MEM01K6 OF THE NATIONAL ACADEMY OF SCIENCES. 11 



of PanorjKi communis, form a series passing around the segment a little in front of its middle. 

 There are no setie corresponding in position to seta ti of lepidopterous larvie; that is, there is no 

 posterior pair of dorsal seta?, but the series of eight or ten setiferous tubercles to a segment form 

 a line passing on the side of the segment considerabh^ behind the spiracle. In the Trichoptera, 

 as Dyar has observed, there are no setiferous tubercles. 



In the larvffi of the more primitive groups of Lepidoptera, and in stage I of the more 

 specialized families, I have always observed that, looking at the animal from above, there are 

 two pairs of setiferous tubercles, or four in all, arranged in a trapezoid, the tubercles of the 

 hinder pair being wider apart than those of the anterior pair. This disposition in a primitive, 

 generalized form is seen in iig. 7, p. 03, of my monograph of Notodontida?, representing the 

 larva of AdeJa viridcUa, also fig. 9. This arrangement of the four dorsal setiferous tubercles 

 is common to all the more generalized lepidopterous families — as the families comprising the 

 Tineina, the Hepialida? (p. 72, 73, figs. 31, 32), the Pyralidte, Noctuidw, etc. If we refer to 

 these setiferous tubercles as arranged in longitudinal .series, such .series may be designated as the 

 dorsal, subdorsal, supra and infraspiracular series, though this classification does not well bring- 

 out the varying position of the posterior dorsal pair. 



A more exact designation is that proposed by Doctor Dyar in 1894, who showed that there 

 are two types of arrangement. "The first, which is by far the more generalized, consists (con- 

 sidering onh' the abdominal segments) of five tubercles above the spiracle on each side, three in 

 a tranver.se row about the middle of the segment and two behind; below the spiracle are two 

 oblique rows, containing, respectively, two and four tubercles. This type is found in Hepialus, 

 and is probably typical of the larva^ of the moths in Professor Comstock's first su!)order, tlie 

 JugatiB." 



"The second type contains two dissimilar lines of modification of the first type; Ixit, as thev 

 agree in number of tubercles and in other characters, I will consider them together. The fundi- 

 mental arrangement of the tubercles is as follows: On each side above the spiracle three tubercles; 

 below, or behind the spiracle and above the base of the leg, three more; on the ba,se of the leg, 

 three (or four) on the outside and one on the inside near the midventral line. These I propose to 

 designate thus, counting from the dorsal line down the side: Tubercles /, //, ///, above the spiracle. 

 iv, V, vi below it, the group of three on the outside of the leg as vii, and the single one on the 

 inside of the leg as viil. Tubercles vii and viii are present also on the legless abdominal seg- 

 ments (viz, segments 1, 2, 7, 8, and 9 of the abdomen) in the corresponding position."' 



Doctor Dyar has skillfully used these seta? or setiferous tubercles in classifying the families 

 of Lepidoptera, especially in separating the more generalized from the more specialized; yet in 

 some cases they hardly seem the best guide. It should l)e borne in mind that any single character 

 or set of characters do not prove in all cases reliable, since they shift or disappear in specialized 

 types and other characters are substituted and become prominent by })eing more useful and adap- 

 tive, and, as we shall see, the tubercles // are the most variable in position and development of 

 any on a segment during the ontogeny of a species or group, and are thus important in working- 

 out the phvlogeny of the Lepidoptera or .some .special group. 



The arrangement of the tubercles i-^vii on the trunk segment of a .syssphingine moth, is 

 illustrated by Fig. 1. It will be observed that tubercle v is below, while in the Sphingidje it is 

 situated directly in front, of the spiracle. 



Presence or alsence of dorsal tulerdes a. — The presence, reduction, or the atrophy of this 

 pair of tubercles in the course of the ontogeny has an intimate bearing on the phylogeny of the 

 subfamily under discussion. 



In the first stages of AdeJocampa hicolor this pair of dorsal tubercles are present (PI. XLV. fig. 

 2. //), the four doi'sal tubercles being arranged in a trapezoid on abdominal segments 1 to 7, and in a 

 modified form on segments 8 and 9. It persists through stages II to III. Their presence in the 

 fourth and final stages is more doubtful, since the surface of the body is scattered over with 



« " I have not had Micropteryx for examination.' 



