376 



PSYCHE. 



[August, IQ02 



formation, when the case is attached 

 flatwise, he on their side; but these 

 larvae lay on the venter, and the de- 

 pressed abdomen with its short recurved 

 hairs seemed to indicate that as the 

 normal position. Prepupa and pupa lie 

 likewise with the ventral surface to the 

 stone, the former with the head and 

 thorax and terminal abdominal segments 

 twisted to the left side. At this stage 

 the most remarkable feature of the 

 animal is the presence of five pairs of 

 appendages jointed upon the sides of 

 abdominal segments 3-7. Such append- 

 ages are wholly unknown in this order 

 (although cjuite characteristic of several 

 aquatic genera of the Neuroptera) ; and 

 if, as I suppose, these larvae, prepupae, 

 and pupae form a single series represent- 

 ing one species, their transitory appear- 

 ance at the end of larval life is a devia- 

 tion from the normal course of meta- 

 morphosis wholly without a parallel. 

 The nature of these appendages is 

 shown in detail in figures 4 and 5. 

 Figure 4 is an external view and shows 

 how completely the appendage is articu- 

 lated to the side of the abdominal seg- 

 ment. Figure 5 is a frontal section of 

 the same and shows its internal structure. 

 There is, as everywhere, an outer layer 

 of hypodermis (shown as a line of dots) 

 to which the old and loosened cuticle, 

 which has not participated in the articu- 

 lation, does not conform ; there are two 

 tracheae passing out from the body cav- 

 ity into the appendage and repeatedly 

 branching there, and penetrating the 

 mass of fat tissue which fills the entire 



cavity, and which is entirely similar to 

 the fat filling the lateral extension of the 

 abdominal segment which supports the 

 appendage: there are no other tissues 

 entering into its composition. 



Two fairly distinct types of hyper- 

 metamorphosis have been distinguished 

 hitherto : one of these occurs among 

 the Hymenoptera and is apparently re- 

 stricted to egg parasites {Platygastcr, 

 Palyncma., Tclcas, cU.) : among these 

 the hatching of the egg occurs appar- 

 ently very early, and the earliest free 

 stages are regarded as embryonic rather 

 than larval ; obviously, our caddis-fly 

 has nothing to do with this type. 



The other type occurs among the 

 Neuroptera i^Mantispa) and Coleoptera 

 (Ep/caiifa, Mch)c, Sitaris, etc.). Among 

 these the larva hatches six-legged and 

 active (Campodeiform) and during its 

 larval life becomes footless, sluggish 

 and distended with accumulated fat 

 (Eruciform). These forms are held to 

 furnish the best of ontogenetic evidence 

 as to the course of development of com- 

 plete metamorphosis among insects. 

 The striking change of form, which 

 here con.stitutes hypermetamorphosis, 

 occurs during the period of larval 

 growth, and is therefore unlike that 

 which we are describing. 



There is also in those msects whose 

 transformation is most complete and 

 rapid, after the cessation of feeding on 

 the part of the larva, a period of making 

 over into the pupal form : this prepupal 

 stage is due to the swelling and shorten- 

 ing of the larva under its loosened cuti- 



