tlte Larva and Pivpa of ihe Genus Sabatinca. 451 



ainoimt of fusion or soldering of external parts is very- 

 slight, being confined to the tips of the wings and the. 

 distal portions of the hind-legs, with possibly a slight fusion 

 of the other legs, though these were free in the macerated 

 specimen. This is about the amount of fusion noticeable 

 in Mecopterous and Planipennian pupae. Judging from 

 the macerated specimen, all the segments of the abdomen 

 of the pupa are freely movable in Sabatinca, as in the case 

 of the pupae of the other Orders just mentioned. 



Much has been made by various authors of the remark- 

 able mandibles found in the pupa of Eriocraniay and their 

 resemblance to similar hypertrophied mandibles found in 

 certain Trichopterous pupae. A wider knowledge of the 

 more archaic types of pupae within the Holometabolous 

 Orders would surely have convinced these authors that it 

 is a very far-fetched argument to try to draw from this any 

 grounds for assuming that the Lepidoptera are descended 

 from the Trichoptera. It is only certain specialised pupal 

 types witliin the Trichoptera which have the mandibles thus 

 hypertrophied and crossed, and such a development is 

 only a secondary one, correlated with some special diffi- 

 culty in cutting a way out of a particularly tough cocoon. 

 In the case of Eriocrania, the cocoon is placed underground, 

 and is made of silk with ]:)articles of sandy soil closely 

 interwoven. Pupal mandibles of the normal type would 

 not open such a cocoon ; and hence, undoubtedly, occurred 

 the evolution of the hypertrophied form found in that 

 genus. The pupa of Sabatinca lives in a softer cocoon, and 

 exhibits a normal archaic type of mandible, very similar to 

 the mandibles found in the pupae of Mecoptera, most 

 Planipennia and a fair number of Trichoptera also. They 

 help to prove its relationship, undoubtedly, to all of these 

 Orders, but not its descent from any single one of them. In 

 the same way, the complete form of the pupal maxilla, 

 with its separate galea and lacinia, and five-segmented palp, 

 does not help to prove descent from any one of these Orders, 

 but only shows that this httle Lepidopteron has kept the 

 original archaic form of pupal mandible common to all 

 Panorpoid Orders, and is closely related to all of them. 



Taking the total evidence of the larva, cocoon and pupa, 

 we may reasonably conclude that Sabatinca is, on the whole, 

 more archaic in its early stages than any existing Panorpoid 

 insects excepting only the Mecoptera. The larva shows on 

 the whole a preponderance of Mecopterous characters, but 



TRANS. ENT. SOG. LOND. 1922.— PARTS III, IV. (FEB. '23) HII 



