ORYCTOS,Vol. 1 : 113-120,Octobre 1998 NOT CRANESOR MASTS, BUT BEAMS : THE BIOMECHANICS OF SAUROPODNECKS John MARTIN', Valérie MARTIN-ROLLAND2 and Eberhard (Dino) FREY3 lleicesterCity Museum, New Walk, Leicester LEI 7EA, and Department of Geology, Leicester University, Leicester LEI 7RH;U. K. 2Muséedes Dinosaures, 11260 Espéraza, France. 3Museumfiir NattirkundeKarlsruhe, Posdach 6209,D-76042, Germany. Abstract : The mechanicsof sauropodnecks are still poorly understood,judging from many recentlife reconstructions. Only sevenor eight sauropodtaxa have necks well-enough known for their mobility and posture to be reconstructedfully and reliably. In theseanimals, the limits of mobility imposed by the zygapophysealand central articulation structuresmay be calculated. Simple biomechanicsshows that sauropodnecks were segmentedbeams, and that the way in which seg- mented beamsmust be braced adds further limits to mobility. The bracing systemsimplied by the vertebral anatomy can be reconstructed,based partly on an appreciation,from extantanimals, of what soft-tissuestructures were possiblegiven a particularset of bone geometries.Sauropod necks were dorsally-braced,ventrally-braced or a combinationof the two: eachsystem was associatedwith a distinctive,exclusive, group ofvertebral features- high, low or divided neural spines, largeor small transverseprocesses and shortor elongatedcervical ribs. The two setsof mobility-controlling systems(oint morphology and bracing) madethe necks of all "popular" sauropodsrelatively, or very rigid structures,and also show that they were habitually carriedwith the ventral aspect"down" with respectto gravity - in other words as beams,not masts. Key words : Dinosaur,sauropod, biomechnnics, necks, Iift reconstructions. Ni gruesni mâts, mais timons - Biomécaniquedes cousde sauropodes Résumé : D'aprèsde nombreusesreconstitutions récentes, la mécaniquedu cou des sauropodesdemeure méconnue. Au seindes sauropodes, seuls sept ou huit ont le cou suffisammentconnu pour permettreune reconstitutioncomplète et fiable de sa position et de sa mobilité. Chez ces animaux,les limites de la mobilité (imposéespar les zygapophyseset I'articu- lation du centrum)peuvent être calculées.Des principessimples de biomécaniquemontrent que le cou peut être assimilé à une poutresegmentée. La manièredont ces segmentsdoivent être soutenusajoute des limites supplémentairesaux pos- sibilitésde mobilité. Le systèmede soutienimposé par l'anatomievertébrale peut êtrereconstitué, partiellement fondé sur une appréciation(à partir d'animaux actuels)des dispositions particulières des structuresosseuses qui permettentde défi- nir le type de partiesmolles correspondantes.Le cou dessauropodes était soutenudorsalement, ventralement ou par une combinaisondes deux, chaquesystème étant associéà un groupeunique et distinct de caractèresvertébraux (épine neu- rale haute,basse ou divisée,grands ou petitsprocessus transverses, et côtescervicales courtes ou allongées).Les deux sys- tèmespossibles du contrôlede la mobilité (morphologiearticulaire et type de soutien)font du cou de tous les sauropodes < populaires >>une structure assezou très rigide et montre égalementque le cou était habituellement porté horizontale- ment, en d'autrestermes comme un timon et non colnmeun mât. Mots clés : Dinosaure,sauropode, biomécanique, cou, reconstitutions. INTRODUCTION as a result of their complexintervertebral articulation systems.The limits imposedby the zygapophysealand Popularlife reconstructionsregularly show sauro- centrapophysealmorphology of sauropodnecks may be podswith the neckheld almostvertically, or in an artis- calculatedquite accuratelyin the few taxathat arewell- tically flexed 's' curve, suggestingthat the animals enoughknown (Martin, 1981 ; Stevens& Parrish,1996). could browsehigh in the treesor fight by "necking" This paperapplies the principlesof biomechanics (e.g.Bakker, 1986; Paul , 1987). On the otherhand, to sauropods'necks to help resolvethe dilemma.An somestudies (e.g. Coombs, I9l5 ; Martin,I98l ; Dod- important limitation of the project is that there are son, 1990)have arguedthat there were rather severe very few reliable specimenson which to base our mechanicallimits to the flexibility of sauropods'necks conclusions. 113 ORYCTOS.Vol. 1. 1998 LIMITS OF KNOWLEDGE be made. In most cases,too (exceptionsbeing Pla- .SAUROPOD NECKS teosaurus and, perhaps, Shunosaurusin which several or many individuals belonging to one species are Only one or two prosauropod and perhaps eight known) there is only one individual animal known. sauropodtaxa arewell-enough known for anear-com- Does this animal representthe constructional situation plete, accurate reconstruction of the neck skeleton to in a species,genus or higher taxon? TAXON AGE NECK known... SITUATION PROSAUROPODA Thecodontosaurus. .. .Norian-Camian. .incompletely. .? Anchisaurus ... ...Pliensbachianor Toarcian... ..incompletely ? ........ .? Massospondylus.............Hettangian-Pliensbachian ......nearly completely. ......?10 low, undivided neural spines; "long ribs" Yunannosauru,r. .Hettangian-Pliensbachian. .incompletely. .? Mussaurus .Norian ........?compleûely .... ... .. ...? Ammosaurus . ......Pliensbachianor Toarcian.. ...incompletely .... ... ..? Plateosaurus ........Norian ........many,completely........l0 low, undivided neural spines; ribs 2-3 segments Sellosaurus .Norian ........incompletely.............? Euskelosaurus ......CarnianorNorian .....incompletely..........? Lufengosaurus . .Hettangian-Pliensbachian. .incompletely.. .? Melanorosauru,r.............Carnian or Norian .....incompletely...........? Riojasaurus .........Norian ........completely? .... .........? SAI.JROPODA Vulcanodon .........?Hettangian ..?...... ...........? Barapasaurus .......?"early Jurassic" ......?. ................? Cetiosaurus .........Bajocian-Bathonian ...nearlycompletely .......14 medium,undivided neural spines; ribs 1-2 segments Haplocanthosaurus..........Kimmeridgian-Tithonian.......nearly completely .......14 medium,undivided neural spines; ribs 1-2 segments Patagosaurus .......Callovian .....?partially................? Shunosaurusv .......Bathonian-Callovian ..?completely ....?15medium, undivided neural spines; ribs ? Omeisaurus .........Bathonian-Callovian ..?partially ........? Mamenchisaurus.............late Jurassic ..incompletely ....8low, undivided + ?14medium, undivided neuralspines; ribs 2-3 segments Phuwiangosaurus............?. .....?.. ...............? Euhelopus ..?Kimmeridgian .... ...Nearlycompletely .......?l21ow-medium, undivided + 3 medium-high dividedneural spines; ribs 2-3 segments Diplodocus .Kimmeridgian-Tithonian... ....? completely... .........7 medium,undivided + 8 medium,divided neuralspines; ribs l-2 segments Barosaurus .Kimmeridgian-Tithonian.......nearlycompletely .......?15 medium, undivided + ?2medium, divided neuralspines; ribs 1-2 segments Apatosaurus ........Kimmeridgian-Tithonian.......nearlycompletely .......7 low,wide but undivided + 8 low,widely dividedneural spines; ribs <l segment Dicraeosaurus ......Kimmeridgian... ......?nearlycompletely ......74 high, undivided + ?7high, divided neural spines;ribs 1 segment Brachiosaurus . .. ...Kimmeridgian-Tithonian.......?nearly completely . ...?13low, undivided neural spines; ribs 1-2segments TABLE l. Completenessof well-knownsauropodomorph dinosaurs'necks rt4 MARTIN , MARTIN-ROLLAND& FREY - SAUROPODNECKS BIOMECHANICS . SAUROPODNECKS \ryEREBRACED BEAMS A simple beam supported at just one end, like a FIGURE I - A simple,unbraced beam supported at cranejib, has its upper part stressedby tension while one end. The effect of gravity on the unsupported its lower part is under compression (fig. 1). Most structureis to placethe upper part under tension the lower part. engineering beams are braced, either by their cross- while compressing sectional design, by internal struts or by external braces, to distribute and reduce these stresses.The Forth Bridge, near Edinburgh, Scotland, shows how this works. It is a cantilever bridge; each half-section is a beam braced above and below and supported at one end by a leg. Civil engineering does not make segmented,mobile-jointed beams (except for those wooden toy snakes), but long-necked animals (like sauropods)have them. A segmented,flexible beam cannot be unbraced. Becauseeach segmentneeds to be free to move againstits neighbours,the tensile and compressivestresses must be redistributed away from the intersegmental articulations by bracing systems running the whole length of the structure. knees so that we can stand for long periods without Just like the Forth Bridge, a horizontal beam has tiring. to be braced above and below; in living structuresthis Sauropod neck ligament systemshave been des- means dorsal and ventral bracing systems. If the cribed before, particularly in discussionsof the func- beam is tilted, the detailed geometry (including the tion of divided neural spines (e.g. Alexander, 1985, orientation of the bracing members) changes,but the and references therein), as "cables" supporting a principle remains the same. For stability in all orien- "crafiejib". Recognition of a ventral system, and of tations, the beam has a minimum of three bracing its function in combination with the dorsal structures members at least one dorsal and two ventral. The as bracing for the neck, is new in this paper. In sau- dorsal braces are tension members while the ventral ropods the long, forward and backward directed pro- set work in compression (again, like the Forth Bridge) cessesof the cervical