64 Proceedings of the Biological Society of Washington. 



ters that appear to me quite sufficient to warrant the generic 

 separation of Chrysanthrax from Hyalanthrax, and very probably 

 there are just as good characters available in the pupae of the 

 other so-called subgenera which are as yet unknown in the 

 pupal stages. It is necessary to indicate that in my paper, 

 recently published, on this group I retained the generic name 

 Anthrax for the two species of Hyalanthrax therein described — 

 hypomelas and lateralis.* These two species and another that 

 I have since obtained are primary parasites, and in this respect 

 differ from fulvohirta, but I do not purpose suggesting that the 

 species be separated generically on this account. There is in 

 the structure of the head capsule of the two species a very great 

 difference, and it is upon this character that the groups are 

 entitled to generic and not subgeneric separation. The num- 

 ber and strength of the thorns which are present upon the head 

 capsule of pupae of different genera of Nematocera, or the 

 armature of the thoracic and abdominal segments are indices 

 in great measure of the pupal habitats of the species that bear 

 them. The cephalic spines are not only shields or sheaths for 

 the antennae or other cephalic appendages, but are retained 

 and used as instruments to assist in the work of emerging from 

 the ground or other habitat — a process undertaken by the 

 imago of this group before it leaves the pupal skin. In the 

 paper just cited I have stated that I consider the absence of 

 cephalic armature in the Tahanid* and its presence in the 

 Asilidse to be due to the fact that the former are normally 

 present only in soil that is damp or sandy and easily pene- 

 trated, while the latter iAsihis et aL) are almost invariably 

 found in soil that is dry and much more compact. There is, 

 therefore, a greater need for strong armature on the head in 

 Asilidffi than there is in Tabanida?, though both have abdomi- 

 nal locomotor spines. It is probably unnecessary for me to 

 emphasize this fact further, but it may be of interest to restate 

 the fact that the pupae of the Cyrtidse have neither cephalic 

 armature nor abdominal spines, neither being requisite for the 

 emergence of the imago as the larvae are parasitic in spiders; 

 and it is certainly interesting to discover that there is in con- 

 junction with the different larval hosts of the species of 

 Chrysanthrax and Hyalanthrax a coincident difference in pupal 



* Bull. 111. state Lab. Nat. Hist., Vol. XI, Art. 4. 1915, pp. 332-334. 



