ELECTROMOTIVE PHENOMENA IN GLANDS. 523 



water secretion as the process associated with the ingoing- 

 current, and the formation of complex paraplasm as that 

 connected with the outgoing current. 



One cannot doubt that water flow in cells and ducts 

 must give current, for the phenomenon is well known in 

 porous diaphragms, capillary tubes, and also in plants ; but 

 if we accept water flow as the main cause of the outgoing 

 action current, how are we to explain the marked outgoing 

 currents produced in the reversal of the current of the glands 

 of the frog by abstraction of water (and be it noted abstrac- 

 tion from the basal ends of the cells) by injection of glycerine, 

 when it has been granted that explanations of "action" 

 and " rest currents " must stand or fall together ? 



The events in the protoplasm of the gland cell are as yet 

 obscure, beyond the fact that it builds secretory substances, 

 which dissolve in water to become the secretion. By what 

 means the water crosses the cell from lymph to lumen and 

 in so doing produces the " secretory pressure" is unknown. 

 The " pump theory " of Strieker and Spina, elaborated from 

 the well-known movements of the cells of the glands of the 

 membrana nictitans of the frog, will not hold according to 

 the more recent and exact work of Drasch, let alone the 

 fact that we have no clear evidence of anything of the kind 

 in the case of mammals. 



The older theory of Hering of the formation of colloid 

 substances by the protoplasm, their swelling by water 

 abstraction from the lymph (in order to account for the 

 secretory pressure), and their final extrusion into the lumen, 

 is perhaps as much in accord with the histological story as 

 any other. Granules do dissolve in cells, according to 

 Langley and Biedermann, -who have watched the process 

 intra vitam, but the chemical solution of such granules is 

 not a likely source of electromotive force, and certainly a 

 streak of mucin granules from Myxine, connected at each 

 end with a sensitive galvanometer, gives no constant result 

 upon causing solution of the granules to proceed in the 

 region of one or the other electrode. 



Against the water flow theory, it must also be noted, 

 that, so far as measurements go, the latency of the outgoing 



