GAS GLANDS OF SOME TELEOSTEAX FISIJES. 197 



gland is normally (juit^seent *. To deny the existence of these 

 cytoplasmic bubbles simply because cei-tain microscopic pre- 

 parations do not show them is illogical, to say the least, since 

 intracellular gas babbles, unlike intracellular capillaiies, are not 

 pei-manent but transitory structures. Ot" over fifty series of 

 preparations which I have made of numerous types of gas gland, 

 not more than seven or eight show these intracellular gas bubbles 

 in an unmistakable manner. In the type of gas gland under 

 consideration, one only of my six or seven series of preparations 

 of the gas glands of different specimens of Gohius niger and 

 G. paganelliis exhibits a few gas bubbles (fig. 27). I have 

 seen these bubbles best in a preparation of Gohius Tninutus (PI. V. 

 fig. 35), in which the gas gland cells were fixed in the active con- 

 dition. As shown in figure 35, the majority of the gas-producing 

 cells possess large vacuolar spaces in their cytoplasm usually present 

 in the vicinity of the nucleus ; indeed, the nucleus is often so 

 adpressed as to assume a crescentic form. These vacuoles, which, 

 as just mentioned, are assumed to repiesent the moulds, so to 

 speak, which contained gas bubbles, just as liver- and kidney-cells 

 exhibit similar vacuolar spaces which contained liquid globules 

 (text-fig. 60, p, 225), are of various sizes, and some can be seen in 

 the act of being ejected from the cell substance into the bladder 

 lumen, where they are also occasionally found in a liberated 

 condition (see Appendix B). The bursting of these gas bubbles 

 is doubtless accountable for the masses of granular matter always 

 found, when the gland is active, in the bladder lumen and gland 

 ducts just external to the glandular epithelium, this granular 

 matter, of course, having composed the walls of the bubbles. The 

 gland cells, when active, always have that portion of their 

 cytoplasm situated next vascular tissue veiy distinct from the 

 rest, it being, as already described, very dense and often striated 

 in appearance ; the rest of the cj^toplasm usually assumes a 

 " stringy " appearance and is very vacuolate f. It may also be 

 mentioned that the nucleus is not situated in the dense cytoplasm 

 next the blood, in which respect gas gland cells show a mai'ked 

 difi"erence from the cells of the pancreas, e. g. 



In addition to intracellular capillaries and gas bubbles there 

 are also occasionall}' to be seen intracellular lumina or ducts — 

 continuations of the intercellular ducts into the substance of the 

 gland cells. These, again, vary greatly in different types of gas 

 gland and in different specimens of each type. In none of my 

 preparations are they very numerous, and doubtless, like the gas 

 bubbles, the smaller intracellular lumina ai-e transitory structures. 

 I have figured one or two of these structures in connection with 

 the gas gland of Oj^hidmrn harhatnm, a type very effectually 

 studied by Reis k, Nusbaum (62), and they are occasionally to be 

 found in my preparations of Gohius niger and Gobivs paganellus 

 opening into the numerous finer intercellular ducts. Since these 



* See Addenda (D- . . 



t Fov excellent descriptions and figures of the?e gas bubbles in the cytoplasm of 

 the sjland cells, see the papers of Hykowski & Nusbanm (24) and Keis & Nusbauni 

 (55,62,63). 



Proc. Zool. Soc— 1911. No. XIV. 14 



