344 MR. C. C. ROBSOX ox THE EBSPIIIATORY 



It is desirable to know to wliat extent the difierences illustrated 

 by his classification are to be found in other features of the 

 respiratory system. 



In order to test this I have examined the respiratory oi'gans 

 of the following forms, all of which are represented in the 

 material available in the Zoological J)epartment of the British 

 Museum : — - 



1. JIarisa cormt-arieits (Trinidad). 



2. Am/jyullaria i^ermiformis (Pernambuco). 



3. ,, glauca (Demerara). 



4. ,, ins'ularum (South America, ? loc). 



5. Lauistes afflnls (Victoria Falls). 



6. ,, olivieri (Egypt). 



7. Pachylabra sp. (Sikkim). 



8. „ sp. (Perak). 



9. ,, gordoni (Victoria ISTyanza,). 



According toBouvierand Fischer (3) Lanistes is less specialized 

 for siphonal respiration than Amjndlaria. The siphon is short 

 and does not appear to form a complete tube. On coming to the 

 surface of the water it apparently places its respiratory orifice 

 in direct communication with the air, and the siphon is not an 

 indispensable part of the mechanism. A. {=Pachylahra)glcmca,to 

 judge from Ramanan's account, represents a more advanced con- 

 dition, but not yet approximating to Ampullaria s. s. There 

 is true siphonal respiration and apparently, as in the South 

 American -vermiforinis, the base of the siphon assists in closing 

 the left hand side of the pallial aperture. But the siphon is very 

 short, its orifice very broad, and there is no i)umping-movement 

 of the head such as Bavay and others have described for South 

 American forms and as Miss Cheesman and I have witnessed in 

 A, vermiformis. We finally have complete siphonal breathing in 

 AonpuUaria s. s. achieved by a very long flexible siphon. 



In Ampullaria and Marisa there is a rather small and sub- 

 circular or ovoid respiratory orifice situated on the anterior end 

 of the lung-cavity rather towards the left. Its edges are some- 

 times thickened slightly and tend to overlap each other. It is 

 difficult to see whether closure is effected by a sphincter or by the 

 overlapping edges. 



In the two Asiatic species examined and in P. gordoni from 

 East Africa a very different condition is found. The respiratory 

 aperture is large and gaping, and is not a simple orifice in the 

 floor of the lung. Genei'ally found more in the middle of the 

 latter, it extends backwards a.nd to the right — as far as the right- 

 hand wall of the- pulmonai-y cavity — as a wide longitudinal slit 

 l)etween two maikedly overlapping edges, the inferior of which 

 (left-hand) is inserted into the floor of the pallial cavity. 



In Lanistes the orifice occupies more or less the same position 

 as in Ainpidlaria, i.e. anterior, to the left and close to the osphra- 

 dium. It shows, however, the same general structure as in 



