OF THE AUSTEALIAN MAESTJFIALIA. 85 



The precise distinction of Huxley between the ancestral or metatherian cliaracters of 

 the Marsupialia and their special characters definitely removes the latter animals from 

 an ancestral to a collateral position with reference to the Placentals. The principle is 

 essentially one of reduction in number of the characters in which the Marsupials appear 

 to occupy a prototypal position. For example, the single-tooth change of these animals 

 ATas regarded by Flower (1867) and afterwards by Thomas (1887) as prophetic of the 

 more complete tooth-change of Placentals, but was pointed out by Huxley as a special 

 character of the Marsupials and as derivative of a former diphyodont condition common 

 to the Metatherian ancestors of both Marsupial and Placental groups, but more com- 

 pletely retained in one of them than in the otlier. The correctness of this position has 

 been amply confirmed through the embryological investigations of several zoologists, 

 including Rose, Leche, Kiikenthal, Woodward, Wilson and Hill, and Dependorf. 



Recent research has, however, gone still further in the reduction of the recognized 

 prototypal characters of Marsupials, and certain observers, namely, Wilson and Hill, 

 and Dollo, are in favour of removing the most essential one, namely, the possession of a 

 non-placental allantois, which they regard as a secondary feature of the Marsupials and 

 as derivative of a former placental condition. This view is based on Hill's discovery of a 

 placental connection in JPerameles ; and the reduction of the allantois in other Marsupials 

 is supposed by Dollo to have been associated with premature birth and fundamentally with 

 arboreal habit. In a former paper (1901 b), however, the present writer has pointed out 

 the probability that the aplacental condition of most Marsupials is actually jirimitive, 

 and that the placental connection in Perameles, like a multitude of other characters in 

 which Marsupials resemble Placentals, has been independently acquired ; in other words, 

 that it represents a convergent or homoplastic * development. 



However this may be, we have the more definite fact that the Marsupials and 

 Placentals are collateral and, in a certain sense, equivalent groups, of common parentage ; 

 and this conception may be welcomed as clearing the way for a better perception of the 

 d' tails of their secondary evolution or adaptive radiation. Especially is this true of 

 the former: any attempt to explain their secondary differentiation on the basis of a 

 Marsupial ancestry of the Placentals must naturally result in confusion, because of the 

 lack of distinction between those of their ada^itive characters which are, generally 

 speaking, common to all of the members of the group and those distinctive of minor 

 divisions, leading to the doubly erroneous conclusion that tiie evolution of the Marsupial 

 group is not finite, and that characters of both kinds have been carried over tVom 

 Marsupial to Placental stages. At the present time, while the evidence at our disposal 

 may not be of sufficient extent to furnish a complete plan of the whole Marsupial 

 radiation, we at least have the advantage of being able to form a clear conception as to 

 what the problem involves. Referring to the plan above presented, the whole situation 

 in Marsupial phylogeny may be summed up in three questions. In the first place. 

 What are the characters of the ancestral Marsupio-placental or Metatherian forms ? 

 Some of these characters, as, for example, the narrow cranium and projecting zygomata, 



* Cf. recent paper on Ilomoplasy by Osborn (1902). 



13* 



