BY R. J. TILLY ARD. 319 



Fusion of arcidus-sectors. — It is somewhat unexpected to find 

 this character, an evidence of considerable specialisation, well 

 ■developed in all the species of Synthemis, except perhaps in *S'. 

 ■cyanitincta,(P\a,te vi., fig.o) where the sectors fuse quite close to 

 the arculus. The same species, curiousl}' enough, shews the 

 greatest specialisation in the recession of the hind wing-triangle ! 

 Anal loop. — The form of this area is of the greatest importance 

 to a correct understanding of Synthemis wing-venation. The ten 

 tigures of portions of wings of different species( Plate vi., figs. 1-10) 

 are arranged in order of development of their anal loops. In .S". 

 ■olivei{fig.l) we have a scarcely formed loop of six cells; in ^S*. 

 Jlavoterminata{fig.2) the same six-celled loop is very distinctly 

 formed. Six-celled loops occur in several other species, but 

 always in males only (see figs. 5, 7, 9). A characteristic of the 

 females is that their venation is always more reticulate, and 

 hence the number of cells in the anal loop is greater. In S. 

 eustalacta{i^gii.3 and 4) we have a slightly deeper and larger anal 

 loop, as also in S. guttata{G.g.6). All these are of Cordulegas- 

 terine or Macromian form. But in S. regina J" we find a double 

 looping(fig.9), which I find is constant throughout my series, 

 except for one specimen, which has a large single loop. A double 

 loop is also the constant characteristic of S. macrosiigma 2 (fig. 8), 

 but in this species the two loops are very unequal, a basal loop 

 of 3-4 cells being followed by a distal loop of twice the depth, 

 more than twice the width, and 12 or more cells. In *S'. regina 2 

 we have the remarkable form of three loops(fig.lO), a basal one of 

 3-4 cells, a slightly larger middle one of 6-7 cells, and a distal 

 one of greater breadth, containing 12 or more cells. These three 

 loops occur in all my S. regina ^ except two; one of which has 

 a small basal loop of four cells, followed by a very large loop of 

 nearly 20 cells (evidently the two other loops fused together), and 

 another has one very large anal loop consisting of all three loops 

 fused together. 



This interesting variability of the anal loop throws a strong 

 light on the problem of Synthemis wing-veuation. If we look at 

 the anal area of the hindwing of an archaic Aeschnid, say 



