108 



FOSSIL MAMMALIA OF THE 



Messrs. Krefft (fig. 25) and Flower (fig. 26) restore the mandible of Thjlacoleo, in 

 reo-ard to these hght-giving structures, according to the analogies of the carpophagous 



Fig. 25. 



Fig. 26. 



Restoration of the skull and teeth of Thylacoleo, by Mr. KreflFt, 

 on the herbivorous hypothesis (' Ann. and Mag. 

 of Nat. Hist.; 1866, pi. xi). 



Restoration of the skull of Thylacoleo, by Professor Flower, 



on the herbivorous hypothesis (' Quart. Journ. Geol. 



Soc.,' 1868, vol. xxiv, fig. J, p. 312). 



Phalangers and Koalas, and the poephagous Potoroos. But this restoration is shown by 

 the specimen before me (fig. 15) to be erroneous. The dentition of the upper jaw 

 restored on the same analogies is proved by the specimen figured in my last Memoir on 

 Thylacoleo, communicated to the Royal Society, to be equally unfortunate. 



The labours of Dr. Falconer are worthy of highest praise ; but loyalty to our common 

 science compels me to say that he fell into his mistake as to Flagiaulax by neglecting its 

 fundamental principle. (See Note, p. 96.) 



Observing a general community of trenchant character, with grooves and ridges, 

 between the premolar in certain Poephaga and one of the premolars of Flagiaulax, Dr. 

 Falconer jumped at the conclusion that the Flagiaulax was an herbivorous or ' Rodent ' 

 Marsupial like the Kangaroo-rat [Hypsiprymnus). 



The one-and-twenty pages of the ' Quarterly Journal of the Geological Society,' and 

 the forty-three pages of the posthumous volume of his ' Palaeontological Memoirs,' devoted 

 to a defence of this conclusion, have nevertheless seemed ' masterly ' to some,^ ' amply 

 demonstrative ' to others.^ Something, therefore, needs to be said, by way of warning to 

 beginners in Palaeontology and in vindication of the noble science itself. 



A few words, first, as to the extent to which affinity or conformity of organization, 

 involving proportions and structure of limbs, resultant specialities of locomotion, habits, 

 and food, may be predicated or inferred from the occurrence of secondary characters, such 

 as grooves and ridges, in homologous teeth, not being molars. Certain Qiiadrumana have 



1 Prof. Flowek, &c., op. cit., p. 307. 



2 Mr. Boyd Dawkins, as quoted at p. 95. 



