BY R. J. TTLLYARD. 623 



evolution, from tlie archaic Saccoid Type to the liiglily specialised 

 Triquetro-quadrate Typ®? characterises tiie family Calopterytjida' 

 as a whole. 



From the Triqiietro-<iuadrate Type of the Calopteryginct^, tlie 

 Lestid form of Lamellar Gill-Type is undoubtedly derived. This 

 can be seen in 'J"'ext-fig.45, e-y, where I have shown diagrammati- 

 cally the probable stages in the reduction-process. In connection 

 with this, the following points are of importance : — 



(1) There is no evidence of the LesfAdcE ever having possessed 

 the division of the blood-canals into two, in their lateral gills, as 

 seen in these gills in the CalopteyygiiuE. This must be regarded 

 as a specialisation confined to that subfamily. 



(2) There is strong evidence that the Lestid gill-form is derived 

 from a form in which the tracheation was the same as in the 

 Triquetro-quadrate Type : — 



For {a), regenerated gills of Lestid larvje always show an 

 increase in the number of tracheae: and, in most cases, this in- 

 crease takes the form of a doubling of the original number of 

 tracheae, and therefore agrees with the original Triquetro-quad- 

 rate arrangement; and {b) the reduction of the number of main 

 nerves in the median gill in the LestidcE is unique in Zygoptera, 

 and only to l^e explained as a specialisation from the original 

 arrangement of four nerves in the median gill and two in each 

 of the laterals. 



Within the Leslid(n^ it is clear that the gills of the St/nUmtincf 

 are the more primitive type; for (a), regenerated gills of the sub- 

 family LestitKe assume the SyidesH^ie form, with the branch- 

 tracheic obliquely placed in the main stem; and (b), the arrange- 

 ment of these branch-tracheie at right angles to the main stem 

 in Lesliiid', and the development of the pigment-bands, are evi- 

 dently high specialisations, the forjiier being unique in the Sub- 

 order. 



The evidence of the structure of the Caudal Gills, then, tends 

 to bring the LestidcB into closer relationship with the Caloptery. 

 gidcE than a study of the imaginal characters alone would warrant. 

 In particular, it would seem that the larvae of the Caloptery yincfi 

 and Synlestitue are very closely allied. 



