8 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



humps, and spines, and thus more nottuit'oi'ui. At the same time the yellow and whitish stripes 

 and pinlc blotches become indicated at an earlier stage tiian usual, as if the aim were to adapt 

 the caterpillar to the hues of the ribs and parallel greenish and yellowish lines and redtlish spots 

 of the leaf on which it feeds. 



A variety of TR'terocampa manteo (Monogr., PI. XXIX. figs. 5, ~Mt)^ with a ground color 

 like the green of birch leaves, is beautifully colored with two subdorsal white lines, which in the 

 thii'd and sixth abdominal segments expand so as to 'contain a conspicuous deep blood-red sjjot 

 on each side: there is also a lateral yellow spiraeular line. As I state in the monograph (p. \L^1%): 

 ''It is plainly derived from normal ;«(/«?'<■«, and is adapted for existence on the pale yellowish 

 green underside of the birch leaf, while the deep l)lood-red spots are similar in color to those of 

 the birch twigs or leafstalks." 



The evolution of the red dorsal spots on IL'ttrocuiiipa (jtdtivitta is shown on PI. XXXI. In 

 stage I there are no lines, the body being of a uniform reddish hue, and armed with nine pairs of 

 liorns, those of the tirst pair being really collossal in size, like the antlers of a stag; but after 

 casting its skin, when the horns are all dropped, onl\' slight vestiges of the prothoracic pair 

 surviving, two parallel paler reddish lines arise and extend along the sides of the Ijody. These 

 lines widen, become diffuse and are lost or at least the lower line, just before the next ecdysis, 

 leaving two medio-dorsal closely contiguous lines. These two lines after the second molt (stage 

 III) become broken up into large deep blood-red jjatches on the thoracic and third, fourth, and 

 seventh and eighth abdominal segments. On the prothoracic segment the two lines diverge, 

 Ijeing forked, and this forked line persists to the end of stage IV. Already in stage III there 

 are the indications of the three short lateral oblique lines, which are more fully developed in the 

 fourth stage and continue to the end of larval life. They are deep red and are shaded al)Ove with 

 white. With little doultt the dorsal red spots of the young larva of //. lumitio (PI. XXIX, 

 fig. 2, 2(/) arise in the same way, the spots being on the same segments, while the forked bright 

 red thoracic line is like that of JT. tjiifflriffii. 



In //. hinndatu also the freshly hatched larva is red, the green hue appearing after the worm 

 has begun to eat. In this beautiful insect toward the end of larval life it is readily seen from 

 PI. XXXII that it is amply "obliterated" })y its coloration, and stands in no need of horns, 

 which are, rather late in life, however, discarded. 



The caterpillar when at rest usually stands on the midrib on the underside of the leaf, where 

 its green and white shades and lines well protect it from the prying eyes of insectivorous birds. 



The figures of this species on 1*1. XXXII well illustrate the wonderfully close resend)lance 

 in color and markings of this caterpillar to the hues of the leaf and the rust-brown and darker 

 brown blotches on either surface of the leaves, in whatever position the larva assumes. 



ORIGIN OK THE LINES AND SPOTS IN DATANA. 



Origin of the transverse spots of Datana iiiajcr from longitudinal stripes. — The larviu of all 

 the species of Datana, with the exception of I), major, are longitudinally striped. The ground 

 color of the body in all the species is usually some shade of reddish brown or dark brown, 

 while the sti-ipes are reddish, yellowish, or whitish, according to the species, as may be seen by 

 reference to Pis. X to XIV of my monograph. Part I. As is well known, these conspicuously 

 marked caterpillars feed in an exposed maimer in clusters cm the apple and certain forest trees 

 and shrubs, assuming grotesque attitudes; it is most probable, though no experiments have 

 been made with them, that these caterpillars are distasteful to birds, since the European Pygivra 

 huci'pluila is distasteful to lizards (Poulton). These larvae are abundant, there are numerous 

 species, and the genus in point of numbers of individuals as well as species is a successful one. 



In my work I described in some detail the diffei'ences in the markings of the larviB at different 

 stages, but failed to draw, or rather deferred to a later occasion, some conclusions as to the mode 

 of origin of the lines, and particularly^ the transverse series of spots of D. major. 



In D. major the larva has taken a wide departure from the style of markings of its congeners. 

 As may l)e seen in PI. XII, Part I, the young larva, up to the time when it is about half grown, 



