Morphology and Phylogeny of Insects. 441 



the tlioracic legs, but somewhat to tlic side of or inwards from 

 it, in order to reject the homology between these appendages 

 and the tlioracie legs. It appears to me, however, that the 

 acceptance or rejection of homologies is, in the great majority 

 of cases, absolutely impossible without reference to embryo- 

 logy, which alone can show us whether the appendages in 

 question proceed or not from abdominal legs, of which rudi- 

 ments are formed in the embryo. If it is pi^ecisely the 

 embryological facts that are wanting the question must remain 

 undecided, and all conclusions are premature. Thus, for 

 example, many Tenthredinid larvje possess numerous abdo- 

 minal appendages which appear in the liighest degree similar 

 to those of lepidopterous caterpillars. But since the embryo- 

 logy of the saw-flies, apart from a few extremely incomplete 

 statements by Packard *, is as yet unknown, the nature of 

 the appendages in question, notwithstanding their great simi- 

 larity to those of caterpillars, cannot be precisely determined. 

 Of great interest is the question as to the morphological 

 value of the so-called gonapophyses, i. e, the male copulatory 

 organs, ovipositors, the sting of the Hymenoptera, &c. 

 Under the head of gonapophyses I also include, inter alia^ 

 the male appendices copulatarii of the Lepidoptera. Some 

 authors, such as Huxley and Dewitz, consider tlie gona- 

 pophyses to be the homologues of legs ; others, such as Glaus, 

 do not venture to express themselves positively upon this 

 question, and merely remark that the homology of the gona- 

 pophyses with the legs is not proved ; finally, certain authors 

 decisively reject this homology. Against the horaologization 

 of the gonapophyses with legs various evidence can be adduced. 

 Thus, for instance, in Machilis ventral styles are also found 

 upon the abdominal segments which bear the gonapophyses, 

 so that these segments are each provided with two pairs of 

 appendages. We have already seen that the question of the 

 value of the ventral styles of the Thysanura cannot be decided, 

 on account of our ignorance of their embryological develop- 

 ment. But apart from this, the fact of the presence of two 

 pairs of appendages upon one segment is not in itself an 

 argument against the homologization of these appendages 

 with the legs. In the first place it has been shown by the 

 beautiful investigations of Uljanin upon the post-embryonic 

 development of the bee f that it is possible for ap[)endage3 



* Packard, " EmbiyoLigical Studies on Hexapodous Insects,"' Memoirs 

 of tlie Peabody Academy of Science, v. 1, no. 3, 1872, 17 pp., ii plates. 



t B. Uljanin, " Zur postembryonalen Entwicklung der Biene," Proto- 

 kolle der Sitzungen der ^loskauer Gesellschaft der Liebhaber von Natur- 

 wissenscliaft, Anthropobigie uiid EtlmogT.iphic, Jahrg. ix., 1872, Moskau, 

 pp. 17-.'j2, Taf. ii. v. (in I{ii.-^ian). 



