EARLY ONTOGENETIC PHENOMENA IN MAMMALS. 157 
presence or absence of amnion and allantois is to a great 
extent artificial. 
When embryology no longer forces us to go on extending 
the distinction between the so-called Amniota and Anamnia 
into the palzozoic period, certain lines in comparative anatomy 
may perhaps suggest a new grouping in which also that 
other inadequate test, the double or single occipital condyle, 
is relegated to its real value. But then the paleontologist 
who will go deeper into this matter should bear two other 
points in mind which both this investigation and numerous 
other researches in comparative anatomy have brought to 
light of late years, viz. that the mammalian characteristics 
bring us down to a point where comparison with the lower 
Amphibia—as Fiirbringer (’00) has more especially advocated 
—is more ad rem than comparison with the more specialised 
reptiles; secondly that the Ornithodelphia should be looked 
upon as a sub-class by itself, small at present, but perhaps 
more extensive long ago (Multituberculata), in which 
sauropsidan and mammalian characters are curiously com- 
bined but which was never in the direct line of descent of 
Mono- and Didelphia.' Then, again, that these latter may be 
said to bea very specialised side branch of ancestors that were 
1 I wish here to refer to a passage in an interesting article by Wortman 
(°03) on the origin of mammals (I. c., p. 429). He says, “ Early in the meso- 
zoic there appeared smail, mammal-like forms, which were widely distributed 
over both the northern and southern hemispheres. Representatives of these 
species continued throughout the Cretaceous, and finally disappeared in the 
early stages of the tertiary. . . . . Many of them are classified in the 
group Multituberculata, which, without much doubt, finds its nearest living 
representative in the Duckbill of Australia. . . . . In one instance a 
fairly complete skull is known (Tritylodon) from the Karoo-beds of South 
Africa. The teeth of this species are astonishingly like those of many types 
in the northern hemisphere, and hitherto it has always been classified in this 
group. Seeley has shown that the organisation of the skull presents so many 
reptilian characters as to cause him to refer it to the Reptilia. If this 
reference is correct, then, in the absence of any fact to the contrary, it is 
highly probable that all the multituberculates are as much reptile as mammal. 
Indeed, it is not easy to say, at first glance, upon which side of the line living 
monotremes should be placed. There can be Ilttle doubt that, when more 
