleren und Oberen Saurier-mergel sowie aus den Übergangsschichten über der Trigonia smeei-Schicht dokumentiert . Die Lage und der Erhaltungs-zustand der Funde lassen auf erheblichen Zerfall der Kadaver und post-mortalen Transport von Skelettelementen vor derEinbettung schließen . Das Vorkommen von Saurierknochen in mehreren übereinanderliegenden Profilabschnitten der Tenda-guru-Schichten zeigt, daß Skelettreste während der späten Jura-Zeit über einen längeren Zeitraum hinweg akkumuliert wur-den . Die Ansammlungen von Skelettresten gehen wahrscheinlich auf �normale" Sterbe-Ereignisse zurück, wie z . B . Ver-hungern, Verdursten, Kankheit, Altersschwäche und jahreszeitliche Dürre . Als Ablagerungsraum der Mittleren und OberenSaurierschicht kommt ein küstennaher limnischer, zeitweise wohl auch brackischer Küstenstreifen in Betracht. Die knochen-führenden Übergangsschichten unter- und oberhalb der Saurierschichten sind randlich marine Ablagerungen.

Schliisselwörter Dinosaurier, Taphonomy, Oberjura, Tendaguru, Ostafrika .

1. Introduction

Tendaguru in southeastern Tanzania, East Africa,has yielded one of the most important Late Jur-assic dinosaur assemblages . The first dinosaurs

were discovered by B. Sattler in 1907 (Hennig1914a, Wild 1991), and E. Fraas was the first toexcavate dinosaurs in the vicinity of TendaguruHill in September 1907 (Wild 1991). Shortly after-wards, Fraas (1908) described and figured some

The paper is dedicated to Professor Dr. W. Janensch (1878-1969), leader of the German Tendaguru(1909-1913), in appreciation of his outstanding contributions to the understanding of African dinosaurs .

z Museum fiir Naturkunde, Institut fiir Pal5ontologie, Invalidenstr . 43, D-10115 Berlin, Germany.Received January 1999, accepted May 1999

expedition

26

of these bones as "Gigantosaurus" africanus(= Barosaurus africanus) and "Gigantosaurus"robustus (= Janenschia robusta) . Extensive col-lections of dinosaurs were made by the GermanTendaguru expedition between 1909 and 1913(Janensch & Hennig 1909 ; Hennig 1912a, b ; Ja-nensch 1914a, c), followed by field seasons ofthe British Tendaguru expedition between 1924and 1931 (e.g., Migeod 1927, 1930, 1931 ; Parkin-son 1930) . The German and British expeditionsyielded spectacular collections of dinosaurs andsince that time Tendaguru has been known asone of the world's most important Upper Juras-sic dinosaur localities.

Whereas the British dinosaur collection re-mained undescribed (Russell et al . 1980), theGerman dinosaur material from Tendaguru hasbeen well studied . Sauropods (e.g ., Brachio-saurus, Dicraeosaurus), theropods (e.g ., Elaphro-saurus, Ceratosaurus, Allosaurus), and ornitho-pods (Kentrosaurus, Dryosaurus) were identifiedand described in a series of monographs. Most ofthese works concentrated on the anatomical de-scription of the Tendaguru dinosaurs (e.g., Hen-nig 1925; Janensch 1914b; 1929a, b ; 1935, 1950a,1955, 1961a, b) . By contrast, the available tapho-nomical accounts are rather limited (e.g., Hennig1912a, b ; Janensch 1914a, c, 1929a, b) and de-tailed excavation plans have not been publishedso far .

The depositional environment and the originof the bone accumulations recovered from theTendaguru Beds are still under debate. Severaltaphonomic models have been suggested, amongthem the interpretation of Janensch (1914a, c),according to which the Tendaguru Saurian Bedsare marginal-marine in origin . The attritionaland catastrophic bone accumulations and re-markable pre-burial taphonomic biases havebeen taken into consideration in this model . Ja-nensch (1914a, c) argued that dinosaurs hadbeen trapped and killed in the mud or 'mire-holes' of lagoons that were temporarily exposedby ebbing tides (see also Hennig 1912a, b) . Thiswas followed by carcass flotation and decay inshallow water, prior to burial . Bone beds mainlyconsisting of disarticulated skeletal elements ofDryosaurus or Kentrosaurus were thought to re-sult from sudden, catastrophic, mass-mortalityevents (e.g., Hennig 1912a, 1925; Janensch 1914a,c) . Other workers argue that an estuarine en-vironment existed in what is now the Tendaguruarea (e.g ., Abel 1927, Kitchin 1929, Parkinson1930, Colbert 1984), and that dinosaur carcasseswere washed in by rivers. They were transported

Heinrich, W.-D., Taphonomy of Dinosaurs from Tendaguru

downstream and buried close to the mouth ofrivers that flowed into the sea . Reck (1925),however, doubted catastrophic mass mortalityand interpreted the depositional environment ofthe Saurian Beds as saline marshes . To datethere is no taphonomic model that convincinglyexplains the enormous accumulations of dino-saur bones in the Tendaguru Beds .

The archive of the Institute of Palaeontologyof the Museum of Natural History of the Hum-boldt University, Berlin, retains unpublishedfield sketches made by the German Tendaguruexpedition (1909-1913) . These field sketches areextremely important because primary field re-cords of the expeditions are very scarce. The un-published field records reveal details of the pre-servation characteristics of Tendaguru dinosaursand enable some reassessment of the tapho-nomic interpretations suggested by previousworkers. The main focus of this paper is to docu-ment and assess these unpublished Tendagurufield records.

Setting

2.1 Geological and palaeontological background

The Tendaguru Beds that have yielded the dino-saur assemblages crop out in southeastern Tan-zania (Hennig 1914b, 1937 ; Janensch 1914a, c ;Aitken 1956, 1961 ; Kent et al . 1971 ; Kapilima1984 ; Zils et al. 1995a, b; Fig . 1) . The sedimen-tary sequence was named after Tendaguru Hill(= steep hill ; Dr . J Kabudi, Daressaalam, oralcommunication), which is located about 60 kmnorthwest of the seaport of Lindi (Fraas 1908,Janensch 1914a, c) . The following paragraphbriefly summarises the original description by Ja-nensch (1914c) .

The Tendaguru Beds represent shoreline de-posits that were laid down near an oscillatingstrandline during Late Jurassic and Early Cretac-eous times (Janensch 1914a, c ; Russell et al .1980) . There are three sedimentary cylces in theTendaguru Beds (Hennig 1914b, Janensch1914c) . The series of sediments begins with sandymarls of the Lower Saurian Bed (Untere ersteSaurierzone, Janensch 1914c ; Unterer Saurier-mergel, Janensch 1961a) which measure morethan 20 m in thickness (Fig . 1) . These strata passgradually into the marine sandstones of the Ner-inea Bed (Untere Sandsteinzone, Nerineenzone,Janensch 1914c ; Untere Zwischenschichten, Ja-nensch 1961a) which are overlain by greenish-

Mitt . Mus. Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

grey and reddish sandy marls of the MiddleSaurian Bed (Mittlere zweite Saurierzone, Ja-nensch 1914c ; Mittlerer Sauriermergel, Janensch1961a) . Janensch (1914c) reports a thickness ofabout 25 m for the Nerinea Bed and some 15 mfor the Middle Saurian Bed. The succeedingfine to coarse grained marine sandstones ofthe Trigonia smeei Bed (Mittlere Sandsteinzonemit Trigonia smeei, Janensch 1914c ; ObereZwischenschichten, Janensch 1961a), which mea-sure about 20 m in thickness, separate the Mid-dle Saurian Bed from the Upper Saurian Bed(Oberste dritte Saurierzone, Janensch 1914c;Oberer Sauriermergel, Janensch 1961a) . Thelatter consists of greenish-grey and reddish sandymarls which reach some 40 m in thickness. TheUpper Saurian Bed is overlain by sandy shallowmarine sediments of the Trigonia schwarzi Bed(Obere Sandsteinzone mit Trigonia schwarzi,Janensch 1914c), the thickness of which was esti-mated at approximately 5 m.

Sandy or oolitic sediments at the boundariesbetween the Saurian Beds and the marine sand-stone units (Zwischenschichten, Nerinea Bed,Trigonia smeei Bed) have been described asTransitional Beds (Obergangschichten), sincemarine invertebrates (e.g, belemnites) are asso-ciated with skeletal remains of dinosaurs (Hen-nig 1914b, Janensch 1914c) . Transitional Bedscontaining marine invertebrates and dinosaurshave also been reported from the Nerinea Bed(Janensch 1961a) . Sandy deposits below andabove the Trigonia smeei Bed have been distin-guished as the Lower and Upper TransitionalSands by Russell et al . (1980) .

The age of the Tendaguru Beds has been con-troversial since the first dinosaurs were discov-ered in the vicinity of Tendaguru Hill (e.g ., Ja-nensch 1914c; Dietrich 1925, 1933; Kitchin 1929 ;Hennig 1937; Aitken 1956, 1961) . The age of theTendaguru Beds ranges from Late Jurassic toEarly Cretaceous (e.g., Hennig 1914b ; Janensch1914c; Schuchert 1918, 1934; Dietrich 1933 ; Hen-nig 1937 ; Aitken 1956, 1961 ; Kapilima 1984),contrary to previous views that placed the se-quence in the Cretaceous (e.g., Fraas 1908,Kitchin 1929), and there is general agreementthat the Trigonia schwarzi Bed, above the UpperSaurian Bed, correlates well with the Early Cre-taceous (e.g ., Janensch 1914c, Lange 1914,Zwierzycki 1914, Dietrich 1933, Aitken 1961) .

Hennig (1914b, 1937) believed that the Ten-daguru Beds represent a nearly continuous se-quence without major breaks . By contrast, Par-kinson (1930) suggested a disconformity below

27

the Trigonia schwarzi Bed, a view that is stronglysupported by palaeontological (Dietrich 1933,Spath 1927-1933, Aitken 1956, 1961) and geolo-gical evidence (Aitken 1956, 1961) . The inverte-brate fossil record indicates an Upper Valangi-nian (Oberes Valendis) to Aptian age (Dietrich1933) or Hauterivian to Aptian age (Spath1927-1933) for the Trigonia schwarzi Bed (Ait-ken 1956, 1961) . The Jurassic-Cretaceous bound-ary is placed above the Upper Saurian Bed,which is thus regarded as entirely Late Jurassicin age (e.g., Dietrich 1933, Aitken 1956), and be-low the Trigonia schwarzi Bed, contrary to theopinon of previous workers who positioned thisboundary either at the base (e.g ., Hennig 1914b,Janensch 1914c), or within the Upper SaurianBed (e.g., Hennig 1937) . The stratigraphic workin the Tendaguru region is concisely summarizedby Aitken (1961) .The present report is especially concerned

with the Middle and Upper Saurian Beds andthe Transitional Sands above the Trigonia smeeiBed. These units of the Tendaguru Beds are nowgenerally considered to be Kimmeridgian toTithonian in age (Aitken 1956, 1961; Russell etal . 1980; see also Schrank 1999, Schudack et al .1999) . Moreover, charophytes suggest a Kimmer-idgian age for the Middle Saurian Bed (Schu-dack 1999).

The majority of dinosaur remains were recov-ered from the Middle and Upper Saurian Bed,occuring mainly as incomplete skeletons and soli-tary bones (Janensch 1914a, c ; 1929a, 1961a) . Bycontrast, the Lower Saurian Bed has onlyyielded a small number of isolated sauropodbones and theropod teeth (Janensch 1929a,1961a) .

The Middle and the Upper Saurian Bed showclear evidence of redeposition (Janensch 1961a :179) . This is consistent with the lithological char-acters of the matrix from the Middle SaurianBed from Site Jg (WJ) which was searched formammals (Heinrich 1998, 1999) . The matrix oc-casionally contains assemblages of oncoids, mudclasts and lithoclasts consisting of reworked ca-liche nodules (Dr . M. Aberhan, Prof. Keupp, Dr.Schudack ; oral communication) . The latter werederived from palaeosoils that had developed inthe Tendaguru palaeoenvironment, under awarm and seasonally dry (arid or semiarid) cli-mate . In this connection it is interesting to notethat the known invertebrate fossil record fromthe Middle and Upper Saurian Beds, which in-cludes records of freshwater gastropods (e.g.,Dietrich 1914, Hennig 1914c), evidence of As-

28

mussia (= Estheria) (Janensch 1933), and limnicto brackish ostracods (Schudak et al . 1999) isconsistent with the sedimentological data .

Dinosaur bone assemblages, including concen-trations of water worn limb bones of sauropodswere also frequently found in the TransitionalSands above and below the Trigonia smeei Bed,and some of them show evidence of reworking(Janensch 1961a) . Dinosaur remains have alsobeen reported from the Transitional Sands abovethe Nerinea Bed (Janensch 1961a). The co-occur-rence of dinosaur bones and marine to brackishinvertebrates in these beds suggests a shallowmarine to brackish depositional environment forthe Transitional Sands (Janensch 1914c, 1961a) .

The marine depositional environment of theNerinea Bed, Trigonia smeei Bed, and Trigoniaschwarzi Bed is indicated by the presence ofmarine invertebrates, including ammonites (e.g .,Zwierziecky 1914, Dietrich 1933, Zeiss 1975),gastropods (e.g ., Dietrich 1914, 1933; Hennig1914c), lamellibranchs (e.g ., Lange 1914, Dietrich1933, Aitken 1961), and brachiopods (e.g ., Lange1914), but these all require further study. More-over, with regard to the stratigraphic terminol-ogy, Aitken (1961) doubted that Trigonia smeeidescribed from the Tendaguru Beds (e.g ., Lange1914, Dietrich 1933) was identical to Trigoniasmeei from the type locality of Shapoor in Cutch(India), contrary to Cox (1952) who acceptedtheir synonymy.

The field scetches of the German Tendaguruexpedition cover sites in the Middle SaurianBed, the Upper Saurian Bed, and the Transi-tional Sands between the Trigonia smeei Bedand the Upper Saurian Bed that yielded a di-verse vertebrate fauna . As far as the land verte-brate assemblage of the Middle Saurian Bed isconcerned, the dinosaurs are the most commonelement and include among others Brachio-saurus brancai, Barosaurus africanus, Dicraeo-saurus hansemanni, Kentrurosaurus aethiopicus,Dryosaurus lettow-vorbecki, Elaphrosaurus bam-bergi, Labrosaurus (?) stechowi, Ceratosaurus (?)roechlingi, and Allosaurus (?) tendagurensis (Ja-nensch 1914b, 1925b, 1929a, 1955, 1961a; Hennig1925). Pterosaurs (Janensch 1914b, Reck 1931,Unwin & Heinrich 1999) and lizards (Broschins-ki 1999) and mammals (Heinrich 1998, 1999)also occur .

The land vertebrate fauna of the Upper Sau-rian Bed is also dominated by dinosaurs, namelyby Brachiosaurus brancai, Barosaurus africanus,Janenschia robusta, Dicraeosaurus sattleri, Ken-trurosaurus aethiopicus, Elaphrosaurus bambergi,

Heinrich, W.-D ., Taphonomy of Dinosaurs from Tendaguru

Labrosaurus (?) stechowi and Ceratosaurus (?)roechlingi (Janensch 1914b, 1929a, 1961, Hennig1925) . Pterosaurs (Reck 1931) and a mammal(Dietrich 1927, Heinrich 1991, 1999) were alsoreported .

The Transitional Sands below the Upper Sau-rian Bed and above the Trigonia smeei Bedyielded among others Brachiosaurus brancai,Barosaurus africanus, Janenschia robusta, Di-craeosaurus sattleri, and Kentrurosaurus aethiopi-cus (Janensch 1961a) . Pterosaurs were also de-scribed (Reck 1931, Galton 1980) .

