TUK CATKIJPIIJ.AK: STRUCTURE OF THE UODY. 13 



roimdcd head whose siiinmit hears a pair of excessively produced conical 

 ap[)endai2;es ; or aii'ain, hairs seated on papillae arranged in definite series 

 in a juvenile caterpillar ni;iy he supplanted hy thorny spines, arranged also 

 in definite rows In the mature larva, but <)ccu[)ying an entirely different 

 position from the series seen in the young. An examination of the descrip- 

 tions in the body of this work will afford many other instances quite as 

 striking as those given ; and from the observations of others on many dif- 

 ferent groups of insects 1 am inclined to believe that this is but a too partial 

 statement of a general law. (Compare plates 70-73 with 74-77; see also 

 plate 86.) 



besides the clothing or armature of the body ])ro})er, there is another 

 feature which may draw our attention foi- a mcmient. Each segment is 

 divided into sections or sub-segments, as perhaps they may be called, l)y 

 transverse creases which extend around the entire body excepting the ven- 

 tral surface ; it is as if the division lines between adjoining segments were 

 not sufficient for the free motion of the creature. These creases, which 

 have greater or less distinctness and depth, are found in allied insects at 

 the same place, so that even some gronps of considerable size may be char- 

 acterized by the number of the sections into which the princi[)al body seg- 

 ments are divided. ^Vs a general rule these are nearly the same on 

 successive segments of the body, but the thoracic segments often differ 

 from the abdominal to a greater or less extent, and the first and last seg- 

 ments of the body rarely agree with the others. In the following, note is 

 made only of the distinction seen in the abdominal segments. 



In a few caterpillars these segments do not seem to be divided at all. 

 Such are our native Papilioninac with Avhich also the gerontogeic Thais 

 agrees, although some of the allied genera, Parnassius, Doritis and Ismene, 

 have three or foiu' sub-diNisions. The same simplicity is found in the 

 Chrvsophanidi ; and indeed the same is the case to a nearly complete extent 

 in all the Lycacninae though, excepting in the Chrysophanidi, they are 

 often faintly separable into two subsegments. The only genus of Lemonii- 

 nae which I have been able to examine, the European Xemeobius, also has 

 two subsegments, though the hinder of the two is again subdivided equally. 

 So, too, all the Argynnidi and ]\Ielitaeidi have too subsegments, the anterior, 

 as is nearly the universal case, considerably larger, excepting that in our 

 species of Brenthis the hinder sul)segment is again divided. The Ilelicon- 

 inae agree with the xVrgynnidi, while in the Euploeinae not only is the 

 hinder half subdivided into two subsegments, but the anterior half as well, 

 though somewhat obsciu-ely. The next degree of complication is found in 

 the remaining Xymphalinae which have four subsegments, the antericn- the 

 larger and always bearing the i)rincipal armature, the hinder portion being 

 subdivided into three smaller subsegments. Anaea, however, has the 

 hinder portion subdivided into four subsegments. It is curious to note in 



