260 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



Stage 1 T.— Length of body, 12 min., and of tails, mm. ; total, 18 mm. Head yellowish greeu, 

 iiairowins- toward the vertex, flattened in front, with two broad pink lateral bands, which meet 

 above, and are broadly bordered behind with pale straw-yellow. Body ])ale, distinct green, 

 speckled in parallel lines with reddish brown, there being ten oblique, parallel, pale yellowish, 

 lateral stripes passing downward ami backward, the first one on the second thoracic segment 

 very short, being one-third as long as the next one. Two high, slender, conical, pinkish red 

 tubercles on the prothoracic segment. A dorsal broad yellow band, becoming pinkish on the 

 suranal plate; the line is broken into a series of surootii yellow swellings, four or five on a 

 segment. The tails are greenish, with a slender internal pink thread line, the ends with the 

 flagella (or eversible portions of the stemapods) deep coral-red. 



(In some alcoholic specimens in the United States National Museum the length of body, not 

 including the tails, is 22 mm.; of the tails, 8 to 9 mm.; length just before molting, 28 mm.; of the 

 tails, 10 mm.; total, .38 mm. The larva is much as in the third stage; the protlioracic tubercles 

 as before, but slightly smaller in proportion to the body. The tails are as in Stage III, the iiagella 

 nearly as long as the sheath, which is red at the end.) 



Stcujc V and last. — Length of body, 40-12 mm.; of tails, 1-5 mm.; total, 41-17 mm. 



It will now be seen that the tails are only about one-tenth as long as the body, while in Stages 

 I and II they are about two-thirds as long as the body. 



It is a large-bodied, pale green caterpillar, thickest in the middle, being somewhat spindle- 

 shaped. The head is moderately large, flat in front, snbconical, with the vertex high and conical, 

 pale green, edged very irregularly with roseate on the sides. A small, double reddish tubercle on 

 the top of the prothoracic segment, from which a median white or yellow dorsal stripe, here and 

 there marked with roseate spots, passes back to the suranal plate. The anal legs are represented 

 by two slender filaments held outstretched, which are nearly as long- as the body is thick. There 

 are seven jiairs of oblique, lateral, faint yellowish, slender stripes, the last pair extending to the 

 sides of the anal filaments. All the legs are pale green and concolorous with the body. 



A great change has occurred in the prothoracic tubercles which are now two low, flattened, 

 inconspicuous warts on the upturned or flaring edge of the segment. The anal legs are much 

 shorter in proportion and not so long as the body is thick, being about one-third as long in proportion 

 as in the third and fourth stages. 



This caterpillar we have observed when disturbed to send out from near the head a coi)ious 

 •shower of spray or vapor, being in this respect like that of Cerura, so carefully worked out by 

 Prof. E. B. Poultou. The opening is hard to find. The opening of the median prothoracic gland 

 is exactly like what we have observed in Cerura horeaUs. It is a transverse slit situated in the 

 median line of the body, between two transverse folds directly behind the head, but yet a little 

 way behind the front edge of the segment. It has slightly develojjed lips. 



The points of interest in the ontogeny known to us are as follows: 



The congenital characters are the enormously long stemapoda, in proiwrtioii to the body, and 

 the pair of long, prominent prothoracic tubercles. 



The acquired characters are the dorsal line, the oblique yellow bars, and the gradual reduction 

 in the length of the tail. 



Other features are : 



(1) The presence of fllamental anal legs exactly homologous with those of Cerura, and nearly 

 as long, and the fact that they are much longer in the early stages than in the final one, which 

 seems to suggest strongly the view that this genus is the ancestor of Cerura, and that tlie very 

 long lashes were of more use to the ancestors of the i^resent species than to the form we now 

 have. It will be remembered that M. marthesia ranges as far south as Brazil, and that it may have 

 originated in South America and spread northward; it is also possible that it had a set of enemies, 

 probably ichneumons, which it has not had to contend with in temperate North America, and 

 that the filaments have begun to diminish in size from partial disuse. On the other hand, the 

 spraying apparatus lodged in the first segment next to the head seems to perform its function 

 in undiminished vigor. Experiments like those made by iNIr. Poulton on the fluid secreted by 

 Cerura should be conducted with the present insect. 



(2) The second point is the complete reduction in size of the two high iirothoracic spine-like 

 tubercles which takes place at the last exuviation. 



