34 
dorsal wall, very narrow in the ends, where they run parallel with 
the cutting plane in a sagittal section, consequently they must be 
examined on a transverse section of the bladder. The heads of the 
pink stripes diminish and so far as I can see finally disappear at 
the ends of the bladder, however they are distinct both in the ven- 
tral and the dorsal wall. For details and measures of the various 
parts vide the accompanying figures. 
In the lake form the differences between the various thicknesses 
of the wall are still more marked than in the pond form, although 
the general structure is the same, still it must be noted that there 
are two dark layers, thin but distinct, the one at the distal side, the 
other at the proximal side of the chitine wall (Fig. 4). Apparently they 
are not of any chitinous matter; in the pond form. I have found no 
trace of them, and their nature seems rather puzzling. But about 
the nature of the other parts of the chitine wall there can hardly 
be any doubt; the pink "heads" correspond to the thin continuous 
wall in normal tracheæ, the purple stripes are homologous with 
the spiral ribs, and the narrow pink stripes are merely outgrowths 
from the original thin wall to unite the very long spiral ribs in 
order to give the wall sufficient strength. And it is very easy to 
understand why these changes have taken place; the air-bladders 
need a certain strength to stand a varying pressure without col- 
lapsing, consequently their walls must be thick and massive. And 
as the slightly curved sides are more exposed to attacks from the 
pressure than the strongly curved, hemispherical or conical ends, 
their walls must be thicker; and the concave ventral wall being 
still more exposed than the convex lateral and dorsal walls con- 
sequently must be still more massive (Fig. 5). Thus we see that 
the structure of the air-sacs is in perfect harmony with the phy- 
sical claims of the surroundings. And in correspondence with this 
harmony between milieu and structure we find that the thickness 
of the walls of the lake form is on an average twice as great as 
that of the pond form, as we were entitled to expect. The thick- 
ness of the lateral walls of the air-sac of a full-grown pond larva, 
measured on a transverse section, is about 8 u, but of a lake 
larva it is about 16 w (Fig. 6). This gives the explanation of the 
different resisting power of the two kinds of larvæ against increa- 
sed water pressure, The different surroundings have developed 
