314 



PSYCHE. 



[September 1892. 



that the related groups of the Orthor- 

 rapha, and even the Cyclorrapha of 

 older origin, as the Syrphidae, are 

 much more sharply separated from each 

 other, and easier to establish by definite 

 characters, than those of the Muscaria. 

 The Muscaria schizometopa, which, 

 exclusive of the Anthomyidae, are here 

 treated, have been repeatedly divided 

 into differently related groups, but no 

 author has been able to establish his 

 groups upon characters which apply to 

 all cases. In the description of genera 

 almost every one has overstepped, 

 either inadvertently or purposely, the 

 limits he had previously set. The 

 characters applied change in their con- 

 bination with such complexity that it is 

 almost impossible to discover one which 

 continues constant through several or 

 many genera, much less to discover a 

 harmony of characters indicating a nat- 

 ural system. After having labored 

 long in vain to define the previously 

 formed groups by constant characters, 

 we have come to the conclusion 

 that our object can only be attained in 

 another way, — by ignoring the previ- 

 ously established groups of Tachinaria, 

 Dexiaria, &c, and designating a much 

 larger number of smaller groups, or 

 related series. Within such a small 

 group it is possible to determine har- 

 moniously combined characters for a 

 series of several genera. How these 

 groups are to be arranged with refer- 

 ence to each other can, however, be 

 settled only very impeifectly. We refer 

 to this particularly, as it might other- 

 wise be supposed that we wish in the 

 sequence observed to indicate a serial 



relation. To us the relations seem 

 rather to extend in many directions from 

 every group, and to allow no simple 

 linear arrangement. 



The question may arise, why we do 

 not rather combine all the forms of a 

 group into one genus with many sub- 

 genera. On this question we take the 

 ground occupied by almost all syste- 

 matic writers at the present day. In the 

 first place, it is more intelligible to fix 

 our stations by names than by numbers ; 

 second, our genera, with few excep- 

 tions, show only plastic characters ; 

 third, we possess several species for 

 almost every genus, which are distin- 

 guished from each other by many sub- 

 ordinate differences ; and fourth, we 

 would so treat the Muscaria, and so 

 comprehend the value of genera, as to 

 agree with what Rondani, the best re- 

 viser of the group, has already done. 

 Whoever begins to divide the genera 

 Tachina, Dexia, &c, is compelled to 

 continue the work until only those 

 characters which are altogether subor- 

 dinate remain for the species, — the ar- 

 rangement of the macrochaetae, the 

 form of the sexual organs of the male 

 and female (not the typical develop- 

 ment for a group, but lesser characters 

 than these ; for instance in Phaniidae), 

 the color, hair, width of face, <.y.c. 



If we adhere to the proposition that 

 a genus comprises only a series of 

 species, disunited from other series by a 

 lack of connecting forms, then we must 

 mark off a new genus whenever, among 

 the slight subordinate differences of the 

 various parts of the body, new charac- 

 ters are found, lacking in the previous 



