J^mhrijology of certain of the Lower Fishes. 203 



the lowest fishes where it exists (Crbssopterygians, Dipnoans) 

 it does act as a lung. 



3. With the adoption of a deep-water habitat, the lung 

 began to function mainly as a hydrostatic organ, and, 

 associated with this, its connection with the pharynx 

 migrated to the mid-dorsal line. An important clue to how 

 this migration may have taken place is given by the Dipnoi. 

 In them the glottis is ventral, but the pneumatic duct 

 passes up round the right side of the pharynx, the lung 

 being situated dorsally. This dorsal position of a lung which 

 has not yet lost its original undivided character is rendered 

 practically necessary by the aquatic existence in which the 

 lung, besides its primary respiratory function, necessarily 

 exerts a secondary hydrostatic function. When once the 

 lung itself had become dorsal in position, it would seem 

 only natural that the glottis should then gradually migrate 

 dorsad so as to make the communication between lung and 

 pharynx as direct and uncomplicated as possible. 



The Origin of the Limbs of Vertebkates. 



Before leaving the subject of external gills, a few words 

 may be said in regard to their possible connection with the 

 origin of the limbs of vertebrates. I may say that with 

 more extensive knowledge of the development of external 

 gills and of limbs, I see no reason to give up or to modify 

 my opinion that the only feasible working hypothesis of the 

 origin of limbs is that which derives them from external gills. 

 The guiding principle in all speculations as to the origin of 

 new organs must be that a pre-existing organ, with a different 

 dominant function, must be found, out of which, by modifica- 

 tion in structure and function, the new organ may have been 

 evolved. In the case of the limb, the pre-existing organ 

 must have possessed two qualifications — (1) it must have 

 projected beyond the general surface, and (2) it must have 

 been movable by means of voluntary muscles. The only 

 paired organs of the more lowly organised vertebrates which 

 fulfil these conditions, and are not limbs, are the external gills. 



It has been suggested that a forerunner of the paired limbs 

 once existed as a lateral fold running along each side of the 



