GAS GLANDS OF SOME TELEOSTEAN FISHES. 189 



mechanii^m must he for some definite purpose, especially since it 

 is found in all fishes possessing gas glands. What that purpose 

 may be is a question I shall discuss later (see Part II.). 



The epithelium lining the ductus pneumaticus, Avhich is sup- 

 plied with a few branches direct from the bladder artery and is 

 therefore not supplied by the vessels connected with the retia 

 mirabilia, is quite simple in character. It merely consists of a 

 single layer of flattened cells raised up at intervals by the large 

 capillai'ies running in the basis of connective tissue underlying 

 the epithelium (PL II. fig. 7). These capillaries in the wall of the 

 duct apparently have much the same surface ari-angement as those 

 found in the lungs of certain reptiles (Quekett, 61), but it is 

 hard to suppose from this fact alone that the epithelium of the 

 duct is normall}^ respiratory in function, since the blood supplied 

 to this has already been oxygenated in the gills ; on the other 

 hand, it is equally improbable that the ai'terial blood in the wall 

 of the duct serves to alter the composition of the adjacent gas 

 since there is reason to suppose that the duct normally serves as 

 a mere exit for the superfluous gas of the bladder (Jaeger, 45). 

 It is possible, however, that this large vascular supply of the duct 

 epithelium in the Eel is really correlated with the terrestrial 

 habits of the animal, since when an Eel travels across a meadow 

 the gills are useless for obtaining oxygen, and under these circum- 

 stances the animal doubtless draws upon the oxygen in the 

 bladder — the duct acts as a lung, Moreau proved that Perch, 

 when placed in water previously boiled so as to deprive it of 

 dissolved air, utilized the bladder- oxygen for purposes of respira- 

 tion. An interesting confirmation of this view is that in other 

 physostomes {Myrus, Ojyhichthys, Esox) the dtict epithelium is, as 

 we shall see, practically devoid of capillaries *. 



The lining epithelium of the bladder proper is very difierent 

 in chai'acter from that of the duct. It consists of a more or less 

 folded single layer of cells which are large in size and short 

 columnar in form, slightly gTanular, and, when stained by the 

 picro-indigo-carmine method (Appendix A), usually rather opaque 

 (PL II. fig. 8). The folding of this layer is due to the presence of 

 large capillaries which push their way inwards towards the bladder- 

 cavity from the connectiA^e tissue base, and in so doing cause the epi- 

 thelial layer to form corresponding emergences. Thus the iutei'ior 

 of each projection of the glandular epithelium into the bladder 

 cavity is occupied by an extension of connective tissue enclosing 

 capillaries. The spaces situated between such emergences adjacent 

 to each other are also regarded as forming gland ducts. The 

 folding of the epithelium is very vaiiable in difierent parts of the 

 bladder wall, in some parts almost disappearing and in others 

 being more accentuated than that depicted in figure 8 ; this 

 folding is much more marked in the anterior region of the 

 bladder than in the posterior. It might appear to be significant 

 that immediately the lining epithelium of the duct becomes sup- 

 plied with blood connected with the rete mirabile system of vessels 



* I iiiul that Jacobs (42, 43) status tliis view as a fact. 



