349] LARVAE OF THE TENTHREDINOIDEA-^YUASA 31 



The exact limit of the somite in a larva is not easy to determine. 

 Entomologists seem to have paid little or no attention to this point. 

 Castle (1900) has made an excellent study of the metamerism of the 

 Hirudinea, but his conclusions are not directly applicable to the case of in- 

 sect larvae tho they are pregnant with important suggestions. He found 

 that the natural and true limits of a somite coincide with the limits of the 

 neuromere, and that both reduction and increase in the number of rings, 

 which correspond to the annulets in the saw-fly larvae, take place at the 

 ends of the segment. The classical work of Lyonet (1762) and the recent 

 study of Forbes (1914) on the musculature of lepidopterous larvae, as also 

 a study of Boving (1914) indicate that the musculature affords a reliable 

 criterion for determining the limits of a somite. From a careful examina- 

 tion of the musculature of various types of tenthredinoid larvae (Fig. 129), 

 the author has come to the conclusion that in these larvae the natural 

 and accurate determination of the extent of the somites composing the 

 body is best based upon the musculature. A detailed discussion of this 

 subject is out of place here. It is sufficient to say that a majority of the 

 longitudinal muscles, including the dorsal, lateral, and ventral retractor 

 muscles, originate on the cuticular fold, or coria, which, on the exterior, is 

 usually indicated by a deep depression (is). Only a few muscles of impor- 

 tance cross this fold, nearly all the muscles being attached either to the 

 cephalic or caudal part of the coria. This cuticular fold, therefore, is con- 

 sidered as the cephalic limit of the somite, and the annulets into which 

 the somite is subdivided are numbered consecutively, commencing at its 

 cephalic end. To assume, a priori, the spiracular annulet, or the annulet 

 which bears the spiracle, to be the first annulet is arbitrary and inaccurate 

 inasmuch as this annulet, according to the criterion of musculature, cor- 

 responds to any one of the first three annulets of the somite. The position 

 of the spiracular annulet is constant and definite within a genus or sub- 

 family as long as the number of annulets of the somite is constant. Some 

 of the annulets are usually setiferous and often bear in addition transverse 

 rows of glandubae. The position of such setiferous annulets is constant within 

 the genus or subfamily when the annulation is constant. The number of 

 annulets on the ninth abdominal segment is always smaller than that on 

 the preceding segments. Since the setiferous annulets have a definite 

 order, it is possible to determine which annulets are obsolete on the penul- 

 timate segment. The first annulet to disappear is a caudal one, and 

 ordinarily it is only the caudal annulet that is missing. The number of 

 annulets of the sternum is less than that of the tergum. The length of 

 the annulets varies but their relative size is constant within a species. 

 The primitive number of annulets of the abdominal segments is unknown. 

 If the annulation of the generalized Tenthredinoidea is assumed to be 

 representative of the primitive condition, then four is the primitive num- 



