ftV p.. J. TILLY ART). 62 V 



especially near tlie base of the gill. Text-fig. 47 shows diagram- 

 matically tlie derixatioii of the Horizontal Lamellar Type from 

 the Saccoid Type. 



The gills of Argiolentes are composed of a single joint. There 

 is nothing in their structure to suggest that they are derived 

 from two-jointed, ancestial forms.* I think, therefore, that the 

 archaic subfamily Megapodagrionhue, to which this genus 

 belongs, will prove (when its larva? are known) to possess, 

 throughout, one-jointed gills, and that most of these are more 

 likely to be saccoid, triquetro-quadrate, or vertical lamellar 

 types, rather than of the highly specialised form found in our 

 Austi'alian Arg'wlesles. 



O 



a. 



Text-tig. 47. 

 E\()luti(ni of the Hoiiz(jiital Lamellar Type; «, saeeoid type; h, r, d, suc- 

 cessive .staj:;es in the hori/ontal tlattciiing: r, hurizuutal lanielkir 

 type. 



The number of forms studied in this paper, particularly in the 

 case of the extensive family AgrionvUn (within which, not a 

 single representative of either of the two subfamilies Fseudv- 

 stigmatluff, and Flatycuevdiui: is available in Australia for study;, 

 is far too small for us to draw any very definite conclusions as 

 to the course of evolution of the various groups within the Sub- 



Except, perhaps, the peculiar form of the regenerated gill in Aryio- 

 (et<tts (frlffVf<: see Part ii. of tliis papei'. Plate i.. lig. 1 1. 



