MEiMOlKS OF THE NATIONAL ACADEMY OF SCIENCES. 15 



III Adelocephiila, after the tirst molt (PI. XLV, tig-. 3). thoracic horns diti'cr in the reduction 

 of the bulbs, which lose their .swollen shape, becoming simply forked, and the rods becoming 

 mere setiB. In the third and later stiiges there is a further reduction of the two terminal spines, 

 the horns ending in a double point, with little sigiiiticance, while the horns themselves become 

 shorter. On the other hand, a new and striking feature comes into prominence, i. e., the silvery 

 color associated with pink of the larger tubercles on abdominal segments two, four, and six 

 (PI. XLIX, tig. 2b), as if the welfare of the species depended on this new colorational feature. 

 Just how this is to be explained depends on further special observation of the living caterpillar 

 amid its surroundings of leaves and spines. 



In Eacles, stage II, a profound change takes place in the armature, the reduction of the size 

 of the spines being marked. From the prothoracic to the tenth abdominal segment thej' are cut 

 down, reduced. New influences have begun to be exerted, and the formidable set of complex 

 spines, inherited from some extinct ancestor, no longer of use, become discarded, and a new 

 kind of protection arises, these perhaps being the hairs so peculiar to this genus. In the third 

 stage the thoracic spines are as long as the body is thick, and taper at the end, which is forked; 

 in stage IV the horns are now considerably shortened, not so long as the body is thick, while 

 those on the prothoracii' segment are scarceh' higher than broad. After the tinal molt the 

 prothoracic spines (PI. XL VIII, figs. 3, i) are reduced to broad, low, flattened rough tubercles; 

 those of the two following segments are reduced to short, stout, conical tubercles, about half as 

 long as those in the fourth stage, and they are now shorter than in any other genus excepting 

 S.vssphinx. 



In the second stage of C. rcgaJl.x the bulb of all the horns have become moditied into two 

 simple branching spines, each ending in a simple seta (i*!. LIV). The larva in this stage bears a 

 forest of twelve spinulated, long, slender spines, the longer ones equaling half the length of the 

 body, while at the end of the l)ody is a less dense forest or thicket of repellant branching thorny 

 spines. 



In the third stage the iiorns are still slender, and the spinules, being large and long, they 

 appear to be much branched. In the fourth stage they have become three times as stout as 

 before and of the same thickness and length, but a little longer in proportion than in the last 

 stage. Their efl'ect is in the two last stages heightened by the black thoracic and the abdominal 

 oblique liands. 



It still remains to be seen whether these spines secrete at their base a poison; and it is to be 

 hoped that close and repeated observations will be made on the larvse while feeding to see how 

 far thej' are protected by their armature and dark conspicuous stripes from the observations of 

 birds, etc. 



It should be noted that while C. regalia has the longest spines of any species of the genus, in 

 C. sej)uIcraU.'i they are shorter, while in the Brazilian C. j)'^'''^'^'^'']"- f^^.v are scarcely longer than 

 in Euclen imperalis. 



TTie hairs of Eacles vs. the reduction in the spines. — The bodies of all the Ceratocampinae 

 except Eacles are destitute of hairs. In Eacles the long irregular hairs are a generic feature. 

 They tirst appear in stage II, Ijut in stage IV these long white hairs are frecpientlj' as long or a 

 little longer than the body is thick. They are longer in the penultimate than the last stage, and 

 form a more dense clothing. Now, it is well known that birds and ichneumon flies of various 

 kinds avoid haiiy caterpillars, and the question arises whether the hairs of Eacles in its last stage 

 do not to some extent protect the creature from its natural enemies, and whether, since it is thus 

 protected, the spines have undergone reduction from being no longer useful in defense. On the 

 other hand, it is only fair to state that L'itlieronia penelope, with horns scared}- longer than 

 those of Eacles, is destitute of long hairs, though the vestigial dorsal tubercles of the abdominal 

 segments are represented by Burmeister as giving rise each to three or four short sets. 



The d<jrs<il spines of Anisata. — In this genus the armature is so reduced that there is but a 

 single pair of horns on a single segment, the second thoracic, and this reduction is associated in 

 the moths with the atrophy of the maxilla?. In each of the four species whose ontogeny is 

 Vol. 9—0.5 2 



