518 PROCEEDINGS OF THE NATIONAL MUSEUM vol. loi 



Group H. Edwards italicized any scientific names appearing in the 

 group titles, but since these titles are vernacular in nature, this policy 

 has not been followed here. In addition to the above, many species 

 have been removed to other groups. 



The greatest difficulty has been encountered in defining the groups 

 so that the definitions would be sufficiently diagnostic to permit the 

 preparation of a key, and it is quite plain that the results obtained have 

 not been successful in all cases. 



For example, the Madagascar species monetus Edwards and philUpi 

 van Someren, while unquestionably related to the other members of 

 Group C, break down the group definition in the possession of all 

 dark hind tarsi and will normally key to Group H (where on the basis 

 of scutal pattern we find monetus amazingly similar to leucocelaenus 

 Dyar and Shannon and leiicotaeniatua Komp) . This striking parallel 

 development of similar ornamentation patterns in species from differ- 

 ent groups and different geographical regions suggests that the poten- 

 tiality for the development of at least some of such characters is inher- 

 ent within the subgenus as a whole rather than within any particular 

 group. 



Another type of problem encountered in defining and keying the 

 groups is illustrated by such a species as qudsiruhithorax (Theobald). 

 This species usually possesses a linear scutal marking pattern and 

 on the basis of this and other characters appears to belong to Group 

 D. However, in some specimens this scutal pattern either may be 

 obliterated by general pale scaling or it may be absent altogether; as 

 a result such specimens key to Group F. 



The policy followed with marginal species has been to include them 

 in the group to which they appear to be most closely related on over-all 

 general appearance. 



The question has been raised of whether or not the groups of Fin- 

 laya could be considered subgenera if Flnlaya were to be given generic 

 rank once again. It is true that names are already available for all 

 but Group C, and that in almost all cases one has no difficulty in recog- 

 nizing the group to which a species belongs ; yet a serious handicap to 

 such a plan lies in the fact that all the groups are erected on ornamen- 

 tation characters solely, and consequently, in many cases, unrelated 

 species are included. Additional and more relevant arguments for 

 the retention of the larger generic concept have been aptly phrased by 

 Edwards (1932) as follows: "The advantages of employing larger 

 generic concepts are, firstly, that the wider relationships of the species 

 are more clearly indicated; secondly, that limits can more readily be 

 assigned to the genera than in the case of more numerous and smaller 

 groups; and, thirdly, that it ensures the avoidance of duplication of 

 specific names; the use of subgeneric terms enables those who wi -h to 

 do so to make use of the smaller divisions." 