2.2 Sites

Dinosaurs were first discovered at a few sites inthe vicinity of Tendaguru Hill (Fraas 1908), andwere subsequently collected from approximatelyone hundred sites during the field seasons of theGerman Tendaguru expedition between 1909and 1913 . Most of these sites are located close toTendaguru Hill (see Janensch 1914a : 45 ; 1925a :18), except for a small number of localities inthe Mchuya and Makangaga area that are notconsidered here . The former is about 40 km, thelatter approximately 80 km north of the Tenda-guru region (Janensch 1914a, Fig . 1) .

The German Tendaguru expedition wasrun by W Janensch (1909-1911), assisted byE. Hennig (1909-1911), H. von Staff (1911), andH. Reck (1912-1913) who was supported by hiswife Mrs . Ina Reck nee von Grumbkow. TheTendaguru sites were worked intensively, and atimportant localities, the collecting period ex-tended through three field seasons from 1909 to1911, for example at Site S (Fig. 1) which yieldedthe type specimen of Brachiosaurus brancai (Ja-nensch 1914a, c) . In 1911, about 500 Africanswere involved in the excavations . Together withtheir families about 900 people temporarily inha-bitated the region of Tendaguru Hill (Jaeger1971) .The excavations met with outstanding success .

African field crews supervised by the leaders ofthe German Tendaguru expedition and nativeforemen uncovered the bones, which weremapped, labelled, catalogued, packed, carried tothe seaport of Lindi, and shipped to Germany(Janensch 1914a) . Each site was indicated withlower case or capital letters, Roman numerals orother symbols (e.g ., Nr.) combined with Arabicnumerals (Janensch 1914a) . Janensch (1914a,1925a) published maps indicating the topo-graphic and stratigraphic position of the princi-pal sites of the Tendaguru area, most of which

Mitt . Mus. Nat.kd. Berl., Geowiss. Reihe 2 (1999)

are from the Upper and Middle Saurian Beds .Only a few sites are known from the TansitionalSands at the base and top of the Saurian Beds.A detailed account of the diggings at Tendaguruand the recovery of dinosaur bones is given byJanensch (1914a) .

2.3 Excavation records

Documentation of the German Tendaguru expe-dition is largely incomplete . Records include adetailed field catalogue, containing lists of thedinosaur bones recovered with hand-writtennotes by Werner Janensch, Edwin Hennig, andHans Reck. Sketchbooks of the German Tenda-guru expedition with pencil and ink drawings byWerner Janensch, Edwin Hennig and HansReck, documenting single bones and skeletons,and watercolours of individual skeletal elements,possibly by Ina Reck, are also available . A fewtopographic quarry plans are also present in thesketchbook . The collection of records also con-tains Werner Janensch's photo album containingblack and white photographs documenting thefield seasons from 1909 to 1911. Most importantare photographs of uncovered sauropod skele-tons including Dicraeosaurus hansemanni (Sitem) and Brachiosaurus brancai (Sites D, S, no) .Three oil paintings completed by Ina Reck atTendaguru Hill in 1912 were recently rediscov-ered in the Institute of Palaeontology of the Mu-seum of Natural History of the Humboldt-Uni-versity, Berlin . These paintings give some idea ofthe size of the quarries and field work at Tenda-guru, possibly of Sites Jg (WJ) and St . This doc-umentation, detailed above, forms the basis forthis study.

Nevertheless, taphonomical analysis of theTendaguru dinosaurs are greatly hampered bythe loss of field records. There is, for instance,only a partial field book of the German Tenda-guru expedition and this covers the time spanbefore the expedition's arrival at Tendaguru Hillon April 16, 1909 . Missing parts of the fieldbook, quarry maps and excavation plans appearto have been destroyed by fire during WorldWar II, possibly as a result of an air raid on No-vember 12, 1943 that completely destroyed a col-lection room containing dinosaurs from Tenda-guru . The type material of Dryosaurus lettow-vorbecki has been missing since that time .

As mentioned previously, almost 100 dino-saur-bearing sites were discovered by the Ger-man Tendaguru expedition between 1909 and1913 . The available field sketches, however, cov-

29

er not more than 15 sites. Skeletal elements areroughly sketched in the quarry plans, partly inpencil, partly with ink . The bones are mostly la-belled with arabic numerals combined occasion-ally with the symbols of the sites (e.g., Cl, C2) .Preliminary anatomical identifications of thebones are given in the field sketches and theTendaguru field catalogue . Hand-written notesby W. Janensch indicate that some of these iden-tifications were later reassessed . With a few ex-ceptions, the length and width of the bones arenot indicated in the field sketches.

Scale drawings of bone accumulations madewith grids were probably lost during World WarII, and arrows indicating compass directions areoften lacking . Discrepancies between the speci-men count given in the Tendaguru field catalo-gue and bones documented in the excavationsketches indicate that the quarry plans often re-present only "snapshots" of the diggings, and donot completely reflect all details of the fossil ver-tebrate record encountered during the excava-tions. Moreover, the field sketches lack detaileddata with respect to the contemporary fauna andsedimentology of the dinosaur-bearing beds ex-posed by the digging activities.

Despite these gaps, there is no doubt that theunpublished field records of the German Tenda-guru expedition can contribute to a better under-standing of the taphonomy of the Tendaguru di-nosaurs. The field sketches record approximatelythe original position of skeletal elements andtheir position relative to one another. A broadspectrum of disarticulation stages can be recog-nized, and reveal interesting details of the tapho-nomic history of the Tendaguru dinosaurs thathave not been documented so far. It is note-worthy that much of the skeletal material docu-mented in the field sketches is housed today inthe collections of the Palaeontological Instituteof the Museum of Natural History, HumboldtUniversity, Berlin . Therefore, some of the pre-served field sketches could be "updated" and,for example, augmented with scales and ontoge-netic data obtained from palaeohistological ex-aminations of sauropod limb bones (Sander1999, in press) .

3. The field sketches

The majority of field sketches document dino-saur bone assemblages from the Upper SaurianBed. Only a few quarry maps cover bone accu-mulations from the Middle Saurian Bed and the

Trigonia schwarziBed

Upper

Saurian

Bed

Lower

Saurian

Bed

30

Heinrich, W-D., Taphonomy of Dinosaurs from Tendaguru

w

Geographic location and geologicalstructure of Tendaguru, Tanzania

2 30

500

1000

1500m

+ ,-

+

+ + ++ r + + + + + + + + + + + + + + _+ +

+ + + + + + + + + + + + + + + ++ r

+ + + + + + + + + + +

+ + +

E-~_

TSBSMB

-MSB~NB~LSB

-G

E0

Fig. 1 . Map of the Tendaguru area with geological sections and the location of the sites considered in the present account .Data from Hennig (1914b) and Janensch (1914c, 1925a) . Note that the marine Nerinea Bed, the two Trigonia Beds, and theTransitional Sands were not mapped in the original site plan (Janensch 1925a : 18) and the Transitional Sands were not distin-guished as separate stratigraphical units by Janensch (1914c) . 1, Lower Saurian Bed ; 2, Middle Saurian Bed; 3, Upper Saurian

_TSTS

Trigonia

smeei

Bed

TSTS

Middle

Saurian

BedTSTS

Nerinea

Bed

TSTS

Mitt. Mus. Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

Table 1Stratigraphic distribution of selected dinosaur-bearing sites in the Tendaguru area .

Stratigraphy

Upper Saurian BedTransitional Sands between Trigonia smeei Bed and Upper Saurian BedTrigonia smeei BedTransitional Sands between Trigonia smeei Bed and Middle Saurian BedMiddle Saurian BedTransitional Sands between Middle Saurian Bed and Nerinea BedNerinea Bed

Transitional Sands overlaying the Trigonia smeeiBed. No records are available from the LowerSaurian Bed, the Transitional Sands above andbelow the Nerinea Bed, and the TransitionalSands below the Trigonia smeei Bed. The topo-graphic and stratigraphic position of the sites de-tailed in the present taphonomic account is indi-cated in Figure 1 and Table 1 .As in other vertebrates, the Tendaguru dino-

saurs began their taphonomic history as car-casses with complete skeletons, followed by abroad spectrum of disarticulation stages and scat-tering of individual skeletal elements . Analysis ofthe field sketches and data given by Janensch(1929a) reveal that four main preservational ca-tegories occur in the Tendaguru Beds (Table 2) :(1) skeletons, (2) partial skeletons, (3) bone ac-cumulations dominated by disarticulated indivi-dual elements, and (4) solitary bones. Thesemain preservational types can be further subdi-vided by using additional taphonomical criteriasuch as the degree of disarticulation and loss ofbones (see Table 2) .

Table 2Disarticulation stages of a dinosaur skeleton (simplified). For explanation see text.

Disarticulation stages

Designation

A

Complete skeletonB

Complete skeletonC

Complete skeletonD

Incomplete skeletonE

Incomplete skeletonF

Incomplete skeletonG

Complete partial skeletonH

Complete partial skeletonI

Complete partial skeletonK

Incomplete partial skeletonL

Incomplete partial skeletonM

Incomplete partial skeletonN

Bone fieldO

Bone fieldP

Solitary Bone

A

Continued legende Fig. 1

The assignment of bone accumulations to dis-articulation stages such as incomplete skeletonsor partial skeletons is based on a simplified de-scriptive scheme that provides unambiguouslydefined taphonomic categories (Table . 2) . Ac-cording to this classification, partial skeletonsonly consist of one discrete skeletal unit (e.g .,tail or hind limb), in contrast to skeletons thatalways comprise more than one of those units(e.g., tail and hind limb; fore limb, shoulder gir-dle and neck) .

The skeletal elements are annotated with sym-bols used by the German Tendaguru field crew .With the exception of Fig . 11, triangles indicatespecimens documented either by arabic numeralsor by hand-written notes in the Tendaguru fieldcatalogue and descriptions in the literature. Theprecise position and orientation of these bones isuncertain . The assessment of quantitative dataand the azimuth orientation of bones must betreated with caution because of the incomplete-ness of the excavation records. Reference speci-mens that were used to calculate the scale for

Skeletal elements

Sites

C,D,N,0,P,k,ab,cc,XH, IX

S, Y,Jg(=WJ),m

articulated, no loss of bonespartly articulated, partly disarticulated; no loss of bonesdisarticulated, no loss of bonesarticulated, loss of bonespartly articulated, partly disarticulated ; loss of bonesdisarticulated, loss of bonesarticulated, no loss of bonespartly articulated, partly disarticulated ; no loss of bonesdisarticulated, no loss of bonesarticulated, loss of bonespartly articulated, partly disarticulated ; loss of bonesdisarticulated, loss of bonespartly associated, partly disassociateddisassociatedisolated

31

Bed ; TSB, Trigonia schwarzi Bed ; USB, Upper Saurian Bed ; TSMB, Trigonia smeei Bed ; MSB, Middle Saurian Bed ; NB,Nerinea Bed ; LSB, Lower Saurian Bed ; G, gneiss; TS, Transitional Sands.A - Sites in the Middle Saurian Bed, " - Sites in the Upper Saurian Bed, 0 - Sites in the Transitional Sands between theTrigonia smeei Bed and the base of the Upper Bed

32

several quarry plans are quoted in the captionsof the textfigures.

The definition of taphonomic variables usedin the description is as follows (Behrensmeyer1991, Lyman 1996): MNI - minimum number ofidentified individuals; NISP - number of speci-mens documented in the field sketches ; size ofaccumulation - area (m

2) over which skeletal

remains are recovered ; density - number ofspecimens documented in the field sketchesper m2 .

The following anatomical abbreviations areused in the quarry maps: as - astragalus, ca -caudal vertebra, c - rib, cc - cervical rib, cd -dorsal rib, co - coracoid, cr . - skull, cv - cervi-cal vertebra, do - dorsal vertebra, fe - femur,fi - fibula, he - haemapophysis, hu - humerus,il - ilium, is - ischium, me - metacarpal, mt -metatarsal, pb - pubis, pm - phalange (manus),pp - phalange (pes), ra - radius, sa - sacrum,sc - scapula, st - sternal plate, tb - tibia, ul -ulna, v - vertebra, ~) - specimen of known on-togenetic age. The right side is indicated by dand the left by s. Capital or lower case letterswith arabic numerals are catalogue numbers ofthe German Tendaguru expedition .

Tendaguru, Site CUpper Saurian BedBarosaurus africanus

IIIII~IIIIIIIIIIIIII

vertebral column and

shoulder girdle

pelvic girdle

hind limb

indet .

ribs


Further abbreviations : GTE - German Ten-daguru expedition, MNHB - Museum für Na-turkunde der Humboldt-Universität zu Berlin,IPHB - Institut für Paldontologie am MuseumMr Naturkunde der Humboldt-Universität zuBerlin .

3.1 Sites in the Upper Saurian Bed

3.1.1 Tendaguru Site C

Location : Site C is situated approximately0.7 km south of Tendaguru Hill .Stratigraphy : Upper Saurian BedTa xon : Barosaurus africanusFigure : 2Reference : Janensch (1914a, 1929a, 1961a),GTE field catalogue : 3Assemblage dataNISP: 37MNI: 1Number of taxa : 1Designation : Single-individual, single-taxonassemblage .

C20C21 cv C19 C18

.wIONN~i~

'W10111111.0, 11111-111

~ IIOII IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII~h

C22tb

~C23fi

approx . 1 m

Fig . 2 . Incomplete skeleton of Barosaurus africanus from Tendaguru Site C. A few bones are still articulated, but most aredisarticulated . The length of the string of cervical vertebrae C18-21 (2.7 m) was used to calculate the scale for the quarryplan. Redrawn after a field sketch from W. Janensch

Mitt . Mus . Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

Quarry dataSize of accumulation : approxmately24.00m2Density : approximately 1 .54Modification dataDisarticulation stage : EMeasurements : The length of the caudal ver-tebrae was measured as 180 mm (C 2), 200 mm,(C3), and 200 mm (C5) ; that of the femur (C15)at 1320 mm (Janensch, GTE sketch book). Thelength of the left tibia (C13) was measured at890 mm (Janensch 1961a) .Ontogenetic data : UnknownComments : The digging at Tendaguru Site Ccommenced in April 1909 (Janensch 1914a: 43) .Cervical and caudal vertebrae, an ilium, a femur,tibia, fibula, and an astragalus assigned to an in-dividual of Barosaurus africanus were recovered(Janensch 1929a: 5) . Note the dominance of ele-ments of the posterior body region . Most of theskeletal elements were disarticulated, but stillclosely associated (e.g ., tibia and astragalus) . Theimperfect femur (C15), the tibia (C13), and theastragalus (C14) belong to one single hind limb

(Janensch, GTE field catalogue : 3) . The asso-ciated tibia (C22) and fibula (C23) were foundjust below a string of articulated cervical verteb-rae (C18-21) which measure 2,7 m in length . Astring of articulated caudal vertebrae (C4 and 5)is also present. These articulated specimens sug-gest that some skeletal elements were still con-nected by soft tissue (ligaments or muscles) priorto burial . Several caudal vertebrae were disarti-culated, but apparently still close to their properanatomical order (Janensch, GTE field catalo-gue: 3), indicating minimal transport from theposition in which the carcass finally came to rest .Extensive damage occured to the ilium and oneof the femora prior or subsequent to burial .Most of the long bones and the string of cervicalvertebrae show a similar (? west to east) orienta-tion .

3.1.2. Tendaguru Site D

Loc a t i on :

Site

D

is

located

about

1.0 kmnorth-north-east of Tendaguru Hill .Stratigraphy : Upper Saurian Bed

33

Fig . 3 . Incomplete skeleton of Brachiosaurus brancai uncovered from the Upper Saurian Bed at Tendaguru Site D . Photo-graph : W. Janensch

34


Tendaguru, Site DUpper Saurian BedBrachiosaurus brancai

D33

11111111111111111111

shoulder girdle

fore limb

hind limb

Taxon : Brachiosaurus brancai

Figures : 3, 4

assemblage

vertebral column and ribs

Quarry data

Density : approximately 2.27

Modification dataDisarticulation stage : F

Ontogenetic data : Unknown

cv

Size of accumulation : approximately22.00 m2

1111111111101

111111111 IIIIII-iniivnnynIIIIIIIIIIIIIIIIIIIIIillllllll y1111 1IIIP~~

D41

mcD1-31

Reference : Janensch (1914a, Fig. 16, 1929a;1950a), GTE field catalogue : 4-5

D38tb

approx . 1 m

Assemblage dataNISP: 50MNI: 2Number of taxa : 1Designation : Multi-individual but single-taxon

Measurements : The length of the limb boneswere measured as follows (Janensch 1961a: 230,tab. 18): humerus (1339) : 1600 mm, femur (1336) :1550 cm, tibia (D?) : 900 mm, and fibula (D?) :940 mm.

