40 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 93 



consider Carcelia in most respects to be similar to Winthemia. While 

 I admit a strong systematic relationship, it is doubtful if Carcelia 

 and Winthemia are as closely related biologically as perhaps Sisyropa 

 and Winthemia. In the Nearctic realm Carcelia is b'.ologically re- 

 lated to Gymnocarcelia, genotj^pe ricinorum Townsend, a bare-eyed 

 species with Carcelia-like characters that has been removed from 

 Sturmia. Townsend places Carcelia and Gymnocarcelia in the tribe 

 Carceliini and the genus Winthemia in the tribe Sturmiini. 



In a restricted sense the Nearctic species belong to the subgenera 

 Carcelia and Paraexorista, but the lines of demarcation between the 

 two genera as previously defined are clearly lacking. The following 

 treatment is suggested for those workers who prefer or insist on 

 restricted genera or a definition of species groups : 



1. formosa, inftatipalpis, lagoae. 



2. diacrisiae, amplexa, perplexa, olenensis, yalensis. 



3. reclinata (Paraexorista). 



4. malaeosomae, laxifrons, protuberans. 



5. flavirostris. 



6. gnava (Chetoliga). 



7. separata (Scnometopia). 



The combination of groups 1 and 2 equals Townsend's concepts of 

 Oxexorista. The combination of groups 3 and 4 equals the previous 

 concept of Paraexorista. The recognition of Chetoliga would be 

 dependent upon dividing Carcelia (s. s.) into a gnava group and a 

 oombylans group. As no useful taxonomic or biononric result can 

 be achieved at this time by carrying out such a program, the genus 

 Carcelia is retained as proposed. 



The oviposition habits of the genus Carcelia should be discussed to 

 demonstrate its biological unity. Lundbeck states that "gnava and 

 cheloniae belong to Pantel's group 10 as characterized by pedunculate 

 eggs, and the same no doubt holds good for the other species" of 

 Carcelia. Brown mentions that the eggs of gnava are laid on the ends 

 of the hairs of the hosts. Townsend describes the eggs of Paraexorista 

 and Oxexorista as being of the elongate, cylindrical, pedunculate type. 

 The females of formosa, amplexa, laxifrons, malaeosomae, separata, 

 rasa, and cheloniae (if not confused with laxifrons or the so-called 

 "cheloniae 1 "' in American literature) all deposit elongate, cylindrical, 

 pedunculate eggs that are attached either to the hairs of the host or 

 fastened to the body. C. gnava, which has been much confused with 

 separata in literature, has a strange and most interesting habit of 

 projecting its eggs or larvae through the air successfully for a distance 

 from one-half the length of the body of the adult up to at least one-half 

 inch. In agreement with Lundbeck, the writer believes that the eggs 

 of the other species of this group will be of the elongate, cylindrical, 

 pedunculate type. 



