1904. PACKARD — OPISTHENOGENESIS. 291 



After the first molt there arises behind the dorsal hump two, instead 

 of one, median black spots, and two black spots are added on the 

 side of the body near the base of the anal legs, />., two each on the 

 ninth and last segments. 



After the second casting of the skin, the marking of the three 

 last abdominal segments becomes specialized ; what on the body 

 in front are parallel black and red lines being in this region now 

 represented by separate spots. Thus as regards the marking, the 

 anterior part of the body remains ornamented with the primitive 

 parallel lines ; while the process becomes on the three hinder seg- 

 ments accelerated or specialized. It thus appears that the more 

 advanced or ontogenetically later style of ornamentation originates 

 at the end of the body. 



A parallel process takes place with the formation of the caudal 

 horn or hump. Thus in Symmerista, Dasylophia and other 

 horned Notodontidae and members of other groups, the eighth 

 abdominal segment is the theatre of the process of fusion of the 

 two dorsal tubercles of the first larval stage into a single tubercle or 

 horn ; so that this segment appears to be the theatre of a process of 

 specialization which does not take place on any other segments of 

 the body. 



When in other genera it does take place and there is a special- 

 ized single tubercle on the first abdominal segment, as in Noto- 

 donta, Nerice and more especially in Hyparpax and Schizura, the 

 process of fusion of two tubercles into a single specialized one, as 

 on abdominal segments i, proceeds from behind forward, as it were 

 in waves of translation of the specialized growth-force from behind 

 forwards. 



This may clearly be seen in the figures on Plate XXIV, showing 

 the development of the single hump in Hyparpax aurora. In Fig. 

 I, the dorsal tubercles / in stage I are all separated; in Fig. 2, 

 those on the eighth abdominal segment have all begun to unite at 

 their bases before they have on the first abdominal segment ; they 

 seem to be a little behind at first, though later on the hump on the 

 first segment becomes higher and larger than the caudal horn. 



If there were any doubt as to the relative period when the 

 tubercles become fused in Hyparpax, in Schizura leptinoides (PI. 

 XXVI) it is very clearly shown by Fig. i that the fusion of the 

 two tubercles forming the caudal hump as we will call it, i.e., that 

 on the eighth abdominal segment, has taken place before any signs 