Comments : The excavations at Site D com-menced on June 21, 1909 (Janensch, GTE field

sa

D36fe

D34ra

mc

D43sc

D42sc

D32fe

_D35tb

Fig. 4 . Incomplete skeleton of Brachiosaurus brancai from Tendaguru Site D . The majority of the skeletal elements are disar-ticulated, but closely associated . Most caudal vertebrae were found in natural order . Some limb bones rest on top of oneanother. The measured length of the humerus D 39 (1600 mm) was used to calculate the scale for the quarry map . Theorientation of the map is not known . Redrawn after a field sketch from W. Janensch

catalogue : 4) . The dig yielded a skeleton consist-ing of a series of 29 caudal vertebrae, a poorlypreserved sacrum, pelvic bones (ilium, pubis),dorsal ribs, two anterior limbs without feet, twoscapulae, femora, tibiae, and fibulae (Janensch1914a, 1929a: 7 ; 1950a; GTE field catalogue:4-5). Only some of these are documented in thefield sketch . The dimensions of the limb bonesindicate a medium-sized individual of Brachio-saurus brancai (Janensch 1961a: 230, tab. 18) .The series of caudal vertebrae resting next to thebadly preserved sacrum was found almost in nat-ural articulation (Janensch 1950a) . The associa-tion of the remaining disarticulated elements isevident. The scapula (D42) was apparentlyfound lying upon the head of humerus D39, andseveral other limb bones were found one on topof the other (e.g., tibia D38 and femur D36; ra-dius D34 and femur D32). Plate-like elementssuch as the shoulder blades seem to be orientedin a nearly horizontal plane, and the scapulaD43 and ulna D41 are broken . Most of the elon-gated limb bones are oriented nearly parallel,and the two scapulae are positioned transverseto the prevailing compass direction of the heavylimb bones.

Most bones documented appear to derivefrom a single individual preserved as an incom-

Mitt. Mus. Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

plete skeleton . Originally, the leg bones D34 andD35 were tentatively identified as fibula (?) andradius (?) respectively (Janensch, GTE sketchbook). Later, Janensch (GTE field catalogue: 4,hand-written note) identified these bones as fibu-la and tibia. If properly identified, the existenceof three tibiae indicates the presence of at leasttwo individuals of Brachiosaurus brancai . Unfor-tunately, a reassessment of the limb bones fromSite D is not possible, since the location of thecollection is not known.

3.1.3 Tendaguru Site N

Loca t i o n : Site N is located about 0.9 km eastof Tendaguru Hill .Stratigraphy : Upper Saurian BedTaxon : Brachiosaurus brancai

Figure : 5Reference : Janensch (1914a : 44), GTE fieldcatalogue: 19Assemblage dataNISP: 21MNI: 1Number of taxa : 1Designation : Single-individual, single-taxonassemblage

Tendaguru, Site NUpper Saurian BedBrachiosaurus brancai


shoulder girdle

pelvic girdle

fore limb

indet .

N10cv?

Quarry dataSize of accumulation : UnknownDensity : UnknownModification dataDisarticulation stage : FMeasurements : No measurements are avail-able.Ontogenetic data : UnknownComments : The excavation at Site N com-menced in September 1909 (Janensch, GTE fieldcatalogue : 9) . The dig produced a disarticulatedincomplete skeleton, with a scapula, completeand fragmentary humerus, radius, and ribs. At alittle distance to these bones there were twoischia, a dorsal vertebra, a caudal vertebra, ribs,and a large fragmentary bone (N9) tentativelyidentified as a fragmentary scapula (Janensch,GTE field catalogue : 19). The bone accumula-tion at Site N is dominated by elements of theanterior appendicular skeleton . In total, 21 skele-tal elements assigned to Brachiosaurus brancaiwere collected from site N (Janensch, GTE fieldcatalogue : 19). The bones are attributable to oneindividual of Brachiosaurus brancai. Except for afew skeletal elements, most of the long bonesare oriented nearly parallel, possibly reflectingwave action or current action . Some of themsuch as the humerus and ischia are broken .

N9SO

CNP N8do

N7is

N12cd

35

Fig. 5 . Incomplete skeleton of Brachiosaurus brancai from Tendaguru Site N. All skeletal elements are completely disarticula-ted . Scale and orientation of the map are not known . Redrawn after a field sketch from W Janensch

36

3.1.4 Tendaguru Site 0

Location : Site O isnorth of Tendaguru Hill .Stratigi- aphy : tJppcr Saurian

Ta x o i1 : Dlci-acostim-its Saidei -i

Figure : 6 (colour figure on p. 36)

Reference : Janensch (1914x,GTE field catalogue : 20Assemblage data

NISP: 17MNI : I

Number of taxa : 1

Designation : Single-individual,assemblageQuarry data

located about

Sire of accumulation : approximately7.00 i11 2

Density : approximately 2 .43

Modification data

Disarticulation stage : F

Tendaguru, Site OUpper Saurian BedDicraeosaurus sattleri

013ca .

vertebral column

shoulder girdle

pelvic girdle

fore limb

hind limb

indet .

02 0+fe .d

Bed

1929x,

014ca

(I .S km

1961a),

single-taxon

015mt

Heinrich, W.-D. . Tiplionomy of Dinosaurs from TLndaguru

M c a S u r e in e n t s :

The length of tllc right limbbones were measured as follows (Jancnsch 1961a) : humerus (03) : 610 mm, ulna : 410 min, femur(02) : 9ti0 mn1, tibia (04) : 580 mm, fibula (OS) :620 mm.Ontogenetic data : Janensch (1929x : 10) be-lieved that the skeletal elements ori �̀inatcd froma subadult individual of Dicr-aeosniu-its snulcri,howcvei, hislological examinations of' a right fe-mur (02) and a right 11umeru5 (03) Su0`11est thatboth limb bones belong to in adult individual ofthis species (Sander 1999, in press) .

Comments : Digging by the German Tcncla-guru expedition's crew at Site O commenced atthe beginning of October 1909 (Janensch, GTEfield catalogue : 47) . Janensch (1914x : 44, 1929x:10) reported a scapula, pelvic bones, cents ofcaudal vertebrae, and an imprecisely determinednumber of fore and hind limb elements, all as-signed to Dicrncosaurtrs siadcri. All bones aredisarticulated, but closely associated . There aretwo groups of long hones oriented with theirlong axes in two perpendicular directions. Theboric accumulation at Site O is dominated by

011ca

approx . 1 m

08sc .d

Fig . 6 . Incomplete skeleton of Dicrucosuurus eullleri from lcndaguru Sitc 0 . The spccimcn is completely disarliculated . Notetlic dominance of large liml, hones . The measured lem0th of the humerus 03 (610 inm) was used to calculate tlic scale for thequarry mali . The orientation ol tlic oral) is not known . Redrawn after a field sketch from W. Jancnsch

Milt . Mus. NatId. Berl- Geowiss. Reihe 2 (1999)

heavy linlh hones (11un1erus, ulna, femora, tibiae,fibula) . This may indicate that most smallerhones had been almost completely removedfrom the carcass by ctu -rents prior to burial .

3.1.5 Tendaguru Site P

Location : Site P is located at Nterego, ap-proxinuttely 1 .2 km north-north-east of Tenda-ouru Hill .

S t r a t i o r a p h y :

Upper Saurian Bed

Ta x o n : Jruw»schia mbusiaFigure: 7 (colour figure on p . 37)

Reference : Janensch (1914x, 1922, 1929x,1961x), Bonaparte et al . (in press), GTE fieldcatalogue : 21-22Assemblage data

NISP: 65MN1 : 2

NLinlhcr of taxa : 1

D e s i g n a t i o n :

Multi-i11dividttal

but

sin,le-tax-on assenlblage

Ouarrv dataSize of accumulation : appoxilllatCly24 .(1(1 n1 -

D c n s i t v : approximately 2.74

Modification data

Disarticulation stage : H (find I), 1< (findIII, IV) and L (find II)

Measurements : The length of the left limbhones were measured as follows : 1lttmerus (Pty) :890 111111 .

radius

(Pl l ) :

620 mn1,

ulna

(P12) :670 mn1,

tibia

(P5):

850

111,

ilstragaltts

(P6) :210 mm.


Comments : Dioging at Site P commenced inOctober 19(19 (Janensch, GTE field catalogue :

approx . 1 m

Tendaguru, Site P

Upper Saurian Bed

Janenschia robusta

vertebral column

pelvic girdle

fore limb

hind limb

indet .

)7

Fig. 7. Partial skeletons of hmen.schia rohnsra from Tendaguru Site P The hone asscmhlawc consists nmink , of two articukiledleft fore limbs and two articulated left hind limbs. Pelvic elements rest neat to the sacrum, and some vcrtchrac are scatteredover the llltat- I-N flour. The measured length of the 1111n1ct1IS Pi; (890 nun) wns used to calculate the scale for the yuKtrrN mat) .Tlic oricnlation of the mat) is not known. Redrassn after a field sketch lioin W. Janensch

38

21-22) Four partial skeletons of Janenschiarobusta were uncovered, along with additionalscattered postcranial bones (GTE field catalo-gue; Janensch 1922, 1929a; Bonaparte et al ., inpress) . According to Janensch's unpublished ex-cavation sketches and hand-written notes (GTEfield catalogue : 21), partial skeleton I consists ofa reasonably complete left hind limb, with femur,tibia, fibula, and the foot, including astragalus,metacarpals, and phalanges . The foot bones werestill close to their natural anatomical position,while femur, tibia and fibula were slightly disarti-culated. Close to this bone accumulation lay par-tial skeleton IV, represented by an incompleteleft fore limb, with radius, ulna, carpal, metacar-pals, and phalanges. Nearly all these bones werefound in their proper anatomical order. Partialskeleton II consists of an incomplete hind limb,including tibia, fibula, and slightly disarticulatedand scattered foot bones (astragalus, metatarsals,phalanges). Partial skeleton III includes the ra-dius, ulna, and the manus of an incomplete leftfore limb. In addition to the four partial skele-tons, dinosaur bones found at site P include,among others, a left humerus, a sacrum, and sev-eral disarticulated vertebrae scattered over thequarry floor. Janensch (unpublished sketchbook)also noted ribs in his excavation map but no lo-cation data are given. The association of thesebones with the four partial skeletons remains farfrom clear. The long axes of the two anteriorlower legs (partial skeletons III, IV), the femurof partial skeleton I, and the single humeruswere similarly oriented . The two posterior lowerlegs (finds I, II) are deposited with their longaxes perpendicular to the preferred orientationof partial skeletons III and IV

The low degree of disarticulation, especiallyof the feet, suggests that soft tissue must haveremained attached to the limb bones when theywere deposited . The taphonomic evidence alsosuggests that the partial skeletons were buriedclose to the point where the original corpsescame to rest . In addition, several limb bonesshow signs of strong weathering due to pro-cesses subsequent to burial (Janensch 1914a: 44 ;1961a) .

Taphonomic and anatomical analyses showthat the four partial skeletons belong to a singlespecies, and represent two individuals of Ja-nenschia robusta (Janensch 1922, 1929a: 7; GTEfield catalogue : 21-22) . Tendaguru Site P is theonly known site, so far, that has provided spe-cific information on the fore and hind limbs ofJanenschia robusta .


3.1.6 Tendaguru Site k

Location :

Site

k is situated

about 0.7 kmsouth of Tendaguru Hill .Stratigraphy : Upper Saurian BedTaxon : Barosaurus africanusFi gu r e : 8 (colour figure on p. 39)Reference : Janensch (1914a, 1929a, 1961a),GTE field catalogueAssemblage dataNISP: 50MNI: 1Number of taxa : 1Designation : Single-individual, single-taxonassemblageQuarry dataSize of accumulation : approximately19.00 m2Density : approximately 2.62Modification dataDisarticulation stage : EMeasurements : The length of skeletal ele-ments is as follows (Janensch 1961a) : left scapula(k34) : 1340 mm, right humerus (k37): 970 mm,left ulna (k38) : 740 mm, left ischium (k44):880 mm, left tibia (k41): 860 mm.Ontogenetic data : UnknownComments : Skeleton k was found close to theremains of Barosaurus africanus from Sites Aand B (Janensch 1914a, 1929a) . The collectingsite was discovered in 1910 (Janensch 1914a) . Inspecimen k, all the main body regions (e.g.,skull, axial skeleton, appendicular skeleton) arerepresented. Janensch (1929a : 6) reported abraincase, maxilla, cervical and dorsal vertebrae,anterior caudal vertebrae, ribs, scapula, coracoid,humerus, ulna, ischium, femur, tibia, and fibula,all assigned to Barosaurus africanus. There aretwo clusters of skeletal elements, one of whichcomprises mostly bones of the anterior body re-gion (e.g ., braincase, cervical vertebrae, scapula,coracoid, humerus), whereas the other containsbones from the posterior body region (e.g .,ischium, femur, tibia, fibula, caudal vertebrae) .Only a few skeletal elements are articulated: twostrings of dorsal vertebrae (k12 and k13, k14 andk15) and two articulated caudal vertebrae (k 20,k21) have been found (Janensch GTE field cata-logue: 48). The remaining skeletal elements weredisarticulated but still closely associated . Most ofthe cervical vertebrae and caudal vertebrae were

Mitt . Mus. Nm.kcl . Beil . . Geowiss. Rcilic 2 ( 1999)

apparently found in their natural anatomical or-der (Janensch, GTE field catalogue : 48), andhaemapophyses were associated with the centraof the caudal vertebrae .

The excavation plan (Fig . S) shows that thelimb hones, ischia, scapula, and most of dorsalrips are embedded with their long axes trendingapproximately northwest to southeast . The headsof all the dorsal ribs are directed southwards .The fibula is water-worn (Janensch, GTE fieldcatalogue : 49) .

The arrangement of the skeletal elements ap-pears to have been modified by water currents.However, the presence of articulated and disarti-culated vertebrae found in almost natural anato-mical order, the orientation of the dorsal ribs,and the presence of a bralncase indicate that thecarcass was embedded and decayed close to theplace where it came to rest .

Tendaguru, Site k

Upper Saurian Bed

Barosaurus africanus

skull


shoulder girdle

pelvic girdle

fore limb

hind limb

CJ k2

k1cr

approx . 1 m

3.1.7 Tendaguru Site ah

Location : Site ab is located approximately1 .2 km nortllnorthcast of Tendagtlru Hill .Strati`l'ra1'11y : tJpper Saurian Bed

Ta x a :

Brrrosata -us a/i- icamis, Dicracosaunts ,sat-tlcri, Brachiosaurus hralwai '?, Sauropoda indet .Figure : 9 (colour figure on p . 40)Reference : Janensch (1961a), GTE field cata-logue : 40-41

Assemhlaac dataNISY: 37MNI : 5

NLtt1lhCr of taxa : (4)Designation : Multi-individual, n1ttlti-taxonasscmhlaocQuarry clata

;9

Pig. S. Incomplete skeleton of Rewo.wun -u.s rr/iieluno from Tend,iIIuru Sitc k, in an advanced staoe of disarticuLitirm, with hrain-cnsc and cervical vertebrae resting next to the anterior appendicular MW trunk element, . The eoudal vertebrae arc close lc,the pelvic and hind limb dements. Most skeletal elements are dISPINICUIatCCI, but still closely associalrd . The measured lengthof the scalnila k?4 (1 3,4(1 mm) was used to calculate the scale lür tltc quarry map. Redrawn after a field sketcli fromW. .lanensch

40

Size of accumulation : approximately10.00111 ,

D e n s i t y : approximately 3.61

Modification dataDisarticulation stage : O

M e a s u r e m c n t s :

Dicraeosam-rrs saideri, lengt11of three

left

humcri :

620 mm

(ahl),

620 mm(ah2), SRO1nin (abl0) ; Barczsarrrus all-icunus,length of a left tibia (ab4) : 650 mnl .

