MiaitAioiir risj££S or as/a. 209 



just out of, or immediately under, the surface. Here they respire 

 with ease, and are ready to capture any incautious frog that passes 

 that way. 



It is tlie same with the amphibious Siluroids. It has appeared 

 to me that they are mostly found in irrigated fields, ditches, and 

 the like, that they push up these to deposit their ova, and are thus 

 constantly running the risk of having the water-supply cut off. 

 This possibility of respiring air permits amphibious fishes to ascend 

 small streams and deposit their ova in situations where the young 

 may be hatched and remain in safety until large enough to take 

 care of themselves, but where their great enemies, especially carp, 

 are unable to attack them. 



We now arrive at the consideration of my fourth propositioa 

 — that the respiratory sac is certainly not homologous with the svdm- 

 hladder of fiahes, hut is more probably vMh the respiratory sac of 

 amphibious reptiles. 



Cuvier observed that " the gills are the lungs of animals abso- 

 lutely aquatic." Geoffroy 8t.-Hilaire regarded the branchial 

 arches of fishes as the modified tracheal rings of the air-breathing 

 classes. If this view is correct, these respiratory sacs, which are 

 merely formed by a continuation of the mucous membrane lining 

 the branchial cavity, are a part and portion or an addition to the 

 branchiae. 



Professor Owen denies the homology of branchial apparatus 

 with the lungs, and tests his opinion by considering the homolo- 

 gies of the air- or swim-bladder in fishes. First pointing out that 

 the totality of the organization of Lejjidosiren is ichthyic, he con- 

 trasts it with the Granoid Polypterus. The Lepidosiren has a cel- 

 lular lung formed by the partition of the bladder into two elon- 

 gated sacs, with a supply of venous blood from a true pulmonary 

 artery, and also a pneumatic duct going to the ventral surface of 

 the oesophagus. In the two forms compared, the arteries of the 

 swim-bladders are derived from the returning dorsal portions of 

 the branchial vascular arches before their union to form the aorta, 

 venous in Lepndosiren, partially arteriahaed in JBolypterus. 



Without doubt this negative argument is good, and might be 

 considered conclusive, were it not that we have in Saccohranchus a 

 respiratory sac {h, b) and also a swim-bladder {a). This swim- 

 bladder, although partially enclosed in bone, has its pneumatic duct 

 leading into the alimentary canal. This duct, of which so much 

 has been made as establishing a communication with the ventral 



