29 
2. Tlie preceding ex])ei iment was repeated, excepting that the animal was placed in normal 
salt solution instead of water. The result was that the chamhers contained hlood plasma, but per- 
haps less in quantity than in the previous case. 
In this case the two fluids were of approximately e(j^ual densities and, it may he inferred, 
tended to replace the air in equal measure. Allowance must be made, however, for the greater 
thinness of the wall l)ounding the hlood-chamher than the outer chitinous wall. 
3. The experiment was repeated using a concentrated salt solution. The result was that 
the chambers appeared as empty, or nearly so, little blood plasma having entered them. 
In this case it woiild appear that the more concentrated fluid replaced, in the main, the air. 
In all three cases the results become intelligible on the supposition that the chambers con- 
tain air. Thus these experiments wai’rant the inference tliat the chambers normally contain air 
and not blood. 
In considering more precisely the function of the air chamhers we must first take account 
of the fact that they do not serve to increase the area of respiratory sm-face. They are not, as 
in the case of PorccMie and its congeners formed by an inward fold of the wall of the gill but are 
simply spaces tilled with air, lying between the parallel walls of the blood cavity and the gill. But 
it is to be noted that a provision for exposure of a large surface of hlood to air is made by the 
expandet dimensions of the s]jecial part of the gill and by the presence of these parts in all five 
pairs of the outer gills. 
In studying the outer gills of PorceJlio we reached the conclusion that the respiratory tree 
is an adaptation having a two-fold purpose ; namely, first, to provide for a large surface of exposure 
of hlood to air and, second, to secure the protection of the Iflood against dessiccation from exposure 
to air in the process of respiration. Now, in Oniscus, the former of these purposes hehig provided 
for by the means just noted — an increase of the total area of the special respiratory surface — 
it is left to the air-chambers to prevent a too great loss of water from the blood in the process 
of respiration. One may easily conceive that if nothing intervened between the blood and the out- 
side atmospheric air but the two very thin membranes of the wall of the hlood cavity and the chi- 
tinous wall of the gill, the two being in contact, dessiccation would follow. Again, phylogenetic 
considerations lead us to expect that the respiratory siu'faces must he maintained in a moist state 
as a condition of their functional action. 
In the respiratory process the air of the chambers becomes charged with water of respiration. 
The only means of escaj)e of this water to the outer atmospheric air is by passage through the 
chitinous wall. This process takes place slowly and secures tlie maintenance of the moist condi- 
tion of the air in the chambers. This air is thus separated from the blood by only a very thin 
moist membrane. Thus conditions most favorable to respiration are provided ; the blood is both 
shielded from dessiccation and at the same time separated from air by only a very slight barrier. 
I reach the conclusion that through the possession of these special modifications, the outer 
gills of Oniscus constitute organs for breathing ordinary dry atmospheric air. The several collateral 
reasons in support of this conclusion that have already been given for Porcellio apply also for Oniscus. 
The Inner Gills. The inner gills of Oniscus agree in all essential features with the 
corresponding parts in Porcellio and the other genera already described. 
