The best test of the relations in density between molten and 
solidified rock is apparently to compare the density of the rock — 
just before fusion, or at least near that point, with the density o — 
the same rock after melting. This would give a comparison be 
tween the crystalline and liquid states, while the usual method 
only affords a comparison between the liquid and the glassy or 
semi-glassy states. It would also save any error arising from 
cells in the cooled rock, if a solid mass was chosen in the first 
place. Again the fresh unaltered varieties of a rock should be 
chosen instead of such old and altered ones as those usually et 
perimented upon. _ , 
In all discussions relating to the question of the liquidity of 
the earth’s interior, it is to be borne in mind that the chief portion 
of our knowledge of the properties of liquids is derived from the 
study of water, a mobile liquid—while liquid rock, as lava of 
melted iron, is viscous, and its laws and properties may om & 
periment be found to differ considerably from those of ye 
under like conditions. Also in these solids the passage gory 
solid to the liquid state or the reverse is not abrupt as ® the : 
case with water, for every grade of viscosity exists rsa 
normal solidity and the approximately perfect liquid eer 
This is especially the case with iron and seems to be so 
common rocks. 
686 The Tertiary Marsupialia, 
(Zo be continued.) 
:0: 
THE TERTIARY MARSUPIALIA. 
BY E. D. COPE. 
ti Cuvier discovered an opossum in t ; 1 
the knowledge of the Marsupials of the Tertiary bb 
Europe and North America has been gradually sigs ee they 
Europe they have been traced to the Middle wre thet 
disappear from that continent. In North America and are 
from Oligocene beds (White River), when they was gust 
only known as yet thereafter as members of the pe j 
Descending the scale we have them in the Ma Tisi South a 
and Jurassic in America and Europe, and in the per " 
Africa. Whether the Triassic Mammalia of the Under th head 
sphere belong to this order or not is pee ni 
of Creodonta? I have discussed the marsup! 
1 NATURALIST, March, April and May, 1884. 
he gypsum of be : 