0ntoQenetic data : Histological data ob-tained from two left humcri (ab2, abIO) indicatethat these limb hones are froth two adult indivi-duals of Dicraeosaunrs ,scrttlcri (Sander 1999, inpress) .C o mm e n t s : There arc no excavation data giv-en by ,Jancnsch (1914x, 1929x, 1961 a), except forthe GTE field catalogue. The dig commenced atthe beginning of October, 1909. Remains of atleast five individual sauropods were recoveredfrom the Upper Saurian Bed at Site ab . Allbones are disarticulated and scattered, appar-

Tendaguru, Site abUpper Saurian Bed

Barosaurus africanus

Brachiosaurus brancai?

L Dicraeosaurus sattleri

Sauropoda indet .

entl_v randomly, across the excavation floor . Sev-eral bones are imperfect and show distinct signsof breakage . Scattering and mixing of skeletalelements from different taxa and individuals su1-gests that there Was mach transport of honesprior to burial . Some of the heavy limb bonesarc oriented northwest to southeast, others areembedded with long axes trending approxi-mately west to cast . Tlie presence of three lefthUtllerI (abl, ab2, abIO) reveals that the diplodo-cid Dicracosourus .scrttlcri is represented by atleast three individuals. Two tibiae (ab4, ah5) be-long . apparently, to one individual of Barosaurtrsall -icanus (Jancnsch, GTE field catalogue : 4()),and two large femora arc tentatively assigned toBrachiosarrrus hrarwai (Jancnsch, GTE field cat-alogue : 4()) . A huge uncatalogued limb bone thatwas not collected (Jancnsch, field sketch, hand-Written note) Was tentatively Identified as ahumerus of Brachiosaurits branwai .

Heinrich .W.-D., haphonomy of Dinosaurs from Iendaouru

approx . 1 m

Fig. d. Bone field from TCI1Chlgurlr Site ab . The accumulation consists nr<ll111y of disarticulated smiropod limb bones and vcrtc-

hrae . Scattering is evident . Note the prcscnce of three ICft humCrl of Dicrcreu.~rrrrrrr .c .sable°ri . one of which. I1t1111Cr -LIS ahl

(020 111111), was used 10 calculate the scale (or tlic quarry map. Redrawn after a field sketch from W. Jancnsch

Mitt . Mus. Nat.kd. Berl ., Gcowiss. Rcihc 2 (1999)

3.1 .8 Tendaguru Site cc

Location : Site cc is situated approximately2.9 km eastnortheast from Zenc!aguru Hill .

Stratigraphy : Upper Saurian Bed

Ta x oil : Brachiusaurus brancai

Figure : 1() (colour fiL0ure Oil p. 41)

R e fe r en c e : GTE-field catalogue: ! 42

Assemblage clata

NISP : 15

MNI : 1

Number of taxa : I

Designation :assemblage

Quarry data

Density : approximately 1 .11

Modification data

Measurements : Length of a(cc2 ) : 1080 111111 .

DlsartICttlation stage : F

Single-individual, single-taxon

Size of accumulation : approximately24.(1(1 n1~

Ontogenetic data : According to Sander11999, in press), hurnerus cc2 belongs to an adultindividual of Brachinsaur-it .s brancai.

Tendaguru, Site ccUpper Saurian BedBrachiosaurus brancai


shoulder girdlefore limb

indet .

approx . 1 m

cc14do

right hutnerus

Comments : Tapllonomic data for Site cc arepoor . Remains found here include cervical ver-tebrae, a dorsal vertebra, several dorsal ribs, ascapula. a 11tt111CruS, and hone fragments (,la-nensch, GTE field cataluouc: 142-1 ) . All elementsare clisarticttlatecl . but still associated. A pre-ferred orientation of the hones is not discern-able .

The documented skeletal elements are mainlyfrom the anterior body region (neck, anterior ap-penclicular skeleton) of a carcass that underwentmaceration and decay prior to hurial .

3.1 .9 Tendaguru Site X

Location :

Site

X

is

located

about

1 .t) kn1south-south-east from 'Tendagttru Hill .

S t r a t i ,o r a p h y :

LJpper Saurian Bed

Ta xo n : Kciarosam"its acthiopicus, Sauropocla in-clet .

Fi 0 urc : Il

Reference : Hennig (1925), GTE field catalo-gue : 113-114

Assemblage clata

NISP : 115

MNI : (7)

Number of taxi : (2)

ce2L11u .d

Fig. 10 . Incomplete skeleton of Brachin.saurus hrcuwai from Site cc in an advanced state of disarticulatiun . All hones arediSM'tiCUlatCd . The measured length of the Itumertü CC" (108(1 nun) was used to c~dculalc the sctde Ior the quarry map. Theorientation of the map is not known. Redrawn after a field sketch from W. Jancnsch

42

Des i gna t i o n :

Multi-individual,

multi-taxonassemblageQuarry dataSize of accumulation : approximately 30.00 m2

Density : approximately 3.83Modification dataMeasurements : See Hennig (1925)

Disarticulation stage : N (O?)


Comments : Digging at Site X commenced inAugust, 1909 (Janensch, GTE field catalogue :

Tendaguru, Site XUpper Saurian Bed

Site plan and dispositionof bones in the central partof the excavation

Kentrosaurus aethiopicus

+

+

sac sacrum

tb tibia

" D?.+

_L 0

0?A

Sauropoda indet .

0 caudal vertebra

do

dorsal vertebra

approx . 1 m *AA


113) . The quarry map (Fig. 11) is based on afield plan preserved in the GTE sketch bookand a detailed list of skeletal elements assignedto Kentrosaurus (Hennig 1925 : 228-229). Bonesof Kentrosaurus aethiopicus are the most com-mon skeletal elements of dinosaurs recoveredfrom the Upper Saurian Bed at Site X, constitut-ing more than 95% of the total dinosaur finds.The documented bone cluster is dominated byisolated hand and foot bones which formed a pa-vement on the quarry floor. Scattering and mix-ing of the skeletal elements is evident. Judgingfrom the data by Hennig (1925: 228-229), theskeletal remains belong to at least 6 individuals

40

N

sac

X140do

sac

Fig . 11 . Bone field in the cen-tral part of Tendaguru Site X .The accumulation is dominatedby manus and pes elements ofKentrosaurus aethiopicus . Scat-tering is evident . Because ofspace limitation the cataloguenumbers could only be shownfor some bones . Redrawn aftera field sketch from W Ja-

0

nensch or E . Hennig

radiale" intermedium

ulnare

metacarpal

phalange (manus)A metatarsal +

" phalange (pes) ?+++

+ findet . (manus or pes)

' cervical vertebra +caudal vertebra

vertebra++

fe femur

hu humerus 0

Nlilt . Mus. Nat .kd . Berl ., (icowiss . Reilic 2 (I`)`)`I)

of Kcml-oscnlnls acthinlficus . Moreover, isolatedsauropod vertebrae are also present.

The dominance of scattered and mixed handand foot hones confirms the impression thatthere was substantial transport and sorting ofbones prior to burial . However. the origin of thennass accumulation of kentrosaurian bones is dif-ficult to explain owing to the lack of precisedata . It is possible that carcasses of mired anddrowned kentrosaurs floated or were washedaway, whereas the hands and feet rennainccl inthe mud, which was subsequently reworked bycurrents or waves to form a pavement of bones.This interpretation would agree well with the ta-phonollnic model proposed by Janensch (1914c).However, it is also possible that the hand andfoot bones found at Site X were winnowed awayfrcmn the decaying carcasses of Kcim -osaurllsacihiopicus and thus resulted from current sort-ing. Whatever happened, the place of hurial isapparently not the place of death .

Tendaguru, Site HTransitional Sands betweenTrigonia smeei- Bedand Upper Saurian Bed

hind limb

indet .

3.2 Sites in the Transitional Sands above the Tri-gonia snneei !lied

3.2 .1 Tendaguru Site H

Location : Site H is located 1 .3 km northeastof Tendagurul Hill

S t r a t i g r a p hy : Transitional Sands between the7i-i,golliu stncei Bed and illc base of the UpperSaurian Bed

Ta x a :

Ba1'os(llll"11 .~ 11111CY11111S,

Figure : 12 (colour figure on p. 4 3))

Reference :

~Janensch (1929,1), GTE field cata-logue: 1 1

Assemblage data

NI SP: 24

MNI : 1

Number of taxa : 1

Designation :

Multi-individual but Sill" 1C-taXOllasselllblage

Ouarry data

Size of accunnulation : approximately14 .00 In -

approx . 1 m

Fig . 12 . IncompIcle skeleton of Karo.sauru .s africanus Irom Tendamuru Sitc H . The acetunulatiun is dominated hV limb homesMid pelvic elements . Some limb hones rest on lop of one another . The orientation of the map is not known . Redrawn after afield sketch from W . Janensch

44

Density : approximately 1.71Modification dataDisarticulation stage : FMeasurements : Right radius (H6) : 700 mm;left ulna (H7) : 750 mm; right femur (H1) :1200 mm; left femur (H4) : 1290 m; right tibia(H2) : 880 mm; left fibula(H3) : 920 mm; right fibula (H5): 960 mm.

Ontogenetic data : Unknown.

Comments : The dig at Site H commenced inAugust, 1909 (GTE field catalogue : 11), andproduced disarticulated and scattered bones ofBarosaurus africanus, including ribs, scapula,ulna, radius, ischium, pubis, femur, tibia, and fi-bula (Janensch 1929a: 6) . Some of these bones,

Tendaguru, Site IXTransitional Sands betweenTrigonia smeei- Bedand Upper Saurian Bed

Site plan and disposition of bonesin the excavation

1111111111111111

Kentrosaurus aethiopicusBarosaurus africanusBrachiosaurus brancaiJanenschia robusta

indet .

2m

IX01


such as the right femur (H1) and the right tibia(H2), are still associated, but much of the mate-rial recovered is imperfect (e.g . scapula H13,pubes H10). The dominance of heavy longbones and the remarkable scarcity of small ske-letal elements suggests that there was muchpost-mortem transportation of bones from thecarcasses prior to burial . Because of the size dif-ferences between the two femora found (H1:1200 mm; H4: 1290 mm), it is concluded that thebone accumulation from Site H contains skeletalremains of at least two individuals of Brachio-saurus brancai .

In addition to these skeletal remains, Site Hhas yielded non-sauropod bones not documentedin the field sketch, among them a coelurosauriantibia and a metatarsal of Kentrosaurus aethiopi-

Fig. 13 . Bone fieldfrom Tendaguru SiteTX . The accumulationconsists mainly of dis-articulated limb bonesfrom sauropods scat-tered over the quarryfloor . Redrawn aftera field sketch from W.Janensch

Mitt . Mus . Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

cus (GTE field catalogue : 11). Their precise pro-venance is uncertain . Other discoveries from SiteH include marine invertebrates (e.g ., belemnites)indicating a marginal marine palaeoenvironmentfor the dinosaur-bearing Transitional Sandsabove the Trigonia smeei Bed, as formerly sug-gested by Janensch (1914c) .

3.2.2 Tendaguru Site IX

Location :

Site IX is situated about 1.4 kmnortheast of Tendaguru Hill .S t r ati g r aphy : Transitional Sands between theTrigonia smeei Bed and the base of the UpperSaurian BedTa xa :

Barosaurus

africanus,

Brachiosaurusbrancai, Janenschia robusta, Kentrosaurus aethio-picusFigure : 13Refe r e n c e : Janensch (1914c), GTE field cata-logue: 108-112.

As~"~"v e

Assemblage dataNISP: UnknownMNI: UnknownNumber of taxa : (4)Desi gn a t i on :

Multi-individual,

multi-taxonassemblageQuarry dataSize of accumulation : The precise size ofthe excavation floor is not known. The exposedquarry floors a-f measure, in total, roughly13.0 m by 2.5 m (s . Fig. 13)Density : UnknownModification dataDisarticulation stage : O (N?)

45

Measurements : Brachiosaurus brancai,length of a right femur (IXal), 1180 mm; lengthof left femur (IX1) : 880 mm; Barosaurus africa-nus, length of left humerus (IX94) : 640 mm;length of right humeri: 660 mm (IXkl1), 730 mm(IXvl), 380 mm (IXx9) .

Fig. 14 . Incomplete skeleton of Dryosaurus lettow-vorbecki from Tendaguru Site Jg (WJ) . Original field sketch from 1. Reckor H . Reck

46

O n t o g e n e t ic

data :

According

to

Sander(1999, in press), humerus IX94 is from a juvenileindividual of Barosaurus africanus, while femurIX1 is from an adult individual of Brachiosaurusbrancai .

Comments : Excavations at Site IX com-menced on August 17, 1909. About 150 indivi-dual bones of dinosaurs were collected from SiteIX which was located between the Trigoniasmeei-Bed and the Upper Saurian Bed (GTEfield catalogue : 108), although only some ofthese were mapped (Fig. 13) . Partial skeletonswere not found, and the majority of bones weredisarticulated and scattered over the excavationarea, though some of them may have still beenassociated . Janensch (1914c : 44) reported thatmost of the bones were embedded horizontallyand concentrated in a distinct bone-bearinglayer . He also noted the dominance of bonesfrom the appendicular skeleton (Janensch 1914c:44) . A closer study of the 130 identified bonesfrom Site IX reveals that 85 (= 65,4%) are legelements, including humeri, radii, ulnae, femora,tibiae, and fibulae (GTE field catalogue :108-112) . Other elements (e.g ., vertebrae, pha-


langes, ribs) are under-represented or completelymissing (e.g ., skull elements, lower jaws) .

Some bones are well preserved, others in poorcondition . The epiphyses of several limb bones areconsiderably worn due to rolling and abrasion bythe embedding sands (Janensch 1914a, c), andthere is no doubt that the skeletal remains werecurrent modified. Moreover, the scattering and thestate of preservation of the skeletal elements sug-gest a long period of time during which the boneswere transported by waves and/or currents fromthe carcasses to the place of permanent burial .

Only a few catalogued limb bones have beenidentified taxonomically. Most of them appar-ently belong to sauropod dinosaurs. A minimumof four individuals of Barosaurus africanus areindicated by four right humeri . Skeletal remainsof Brachiosaurus brancai, Janenschia robusta,and Kentrurosaurus aethiopicus are apparentlyless common than those of Barosaurus africanus,but precise specimen counts are not available .Belemnites have frequently been found at Site IX(GTE field catalogue : 108) .

Fig. 15 . Tendaguru Site S, the type locality of Brachiosaurus brancai . Photograph : W. Janensch

Mitt . Mus. Nat.kd. Berl ., Gcowiss. Reihe 2 (1999)

3.3 Sites in the Middle Saurian Bed

3.3.1 Tendaguru Site Jg (= WJ)

L o c a t i o n : Tendaguru Site J O (= WJ) is locatedapproximately 2.3 km north-north-west of Ten-daguru Hill .S t r a t i o r a p h y :

Middle Saurian BedTax on : Diyosenn-us lcttotv-vorhecki

Figure : 14R c f e r c n c c : GTE sketchbookAssemblage dataNISP: UnknownMNI : 1Number of taxa : 1Designation: Single-individual, single-taxonassemblage

Quarry dataSize of accumulation : UnknownDensity : UnknownModification dataMcaSttrcmcnts : Not available

Tendaguru, Site SMiddle Saurian BedBrachiosaurus brancai


shoulder girdle

pelvic girdle

fore limb

hind limb

indet.

approx . Ilm

cd

DisartiCttlatIOn Stage : '? (complete or in-complete skeleton)

Ontooenetic data : Unknown

Comments : Excavations at Site J o (WJ ) com-menced in 191() and continued through the years1911 and 1912 (Branca 1914 ; Janensch 1914c ;GTE field cataloouc : 73) . Quarry Jo (WJ) is oneof the richest dinosaur-bearing sites in the Sur-roundings of Tendaguru Hill . It produced thou-sands of hones of the small ornithopod dinosaurDl-yusnurus blow-ivrrheeki (Janensch 195(lc,1955, 196Ib) . Jttdgino from matrix kept in theBerlin Tendaluru collection, isolated vertebraeand limp hones of Drvosaiwus lcftow-volheckiare the most common elements found in theMiddle Saurian Bed at Site Jg (WJ) . Skull frag-ments and teeth are fairly rare . Site Jg (WJ) hasnot only produced isolated bones but also partialand fairly complete skeletons (Flo . 14) . Strings ofVertebrae with closely associated chevrons alsooccur. The dryosaurs from Site Jg (WJ) arethought to have died close to the places wherethey were buried (Janensch 1914a, c) .

