178 REPORT—1905. 
This is clearly not due to the increase of osmotic pressure ; the same 
must be said of all the substances of Class A. 
As a matter of fact, there is not any evidence that the egg does 
normally absorb water to any extent while the closure of the blastopore 
is taking place ; it is only after the blastopore has become reduced to a 
small circle, and when the medullary folds are being formed, that the 
innermost membrane of the jelly begins to expand and the embryo to 
elongate. Prior to this, therefore, it does not seem likely that the egg 
would be sensitive to a loss of water. Afterwards, however, we do know 
that the embryo absorbs water very rapidly, for Davenport has shown 
that during the first fortnight after hatching the percentage of water rises 
from 50 to nearly 90. By observing the effect produced on tadpoles by 
these solutions it seemed, therefore, possible to ascertain the permeability 
of the tissues to the various substances, and so obtain a sounder basis for 
judging of the effects of these solutions upon the earlier stages. 
The tadpoles were accordingly placed, shortly after hatching, in solu- 
tions—of the same concentrations as those already employed—of cane 
sugar, sodium chloride, urea, and sodium sulphate. 
Tn the first two the tail becomes curled up over the back ; the mouth 
becomes, as in the controls, transversely elongated, the suckers oval, the 
operculum begins to grow back over the three external gills ; then the 
larvee die. 
In urea the tail remains straight ; in other respects the tadpoles 
resemble the last ; these also die in about five or six days after immersion 
in the solution. 
Microscopical examination shows that differentiation has been going 
on at nearly the same rate as in the controls; the cavities of the gut, 
brain, pronephros and ducts, lungs, ccelom, blood-vessels, are all, how- 
ever, much reduced ; they are most reduced—in fact, almost obliterated— 
in cane sugar, less so in sodium chloride, least in urea ; in sodium 
sulphate they do not appear to be reduced at all. 
This result seems to show pretty clearly that the embryonic tissues 
are less permeable to cane sugar than to sodium chloride, less to sodium 
chloride than to urea, and less—a very little less—to urea than to sodium 
sulphate ; the two latter, however, seem to be nearly harmless, and, though 
free to permeate the body of the larva, do not produce, at any rate at 
first, any marked deleterious effect. 
That these solutions do penetrate is shown, possibly, in another way. 
When a normal tadpole is preserved, as these were, in picric acid, the 
chordal sheath becomes crumpled, owing to the extraction of water from 
the notochord ; the sheath of the cane sugar tadpoles is also crumpled, 
but not in sodium chloride, sodium sulphate, or urea. 
It may be noticed that in cane sugar and sodium chloride the 
diameter of the notochord is greater than the normal ; apparently this 
organ continues to absorb water from the tissues and to grow in length 
(as well as in breadth) ; the dorsal caudal is wider than the ventral 
caudal fin; the dorsal curvature of the tail in these two solutions may 
thus possibly be accounted for. 
Assuming that the permeability of the embryo to the different sub- 
stances is the same in earlier stages as in the tadpole, these results may 
be applied to the interpretation of those obtained before as follows :— 
1. The embryo is probably impermeable to dextrose ; the effects pro- 
