Sept. 1897.] Dyar: Study of Young Arctians. 131 



caryce this process is foreshadowed on the metathorax, but while iia is 

 partly united to i it forms a distinct wart on the mature larva. 



This parallelism between the relative advance of larval and imaginal 

 characters is worthy of notice in view of the numerous cases of the re- 

 verse tendency. 



The details of the seven selected species are shown in the accompany- 

 ing plate. 



Spilosoma virginica. 



The setae are perfectly normal for stage I, all the subprimaries 

 absent. Of the six primary set£e on the cervical shield, four remain on 

 the shield, the others are detached and reduced, so that I detect only 

 one seta on the small detached piece. The setae of the prespiracular 

 tubercle are also less than in the primitive Tineid stock. On the other 

 thoracic segments, ia and ib are united, lib separate and reduced, all 

 characteristic of the Bombycine type of wart formation. On the abdo- 

 men tubercle i is small, the rest large, vii with two setge on segments i 

 and 2, one on 7, 8 and 9, instead of the primitive three set?e; leg 

 plates well marked. Tubercle viii present next the midventral line (not 

 shown in the figure). "Joint 13" is evidently composed of two seg- 

 ments, on the anterior portion (9th abdominal) tubercles i to iii on one 

 large wart, iv and v on another; on the anal plate (loth abdominal) all 

 the five setee on a single disk. 



This is the type from which we start — an Arctian in the primitive 

 first stage. 



Spilosoma antigone. 



The detached piece of the cervical shield is rudimentary. General 

 tubercles as in virginica, except for a peculiar modification whereby 

 tubercle iia on thorax has become three or four-haired (in different in- 

 dividuals) and iii on abdominal segments i to 8, four or five-haired ; 

 on the ninth abdominal iii seems to be only three haired, as I find but 

 five hairs on the large wart composed of tubercles i to iii. No subpri- 

 mary setse ; ventral setae as in virginica. 



This modification is to be interpreted as a partial wart formation, 

 pushed back into stage I, yet unaccompanied by the subprimary setae, 

 which in phylogeny must have preceded any wart formation. 



Not only in stage I is 6'. antigone unusually specialized for it genus, 

 but in the later stages it has assumed the plumage and habits of Arctai 

 {Eyprepia), as noticed by Mr. Hulst (Ent. Amer., II, 16). This 

 specialization is not shared by the imago, and is consequently without 



