Reconstructing postures and locomotion of tetrapods are one of the most important issues in the evolutionary studies of vertebrates. Postures and locomotion are highly correlated with the skeletal morphology. One of the most difficult and unsolved questions in reconstructing a skeleton of extinct tetrapods is how the positioin of the shoulder and forelimb area are determined. The shoulder and forelimb do not directly articulate with the axial skeleton. Thus, there have been quite a number of different forelimb reconstructions suggested even for a single species of extinct tetrapods. I hereby propose a new method to securely identify the craniocaudal position of the scapula on the presacral ribs. I would also like to suggest a new method to identify the angle of the elbow joint at its propulsive phase based on the orientation of an olecranon process.

Determining of scapular position is a major issue in reconstructing the skeletal system of extinct quadruped archosaurs and mammals, because the pectoral girdle has no direct skeletal joint with the axial skeleton. In standing or walking, the trunk of quadrupeds is suspended between the forelimbs by the m. serratus ventralis, which originates from the lateral sides of the "thoracic" ribs and inserts to the proximal portion on the costal surface of scapula. Therefore,the "thoracic" ribs are subjected to a static or dynamic vertical compression between the lifting force from the muscle and the gravity force from the vertebral column. To elucidate the body support function of the ribs, I analyzed the mechanical strength of the ribs of extant tetrapods by the finite element method, and compared their degree of strength through their craniocaudal scapular positions. The result of this simulation showed that the "thoracic" ribs of quadrupeds, to which the serratus muscle attaches, have relatively high strength against compaction than the other ribs. In bipeds, however, I could not find such a significant correlation between the strength of ribs and the serratus muscle. It suggested that the location of robust ribs is associated with the arrangement of the serratus muscle, and provides a probable candidate for the scapular position, not only for extant quadruped amniotes but also for extinct quadruped archosaurs and mammals. The correlation between the mechanical strength of ribs and the position of craniocaudal position of scapula is proposed here the first time in this study.

Deciphering the angles of limb joints is another difficult task in the skeletal reconstructions of extinct vertebrates. It is mainly because each joint has a wide range of rotation in the forelimb. Several forelimb reconstructions have been suggested for Oligocene-Miocene desmostylian mammals such as Desmostylus and Paleoparadoxia, and Late Cretaceous ceratopsian dinosaurs, such as Leptoceratops, Centrosaurus, Anchiceratops, and Triceratops. This study also proposed a new method to reconstruct the elbow joint angle, which is formed by the humeral shaft and forearm in quadruped tetrapods with prominent olecranon processes. In the study, orientations of olecranon process to the long axis of forearm were measured in dried skeletons of 25 genera of extant mammals, and the results were compared with the elbow joint angle during locomotion of the living animals. In most cases the olecranon process and the shaft of the humerus are oriented nearly perpendicular to each other in the propulsive phase. By keeping the angle of the olecranon process and the humeral shaft almost perpendicular, major extensor muscles can maximize the lever arm at the elbow joint. Orientation of the extensor lever arm, such as olecranon process, can be used to reconstruct the forelimb posture for quadruped mammals, regardless of taxonomic attritions and body sizes.According to this model, Desmostylus is considered to have had more upright forelimb, with elbow joint angle of approximately 160 degrees, whereas Paleoparadoxia with elbow joint angle of 130 degrees. The difference of elbow joint angles between these two genera may suggest a difference of stance and gait of these animals. Some ceratopsian dinosaurs, including Leptoceratops and ceratopsids, such as Centrosaurus, Anchiceratops, and Triceratops, also possessed the prominent olecranon processes. These animals were supposed to keep their elbow joints in approximately 140 degrees.

Crocodilians do not have the ossified epiphyses nor the prominent olecranon processes at their elbow joints. These animals do not keep their elbow joint angles by relying only on the leverages of olecranon processes. According to an anatomical observation of the bodies and kinetic analyses of three crocodilian species, including Alligator mississippiensis, Tomistoma schlegelii, and Crocodylus siamensis, their elbow joints were suggested to have restricted mobility which is due to surrounding soft tissues. The elbow joint extensions of crocodilians were restricted by skin, muscles, stiff tendons around the joint, and the hooking mechanism on the epiphyses between the cartilaginous olecranon fossa and anconeus process which are firstly reported in this study. Their elbow joint flexions were limited by the skin and muscles around the elbow joint, and the displacement of the radius.

The support mechanisms of forelimbs of ceratopsian dinosaurs are considered as fairly different among taxa. Within the ceratopsian dinosaurs, Psittacosaurus and Protoceratops did not possess the prominent olecranon process. Insertional tendons of triceps muscles of these animals were suggested to be bent along the caudal surfaces of their humeri at the propulsive phases. In contrast to the other ceratopsian dinosaurs with prominent olecranon processes, such as Leptoceratops and ceratopsids, Psittacosaurus and Protoceratops did not rely on the leverage of the olecranon processes.