S14

S35-39mc

N

S43cv

47

Fig. I(, . Incomplete skeletons of Brachiosurrs (run-11i from Tendaouru Site S. Anatomical identifications of skeleton SII arcunderlined . Elements of skeleton SI could not he indentificd with certainty. The mc~isured length of the hun1CIuS S9(213.0011) was used 10 cnICUIAC the scale for the quarry mal, . Redrawn niter a field sketch from W. Jancnsch

A drawing of in in-situ skeleton of Drvo-

srn1r.u.c l(gtonv-vorbccki, preserved in the GTEsketchbook shows, a fairly complete skeleton re-covered in 1912 (Fig . 14) . An articulated skullseems to he present, and most of the vertebralco1L1n111 appears to be preserved in natural ar-ticulation, as do the dorsal ribs. The caudal ver-tebral column is partly preserved . Frotn the forelimbs, nothing but the proximal section of a righthunicrLIS 01' ulna was documented . The left fe-lnur is spread slightly away from the trunk, andthe proximal sections of both femora are appar-ently closely associated with the pelvic region .The lower Icgs were either not found or notcloctunented in the field sketch . Unfortunately,the location of the documented skeleton ofDrvosaürits lclfoiv-vorbccki is not known .

3.3.2 Tendaguru Site S

Location : Site S is located approximately1 .0 km SOLIthsouthwest of Tendaguru Hill

Tendaguru, Site Y

Middle Saurian BedBrachiosaurus brancai

approx . 1 m

Y1

is cr

skull


shoulder girdle

fore limb

hind limb

N

Y27Ca

I Icinrich . W.-D. . Taphonomy ol I)inoseurs from Icndaguru

S t r a t I,,, r a p h _Y :

Middle Saurian Bed

Ta x o n : Brachiosaurus brcuwai

Fi g e I 'C s :15,16 (colour figure on p . 47) and 18

Reference : Janensch (1914a, c ; 1929a,1961a), GTE field catalogue : 25-2t;

Assemblage dataNISI' : Not precisely known

MNI : 2Number of taxi : 1

Designation :assenlbla-e

Quarry dataSize of accurnDensity : '?

Modification dataDlsarticulatlon(skeleton S 1)Measurements : Skeleton S11, length of se-lected elements of the appc:nclICLIlar skeleton

stage : E

1950a, h :

Multi-individual,

Sing lC -1aXOI1

elation :

(skeleton S11), L

Y25

as .d

Fig. 17 . Brochinsaurus brancai from Tendaguru Site Y. The incomplete skeleton consists of disarticulatcd elements from the

anterior 11,11. 1 of the holy. The measured length of the scapula Y8 (154 .11 cm) was used to calculate the scale for Uic quarn

map. Redrawn after a field sketch from W. Janensch

Mitt . Mus. Nat.kd. Berl., Geowiss. Reihe 2 (1999)

(Janensch 1961a) : left sternal plate : 1100 mm;right

coracoid:

ca.

840 mm;

right

humerus :2130 mm; right radius :

1240 mm; right Ulna :1300 mm;

me

I

dext. :

595 mm;

me

II

dext:635 mm; me III dext : 595 mm; me IV dext . :570 mm; me V dext . : 490 mm; right femur : ca .1960 mm, right fibula : 1190 mm.

O n t o g e n e tic

data :

Judging

from

solitarylimb bones that are at least one tenth largerthan those of skeleton SII, it could be concludedthat the SII specimen was not yet fully grown(Janensch 1938) .

Comments : The dig commenced on October11, 1909 and Tendaguru Site S was worked bythe German Tendaguru until 1911 (Janensch1914a) . Site S is the type locality for Brachio-saurus brancai and remains of two individualswere found . Skeleton SI, the type specimen of

49

Brachiosaurus brancai, is a partial skeleton, com-prising the skull, six cervical vertebrae and (?)dorsals (Janensch 1929a: 8) . Specimen SII whichis larger than SI, is an incomplete skeleton ofBrachiosaurus brancai (Janensch 1914a, c ; 1929a ;1950a, b ; 1961a), including skull bones, a nearlycomplete presacral vertebral column (11 cervicaland 11 dorsal vertebrae), cervical and dorsalribs, most of the anterior appendicular skeleton(left scapula, coracoids, sternals, right anteriorlimb, left humerus, ulna, and radius), parts of thepelvic girdle (e.g . both pubes bones), and incom-pletely preserved hind limbs (e.g ., imperfect rightfemur, right tibia and fibula, and foot bones) .The sacrum and the caudal series were alreadymissing when the specimen was discovered, mostlikely due to erosion (Janensch 1914c) .

Janensch (1914a, c) reported that the humerusand a tibia of the SII individual were found in

Fig. 18 . Brachiosaurus brancai from Tendaguru Site S. The articulated series of vertebrae from skeleton S11 includes anteriordorsals and cervical vertebrae . Original field sketch from W. Janensch . The skeletal elements in the figure are tentativelyidentified as follows : S73 - left scapula; S76 - dorsal vertebra 2; S77 - dorsal vertebra 1; S78 - cervical vertebra 13 ;S82-S87 (? S80, ? S88) - cervical vertebra 12 ; S89-S91 - cervical ribs; S92-S93 - dorsal ribs; S96-S97 (? 94) - cervicalvertebra 11

50

an upright position, suggesting that the animalmired in soft mud. The proximal section of thishumerus shows distinct traces of abrasion, indi-cating that the skeleton was affected by movingwater prior to burial (Janensch 1914c) . Thebones of the right fore limb were found in natur-al order, but not articulated, and several bones(e.g ., coracoids, right humerus, radius ulna) wereembedded in a horizontal position (Fig . 16). Thedistal part of the left anterior limb was absentupon discovery, possibly due to post-depositionalerosion (Janensch 1914a, c) .

Sections of the reasonably complete presacralvertebral column of specimen SII are roughlydocumented in two separate field sketches,partly in pencil, partly in ink. One of these twofield sketches is reproduced in the present ac-count (Fig. 18) . According to Janensch (1950a :33), the anterior section of the vertebral column,including presacral vertebrae 3 to 15, were foundin natural order. The three anteriormost presa-cral vertebrae (3-5) are well preserved and donot show signs of abrasion, while the remainingpresacral vertebrae (6-15) are articulated anddisplay close, bone-to-bone contact, except forpresacral vertebra 8 which was slightly dislo-cated. The dorsal parts of presacral vertebrae 9to 15 are missing, possibly due to attrition by thefine sands in which the remains were embedded(Janensch 1950a) . The string of presacral verteb-rae (3 to 15) was recovered from a massive limysandstone bed, unike presacral vertebrae 17 to24 which were found together with other bonesfrom skeleton SII in marls below the limy sand-stone bed (Janensch 1950a: 33) . The majority ofthese posterior presacral vertebrae were disarti-culated, except for vertebrae 23 and 24 whichwere found in bone to bone contact (Janensch1950a: 33).

The sediments exposed at the place of burialshow that low-energy dinosaur-bearing marlspassed into higher-energy sandy deposits . Theshape of most bones recovered from Site S hasnot been substantially changed by taphonomic ordiagenetic processes. The taphonomic evidenceobtained from skeleton SII suggests carcass ma-ceration and decay of a mired individual of Bra-chiosaurus brancai prior to burial (Janensch1914c) .

3.3.3 Tendaguru Site Y

Location :

Site Y is located about 3.1 kmnorth of Tendaguru Hill .


S t r a t i g r aphy : Middle Saurian BedTaxon : Brachiosaurus brancai

Fi g u r e : 17 (colour figure on p. 48)Refe r en c e : Janensch (1929a), GTE field cata-logue: 37Assemblage data

NISP : 23MNI: 1Number of taxa : 1 (2)

Designation : Single-individual, single-taxonassemblage

Quarry data

Size of accumulation : approximately21 .00 m2Density : approximately 1.1

Modification dataMeasurements : length of a left scapula:1540 cm; length of a right astragalus : 160 mm(Janensch 1961a) .

Disarticulation stage : F

Ontogenetic data : The remains of Brachio-saurus brancai recovered from Site Y are from amedium-sized individual (Janensch 1950a: 33).

Comments : The excavations at Site Y com-menced in 1910 (GTE field catalogue : 37) . Thebone accumulation from Site Y represents an in-complete skeleton, dominated by bones from theanterior body half. From the posterior skeletononly a fibula was found. Site Y has yielded amonospecific bone accumulation assigned to Bra-chiosaurus brancai . All bones documented aredisarticulated . Janensch (1929a : 8) noted a brain-case, eight cervicals, ribs, scapula, coracoid, hu-merus, and a dorsal vertebra . The skeleton is ap-parently current-modified, with a scapula restingnext to a humerus and dorsal ribs . The cervicalvertebrae are approximately in anatomical order(Janensch 1929a: 8), and the braincase lies closeto them . The majority of the long bones wereembedded with their long axes oriented west toeast (e .g, right scapula, left humerus, dorsal ribs).Only a few elongated bones (e.g ., left scapula,left fibula) are approximately oriented north tosouth. Some bones were embedded on top ofone another (e.g., right scapula, left humerus) .The taxonomic allocation of a small elongatecaudal vertebra that was found south of the leftscapula is far from clear.

Mitt . Mus. Nat.kd . Berl., Geowiss . Reihe 2 (1999)

3.3.4 Tendaguru Site m

Loc a t i on : This site was situated close to Kin-dope, about 3.2 km north of Tendaguru Hill .Stratigraphy : Middle Saurian BedTax on : Dicraeosaurus hansemanni

Figure : 19Reference : Janensch (1914a ; 1929a, b, 1936,1961a), GTE field catalogueAssemblage dataNISP: UnknownMNI: 1Number o f t a xa :

One-individual assemblageDesignation : Single individual assemblage,single-taxon assemblageQuarry dataSize of accumulation : UnknownDensity : UnknownModification dataMeasurements : The length of the preservedpelvic elements and limb bones is as follows(Janensch 1961a) : right ilium: 840 mm; rightpubis: 795 mm; right ischium: 750 mm; right fe-

mur (m5) : 1220 mm; left tibia (m1) : 760 mm; leftfibula (m3) : 800 mm, left astragalus : 165 mm

Disarticulation stage : EOntogenetic data : Unknown

Comments : One of the most important dis-coveries of the 1910 field season was the typespecimen of Dicraeosaurus hansemanni, recov-ered from the Middle Saurian Bed at Site mnear Kindope (Janensch 1914a) . Unfortunately, aquarry plan is not available. The more importanttaphonomic data published by Janensch (1929a,b) are briefly summarized below.The skeleton is incomplete, but remarkably

well preserved . Much was found in natural ar-ticulation . The skull, proatlas, and axis have notbeen not recorded . A series of vertebrae rangingfrom cervical 2 to caudal 19 lay in natural order.Only the anterior end of the tail was recovered;the posterior tail section beyond caudal 19 waslost prior to discovery. The preserved part of thetail is slightly recurved (Janensch 1929b: 41).Ilium, pubis, parts of the ischium, and the femurof the right body side were discovered in situ.The left femur was found close to the sacrum .The sternal plates, coracoids, scapulae, and ante-

Fig. 19 . Type specimen of Dicraeosaurus hansemanni from Tendaguru Site m . Photograph : W. Janensch

5 1

5 2

rior limbs have not been recorded. Except forthe femora, the hind limbs are represented bythe articulated left lower leg (tibia, fibula, andastragalus), found just above the posterior sec-tion of the neck . Feet are not recorded . Thewell-preserved right dorsal ribs are still articu-lated with the dorsals. The left dorsal ribs areincompletely preserved and scattered, but closeto the skeleton . Some vertebrae suffered consid-erable damage : nearly all the left diapophysesare missing in the section ranging from dorsal 13to caudal 1, and of the posterior dorsals, the leftparapophyses, the left neurapophyses and thecentra are water worn. Three caudals werefound at some distance from the skeleton. Theirassociation with the preserved caudal series ofskeleton m is uncertain .

According to Janensch (1929b: 41), the car-cass floated quite a long distance in quiet waterbefore it came to rest on its right side . After-wards, the anterior section of the neck wasflexed ventrally, rotated and became separatedfrom the posterior cervicals . Subsequently, thissection came to rest in an oblique position adja-cent to the carcass prior to burial . Afterwards,exposed parts of the skeleton, including the leftdorsal and sacral diapophyses, the left rib cage,the left pelvic girdle, and perhaps the left hindlimb were abraded, truncated and dispersedclose to the skeleton . The skull, most of the ap-pendicular skeleton (e.g ., sternal plates, anteriorlimbs, right lower leg, both feet, and ? left hindlimb) and the posterior section of the tail wereprobably lost during post-mortem transport fromthe place of death to the site of burial .

4. Discussion: taphonomic patterns in Tendagurudinosaurs

4.1 Occurrence of bones

Most field sketches preserved in the achives ofthe IPHB document bone accumulations for thesauropod dinosaurs Brachiosaurus, Barosaurus,Dicraeosaurus, and Janenschia . Quarry plansdocumenting bone assemblages for ornithischians(Dryosaurus, Kentrosaurus) are rare . Nothingcan be said about the original position ofskeletal remains of theropods (e.g ., Elaphro-saurus) due to the lack of field sketches forthese taxa .

The documented skeletal remains recoveredfrom quarries in the vicinity of Tendaguru Hillrepresent a broad spectum of disarticulation and


disassociation . They belong to a sequence of dis-articulation stages recognized in the TendaguruBeds (Tab. 2), including (1) incomplete skeletonswith bones partially articulated (e.g ., Dicraeo-saurus hansemanni, Site m; Brachiosaurus bran-cai, Site S), (2) complete or incomplete partialskeletons (e.g . Janenschia robusta, Site P), (3)assemblages of predominantly disarticulatedskeletal elements concentrated in bone fields(e.g. Barosaurus africanus, Site C; Kentrosaurusaethiopicus, Site X), and (4) solitary bones . Vir-tually complete dinosaur skeletons have notbeen found in the Tendaguru Beds so far (Hen-nig 1912b). Partial skeletons show a similarbroad spectrum of disarticulation stages as in-complete skeletons (Table 2), but both may re-flect different taphonomic histories . A partialskeleton, for example, may not only result fromgradual "in-situ" carcass decay at the site ofdeath or final deposition but also from earlypost-mortem truncation and pre-burial transportof carcass parts from the site of death or fromthe original location of carcass deposition.Mono- and multispecific assemblages of dino-

saurian skeletal elements have been recognized.Monospecific assemblages, for instance, aredocumented for Barosaurus africanus (e.g ., SiteC), Brachiosaurus brancai (e.g ., Site S), Dicraeo-saurus hansemanni (Site m), and Dryosaurus let-tow-vorbecki (Site Jg = WJ) . These assemblagesare usually represented by incomplete or partialskeletons, and they are mainly known from theMiddle Saurian Bed (e.g ., Sites m and S) and theUpper Saurian Bed (Sites C, D, N) . Multispecificassemblages have been documented, for in-stance, from Site ab (Barosaurus africanus, Bra-chiosaurus brancai, Dicraeosaurus sattleri), andSites IX and X (Barosaurus africanus, Brachio-saurus brancai, Janenschia robusta, Kentrosaurusaethiopicus) . Multispecific assemblages are typi-cal of the Transitional Sands above the Trigoniasmeei Bed (Site IX), but have also been docu-mented in the Upper Saurian Bed (e.g ., Sites ab,X) .

