CEASSIFICATORY CHARACTERS 
upon as ancestral to both. The method of Paleon- 
tology, however, has at present to be supplemented by 
comparative anatomy. Relationship not only implies 
common ancestry but likeness of structure, which is 
more or less marked according to the degree of relation- 
ship. But here we leave the region of fact for that of 
inference. Paleontology, in proportion to the com- 
pleteness of the records in the rocks, speaks with a 
certain voice as to,relationships. Similarity of structure 
has to be sifted and checked in various ways before it 
can be relied upon as evidence. We will take some 
cases, beginning with extremes, to illustrate the matter. 
The paca is a large brown rodent, with white spots. 
The dasyure is an equally sized Australian marsupial, 
also with a brown coat, diversified with white spots. 
Why don’t we, or do we, put together these two 
animals? We do not admit a close relationship for 
the following reasons :—In the first place white spots 
ona brown or dark ground is a plan of coloration 
found in many diverse kinds of animals between which 
no intimate relationship is at all possible, which are 
of course a priovt conditions between two mammals. 
Molluscs, insects, and other creatures have a host of 
representatives which possess this character. It would 
be tantamount to bracketing together green animals 
and white animals, black creatures and transparent 
creatures, a procedure which would involve too many 
obvious absurdities to be entertained for a moment. 
A second reason forbids us from laying any stress on the 
white spots. In South America, where the paca lives, 
there lives also the agouti. In the Australian continent 
and its outlying islands, where we find the dasyure, 
the Tasmanian devil is also to be met with. Now 
neither the agouti nor the Tasmanian devil are brown 
‘spotted with white; and yet, in every other detail of 
external and internal structure, the Tasmanian devil 
15 
