234 
cover-slip soon turned a deep brown. I sectionized one of these living 
preparations showing the evolution of the spheres and recognized in 
them the vacuolar spaces and extracellular bubbles of Bykowsk1, 
NusßAum and Reis and my own figures (15, fig. 35, pl. V, e.g.). I 
also examined under high magnification the cells of the active glands 
of four of five fish which had been fixed respectively with 1°/) osmic 
acid, Mann’s fluid and absolute alcohol, the fixative being in every 
case pumped into the bladder through a hole made in its wall, but 
again I could detect no gas bubbles inside the cells. 
After making these experiments it occurred to me that possibly 
the very act of pricking the bladder and so releasing the gas gland 
cells from the pressure of the contained gas might cause all the gas 
bubbles inside the cells suddenly to be extruded, the cells in con- 
sequence appearing devoid of gas bubbles on subsequent examination. 
To obviate this objection I carefully peeled off the outer muscular 
layers of the bladder wall covering the region of the gland (a very easy 
matter) and fixed the gland with 1°/, osmic acid from the exterior, 
the bladder still remaining inflated. But subsequent examination still 
revealed no gas bubbles, though all the gland cells stained an intense 
brown, which indicates I presume the presence of lipoid material. 
I determined further to examine the living cells whilst still under 
pressure. To do this I again carefully peeled off all the external 
muscular layers of the bladder wall, leaving practically only the thin 
internal cellular layer strengthened by a small quantity of connective 
tissue. I then cut away the body wall at the sides and placed the 
whole fish on the stage of a microscope strongly illuminated from 
below. On reflecting a beam of light through the gland situated in 
the now transparent ventral bladder wall and (in the absence of a 
water-immersion lens) placing a drop of oil on its exterior I was able 
to focus a 2 mm. immersion lens on to the edges of the gas gland and 
so observe the living cells whilst still subject to the pressure of the 
gases contained in the inflated bladder. Though I performed this ex- 
periment on two fishes, both of which had been weighted for 6 hours, 
I was quite unable to detect gas bubbles. 
As the result of these experiments conducted upon the active gas 
glands of some dozen Pollack, I am forced to the at least provisional 
conclusion that the oxygen abstracted from the blood by the cells of 
the gas gland does not assume the form of intracellular bubbles during 
its transference into the bladder cavity. It therefore appears that 
JAEGER is right in asserting that the spherical intracellular spaces 
figured by ByKkowskı, NusBAum, REIS and myself merely represent 