Calculations of the MIN'S reveals that single-individual assemblages (e.g ., Barosaurus africa-nus, Sites C; Brachiosaurus brancai, Sites D, M;Dicraeosaurus sattleri, Site O) and multi-indivi-dual assemblages (e.g., Brachiosaurus brancai,Site S ; Janenschia robusta, Site P) both occur.The Middle and Upper Saurian Beds haveyielded both single-individual assemblages (e.g .,Dicraeosaurus hansemanni, Middle Saurian Bed,Site m; Barosaurus africanus, Upper SaurianBed, Site k) and multi-individual assemblages

Mitt . Mus. Nat.kd. Berl ., Geowiss . Reihe 2 (1999)

(Brachiosaurus brancai, Middle Saurian Bed,Site S ; Janenschia robusta, Upper Saurian Bed,Site P). The documented bone assemblages fromthe Transitional Sands above the Trigonia smeeiBed are multi-individual in nature (e.g . Baro-saurus africanus, Site H; Barosaurus africanus,Brachiosaurus brancai, Janenschia robusta, Ken-trosaurus aethiopicus, Site IX) . Single-individualassemblages are more frequently documentedthan multi-individual assemblages .

The documented multi-individual assemblagesconsisting of incomplete or partial skeletons ofsauropods are always monospecific . Examples in-clude the skeletons SI and SII of Brachiosaurusbrancai from the Middle Saurian Bed Site S, andthe partial skeletons I-IV of Janenschia robustafrom Upper Saurian Bed Site P In addition, Ja-nensch (1929a: 9) reported two individuals of Di-craeosaurus hansemanni from the Middle Saur-ian Bed at Site dd.

The Middle Saurian Bed has yielded the bestpreserved skeletons from Tendaguru, some ofwhich, notably those of Brachiosaurus brancai(Site S) and Dicraeosaurus hansemanni (Site m),are on display in the Museum of Natural Historyof Humboldt University, Berlin . An unpublishedfield sketch of Site Jg (WJ) indicates that reason-ably complete skeletons of ornithischians (Dryo-saurus lettow-vorbecki) were also found . More-over, the Middle Saurian Bed has producedenormous concentrations of disarticulated skele-tal remains of smaller ornithischians such asKentrosaurus aethiopicus from Sites St, He andX (Janensch 1914c, Hennig 1915, 1916, 1925)and Dryosaurus lettow-vorbecki from Site Jg(Janensch 1914c, 1955, 1961b ; Galton 1981) .

Reported recoveries of theropod dinosaurs in-clude several disarticulated and scattered skele-tal elements from Tendaguru Site dd that are ap-parently derived from a single skeleton ofElaphrosaurus bambergi (Janensch 1925b : 7 ; Gal-ton 1982). Included in this collection, but undo-cumented in field sketches, are 17 presacral ver-tebrae, 18 caudal vertebrae, a chevron bone, twoimperfect ribs, both scapulae and coracoidea,right humerus, two metacarpals, sacrum, leftpubis, two ilia and ischia, left femur, tibia, fibula,astragalus, and two phalanges (Janensch 1925b :7, 1929c : 280) . In addition, several localities ofthe Middle Saurian Bed, such as the collectingsites St and H (Janensch 1920, 1925b), producedisolated limb bones and teeth .

Samples from the Middle Saurian Bed housedin the Berlin Tendaguru collection confirm theimpression that skeletal remains of Dryosaurus

5 3

lettow-vorbecki are extraordinarily abundant atSite Jg (WJ) . Here, the matrix is dominated byvertebrae and limb bones, some partly articu-lated, and some partly disarticulated . Isolatedbones are also very common. By contrast, skullbones and teeth are very rare, as are small handand foot bones. It should be mentioned thatbones from Dryosaurus lettow-vorbecki (Site Jg)and sauropod caudal vertebrae (Site dd) en-crusted with calcite have repeatedly been found .These calcite rinds, measuring 2 mm to 5 mm inthickness, are infillings of post burial shrinkingcracks . Janensch (1925b : 7) also reported calcar-eous rinds that covered bones of Elaphrosaurusbambergi recovered from Tendaguru Site dd.

Skull material from sauropods is extremelyrare, although the Middle Saurian Bed at Site thas yielded a disarticulated complete skull ofBrachiosaurus brancai with 56 teeth in its jaws,and Tendaguru site dd, for instance, producedbraincases of Dicraeosaurus hansemanni andBarosaurus africanus (Janensch 1935).

Janensch (1914c) reported water worn bonesof Brachiosaurus brancai (skeleton SII) fromSite S as well from Dicraeosaurus hansemannifrom Site m. The material from the Middle Sau-rian Bed is evidently current modified . Most ske-letal elements were apparently carried awayfrom carcasses, skeletons or partial skeletons,possibly by currents or wave action .

Bone accumulations from the TransitionalSands above the Trigonia smeei Bed are domi-nated by disarticulated sauropod dinosaur limbbones . Much of the material is in poor condition,confirming Janensch's (1914a, c) interpretationthat there was much destruction, transportation,and sorting of bones. Abraded epiphyses arethought to be evidence of water action (Janensch1914a, c ; 1961a) . Most of the bones rest horizon-tally next to one another, possibly due to waveaction or wave-induced currents (Janensch1914a, c) . There is no doubt, that this materialunderwent sorting through a broad spectrum ofpost-mortem processes prior to burial .

Russell et al . (1980: 172) reported the domi-nance of Barosaurus africanus in the Transi-tionals Sands above and below the Trigoniasmeei Bed and noted "a significantly highernumber of elements from the left side" of thebarosaur body. From Site Ki, Transitional Sandsbelow the Trigonia smeei Bed, Hennig (1925:247) mentioned a complete but somewhat dis-articulated left lower leg of Kentrosaurus aethio-picus, including the tibia, fibula, and bones ofthe pes.

54

Most quarry maps from the Upper SaurianBed primarily document disarticulated skeletalmaterials from sauropods (e.g ., Brachiosaurusbrancai, Sites D, N, cc ; Dicraeosaurus sattleri,Site O) . Incomplete skeletons (e.g ., Barosaurusafricanus, Site k) and partial skeletons (e.g.,Janenschia robusta, Site P) also occur, withboth partly articulated, and partly disarticulatedbones. As mentioned above, most of the boneaccumulations appear to represent single-indivi-dual assemblages, although, multi-individual as-semblages have also been recognized . Janensch(1922, 1961a) reported an articulated manusskeleton of Janenschia robusta from Site Nr. 5and Brachiosaurus brancai from Site R, as wellas two articulated pes skeletons of Barosaurusafricanus from Sites XIII and 28, suggestingthat sauropod dinosaurs became mired in soft

saurus africanus from Site C) and accumula-tions of disarticulated bones (e.g . Barosaurusafricanus, Dicraeosaurus sattleri, and ? Brachio-saurus brancai from Site ab) show that the car-casses underwent considerable pre-burial decay,as formerly suggested by Janensch (1914c,1961a) and Hennig (1925) .

The Upper Saurian Bed produced mass accu-mulations of skeletal elements of the or-nithischian Kentrosaurus aethiopicus (Hennig1925) . A bone field uncovered at Site X, for in-stance, consists mainly of isolated and jumbledelements from the manus and pes (Janensch1914c, Hennig 1925) reflecting a secondary con-centration of dinosaurian material (Fig . 10) . Hen-nig (1915 : 244; 1925 : 247) reported an incom-plete skeleton of Kentrosaurus aethiopicus fromSite bb which is close to the top of the UpperSaurian Bed . The skeleton includes a left scapulaand coracoid, imperfect left pubis, fragmentaryischium, both femora, a caudal vertebra, andfragments of ribs and vertebrae (Hennig 1925:247). Field sketches are not available, the stageof disarticulation is unknown.

Isolated bones and teeth indicate the presenceof theropod dinosaurs in the Upper Saurian Bed,among them Ceratosaurus, Allosaurus andElaphrosaurus (Janensch 1925b, 1929c) . Bonesare very rare, and teeth are more frequently pre-served . According to Janensch (1925b: 61), a leftquadratum, left fibula, and three imperfect cau-dal vertebrae recovered from Site Mw are as-signable to one individual of Ceratosaurus (?)roechlingi .


4.2 Bone representation

The great majority of sauropod bone acculuma-tions either documented in the field sketches orlisted by Janensch (1929a, GTE field catalogue)are incomplete skeletons or partial skeletons .The field sketches examined in this account anddata given by Janensch (1929a) show that someskeletal elements are more frequently recordedthan others . Skull remains for instance, are extre-mely rare, whereas appendicular skeletal ele-ments usually dominate the bone assemblagesexamined . This raises the question, to what ex-tent certain skeletal elements better survived thetaphonomical filters than others. This question isdifficult to answer, however, because no first-hand data are available . According to the re-cords of sites, not all bones found have beendocumented in field sketches. In addition, poorlypreserved bones were not collected at some sites(Janensch, GTE field catalogue) . In order tominimize the affects caused by collector prefer-ences, individual bone representation diagramswere replaced in the present account by a body-part representation diagram . This diagram(Fig. 20) reflects only "summarily" the presenceof body parts recognized in skeletons, but doesnot consider the abundance of individual skeletalelements within the distinguished body parts.

The analysis of incomplete skeletons and par-tial skeletons of Tendaguru sauropods fromabout 30 sites, partly documented in the fieldsketches, partly listed by Janensch (1929a) led tothe following conclusions (Fig . 20) : (1) As ex-pected, the paired elements of the appendicularskeleton dominate the fossil record of Tendagurusauropods . (2) In Tendaguru diplodocids (Baro-saurus africanus, Dicraeosaurus hansemanni andDicraeosaurus sattleri) that have longer hindlimbs than fore limbs, the rear limbs and pelvicelements are dominant . (3) In Tendaguru bra-chiosaurids (Brachiosaurus) that have longerfore limbs than hind limbs, the anterior appendi-cular skeleton with its huge fore limbs andshoulder girdle dominate the fossil record .

Body-part representation is strongly influ-enced by the degree of carcass decay that wasreached before burial . Therefore, the tapho-nomic pattern found is not only an "overprint-ing" of distinctive skeletal traits of the sauropodskeletons. It also suggests a considerable amountof pre-burial sorting of dinosaur bones in theTendaguru Beds . The presence of tiny and verydelicate pterosaurian and mammalian remains inthe Middle and Upper Saurian Bed at Tenda-

mud (Janensch 1914c : 249) . Fore- and hind-limb skeletons (e.g . Janenschia robusta fromSite P) and strings of vertebrae (e.g . Baro-

Mitt . Mus . Nat.kd . Berl., Geowiss. Reihe 2 (1999)

02 Barosaurus, n-11

=Dicraeosaurus,n- 8

guru (Reck 1931, Galton 1980, Unwin & Hein-rich 1999, Dietrich 1927, Heinrich 1998, 1999) in-dicates, however, that this taphonomic patterncannot be due to post-burial destruction of smal-ler sauropod skeletal elements.

4.3 Burial position

The burial position of carcasses of dinosaurs isdifficult to assess because of the relatively smallnumber of reasonably complete skeletons recov-ered from the Tendaguru Beds . Moreover, theskeletons documented in the field sketches dis-play different burial positions.

The specimen of the dryosaurid Dryosauruslettow-vorbecki recovered from the Middle Sau-rian Bed at Site Jg (WJ) was buried on its belly,with the left femur spread away from the trunk,and a cervical vertebral column that was notstraight or only slightly recurved (Fig . 13) . Theskull and neck were apparently still articulated .

Russell et al . (1980: 172) mentioned that Ten-daguru barosaurs "preferred to die laying ontheir left side" . The upright position of both lefthumerus and left tibia of skeleton SII of the bra-chiosaurid Brachiosaurus brancai (Fig . 16) sug-gests that the animal became mired in soft mudand died laying on its belly. In-situ decomposi-tion of soft tissues is indicated by the series ofarticulated cervical vertebrae as well as by paral-lel oriented dorsal ribs resting closely on eachother (Fig . 16) . Except for the cervical vertebrae,the majority of skeletal elements are concen-

Fig.20 . Frequency distributionof main body regions in sauro-pod skeletons from Tendaguru .For explanation see text . 1 -skull and neck, 2 - shouldergirdle and fore limbs, 3 - sacraland dorsal vertebrae, dorsalribs, 4 - pelvic elements and

body region

posterior limbs, 5 - tail5

trated close to these two limb bones, indicatingthat dispersal of bones was minimal . The placeof death must therefore have been the place offinal deposition.

4.4 Individual size and age class distributions

5 5

The following account of the size distribution ofTendaguru sauropods (Brachiosaurus brancai,Barosaurus africanus) is based upon the humeriand femora, and that of the ornithischians(Dryosaurus lettow-vorbecki) upon the femora,because these bones are better represented inthe collection than other limb bones of thesetaxa . Metric data were taken from Janensch(1961a and unpublished) and from Sander(1999, in press), augmented by my own meas-urements.The number of the measured specimens is

rather small, making quantitative assessmentsdifficult . Therefore, sauropod specimens fromdifferent dinosaur-bearing horizons were put to-gether to form time-averaged populations, asexplained in the diagrams (Figs 21 and 22) .Both humeri and femora of Brachiosaurus bran-cai were assigned to 30.0 cm intervals, those ofBarosaurus africanus to 14.0 cm intervals. Life-age data of sauropod individuals are based onhistological examinations by Sander (1999, inpress) .A data set published by Janensch (1961b) as

well as my own measurements can be used toestablish a femur-size distribution-diagram for

56

Dryosaurus lettow-vorbecki . The femora of Dryo-saurus lettow-vorbecki are assigned to 5.0 cm in-tervalls . The diagram (Fig . 23) is based on a col-lection of femora recovered from the MiddleSaurian Bed at Tendaguru Site Jg (WJ) Unfortu-

11

Fig . 21 . Size-class distribution of humeri and femora of Bra-chiosaurus brancai from Tendaguru . Data grouped togetheras follows: A - subset of humeri from the Middle SaurianBed (n = 3), Upper Saurian Bed (n = 12), and TransitionalSands overlaying the Trigonia smeei Bed (n = 2), B - subsetof humeri from the Middle Saurian Bed (n = 3) and UpperSaurian Bed (n =12), C - subset of femora from the MiddleSaurian Bed (n = 11), Upper Saurian Bed (n = 3) and Transi-tional Sands (n = 3), D - subset of femora from the MiddleSaurian Bed (n = 11) . 1 - juvenile individuals ; 2 - adult in-dividuals ; 3 - old adult individuals. The age determination isbased upon life history studies by Sander (1999, in press) .Metric data from Janensch (1961a), Sander (1999, in press)and own measurements

nately, histological data that could calibrate thesize classes with ontogenetic data are not avail-able . Paleohistological studies on femora ofDryosaurus lettow-vorbecki from Tendaguru indi-cate rapid and uninterrupted growth (Chinsamy1995).