絶滅した脊椎動物の姿勢復元に関する研究は多く行われている.しかし脊椎動物の骨化石関節を保ったまま化石化したものは稀で,そのような良好な状態の化石でも、生息時の姿勢は確認することはできない.従来の絶滅脊椎動物は主観的に姿勢復元がなされる例がほとんどで,その復元には理論的根拠の乏しいものが多かった.本論文で藤原君は,まず骨格の力学的なバックグラウンドを重視し,定量的かつ客観的な復元モデルを構築した.さらに現生脊椎動物の骨格標本をもとに,そのモデルが様々な現生脊椎動物に当てはまることを確かめた後,それを絶滅脊椎動物化石に当てはめ,姿勢の復元を行った.つまり充分なデータと力学的な解析手法に基づくモデルを用いて,過去の脊椎動物の姿勢復元を行った.本論文の研究は古脊椎動物学において先駆的な業績で,非常にレベルの高い研究である.

本論文ではまず四肢動物の前肢の位置に関する考察を行っている.従来,絶滅した四肢動物の肩帯の位置,つまり前肢が前後に伸びる胴体のどの部位に位置していたのかを復元することは容易ではなかった.これは前肢が肋骨や椎骨には直接関節せず,単に筋肉(腹鋸筋)を介して両者が繋がっているためである.本論文は,その前肢の位置を腹鋸筋が付着する位置にある肋骨の力学的な特性から明らかにするモデルを提唱した.四足歩行動物の胴体は肋骨を介して,前肢から筋肉でつり下げられた関係にあり,腹鋸筋で体を支えている.そして肋骨は前肢から伸びる腹鋸筋が付着した部位と前肢とは関係しない部位ではその力学的特性が全く異なるはずである.つまり前肢の付着に関与する肋骨は力学的に強く,これに関与しない肋骨は比較的強度が弱いはずである.藤原君は11種類の現生脊椎動物について調べた.肋骨の強度は有限要素法を用い,骨形態を2次元のデジタルデータで示し,実際の不可が肋骨の末端に鉛直方向にかかったとき,肋骨の各部位にかかる負荷の大きさを定量的に示した.その結果,現生の四肢動物では前肢の付着部位にある肋骨の強度は他の部位の肋骨に比べて明瞭に強い強度を示した.そして化石の肋骨形態に対しこのモデルを当てはめ,どの肋骨が末端に与えられた効力に対して最も大きな効力まで耐えることが出来るのかを有限要素法で確かめ,その部位に前肢が付着していたとする推測を行った.この手法を用いることにより,一部絶滅動物の前肢の付着位置を推定することに成功した.

さらに藤原君は化石四肢動物の前肢骨格がどのような角度で関節し,主にどのような角度で歩行時や直立時の姿勢を保持していたのかを類推するモデルの構築を行った.肘関節の角度は可動性に富んでいるため,絶滅動物ではその角度を推定することは容易ではない.この研究では,まず現生25属の前肢骨格標本とこれらの動物の立脚時の姿勢を観察し,これらの動物が尺骨の上面に突出する肘頭突起の上に付着する筋肉が最も大きなモーメントを発揮できる角度で,すなわち上腕と肘頭突起のなす角度が直行するように前肢姿勢を保っていることを確認した.そしてこの関係,すなわち肘頭突起と上腕がほぼ直角に保たれる角度を化石四肢動物に適応し,これらの前肢の姿勢復元を行うことに成功した.特に従来,様々な解釈がなされてきた新第三紀の哺乳動物の中の束柱類(デスモスチルス類),つまりデスモスチルス属やパレオパラドキシア属の生息姿勢を復元したところ,デスモスチルス属ではかなり直立した姿勢を,パレオパラドキシア属では前肢をやや屈曲させた姿勢を保持していたことが類推された.また中生代に繁栄した角竜類のうち四足歩行するレプトケラトプスやケラトプス科のアンキケラトプス,トリケラトプスの肘関節がほぼ140度に保たれていたことが推定された.また本論文では現生ワニ類の解剖学的観察により,肘頭突起を持たないワニ類がこのような突起に頼った姿勢保持をせず,肘関節の可動性が軟組織によって制限されていることも示された.

なお,本論文のうち,第2章は犬塚則久,桑流流理,吉川暢宏との共同研究であるが,いずれも論文提出者が主体となって研究を主導し,筆頭著者としてまとめる予定であり,論文提出者の寄与が十分であると判断される.

以上のように,様々な現生四肢動物の関節形態の力学的解析や関節角度のモーメントの考察からモデルを構築することにより,従来様々に復元されていた化石四肢動物の理論的な姿勢復元が可能になった.本論文は古脊椎動物学の古生態学において先駆的かつ非常にレベルの高い研究である.審査会では,本論文で用いられた新たなモデルにより,理論的に裏付けされた姿勢復元が化石動物に大きく道を開いた点,そして同様の力学的解析が古脊椎動物学における将来的な研究の可能性を与えた点を高く評価し,審査委員全員で博士(理学)の学位を授与できるとの判断を下した.