COCHLIDIDES (EUCLEIDEs). 365 



reading Chapman's observations, that the smooth forms were developed 

 from the spiny ones. We would also note, that from the second skin on- 

 wards, the larva of Cocldidion has a suhspiracular hair on each segment. 

 This subdivision gives us a basis for a tentative arrangement of the 

 superfamily, which may be stated as follows : 



I. Fam. EUCLEIDJE. (1) Subfam. Phobetrinae. Larva in first stage with alter- 

 nate strong and weak segments ; thoracic segments with three (i, ii, iii), 

 abdominal with two (i + ii, iii) tubercles bearing single setae (above each 

 spiracle). After first moult, strong and weak segments persistent ; the 

 tubercles changed to hairy (not spined) warts. 



(2) Subfam. Eucleinae. Larva in first stage not showing 

 strong and weak segments ; thoracic and abdominal segments with two 

 tubercles bearing multiple setae (above each spiracle). After first moult, 

 the tubercles changed to spiny warts. 



II. Fam. COCHLIDID.E. (1) Subfam. Eulimacodinae. Larva in first stage with 

 three warts on thoracic, and two on abdominal segments (above, each 

 spiracle). After first moult, the warts degenerate into single set. 



(2) Subfam. Cochlidinae. Larva in first stage with alter- 

 nate strong and weak segments. Single alternate minute dorsal spines 

 in first stage, double dorsal row in second, third and fourth skins, lateral 

 row in first four skins ; entirely absent after fourth moult. 



We are much in doubt whether the distinctions between the 

 Eidimacodinae and Cochlidinae really exist. We understand that the 

 basis of Dyar's conclusions as to the strong and weak segments was 

 founded on the alternate character of the dorsal spines. It has also been 

 necessary for us to modify Dyar's definition that the Cocldidinae lose all 

 their armature after the first moult, in accordance with Chapman's ob- 

 servations that they retain it in the first four skins. It is possible 

 that the Eulimacodinae' should be merged in the Cm-hlidiuae, there being 

 probably only one structural plan in the larvae of the two groups. 



The urticating property of the spines or setrc of certain Eucleid 

 larvae is well known, but the actual mode of injury has not been 

 explained. Lewin states that the larva of Doratifera rulnerans darts 

 out its fascicles of " stings " whenever it is alarmed by the approach 

 of anything. The larvae of the Javan Parana media and P. lejrida have 

 also this urticating property developed in the highest degree, as, 

 indeed, have the larvae of many other species. Although spoken of as 

 " stings," etc., it is probable that the urtication is caused, as in other 

 lepidopterous larvae, by the mechanical influence of the fine barbed 

 setae that enter the skin. The spines, however, often appear to be 

 constructed like stings, i.e., they are tubular with a brittle tip. 



The larva spins a cocoon which is provided with an easily detached 

 lid, the pupa being provided with a projecting beak, by which the lid 

 is probably forced off. It is not " cut " in the true sense, nor is 

 it prepared by the larva, but fractured somewhat irregularly. It is 

 quite possible that the pupa works round and round, fracturing the 

 lid bit by bit, but the beak has no cutting edge, and anything 

 there is like one is at right angles to the line of fracture of the lid. 

 In this particular, the lid of the Eucleid is different from that of 

 the Megalopygid cocoon, which Lintner has shown to be woven' (in 

 Lagoa crispata] by the larva, separately from the rest of the cocoon, 

 so that the lid, in this latter case, is not a section cut or broken from 

 it after its completion. Generally speaking, the cocoons are brown or 

 ochreous in colour, but those of Calybia slossoniae are chalky-white, 

 conspicuous on the green leaves of the mangrove trees, but hard to 

 detect on the whitish bark. The brown cocoons of CocJtlidion avellana 