Fig . 21 shows the humerus- and femur-size dis-tribution of Brachiosaurus brancai . The histo-grams reveals that the populations are domi-nated by humeri and femora ranging in lengthfrom 144 to 174 cm. Based on life history studiesby Sander (1999, in press), humeri and femoraof these dimensions are probably assignable to

35

30-

25-

20-

15-

10-

5-

0

0/0 -40-

35-

30-

25-

20-

15-

10-

5-0

40-

35-

30-

25-

20-

15-

10-

5 -

0-

0/0


10 24 38 52 66

n=23

A

n=17

B

n=15

Fig. 22 . Size-class distribution of humeri and femora of Baro-saurus africanus from Tendaguru . Data grouped together asfollows : A - subset of humeri from the Upper Saurian Bed(n = 5), and Transitional Sands overlaying the Trigonia smeeiBed (n = 18), B - humeri from the Transitional Sands over-laying the Trigonia smeei Bed (n = 17), C - subset of femorafrom the Upper Saurian Bed (n = 7), and Transitional Sandsoverlaying the Trigonia smeei Bed (n = 8) . 1 - juvenile indi-viduals; 2 - adult individuals ; 3 - old adult individuals . Theage determination is based on life history studies by Sander(1999, in press) . Data from Janensch (1961a), Sander (1999,in press) and my own measurements

Mitt . Mus . Nat.kd . Berl., Geowiss. Reihe 2 (1999)

adult individuals . Adult individuals have alsobeen found in the adjacent humerus size classes(114-143 cm, 174-203 cm), suggesting thatabout 60% of the measured specimens of Bra-chioaurus brancai may belong to adult indivi-duals. Old adult individuals are most likely re-presented in the highest Humeri-size class(204-233 cm).Humeri of juvenile individuals of Brachio-

saurus brancai are distinctly underrepresented,and neonates have not been found so far (Figs .21A, 21B) . The "first" appearance of old adultindividuals in femur-size class 174-203 cm andthe "last" appearance of adults in femur-sizeclass 204-233 cm (Figs . 21C, 21D) indicates thatthe age groups overlap to significant degree, andinclude individuals of different size. A similar in-terval of overlap is to be expected between juve-nile and adult specimens.

Consideration of the size frequency histo-grams for the humerus and femur of Barosaurusafricanus reveals that adult individuals appar-ently prevail (Fig . 22), as in Brachiosaurus bran-cai. Diagram 22A, which approaches a unimodalsize-class distribution is difficult to assess, since

25-

20J

size classes/cm

Fig . 23 . Size-class distribution for the femur of Dryosauruslettow-vorbecki from the Middle Saurian Bed at TendaguruSite Jg (WJ) . Metric data from Janensch (1961b : 164) andmeasurements obtained from recently recovered femora . To-tal number of femora : 28

the sample is based on time-averaged bone as-semblages .

In this connection, it is worth mentioning thathistological data obtained from three Humeri ofDicraeosaurus sattleri from the Upper SaurianBed (Sites ab and O) also suggest that adult in-dividuals predominate in the bone assemblages(Sander 1999, in press), as is the case for Bra-chiosaurus brancai and Barosaurus africanus. Afemur of Janenschia robusta collected from theUpper Saurian Bed at Tendaguru Site 22 is attri-butable to an old adult individual (Sander 1999,in press) .

Dryosaurus lettow-vorbecki (Fig . 23) displays abimodal femur-size-distribution, with two peaks,one at the beginning and one close to end of theplot, suggesting either that juvenile and adult in-dividuals prevail or that sexual dimorphism oc-curs . Sexual dimorphism cannot be excludedwith certainty, but it seems more likely that thebimodality is the result of individual size varia-tion . If this is correct, the pattern of size-distribu-tion may allow conclusions to be drawn regard-ing on the origin of the mass accumulations ofDryosa.urus lettow-vorbecki, discussed below.

4.5 Attritional or mass mortality?

57

The origin of the extraordinary concentrations ofdinosaur bones in the Tendaguru Beds has beenthe of subject of speculations ever since the firstdiscovery of dinosaurs in the Tendaguru area .The interpretation has to consider attributes ofthe recovered bone assemblages, paleoecologicalaspects, and pecularities of the depositional en-vironment that influenced the involved taphono-mical processes.

According to Janensch (1914a, c; 1961a), thedinosaur-bearing deposits were possibly laiddown in a lagoonal environment separated fromthe sea by reefs . The sedimentation was thoughtto have been affected by tides, episodic storms(hurricanes), wind waves, earthquakes, andchanges in the regional climate (Janensch 1914c) .The invertebrate fossil record suggests a brackishto limnic depositional environment for the Mid-dle and Upper Saurian Bed (Dietrich 1914; Hen-nig 1914b, c ; Schudack 1999 ; Schudack et al.1999), and records of Asmussia (= Estheria)from the Upper Saurian Bed may indicate peri-ods of drought (Janensch 1933, Russell et al.1980) .

As mentioned above, Janensch (1914c) sug-gested that the dinosaur bone assemblages from

58

Tendaguru resulted mainly from abrupt cata-strophic killing events. Unfortunately, the preciserelationships of the quarried dinosaur bones tothe enclosing sediments are far from clear, dueto the loss of many records . In addition, the pre-cise stratigraphic position of the dinosaur-bearingsites within the Lower, Middle, and Upper Sau-rian Bed is unknown . This is especially regretta-ble, since the depositional environment of theTendaguru Beds underwent remarkable changesduring the Late Jurassic and Early Cretaceous .The dinosaur-bearing sites in the TransitionalSands that connect the Saurian Beds with theshallow marine sandstone beds are undoubtedlymarginal marine in origin because marine in-vertebrates are associated with the dinosaurs . Ahumerus of Brachiosaurus brancai covered byoysters is reported, for instance, from Site Aawhich is in the Transitional Sands above theMiddle Saurian Bed (Janensch, GTE field catalo-gue : 140) . As previously mentioned, the state ofpreservation of bones recovered from the Transi-tional Sands suggests transport and redepositionof bones prior to burial . Therefore, these enor-mous concentrations of disarticulated dinosaursbones probably represent time-averaged boneaccumulations rather than a short-term massmortality event.

As mentioned above, most of the documentedskeletal accumulations from the Middle and UpperSaurian Beds are single-individual and mono-specific. Multi-individual assemblages are rare,and death assemblages of Brachiosaurus brancai,Dicraeosaurus sattleri, and Barosaurus africanusare apparently dominated by limb bones of adultindividuals. This taphonomic pattern does notnecessarily support the catastrophic mass mortal-ity model, since killing agents that producedmass mortality should have affected all indivi-dual age classes of the sauropod populations .

However, this conclusion should be madewith caution because actualistic studies on largeextant land mammals reveal that patterns ofbone accumulations may result from patterningin behaviour (Haynes 1993). Younger Africanelephants, for instance, sometimes die "muchnearer the inner areas" of die-off sites "than olderindividuals" due to aggressive interactions at ex-cavated wells during severe drought (Haynes1993:131) . As in extant African elephant popula-tions (Haynes 1990, 1993), seasonal droughtcould have resulted in social fissioning andephemeral aggregations of adult and old adultTendaguru sauropods at shrinking water sourcesand feeding sites.


In addition, articulated sauropod limb bonesburied in an upright position (Janensch 1914c)indicate that the body weight of the animalscould have been an important selective tapho-nomic filter for the dinosaurs from Tendaguru .So, for instance, the body weight of the juvenileTendaguru sauropods may have been below thethreshold for becoming stuck in the mud, unlikethe adult animals that passed more frequentlythe "point of no return" due to their heavierbody weight .A similar taphonomic model has been sug-

gested for the Late Triassic (Norian) prosauro-pods Plateosaurus (Sander 1992) and Sellosaurus(Hungerbtihler 1998) . Sander (1992 : 256) arguesthat the absence of juvenile individuals of Plateo-saurus engelhardti in several European Knollen-mergel sites is due to the "lower foot pressure"of juvenile animals that were thus "below thethreshold for becoming mired" in shallow de-pressions.

Hennig (1925) believed that spatially limitedmass accumulations of disarticulated bones ofKentrosaurus aethiopicus from the MiddleSaurian Bed (Site St) and the Upper SaurianBed (e.g ., Sites He and X), containing indivi-duals of various size and age, resulted from asudden mass death event (see also Weigelt1927) . Rich concentrations of disarticulated ske-letal remains of Dryosaurus lettow-vorbecki re-covered from the Middle Saurian Bed at SiteJg (WJ) were also thought to represent an in-stance of sudden mass mortality (Janensch1914c) . Most limb bones (e.g ., femur, tibia, fibu-la) were found parallel to one another, withtheir long axis oriented northwest to southeast,indicating that these bones were accumulatedunder the influence of wave action (breakers)or wave-induced currents (Janensch 1914c : 256) .Janensch (1914c) supposed that herds of smallornithischians that had entered lagoons duringperiods of drought were killed by the returningflood and became embedded close to the placeof death. If this scenario is correct, Dryosauruslettow-vorbecki left rapidly formed mass assem-blages of bones in the Tendaguru Beds that re-sulted from an external catastrophic killingevent . Accumulations of sauropod bones in theMiddle and Upper Saurian Beds are alsothought to have been caused by similar killingevents, since the huge sauropod dinosaurs werealso considered to be gregarious land animals(Janensch 1914c) .By contrast, Reck (1925) who quarried Ten-

daguru Site Jg (WJ) in 1912 doubted the cata-

Mitt . Mus. Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

strophic mass mortality. He argued (Reck 1925:9-10) : "So sehr auch das herdenartige Auftreteneinzelner Arten, und besonders die Herden-zusammensetzung aus Tieren jeder Größe undjeden Alters für sie zu sprechen scheint, sostehen doch auch ihr unüberwindliche Hinder-nisse entgegen, wenn man dem Sprachgebrauchentsprechend unter Katastrophe nur ein seltenes,verheerendes Naturereignis versteht . . . . . Aberzwei Hauptmomente führen . . . . . zur Ablehnungsolcher Theorie : die großzügig gleichmäßigen,außer durch die Transgressionen nicht unter-brochenen, feinklastischen Sedimentationsver-hältisse und die regellose Verteilung der Fundein der Vertikalen innerhalb der Mergelmassen .Dadurch wird die Supposition jeder Katastropheunhaltbar, wenn man nicht jeweils einen ganzenMergelhorizont als Niederschlag und Zeugniseiner Katastrophe ansprechen will . Dagegenaber spricht alles . . . . . Als generelles Entstehung-sprinzip der Lagerstätte kommt . . . . . die Ka-tastrophe nicht in Betracht" . Reck (1925), on thebasis of his studies of the Tendaguru region,concluded that the dinosaur bones were ir-regularly dispersed within the Tendaguru saurianbeds, and that the origin of the bone acumu-lation could not be explained by catastrophicmass mortality. This distribution pattern suggeststhat the bone accumulations resulted from recur-ring "normal" death events of dinosaur indivi-duals over a long period of time rather thanfrom a catastrophic, short-term mass mortalityevent.

In addition to this argument, further supportfor an attritional origin is found in the femur-sizedistribution pattern of Dryosaurus lettow-vor-becki which represents an age profile, dominatedby juvenile and adult individuals (Fig. 23) . Thispattern (Fig. 23) shows a striking similarity to theso-called U-shaped or attritional mortality typeof profile that is due to a long-term input ofbones resulting from "normal", ecologically re-lated deaths in different-aged population mem-bers (Lyman 1996) . This pattern differs remark-ably from the unimodal L-shaped, catastrophicor mass mortality pattern in which younger ageclasses dominate and older age classes becomesuccessively underrepresented (Lyman 1996) . Atthe present stage of investigation, I do not drawthe conclusion, that the mortality profile inDryosaurus lettow-vorbecki is attritional in ori-gin, since possible influence of the sexual di-morphism is far from clear and the interpreta-tion of U-shaped mortality profiles is mainlybased on experience with extant and fossil main-

59

mals. Nevertheless, it is worth noting, that theinterpretation as an attritional mortality profilealso agrees more closely with the data obtainedfrom the Tendaguru sauropods (Figs. 21, 22) aswell as with the distribution pattern of boneswithin the Tendaguru dinosaur beds .The invertebrate fossil record and seasonality

indicated by reworked caliche nodules might in-dicate playa lake-like water bodies close to anoscillating shoreline, and subject to either seaso-nal or long-term droughts. Cross-bedded sand-stone beds within the Middle and Upper SaurianBeds, in turn, might be explained by seasonallyrainy intervals that resulted in water flows, tem-porary lakes, and ponds. These ephemeral waterbodies might have been refuges and die-off sitesfor Tendaguru dinosaurs during seasonaldrought . Most likely, the palaeoenvironment inthe Tendaguru region was much more diversethan generally thought so far .

In the absence of direct taphonomic studiesand substantial parts of the doumentation of theoriginal expedition, it is beyond the scope of thepresent account to establish a detailed tapho-nomic model for the Tendaguru dinosaur accu-mulation. However, the existence of a number ofsuperimposed dinosaur-bearing horizons (Sau-rian Beds, Transitional Sands) shows that dino-saur bones were accumulated over a long timespan during the Late Jurassic . The fossil recordalso reveals that extraordinarily suitable condi-tions for the preservation of dinosaurian bonesoccured repeatedly. The year-to-year input ofbones resulting from short-term intervals of mor-tality due to drought might have been consider-able and "simulates" sudden catastrophic massmortality events. If so, the Tendaguru fossil re-cord may reflect attritional mortality rather thansudden catastrophic mass mortality.

Acknowledgements

I am grateful to Prof. Dr . H.-E Schultze and Dr . D. Unwin(Museum für Naturkunde der Humboldt-Universität zu Ber-lin, Institut für Paläontologie) for much-valued commentsand improvement of the English, and to Dr . E. Cook (Uni-versity of Bristol) for reviewing the manuscript . I extend mythanks, to Dr . E M. Sander (Universität Bonn) for discus-sions and histological data of Tendaguru sauropods, to Dr. B.Curtice and Dr . L. Curtice (Phoenix, AZ, USA) for helpingwith the identification of cervical vertebrae of Brachiosaurusbrancai, to Dr . M. Aberhan (Museum für Naturkunde derHumboldt-Universität zu Berlin, Institut für Paläontologie),Prof. Dr . H. Keupp and Dr. M. Schudack (Freie UniversitätBerlin, Institut für Paläontologie) for comments on the litho-logy of the Tendaguru Beds, to Mrs. E. Glass (Humboldt-Universität zu Berlin, Institut für Mineralogie) for X ray dif-fraction analysis, to Mr . J.-P. Mendau for the redrawing of

60

the site maps, to Mrs . V. Heinrich for the diagrams, and toMrs . W. Harre (all Museum für Naturkunde der Humboldt-Universität zu Berlin) for the photographs . Financial supportfrom the Deutsche Forschungsgemeinschaft (He 2757/1-1) isgratefully acknowledged .

References

Abel, O. 1927 . Lebensbilder der Vorzeit . 637 pp., Fischer,Jena .

Aitken, W G. 1956 . The Jurassic-Cretaceous junction in Tan-ganyika . - East-Central Regional Committee for Geolo-gy, Commission for Technical Cooperation in Africasouth of the Sahara 1956: 67-71 .

- 1961 . Geology and Palaeontology of the Jurassic and Cre-taceous of Southern Tanganyika. - Bulletin of the Geolo-gical Survey of Tanganyika 31: 1-144 .

Behrensmeyer, A. K . 1991 . Terrestrial vertebrate accumula-tions . In Allison, P. A . & Briggs, D . E . G. (eds .) . Tapho-nomy. Releasing the data locked in the fossil record :291-335, Plenum Press, New York and London .

Bonaparte, J., Heinrich, W-D. & Wild, R . Review of Janen-schia WILD, with description of an new sauropod from theTendaguru Beds of Tanzania, and a discussion on the sys-tematic value of the Sauropoda procoelian caudal verte-brae . - Palaeontographica, Abt. A. (in press) .

Branca, W. von 1914. Kurzer Bericht über die von Dr . Reckerzielten Ergebnisse im vierten Grabungsjahre 1912. -Archiv für Biontologie 3(1) : 61- 63 .

Broschinski, A . 1999 . Ein Lacertilier (Scincomorpha, Para-macellodidae) aus dem Oberen Jura von Tendaguru/Tan-sania . - Mitteilungen aus dem Museum für NaturkundeBerlin, Geowissenschaftliche Reihe 2 : (this volume)

Chinsamy, A . 1990 . Ontogenetic changes in the bone histolo-gy of the Late Jurassic Dryosaurus lettow-vorbecki . -Journal of Vertebrate Paleontology 15 : 96-104 .

Colbert, E . H. 1984 . The great dinosaurs hunters and theirdiscoveries. 283 pp., Dover Publications Incorporation,New York .

