GAS GLANDS OF SOME TELEOSTEAN FISHES. 241 



In connection with the two microphotogra.phs of gas bubbles 

 hei'e reproduced as text-figures 61 and 62, I may say that these 

 structures were in my preparations very difficult subjects for 

 photogi'aphy, partly because of the inappropriate staining em- 

 ployed and partly because bubbles shown in optical section of 

 necessity bear but little resemblance to the real thing (this was 

 especially the case with the foam-like mass indicated in text-fig. 61 ; 

 cf. fig. 70). Nevei'theless, these microphotographs, taken for me by 

 Mr. F.J. Pittock, of the Zoological Department, University College, 

 are perhaps of value as affording impai'tial evidence of the 

 appearance of an active gas gland epithelium, also of the " spent " 

 condition of the cytoplasm. Microphotographs were also taken of 

 intracellular bubbles, but, owing to the thickness and staining of 

 the sections, were not suitable for reproduction. 



Addenda. 



(1) The statement that gas glands are normally quiescent may 

 not be strictly nccurate ; it would perhaps be more correct to say 

 that only on occasion do gas glands assume great activity. I 

 make this remai-k in consequence of a communication from 

 Lieutenant G. C. C. Damant, R.ISr., who kindly permits me to 

 repeat it. From observations made under water during diving- 

 operations and from the results obtained in catching fish on the 

 hook at different states of the tide, Lieut. Damant concludes that 

 Pout, e. g., remain fairly constantly at one level, viz., just off the 

 bottom, whatever the state of the tide may be. If this be the 

 case, it follows, as Lieut. Damant suggests, that the gas gland 

 must become functional during each rise of the tide in order that 

 a periodic increase of gas in the bladder may counteract the 

 periodic increase of external pressure. 



(2) It seems probable, from suggestions kindly made to me by 

 Dr. G. A. Buckmaster and Dr. G. 0. Mathison, that the principal 

 factor in the evolution of erythrocytes in Vertebrates has been 

 the necessity for a greatly increased quantity of hfemoglobin in 

 Vertebrate blood as compared with Invertebrate — a necessity 

 corresponding to greater functional activity. It is suggested that 

 the amount of hfemoglobin required to be present in Vertebrate 

 blood would be more than the plasma could possibly hold in 

 solution, and that even if this were possible the viscosity of the 

 plasma produced would seriously interfere with the other func- 

 tions of the plasma, which, like the respiratory, have, in 

 Vertebrates, increased in intensity. This lai^ge quantity of haemo- 

 globin in Vertebrate blood being requisite and its presence in the 

 plasma prohibited, the evolution of erythrocytes in which the 

 hfemoglobin is imprisoned and combined with a proteid has 

 apparently been the only alternative. As I have before remarked, 

 the lack of direct contact between the tissues and the respiratory 

 pigment in Vertebrates is compensated for both by the rapidity of 

 the circulation and b)^ the amount of the respiratory pigment 

 present, and doubtless also by the increase in amount of the waste 



