MEMOIRS OF THE NATIONAL ACADEMY OK SCIENCES. 17 



1, 3. (si). Ill stage 111, tho pioportion l)ctwcen the spines is still the same; now, small sec- 

 ondary setiferous tubereles have arisen. In stage V (Hg. 6) the two lateral tubercles sj are 

 reduced to low flattened four-headed ones, and the plate is well rounded at ihc end; the two 

 terminal black tubercles are minute, while the surface of the plate is without secondary tubercles, 

 not being coar.sely granulated as it is in Eacles. In 6'. sj//e)ulcii-s (PI. LV, fig. la, Ih) the arma- 

 ture is more e.xccssivcly developed than in C. regalls or any species of the genus, and the surface 

 of the plate is covered with transverse rows of secondary tubercles, the two terminal tubercles 

 being distinct. In C. >!ij^>itlri'alis the armature of the surface is intermediate l)etweeM the two 

 species named (figs. 4(?, -iih), while in C. laocoon the plate is nearly smooth, with small tubercles 

 si, Iteing much as in C. rii/dl!.'^. On the otlier hand, in C. pharonra, judging by Burmeister's 

 figure, tubercles si are much longer, but the surface is merely rugose as in C. n-gulls. 



In conclusion, the suranal plate and its armature coincide or are congruent with the other 

 larval characters, so that if found fossil they woukl afford a fair Iwsis for identification and 

 classification. 



III. THE CAUDAL HORN OF THE CERATOCAMPID^E. 



The larv;\? of Ceratocampidiu and Saturniidse, as well as a few other lepidopterous larvfe of 

 other groups, and more especially the Sphingidre, are characterized by possessing the so-called 

 "caudal horn." This modified tubercle is now known to l)e the result of the fusion either before 

 birth, or Ix'fore the first ecdysis, of the dorsal tubercles i of the two anterior tubercles of the 

 eighth abdominal segment. We have designated this double tubercle as the "eighth uromeral 

 tubercle:"" the term uromere being applied to any one of the abdominal segments. We will first 

 describe it in the group Ceratocampina;, and then discuss its mode of origin and occurrence, as 

 well as its phylogenetic significance, in other groups. 



Caudal horn of Ci'rafommjt/'na'. —rln the species of the most generalized genus, Adelocephala, 

 the caudal horn is very large and prominent, as long as any of the thoracic horns in A. fncolor, 

 twice as thick, and considera})ly stouter; it is roughly tuberculated with secondary setiferous 

 tul)ercles; and at the tip has lost the bifid nature of the first two stages, ending in a single 

 tubercle which does not bear a terminal seta. 



We will now trace its history from stage I up to the last. The great thick club-like slightly 

 cui'ved horn ends in two primary tubercles, each bearing a stifi' seta about as long as these 

 tubercles (PI. XLV, fig. 1). In stage II (fig. 3), the caudal horn is slenderer than before (in the 

 specimen figured it had become flattened down and attached firmly along the back). 



In the third and fourth stages the bifid nature is lost, and the hdrn ends in a single 

 bristleless tubercle as in the last stage, the terminal or primary tubercle being l>ut little larger 

 than the secondary ones which render the horn so rough and coarsely spinose. 



In the full-grown larvie of A. loai'dl and ai'gijracantha, as figured by Burmeistei', the caudal 

 horn is slightly longer and twice as thick as either of the thoracic ones. 



In stage I of Siijxspli'in.f inolina the caudal horn, like that of Adelocephala, is entirely unlike 

 in shape and length the thoracic horns, being much shorter and nearly twice as thick (PI. XLVII). 

 In S^'ssphin.x the swollen, squarish end bears on each side a short rounded tubercle, each giving 

 rise to a long bristle, the main tul)ercle showing no signs of its double origin. 



In Eaclt'S imjJer'udis the caudal horn is nearly as large as an}' of the thoracic ones of the 

 same shape, the end Ix'ing deeply divided, with no definite line of division })etween the two 

 terminal tuljercles and the main shaft of the ■"horn;" the seta? are a little longer than the forks 

 or branches. After the first molt the proportions are the same, all the horns being reduced in 

 size. In stage III the horn is still long arid slender, liut oidy about one-half as long as the 

 longest thoracic horns. 



After the third ecdysis there is a decided reduction in size of the caudal horn, which, though 

 still quite large and prominent, is oidy about half as long and lai-ge as the largest thoracic horns. 

 In the fully grown larv:e, where the process of reduction has affected all the armature, especially 

 the dorsal spines, the caudal horn is reduced to a conical spinose tubercle but little higher than 

 broad, though still oidy about half as high as the larger tiioracic ones; and now the somewhat 

 dense secondary hairs rise far above the dorsal spines. 



