The strange features of the skull and dentition of Thylacoleo, along with the scarcity of postcranial material, brought about a fervid and prolonged debate regarding its feeding habits.

    In all of his works, Owen fostered no possibility other than than that of a carnivorous lifestyle for Thylacoleo and interpreted the animal as "one of the fellest and most destructive of predatory beasts" (1859).  Broom (1898) and Woods (1956) concurred with Owen while Glauert (1912) was more of the opinion that Thylacoleo was a scavenger.  The discovery of fractured and incised bones led De Vis (1883), W. Anderson (1889), Zeitz (1907) and Spencer and Walcott (1911) to assert that the carnassial teeth of Thylacoleo functioned as bone crushers.  However, Cope (1882) suggested crocodile eggs or carrion as having been its staple diet.  Bensley (1903) made note that the phalangerid stock had been of an omnivorous-herbivorous nature in which there was an orientation towards the reduction in size of the canine teeth.  He postulated that Thylacoleo had evolved from such ancestors towards a carnivorous diet, which would account for many of the curious features of its dentition.

    However, a number of palaeontologists suggested that Thylacoleo may have been a herbivore.  Flower (1868), Dawkins (1864), Krefft (1866) and Lydekker (1894) all wrote of plant food but, baffled by the down sizing of the molars, added that fruits or perhaps an occasional bird or mammal may have been eaten.  C. Anderson (1929) put forward the idea that cycad pith or fruits of the family Curcubitaceae might have formed the diet of Thylacoleo, while Gill (1954) provided a discerning deliberation of the mystery but put forth no suggestion for what the animal's food source may have been.

    Because of the bewildering group of arguments surrounding the nature of Thylacoleo's diet, a reassessment is necessary of the morphological characteristics which give evidence for its feeding habits and masticatory functions.  A more in depth presentation of Thylacoleo morphology is provided in Finch (1981).
 

    Thylacoleo possesses a brachycephalic skull structure, such as is seen in many other carnivorous mammals.  Its jaws are shortened, bringing the canine teeth closer to the condyle and thus allowing greater force to be applied upon them by the adductor muscles.  Also, the reduction of the rostrum has allowed the eyes to be directed anteriorly, providing binocular vision and thus giving a greater degree of visual acuity which is of assistance in hunting activities.

    Thylacoleo's zygomatic arch is deep and widely divergent from the cranium, supplying an expansive area for muscle attachment and also providing space for a large temporalis muscle.  The size of the latter is also evidenced by the well-devoloped sagittal and lambdoidal crests from which a section of the temporalis originates.
 

The skull of Thylacoleo carnifex.

The postorbital bar, a very rare structure in carnivores, is also found in Thylacosmilus the "marsupial saber-tooth" of South America, and may have served as an area of attachment for temporalis fibres and the aponeurosis which overlays the entire muscle.  It also shields and secures the eyes from pressure by the contracting jaw muscles.  This bar also acts as a supporting structure for the anterior section of the jugal bone which is subjected to compression and tension at the maxillo-jugal suture through piercing by the incisors and slicing by the carnassials respectively (Buckland-Wright 1978).

      No clear boss or bony projection from the anteroventral margin of the zygomatic arch is to be seen, suggesting that the superficial masseter muscle was relatively weak.  Thylacoleo's zygomatic arch is deep and widely divergent from the cranium, supplying an expansive area for muscle attachment and also providing space for a large temporalis muscle.  The size of the latter is also evidenced by the well-devoloped sagittal and lambdoidal crests from which a section of the temporalis originates.  The postorbital bar, a very rare structure in carnivores, is also found in Thylacosmilus the "marsupial saber-tooth" of South America, and may have served as an area of attachment for temporalis fibres and the aponeurosis which overlays the entire muscle.  It also shields and secures the eyes from pressure by the contracting jaw muscles.  This bar also acts as a supporting structure for the anterior section of the jugal bone which is subjected to compression and tension at the maxillo-jugal suture through piercing by the incisors and slicing by the carnassials respectively (Buckland-Wright 1978).  No clear boss or bony projection from the anteroventral margin of the zygomatic arch is to be seen, suggesting that the superficial masseter muscle was relatively weak.
 

The condyle is a laterally-elongated bar some 45 mm in width.  Its dorsal surface is somewhat flattened and it rests in a glenoid fossa, its articulating surface being rather convex ventrally.  A well defined postglenoid process prevents backward movement of the condyle, however, there is no preglenoid barrier.  This condition is comparable to that of the phalangers, particularly that seen in the omnivorous Phalanger maculatus and is not analogous with the typical placental carnivore condition in which only orthal movement is possible.

Right mandible of T. carnifex from the Wellington Caves, New South Wales.

    The condyle is low on the dentary, and on a level with the tooth row.  The condylar neck is robust and short while the coronoid process is wide and high.  This suggests that the animal had a powerful, scissors-like jaw closure.  The thick condylar neck is resistant to force, and the low articulation lengthens the effort lever arm of the temporalis muscle by amplifying the distance between the fulcrum and the centre of the insertion area of the muscle (Scapino 1972).