82 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



He again wrote October 29 that he found a full-grown caterpillar on the Mesquite {Prosopis 

 julifJai'd), which pupated the following day, burying themselves in his garden plot. On the 26th 

 October he again went to the desert and found four more of the same larvie on ParMnsonia 

 microphylla. All were fully grown, and the .smaller one was sent to me. 



I have collected on that desert for nine years, and am positive that this larva did not previously make its appear- 

 ance in this locality. Palo \'erde trees have very scant foliage, and are easily looked over, so that a larva of such 

 beautiful ornamentation could not escape mine eyes. 



In Brownsville, Tex., it was observed b_v Mr. Doll from the end of April until late in the 

 summer. 



I add a description of the San Antonio larva, which appears to be a form of -1. Wmg. Mr. Joutel tells me that he is 

 quite sure that all the larvfe of this species figured on PL III are variations of A. isias. The more specimens I 

 examine the greater seems the range of varieties in the length of the thoracic and abdominally spines, and in the 

 degree of granulation of the skin. Such an amount of variation seems to indicate the primitive nature of this genus, 

 and its comparatively recent adaptation to spiny food plants. 



Larva (PI. HI, tig. 2, 2«; XLIX, figs. 2, 2«, 2h, 2c, 2^/).— I have been able to examine through 

 the kindness of Dr. H. G. Dyar a well-blown larva in the U. S. National Mu.seum from San 

 Antonio, Tex. , the food plant of which is unknown. He thinks it may be that of A. heiUghi-odtli, 

 and this seems probable. Unfortunately no specimens of this species have yet actually been bred 

 from the larva. 



Length, 5(» mm. The head (PI. XLIV, fig. li) is as in A. hicohir. l)ut slightly higher; pale 

 bluish green, extending from in front of the ej^es up to the vertex. On the front edge of the 

 prothoracic segment is a series of low rounded porcelain white tubercles, which are much as in A. 

 hieolor, but larger; thej" are nearer together, crowded, on the sides; and lower down are two 

 small tubercles on each side above the spiracle. 



The dorsal thoracic horns are longer than in A. Vicolor, being nearly as long as the body is 

 thick, much less curved than in the Brazil species, but more so than in ^4. hicoJor; they are not 

 very sharp, and are armed with a few rather sharp spinules on the posterior .side. As usual the 

 subdorsal horns are a little shorter than the dorsal ones: all are green at the base, becoming 

 ro.seate towards the middle. 



The dorsal spine of abdominal segments 1 to 7, all of tht- same she, .sharp, recurved, saber- 

 shaped, porcelain-white, becoming roseate at the tip; they are smooth except those on the 7th 

 segment, which are armed l)ehind with a few fine teeth. The spines of the supraspiracular 

 series (///) a little shorter, and as usual less prominent than those of the dorsal .series. There is 

 a series of small almost minute acute white infraspiracular tubercles {>v-\-i'). 



The body is coar.sely granulated. The porcelain-white tuberculets are larger and more 

 prominent than in A. hicolor. There is a group on the first thoracic segment below the spiracle, 

 and a larger group of still larger ones at the base of the thoracic legs. The sides of the second 

 and third thoracic segments are densely granulated, while on each of abdominal .segments 1 to 7 

 there are two transverse rows above the spiracles, and a group aliove the base of the legs. The 

 abdominal legs are also granulated, there being more on the fourth pair than on those in front 

 of them. 



The caudal horn (fig. 2(') is not so stout as in the Brazilian larva, l)ut longer and thicker 

 than in A. bicolor. The spinules on it are uuich more luunerous than in the Brazilian larva, and 

 each ends in a minute short seta. 



On abdominal segment 9 is a low porcelain-white median tubercle, conical, recurved. 



Snranal ])late (fig. 2e) rounded at the end, not deeply cleft, with a double row of crowded 

 tubercles or granulations around the edge; and at the extremity not a pair of tubercles, but 

 several small crowded granulations. 



Anal legs with setiferous granulations crowded around the edges, the central field smooth 

 except two or three small granulations, some of them long digitiform, but rounded at the tip. 



Body in the drj' specimen pale green, with a yellowish tint. A lateral bright conspicuous 

 deep line, scarlet above and whitish beneath, extends from the hinder edge of the third thoracic 

 to the ninth abdominal segment. 



