344 



margin near the tip. The whole organ is quite variable in its 

 details. 



Pleopod IV is intermediate between III and V in form. 



Pleopod V. The exopod appears to be ovate but is actually 

 four sided. The true posterior margin is bounded by the tip and 

 by an angulation about one-third the way up the lateral margin. 

 A terminal spine seems to be always present, a second, smaller 

 spine sometimes occurring lateral to the first. 



In spite of the variation noted above, certain characters, 

 especially the details of the exopods and the general outlines of 

 the endopods are very constant. These with the form of the 

 antennal lobe may be taken as diagnostic of the species. From 

 my study of T. demivirgo and comparison with available figures, 

 I am strongly in favor of Racovitza's view that it is possible to 

 distinguish the females- of this subgenus in the absence of males. 



The present species adds another name to the list of those 

 which would, according to Verhoeff , fall into the composite species 

 T. caelebs. The caelebs question has been recently discussed in 

 detail by Herold (1929). My own view of the matter, after a 

 careful consideration of the various views, is to avoid the for- 

 mation of a composite species and to restrict T. caelebs to the 

 form described by G. 0. Sars (1898, p. 161, pi. 72, fig. 1) as 

 T. pusillus. It is greatly to be desired that the parthenogenetic 

 females be studied sufficiently to enable them to be placed under 

 definite specific names. 



The subgenus Trichoniscus. — I think that there is now no doubt 

 that this name should be applied to the group which Racovitza 

 designated as Spiloniscus. The subgenus is a rather large one, 

 but I have not found any characters which yield a satisfactory 

 basis for a further division. It is probable that when we have a 

 more detailed knowledge of the general structural characters 

 of the various species this division may be possible. 



The nearly 30 species which may be assigned to this subgenus 

 can be arranged in convenient, if somewhat artificial, groups as 

 shown in the following key. 



1. Ocelli wanting 2 



1 . Ocelli one on each side, dorsum tuberculate ? thielei Verhoeff. 



(This may belong in Miktoniscus.) 

 1. Ocelli 3 (2-5) on each side 3 



