One of the most certain and fundamental points in the classification of the Mammalia is, that all the animals now composing the class can be grouped primarily into three natural divisions, which, presenting very marked differential characters, and having no existing, or yet certainly demonstrated extinct, intermediate, or transitional forms, may be considered as subclasses of equal value, taxonomically speaking, though very different in the numbers and importance of the animals at present composing them. These three groups are often called by the names originally proposed for them by Blainville—(1) _Ornithodelphia_, (2) _Didelphia_, (3) _Monodelphia_—the first being equivalent to the order _Monotremata_, the second to the _Marsupialia_, and the third including all the remaining members of the class. Although actual palæontological proof is wanting, there is much reason to believe that each of these, as now existing, are survivors of distinct branches to which the earliest forms of mammals have successively given rise, and for which hypothetical branches Professor Huxley has proposed the names of _Prototheria_, _Metatheria_, and _Eutheria_, names which, being far less open to objection than those of Blainville, are here used as equivalents of the latter. The only known existing PROTOTHERIA, although agreeing in many important characters, evidently represent two very divergent stocks, perhaps as far removed as are the members of some of the accepted orders of the Eutheria. It would, however, be merely encumbering zoological science with new names to give them any other than the ordinarily known family designations of _Ornithorhynchidæ_ and _Echidnidæ_. Similarly with regard to the METATHERIA, although the great diversity in external form, in anatomical characters, and in mode of life of the various animals of this section might lead to their division into groups equivalent to the orders of the Eutheria, we do not think it advisable to depart from the usual custom of treating them all as forming one order, called Marsupialia, the limits of which are equivalent to those of the subclass. The characters of the six families which compose the group are extremely well marked and easily defined; and since they form a regular gradation between two extreme types, they can be satisfactorily arranged in a serial order. A marked distinction in the dentition enables us to divide them into primary groups or suborders.