218 DE. W. N. F. AVOODLAKB ON THE 



perpendicular to and parallel with the plane of the bladder Avail, 

 but I know of no examples of such. Subsequent embryological 

 enquiry can alone prove which of these views is the correct one. 



The minute structvire of the cells of the gas gland epithelium is 

 doubtless constant throughout the entire series of " red bodies." 

 Gas bubbles are in all cases generated in the interior of the cells 

 (usually in the neighbourhood of the nucleus) when the glands 

 are active, and these bubbles are ejected by the ceils into the 

 gland ducts and bladder cavity where they explode, the shattered 

 walls of these bubbles giving rise to the masses of granidar matter 

 already described. Vincent & Barnes analysed this granular 

 matter (which, as before mentioned, must be caief ully distinguished 

 from the disintegration products of the led blood corpuscles alone 

 found in the blood-vessels) of the Cod and found it to consist 

 principally of a nucleo-proteid ; and since the cytoplasm which 

 forms the wall of the gas bubble is also composed for the greater 

 part of this substance, the result of this single analysis is con- 

 firmatory of the view just expressed, as is also the fact that a 

 lai'ge amount of this granular matter is always associated with 

 the presence of numerous gas bubbles. The iiitracellular ducts, 

 which, with the other cytological features of the gas gland, have 

 been so well figured by Bykowski & ISJusbaum (24) and Reis & 

 Nusbaum (62) within the last few years, are doubtless also constant 

 features of active teleost gas glands, though I am of opinion that 

 some of these intracellular channels are due to the expulsion of 

 bubbles from the cells and not purely nutritive as Reis & Nusbaum 

 suppose. The intracellular capillaries, on the other hand, are 

 permanent structures, and the fact that they occur so rarely as 

 compared with the enormous number of these structures found in 

 connection with the liver cells can only be explained by the 

 relative inactivity of the teleost gas gland. 



I have already described and figured the hsemolytic disintegra- 

 tion of the red blood corpuscles (erythrocytolysis as it may he 

 called) which occurs in the blood-vessels associated with the 

 active gas gland of Syngnathtis actts, and this same phenomenon 

 is to be found in connection with other of my preparations of 

 active gas glands and is doubtless a result of the gland's activity. 

 In this process of haemolysis many of the erythrocytes become 

 entirely broken up, first into lai-ge globules and finally into the 

 granular matter which is present in such large quantity in the 

 capillary lumina, whilst others are distorted in shape and only 

 pa.rtially broken up, looking very much as if influenced by some 

 toxic substance, as Jeeger suggests (see Part II.). This condition 

 of the erythrocytes cannot be due to imperfect fixation since 

 great care was exercised in the fixation and preservation of these 

 gas glands ; moreover sections through other vertebrates — Elas- 

 mobranchs e. g. — fixed by similar methods, exhibit the blood 

 vessels as practically free from granular matter (the small amount 

 sometimes present simply representing coagulated plasma) and 

 the erythrocytes as retaining their form. 



Another fact which I wish to lay stress upon, especially in view 



