1896.] ^UJ [Goddard. 



■what was originally a single sclerite, while the condition in D. rubicun- 

 dula represents an intermediate stage. 



The relation of the parts in these three species are, in the main, toler- 

 ably clear. But when we turn to Libellula the problem is much more 

 complicated. Not only have I not been able to liomologize the parts 

 found in Diplax and tliose of this genus, but I have also found it impos- 

 sible to determine the relations of the parts found in different species of 

 Libellula. I can therefore give little more than a bare description. 



We may begin with Libellula exusta. The general arrangement is 

 much as in Diplax. The genital lobes are short and stout. The ventral 

 sclerite is wide, short, and only slightlj' emarginate caudally (Fig. 11, e). 

 The lateral parts of the free edges are somewhat undulate. The hamules 

 are stout and are membranous proximally, and the tip of the spur is very 

 strongly incurved. The framework is wide and strong. The lateral 

 rods are connected for their entire length to that part of the framework 

 caudad of them by feebly chitinized triangles (x). The triangle («) 

 borne by the middle part of the framework is very long ; its apex lies 

 under the free edge of the ventral sclerite. 



The cephalic part of the bladder is chitinized in a single sclerite with 

 a mesal cleft. The last segment of the penis is made up mainly of a 

 single large sclerite (Fig. 13, p^), much longer on the dorsal than on 

 the ventral surface. Its edges approach but do not quite meet on 

 the ventrimeson. There is a curious dorsal hump on the distal part 

 of the sclerite and the distal edge bears ventrally a pair of small, 

 spine-like projections. If this sclerite is the result of the fusion of the 

 shield and the ring there is no indication of the fact. As to the distal 

 part of the segment, it projects only slightly beyond this sclerite; it 

 consists of two pairs of appendages rising from a full membranous base. 

 The median and dorsal pair are sigmoid rods curved towards the dorsal 

 surface at their distal ends (u). The second pair are membranous at 

 base but strongly chitinized distally (y). 



I would suggest the following as the possible homologies of some of 

 these parts : the large sclerite corresponds to the shield ; the ring is 

 wanting ; the hood is represented in a much less differentiated state 

 than in Diplax, by the full membranous portion of the penis-tip. As 

 to the homologies of the other parts I am entirely uncertain. 



In Libellula pulchella, the blunt division of the hamule lies almost 

 laterad instead of caudad of the spur ; it is moreover reduced nearly to 

 a knob. The spur is long and strong and its point turns laterad. 



The dorsal aspect of the genital bladder, though normally united for a 

 considerable portion of its extent with the abdomen, separates readily 

 therefrom after maceration in caustic potash. In the penis, the first seg- 

 ment is extremely long and bears a dorsal terminal tubercle ; the second 

 segment is very small and triangular ; the third bears distally a large 

 dorsal upgrowth. The edges of this sclerite do not quite meet ventrally, 

 and between the angles projects a small membranous lobe which per- 



PROC. AMEB. PHILOS. SOC. XXXV. 152. 2 A. PRINTED DEC. 18, 1896. 



