GAS GLAXDS OF SOME TELEOSTEAN FISHES. 239 



bubbles. Gas glands must be in a very active condition for the 

 majority of cells to exhibit bubble- structures (see PI. V. fig. 35 of 

 Gobius minuius). These bubbles were quite absent in the sections 

 of non-weighted Perch used for the purpose of controlling the 

 cxpeiiment. In my dacial- Absolute material, structures which 

 are presumably bubljles (very unlike the bubbles sten in most 

 preparations, however) are also to be found in the act of extru- 

 sion, but these instead of being unmistakable foam-like masses 

 ai-e bladder-like structures projecting from the cell-periphery 

 (PI. IX. fig. 71 and text-fig. (32, the latter being a microphoto- 

 grapli) and some are to be found lying free in the bladder-cavity. 

 Possibly the action of the fixative is to cause all the small bubbles 

 of the foam-like mass to run together to form one big bubble. 

 In the surface-view preparations tixed by this method, large gas 

 bubbles are occasionally to be found in the cytoplasm. All these 

 bubble-structures are quite absent in the control material fixed 

 with Glacial- Absolute. 



Apart from the existence of gas bubbles in preparations of the 

 activated gland, another distinction from the non-active or dormant 

 gland is the "worked" or "spent" appearance of many of the cells, 

 a distinction w'hich I have tried to indicate in figures 71 and 72 

 (71 represents cells of the active gland — cf. text-fig. 62 ; 72 repre- 

 sents dormant cells). It v»'as this "spent" appearance, this 

 attenuated and ragged condition of the cytoplasm, which suggested 

 to Nusbaum & Reis that the cell substance actually decomposed 

 into the bladder gases and granular masses, just as the cells of the 

 sebaceous gland disintegrate to produce the secretion of that gland, 

 and, indeed, at first sight of these exhausted cells of the gas gland 

 the suggestion does not seem unreasonable. However, the fact 

 that a cell which posseses but a small amount of protoplasm in the 

 space bounded by its walls is not necessarily in a state of decom- 

 position is shown by the majority of plant-cells, the cytoplasm 

 of which is reduced to a " primordial utricle," by swollen adipose 

 cells, by spicule-cells and by yolk-laden meroblastic eggs, to men- 

 tion a few out of many possible examples, and since no amoiint of 

 cytological technique can demonstrate the actual chemical decom- 

 position Nusbaiim & Reis assume, this attenuation of the 

 substance of the cells of the gas gland is not of very great 

 significance, especially in view of the reasons for rejecting this 

 suggestion of Nusbaum & Eeis alieady advanced in Part II. It 

 must not be supposed that the cells of my activated glands alone 

 wear this "spent" aspect, all the cells of the dormant gland 

 resembling figure 72 ; on the contrary, the only difference between 

 the cells of the two glands is one of degree, cells of the dormant 

 gland being, on the whole, less ragged than those of the active. 

 Only a gas gland completely at i"est (a condition probably rarely 

 attained in a voracious fish like the Perch) would have all its cells 

 in a resting condition. In most marine fish, however, which I 

 have studied, nearly all the cells of the resting gland have been 

 quite solid in appearance, only the cells of the active gland being 

 attenuated. 



