70 THE EEPORT OF THE No. 36? 



atus and Pteronus, the caudal third of the abdomen is narrowed and Wuntly point- 

 ed, the prolegs of the tenth segment are correspondingly reduced or rudimentary.. 



In the Pamphiliidae the tenth abdominal segment bears on each side a dis- 

 tinctly segmented appendage, an anal cercus (Fig. 1, ae). These appendages are- 

 not segmented in any other group of the superfamily and never so large. In a 

 number of genera, though not characteristic for the genus, there is a pair of 

 pointed projections, which have been homologized with the anal cerci (Figs. 27^ 

 and 11, ac). They are of more frequent occurrence in the Nematinae than in any 

 other subfamily of the Tenthredinidse. They are also present in the cephids and; 

 in Tremex. 



There is a distinctly chitinized dorsal plate on the tenth segment in the 

 Pampiliidse (Fig. 1, alO), which is wanting in all the other groups. In the bor- 

 ing larvae the tenth segment bears a fleshy or chitinized mesal spine (Fig. 11)^ 

 The anal opening is a transverse slit located along the dorsal part of the anal pro- 

 legs and its lips usually bear transverse rows of fine setae. In certain specie& 

 where the larvae are opaque white or green in color, the centre of the dorsal aspect 

 of the tenth segment bears a fuscous or black spot. This is of freqnent occurrence 

 in those larvae that curl themselves helix-like on the underside of a leaf. The black 

 spot is placed eye-like in the centre of the coil. 



Each abdominal and thoracic segment is crossed by transverse lines. These 

 lines are interrupted in the region of the spiracles so that the dorsal and ventral 

 lines are not continuous. There is also generally a difference between the thoracic 

 and abdominal segments in the number of these depressed lines. These lines- 

 divide the surface of the segments into elevated ridges. These ridges are known^ 

 as annulets (Fig. 6, and 1-7). The number of annulets is constant for each 

 species. In inflated specimens that are abnormally inflated and in specimens that 

 are dropped directly into alcohol and shrunken, they are difficult to identify, but 

 in living specimens and carefully prepared specimens they are easily followed and 

 undoubtedly will prove of great service in the classification of these larvje. 



If the condition found in the Pamphiliida may be assumed to be the general- 

 ized condition, the primitive number of annulets -would be four (Fig. 1), since the 

 larvge of this family have that number on both the dorsum and the venter. The 

 ventral annulets of the Pamphiliidae are as well marked as the dorsal, but in the 

 groups bearing prolegs the ventral annulets are greatly reduced (Figs. 7, 8, and 12)^ 

 The ventral annulets are usually difficult to identify in the Tenthredinidas and are 

 rarely more than three. While in some species the number of thoracic annulets is- 

 the same as the abdominal, as a rule there is a difference : in the great majority 

 of the species the primitive number of four is retained, but in some species there 

 may be more than four, and in others less. The usual number of annulets on the 

 dorsum of the abdominal segments i&' six, as in the Emphytinae (Fig. 12), Selan- 

 driiufe, and LophyriniE. This number has probably arisen from the primitive four 

 by a subdivision of the second and third. The maximum number is found in 

 Cimbex (Fig. 6), and Trirhiosoma, where the number is seven; the minimum num- 

 ber is found in the Acorduleceriufe, so far as observed, where there are three. In 

 the Pamphiliidae, as already indicated, there are no annulets on the tenth segment 

 and they are frequently wanting in the larvEe of the Tenthredinidae (Fig. 27). 



The annulets of the dorsal and ventral surfaces are not continuous, as already 

 indicated, but interrupted in the region of the spiracles. In the Pamphiliidae there 

 is an elevated longitudinal fold extending the length of the segment just below 

 the spiracle, which has been designated the spiracular area (Fig. 1. sa). This 