Cox, L. R . 1952 . The Jurassic lamellibranch fauna of Cutch(Kachh) . No . 3 : Families Pectinidae, Amusiidae, Plicatuli-dae, Limidae, Ostreidae and Trigonidae (Supplement) . -India Geological Survey Memoires, Palaeontologica Indi-ca, ser. 9, vol. 3, part 4 : 1-128.

Dietrich, W. O. 1914 . Die Gastropoden der Tendaguruschich-ten, der Aptstufe und der Oberkreide im südlichenDeutsch-Ostafrika . - Archiv für Biontologie 3(4) :97-153 .

- 1925 . Über eine dem mittleren Sauriermergel am Tenda-guru äquivalente, rein marine Kimmeridgebildung in Ma-hokondo, Deutsch-Ostafrika . - Palaeontographica, Sup-plement VII, 2 . Reihe, Teil I, Lieferung 1 : 1-24 .

- 1927 . Das Alter der Trigonienschichten am Tendaguru. -Centralblatt für Mineralogie, Geologie und Paläontologie,B 1927(2) : 59-64 .

- 1933 . Zur Stratigraphie und Palaeontologie der Tenda-guruschichten . - Palaeontographica, Supplement VII, 2 .Reihe, Teil 11, Lieferung 1 : 1-86 .

Fraas, E . 1908 . Ostafrikanische Dinosaurier . - Palaeontogra-phica 55 : 105-144 .

Galton, P M. 1980 . Avian-like tibiotarsi of pterodactyloids(Reptilia Pterosauria) from the Upper Jurassic of EastAfrica . - Palaeontologische Zeitschrift 54(3/4) : 331-342 .

- 1981 . Dryosaurus, a hypsilophodontid dinosaur from theUpper Jurassic of North America and Africa . - Pal5onto-logische Zeitschrift 55(3/4) : 271-312 .

- 1982 . Elaphrosaurus, an ornithomimid dinosaur from theUpper Jurassic of North America and Africa . - Pal5onto-logische Zeitschrift 56(3/4) : 265-275 .

Haynes, G. 1990. The mountains that fell down : life anddeath of Heartland mammoth. In Agenbroad, L . D .,


Mead, J. I, Nelson, L. W. (eds .) . Megafauna and man : dis-covery of America's Heartland : 22-31, The MammothSite of Hot Springs, South Dakota Incorporation, HotSprings, South Dakota .

- 1993 . Mammoths, mastodonts & elephants . Biology, Be-havior, and fossil record . 413 pp. Cambridge UniversityPress, Cambridge.

Heinrich, W-D. 1991 . Über Brancatherulum tendagurenseDIETRICH, 1927 (Mammalia : Eupantotheria) aus demOberjura von Tendaguru, Tansania . Vorläufige Mittei-lung . - Mitteilungen aus dem Zoologischen MuseumBerlin 67(1) : 97-104 .

- 1998 . Late Jurassic mammals from Tendaguru, Tanzania,East Africa . - Journal of Mammalian Evolution 5(4) :269-290 .

- 1999 . First haramiyid (Mammalia, Allotheria) from theMesozoic of Gondwana . - Mitteilungen aus dem Mu-seum für Naturkunde Berlin, Geowissenschaftliche Reihe2: 159-170 .

Hennig, E . 1912a . Die Entstehung der Dinosaurier-Lager . -Sitzungsberichte der Gesellschaft naturforschender Freun-de 1912 : 137-142, Berlin .

- 1912b. Am Tendaguru. Leben und Wirken einer deut-schen Forschungsexpedition zur Ausgrabung vorweltlicherRiesensaurier in Deutsch-Ostafrika . 151 pp ., E . Schweit-zerbart'sche Verlagsbuchhandlung, Stuttgart .

- 1914a. Paläontologische Ausgrabungen in Deutsch-Ost-afrika . - Blätter der Mitglieder des wissenschaftlichenCentralvereins 1914: 1-16 . Stuttgart .

- 1914b . Beiträge zur Geologie und Stratigraphie Deutsch-Ostafrikas . I . Geologisch-stratigraphische Beobachtungenim Küstengebiet des südlichen Deutsch-Ostafrika . II.Geologisch-stratigraphische Beobachtungen im Gebieteder Jura-Ablagerungen an der Deutsch-OstafrikanischenZentralbahn . - Archiv für Biontologie 3(3) : 1-72 .

- 1914c . Die Invertebraten-Fauna der Saurierschichten amTendaguru . - Archiv für Biontologie 3(4) : 155-185 .

- 1915 . Kentrosaurus aethiopicus der Stegosauride des Ten-daguru . - Sitzungsberichte der Gesellschaft naturfor-schender Freunde Berlin 1915 : 219-247.

- 1916 . Zweite Mitteilung über den Stegosauriden vomTendaguru . - Sitzungsbericht der Gesellschaft naturfor-schender Freunde Berlin 1916 : 175-182.

- 1925 . Kentrurosaurus aethiopicus : die Stegosaurierfundevom Tendaguru, Deutsch-Ostafrika . - Palaeontographica,Supplement VII, 1 Reihe, Teil 1, Lieferung 2:100-253 .

- 1937 . Der Sedimentstreifen des Lindi-Kilwa-Hinterlandes(Deutsch-Ostafrika) . - Palaeontographica, SupplementVII, 2. Reihe, Teil II, Lieferung 2, 99-186.

Hungerbühler, A. 1998. Taphonomy of the prosauropod di-nosaur Sellosaurus, and its implication for carnivorefaunas and feeding habits in the Late Triassic . - Palaeo-geography, Palaeoclimatology, Palaeoecology 143: 1-29 .

Jaeger, H . 1971 . Werner Janensch 1878-1969 . - Berichteder Deutschen Gesellschaft für geologische Wissenschaf-ten, A, Geologie, Paläontologie 16(2) : 149-154.

Janensch, W 1914a . Bericht über den Verlauf der Tendagu-ru-Expedition. - Archiv für Biontologie 3(1) : 17-58 .

- 1914b. Übersicht über die Wirbeltierfauna der Tendagu-ru-Schichten, nebst einer kurzen Charakterisierung derneu aufgestellten Arten von Sauropoden . - Archiv fürBiontologie 3(1) : 81-110 .

- W 1914c. Die Gliederung der Tendaguru-Schichten imTendaguru-Gebiet und die Entstehung der Saurier-Lager-stätten . - Archiv für Biontologie 3(3) : 227-261 .

- 1920 . Über Elaphrosaurus bambergi und die Megalo-saurier aus den Tendaguru-Schichten Deutsch-Ostafrikas.- Sitzungsberichte der Gesellschaft naturforschenderFreunde Berlin 1920: 225-235 .

- 1922 . Das Handskelett von Gigantosaurus robustus undBrachiosaurus brancai aus den Tendaguru-SchichtenDeutsch-Ostafrikas . - Centralblatt für Mineralogie, Geo-logie und Palaeontologie 1922: 464-480 .

Mitt. Mus . Nat.kd . Berl ., Geowiss. Reihe 2 (1999)

- 1925a. Die Grabungsstellen der Tendaguru-Gegend . -Palaeontographica, Supplement VII, 1 Reihe, Teil 1, Lie-ferung 1 : XVII-XIX .

- 1925b . Die Coelurosaurier and und Theropoden der Ten-daguru-Schichten Deutsch-Ostafrikas. - Palaeontographi-ca, Supplement VII, 1 Reihe, Teil 1, Lieferung 1 : 1-97 .

- 1929a. Material und Formengehalt der Sauropoden in derAusbeute der Tendaguru-Expedition . - Palaeontographi-ca, Supplement VII, 1 Reihe, Teil 2, Lieferung 1 : 1-34 .

- 1929b . Die Wirbelsäule der Gattung Dicraeosaurus. - Pa-laeontographica, Supplement VII, 1 Reihe, Teil 2, Liefe-rung 1 : 37-133 .

- 1929c. Ein aufgestelltes und rekonstruiertes Skelett vonElaphrosaurus bambergi . Mit einem Nachtrag zur Osteo-logie dieses Coelurosauriers . - Palaeontographica, Sup-plement VII, 1 Reihe, Teil 2, Lieferung 1 : 279-286 .

- 1933 . Eine Estheria aus den Tendaguruschichten . - Pa-laeontographica, Supplement VII, 2 Reihe, Teil 2, Liefe-rung 1 : 95-98 .

- 1935 . Die Schädel der Sauropoden Brachiosaurus, Baro-saurus und Dicraeosaurus aus den Tendaguru-SchichtenDeutsch-Ostafrikas. - Palaeontographica, SupplementVII, 1 Reihe, Teil 2, Lieferung 1 : 145-248 .

- 1936 . Ober Bahnen von Hirnvenen bei Saurischiern undOrnithischiern, sowie einigen anderen fossilen und rezen-ten Reptilien . - Palaeontologische Zeitschrift 17:181-198 .

- 1938 . An African Dinosaur. - Research and Progress4(5) : 276-278 .

- 1950a . Die Wirbelsäule von Brachiosaurus brancai. - Pa-laeontographica, Supplement VII, 1 . Reihe, Teil III, Lie-ferung 2 : 27-93 .

- 1950b. Die Skelettrekonstruktion von Brachiosaurusbrancai . - Palaeontographica, Supplement VII, l . Reihe,Teil III, Lieferung 2: 95-103 .

- 1950c. Die systematische Stellung des OrnithopodenDysalotosaurus aus den Tendaguru-Schichten . - NeuesJahrbuch für Geologie und Paldontologie, Mh., 1950(9) :286-287 .

- 1955 . Der Ornithopode Dysalotosaurus der Tendaguru-schichten . - Palaeontographica, Supplement VII, 1 Rei-he, Teil 3, Lieferung 3 : 105-176 .

- 1961a . Die Gliedmaßen und Gliedmaßengürtel der Sauro-poden der Tendaguru-Schichten . - Palaeontographica,Supplement VII, 1 . Reihe, Teil III, Lieferung 4: 177-235 .

- 1961b. Skelettrekonstruktion von Dysalotosaurus lettow-vorbecki. - Palaeontographica, Supplement VII, 1 . Rei-he, Teil III, Lieferung 4 : 237-240 .

Janensch, W. & Hennig, E. 1909 . Erster Bericht über dieTendaguru-Expedition . - Sitzungsberichte der Gesell-schaft naturforschender Freunde Berlin 1909 : 358-360.

Kapilima, S. 1984 . Stratigraphische und paläontologische Un-tersuchungen im Jura und der Kreide des tansanischenKüstenstreifens im Hinterland von Dar-Es-Salaam undBagamoyo. - Berliner Geowissenschaftliche Abhandlun-gen, A, 57 : 1-77 .

Kent, P. E ., Hunt, J. A . & Johnstone, D . W. 1971 . The geolo-gy and geophysics of costal Tanzania . - Natural Environ-ment Research Council, Institute of Geological Sciences,Geophysical Paper 6 : I-VI, 1-101 .

Kitchin, E L ., 1929 . On the age of the upper and middledeinosaur deposits at Tendaguru, Tanganyika Territory. -Geological Magazine, 66(5) : 193-220 .

Lange, E . 1914 . Die Brachiopoden, Lamellibranchiaten undAnneliden der Trigonia Schwarzi-Schichten . - Archiv fürBiontologie 3(4) : 187-289 .

Lyman, R . L . 1996 . Vertebrate Taphonomy. 524 pp . Cambrid-ge University Press, Cambridge.

Migeod, E W. H. 1927 . British Museum East Africa expedi-tion . Progress in the year 1926 . - Natural History Maga-zine 1927(1) : 34-43 .

- 1930. Report on the British Museum East Africa expedi-tion . Season 1929 . - Natural History Magazine 1930(2) :185-198 .

6 1

- 1931 . British Museum East Africa expedition . Account ofthe work done in 1930 . - Natural History Magazine1931(3) : 87-103 .

Parkinson, J., 1930 . The dinosaur in East Africa . An accountof the giant reptile beds of Tendaguru, Tanganyika Terri-tory. 192 pp ., H . E & G. Witherby, London.

Reck, H. 1925 . Grabungen auf fossile Wirbeltiere inDeutsch-Ostafrika . - Geologische Charakterbilder 31 :1-36.

- 1931 . Die deutschostafrikanischen Flugsaurier. - Central-blatt für Mineralogie, Geologie und Paläontologie, B,1931: 321-336 .

Russell, D., B6land, P. & McIntosh, J . S . 1980. Paleoecologyof the dinosaurs of Tendaguru (Tanzania) . - Memoiresde la Societe Geologique de France, Nouvelle Serie, 139 :169-175 .

Sander, E M. 1992 . The Norian Plateosaurus bonebeds ofCentral Europe and their taphonomy. - Palaeogeogra-phy, Palaeoclimatology, Palaeoecology 93 : 255-299 .

- 1999 : Life history of the Tendaguru sauropods as deducedfrom long bone histology. - Mitteilungen aus dem Mu-seum für Naturkunde Berlin, Geowissenschaftliche Reihe2 : 103-112 .

- (in press) Long bone histology of the Tendaguru sauro-pods : Implication for growth and biology. - Palaeobiolo-gy

Schrank, E . 1999 . Palynology of the dinosaur beds of Tenda-guru (Tanzania) - preliminary results. - Mitteilungenaus dem Museum für Naturkunde Berlin, Geowissen-schaftliche Reihe 2 : 171-183 .

Schuchert, Ch . 1918 . Age of the Morrison and East AfricanTendaguru formations. - Bulletin of the Geological So-ciety of America 29 : 245-280 .

- 1934 . The Upper Jurassic age of the Tendaguru dinosaurbeds. - American Journal of Science 27 : 463-466 .

Schudack, M . E . 1999 . Charophytes from the Middle Dino-saur Member of the Tendaguru Formation (Upper Juras-sic of Tanzania) . - Mitteilungen aus dem Museumfür Naturkunde Berlin, Geowissenschaftliche Reihe 2 :201-205 .

Schudack, M . E ., Schudack, U. & Pietrzeniuk, E . 1999 .Ostracods from the Middle Dinosaur Member of theTendaguru Formation (Upper Jurassic of Tanzania) . -(manuscript)

Spath, L. E 1927-1933 . Revision of the Jurassic cephalopodfauna of Kachh (Cutch) . - India Geological Survey Me-moires, Palaeontologica Indica, new ser., vol . 9, mem . 2,pt. 1-6 : 1-945 .

Unwin, D. & Heinrich, W.-D . 1999 . On a pterosaur jaw re-main from the Upper Jurassic of Tendaguru, East Africa .- Mitteilungen aus dem Museum für Naturkunde Berlin,Geowissenschaftliche Reihe 2 : 121-134 .

Weigelt, J. 1927 . Rezente Wirbeltierlleichen und ihre pal5o-biologische Bedeutung. 227 pp., Verlag Max Weg, Leip-zig.

Wild, R . 1991 . Die Ostafrika-Reise von Eberhard Fraas unddie Erforschung der Dinosaurier-Fundstelle Tendaguru . -Stuttgarter Beiträge zur Naturkunde, Ser . C, 30 : 71-76 .

Zeiss, A . 1975 . The Aptian ammonite fauna of Ethiopia : newresults on the biostratigraphy and zoogeography. - NeuesJahrbuch für Geologie und Paläontologie, Mh., 1975(10) :628-639.

Zils, W., Werner, Ch ., Moritz ; A . & Saanane, Ch . 1995a .Orientierende Tendaguru-Expedition 1994 . - BerlinerGeowissenschaftliche Abhandlungen E16 (Gundolf-Ernst-Festschrift) : 483-531 .

Zils, W., Werner, Ch ., Moritz; A. & Saanane, Ch . 1995b. Ten-daguru, the most famous dinosaur locality of Africa. Re-view, survey and future prospects. - Documenta naturae97,1-41 .

Zwierzycki, J. 1914 . Die Cephalopodenfauna der Tendaguru-Schichten in Deutsch-Ostafrika . - Archiv für Biontologie3(4) : 7-96 .

Date post:	14-Feb-2017
Category:	Documents
Upload:	duongnhi
View:	220 times
Download:	2 times

The taphonomy of dinosaurs from the Upper Jurassic of Tendaguru ...

Documents