AXD DEYELOP.AIENT OF COMTS INFELIX. 247 



the remaining (ventral) face of Avhich is completed by the ventral pieces of the petiole. 

 This dark basal portion of the petiole-ring is made up of two separate elements, meeting 

 in the middle line, where they are very narrow; at the sides they are much deeper 

 (PI. 12. fig. 36), so that, in outline, tlie two together are cup-like in mounted specimens, 

 with an almost straight base resting on the segment behind the petiole ; the two sides 

 are prolonged to form blunt points, which articulate with the propodium and make a 

 concavity between, filled by a pallid membrane, like a semicircle between two horns. 

 The (anterioi-) rim of this semicircle is a clear strong chitinous band. The chitin of 

 these parts is smooth, and devoid of surface reticulations such as mark the abdominal 

 plates. Ventrally (PI. 12. figs. 36, 37) the structure of the petiole is very similar 

 to that seen on the dorsal side, but the chitin of which these parts are composed 

 is thin and reticulated, making them obscure and difficult to detect. Another difficulty 

 too in investigating the ventral structures arises from the fact that these almost 

 transparent parts often remain attached to the metathorax, and so get lost. But with 

 careful manipulation it is seen that they are borne betAveen two lateral horn-like projec- 

 tions of the segment abutting on the petiole, and that their anterior edge is not concave 

 as in the dorsal region, but straight or sliglitly convex. The two strong outstanding points 

 from the ventral abdominal segment apparently form the articulation, and in the natural 

 position woiild point towards the dorsal surface on the face that meets the thorax ; they 

 would not point up and forward, as is necessarily the case when the parts ave flattened- 

 out under a cover-glass. This petiole, therefore, differs greatly from what prevails in 

 either the Aculeate Hymenoptera or the Ichneumonidre. Dorsally, behind the petiole 

 there are seven separate plates (figs. 34, 36). Each plate, except the last, is made up of 

 a median area connecting two lateral flaps. In the first of tliese (and what must be 

 looked upon as the true second segment), the lateral jiarts are broad and wing-like, their 

 function being to fold round the anterior face of the abdomen, curving towards the 

 ventral surface ; this segment bears no trace of a spiracle. 



The 2nd, 3rd, 4tb, and 5th plates form a progressive series, each with an atropliied 

 spiracle. The flaps are bent back more in each segment (PI. 12. figs. 34, 36) until in 

 the 6th this curvature reaches its maximum, for the side parts are very slender vmtil 

 near the tips, where they expand. At this point there is the only functional spiracle in 

 the abdomen: it is a large circular orifice (PI. 12. fig. 40) surrounded by a chitinous 

 ring; in the clear centre can be seen a tongue-like structure; the large trachea can be 

 traced running from this spiracle into the body. The median part of the seventh 2>late 

 is slender and deeply concave in front. The last dorsal plate has no lateral expansions ; 

 at its points on either side is situated a remarkable organ, which, for want of a better 

 term, I refer to as the " tactile plate" ; it is over this that the preceding plates have 

 been looped, as it were (figs. 33, 34, 36, 39). This organ consists of a pear-shaped plate 

 (figs. 42, 43) placed transversely to the length of the body, the pointed end being on the 

 inner side; the plane of the plate is at right angles to the dorsal surface. The plate 

 itself consists of a membrane stretclied over a loop-like ring of cliitin and supplied 

 strongly with nerves ; at its pointed inner end there is a solid chitinous rod, which, in the 

 natural position of parts, jioints forwards at right angles to the length of the oval plate ; 



SECOND SERIES. — ZOOLOGY, VOL. IX. 34 



