OF THE MILK-WEED BUTTERFLY. 5 



other Lepidoptera. They are too small for mechanical use in Archippus, and perhaps the 

 papillae are, in this case, organs of taste or touch, as suggested by Fritz Miiller, which 

 appears more likely from the fact that they occur, though in much less number, within the 

 central canal of the proboscis. This canal is not lined with rows of little blocks like the 

 exterior, but with the much narrower solid edges of semiannular plates, which give the 

 canal the appearance of a large trachea (fig. 9). In transverse sections it is seen (fig. 10) 

 that the ends of these semiannular plates form behind hooked teeth, which are snugly 

 dove-tailed together, Avhile their anterior ends are produced into long spines which 

 simply interlock like the fingers of the two hands. These arrangements serve to hold the 

 two maxillae together to form a continuous tube. 



Each maxilla is traversed through its whole length by a nerve and a trachea (figs. 9 

 and 10, n. and tr.\ both giving off numerous lateral branchlets to the muscles which fill the 

 rest of the interior of these oi'gans. The muscles form narrow bands and are arranged in 

 two sets, both running from the anterior to the posterior aspect, and descending in this 

 course. One set, however, is inserted on the outer side of the anterior face, while the 

 other is inserted on the inner, and both converge in their downward and backward course, 

 and thus, when seen through the wall of the maxilla from the front, they form a series of 

 Vs one above the other. Seen laterally, only one set is visible at a time, diagonally cross- 

 ing the maxilla, as shown in tlie section (pi. 2, fig. 9). From the attachment of these mus- 

 cles, their apparent action is to shorten the posterior wall of the maxilla, which of course 

 must produce the spiral coiling of the organ. This unfortunately, leaves the explanation 

 of its extension in the dark, unless we si;ppose elasticity, or possibly the injection of the 

 somatic fluids, to bring this about. Many authors have assumed that the proboscis was 

 coiled by its elasticity, but if we admit the possibility of such action at all, we can as 

 easily require it to account for extension as for flexion. It seems more probable, however, 

 that we fail to see, or to correctly interpret, some proper muscular mechanism for both 

 movements of the proboscis.-' 



MaxAllary ivxlpi are often well developed in Lepidoptera, though in some they are 

 wanting, or at least reduced to a little wart on the base of each maxilla (fig. 3, toix-I).). 



The labial paljji are, as already said, very large, and conceal the proboscis in repose 

 The first joint is short, stout, rather pedicellated, and is curved upwards ; the second is 

 long, tapering somewhat from the base to the tip ; and the last joint is small and pointed. 



Thorax. A slender neck well separates the head from the thorax, whose first somite, or 

 prothorax, (fig. 1, i.), is very small and feebly develo]3ed, and the homologies of its skele- 

 tal parts are not easy to recognize. The scutum (fig. 1, s.^) is divided by a median furrow 

 into two tumid, bolster-like pieces (called by Mr. Scudder the pro-thoraci6 lobes), entirely 

 separated from each other. Behind them is a small triangular plate, representing the 

 scutellum ; and behind this are two rounded and knob-like processes which form an articu- 

 lar surface resting in a corresponding depression in the mesothorax. These probably rep- 



^I have found few references to the myology of the lepi- See also his article in Todd's Cycl. Anat. n; and Gersteldt. 



dopterous proboscis; the most extensive is that of Newport, Ueber die Mundtheile der saugendea Insecten, Dissert. 



Nerv. System of Sphinx ligustri, Phil. Trans., 1834, p. 398. inaug. Dorpat, 8°. 1853. 

 whose whole conception, however, seems to me erroneous. 



