NO. 1140. NORTH AMERICAN NOCTUIDAE— SMITH AND DYAR. 



character corres[)ODded. On these lines our A. fiineralis would be an 

 Agaristid or ISToetuid, since this species and some Agaristids and ISToc- 

 tuids possess in common single spatulate hairs. A. luteicoma would 

 be a Liparid from the brush-like tufts; A.noctivaga an Arctian from 

 the tufts of stiff hairs and black coloration; A. loheliae a Lasiocampid 

 from the weak hairs and flattened form, etc. I mention these views of 

 Dr. Butler's, though they have been long since refuted, not only because 

 they refer specially to Acronycta, but also because they illustrate the 

 difference between a classification based on superficial resemblance 

 versus one on phylogenetic characters. I will briefly review these more 

 fundamental points. They are seen in the position of the hairs (single 

 or multiple) rather than in their modification. The modification is 

 essentially adaptive, is quickly affected by a difference in habit, and 

 hence not reliable in classification. The position, on the other hand, is 

 only slowly affected by evolution and gives characters for large groups. 

 These groups jjrove to be of the grade of superfamilies. The five 

 primitive setae on the abdomen, as in Plate XVIII, fig. 1, are char- 

 acteristic of theTineides (the Micro-lepidoptera and allies), Saturniaus, 

 and Butterflies. Tubercle V moved up before the spiracle, is the 

 condition in the Sphinges (Plate XVIII, fig. 2), while Tubercle IV 

 moved up behind the spiracle is that of the Bombyces (Plate XVIII, 

 fig. 3). The several families of the Bombyces are separated by a num- 

 ber of lesser characters. The lower families have single hairs, as in 

 the figure. The higher ones have a tuft of hairs in place of each 

 single hair. Among these there are two distinct lines of modification 

 of the thoracic hair tufts (warts) which separate two large groups. 

 The group including the Bombycidae proper leads up from the 

 Xotodontidae, culminating in the Lasiocampidae. The other starts 

 from the Xoctuidae and includes the Arctiidae, culminating in the 

 Euchromiidae (Syntomiidae). The conversion of single hairs into 

 warts has taken place several times independently among the Bom- 

 byces, so that the ultimate structure is not an absolute criterion of 

 affinity. Moreover, a return to the single haired condition sometimes 

 occurs, so that these characters have to be used with a certain 

 caution. The result is that family characters are often not strongly 

 marked in the larva, in which they contrast with the superfamily ones 

 mentioned above — characters which are more strongly marked than 

 those of the same grades in the imagines. 



Now there are certain moths, usually classed as Xoctuidae, whose 

 larvae have many-haired warts. They have been recently discussed by 

 Mr. Grote under the term Apatelidae. There are, however, no family 

 characters in the moths, though Professor Smith deems himself almost 

 warranted in separating part of them, as a subfamily. The other ])art, 

 including Acronycta, are true Noctuids. It is this group, the Apateli- 

 dae of Grote or Pantheinae and certain genera of the Noctuidae which 

 form the subject of the present article. 



