198 GENUS ARENIVAGA (bLATTIDAE) 



All of the species of Arenivaga are subject to unusual individual 

 variation. The two specimens recorded by us from the states 

 of Morelos and Guerrero, Mexico, as rehni, we believe will prob- 

 ably be found to represent a distinct species. Until further 

 matei-ial can be examined from that region, how^ever, we do not 

 feel able to diagnose these individuals properly. We were influ- 

 enced in our originally too conservative attitude by the fact that 

 as many features, usually constant in the Blattidae, were subject 

 to considerable variation in the genus, we hesitated to use certain 

 male genitalic differences as bases for further specific division. 

 The fact that females of two of the already recognized species 

 were most difficult to separate gave us further reason for caution. 

 It is unfortunate that females of all the species excepting holliana 

 and floridensis are so difficult to separate. It is hoped that the 

 constant addition of material will eventually make it possible 

 to ascertain features of difference, l^y which this sex can also be 

 determined definitely for all of the species. Due to their -dis- 

 tribution, females of rehni, floridensis and tonkmva can be deter- 

 mined from locality alone.'' The problem of determining females 

 of grata, erratica, apacha and genitalis is, however, most difficult. 

 In atldition to assigning correctly the material previously 

 recorded, and discussing the impoi'tant features of the species 

 now recognized, we have recorded in the present pap(M" the few 

 additional specimens subsequently added to the Philadelphia 

 Collections or received l)y us for study. In preparing the present 

 paper the three hundred and fifty-one specimens of Arenivaga 

 in the Philadelphia C-ollections, as well as a number of individuals 

 sent us for study, have been examined. 



Key to Males of the Species of Arenivaga,-' based on the Concealed 



Genitalia:'' 



■'Though the distribution of tonkaum is hirgely coincident with that of 

 bolliana, the distin(!tive fc^atures of females of the latter species prevent 

 confusion between these. 



'' 80 mucli individual variation occurs in the species of this genus, in the 

 features normally used for specific separation, that we feel it is imperative 

 for the student to examine the concealed genitalia of all males to be recorded. 

 The other features which we consider of some diagnostic value, and the degrees 

 of variation known, are discussed under the species. 



