10 SMITHSONIAN MISCELLANEOUS COLLECTION- VOL. 51 



than an older and well established superfamily whose forms have 

 become fixed through a long period of conformity to their environ- 

 ment. If this be not conceded, it becomes impossible to treat the 

 younger superfamilies by any satisfactory system. 



It will be alleged by some that such plan will result in multiply- 

 ing genera unduly. There is, however, no doubt that the course 

 adopted is warranted by the conditions. This conclusion has been 

 reached after full and mature deliberation. The only possibility of 

 successfully systematizing the superfamily, so that its myriads of 

 forms can be designated definitely by name, lies in the recognition of 

 genera founded upon comparatively slight characters — slight com- 

 pared with those recognized as the standard in the older and less 

 specialized superfamilies. The differences between genera are less 

 pronounced in the more specialized than in the less specialized 

 groups. All are genera, and of equal value systematically ; but, as 

 already pointed out, they can not be measured by a standard gauge. 



The writer has always contended that a proper treatment of the 

 Muscoidea demands the definition of smaller categories and more 

 carefully restricted genera (see Psyche, vol. vi, p. 313, Sept., [892 I. 

 As the characters of the early stages are investigated, more light 

 will be thrown on higher divisions in the superfamily. Such a vast 

 assemblage of closely related forms is not amenable to separation, in 

 the adults, into divisions conceived on lines of mathematical pre- 

 cision. Any system of classification must become more or less arti- 

 ficial if it attempts, in the presence of intermediates and the absence 

 of a knowledge of early-stage characters, to mark off precise lines of 

 division between categories of higher value. When the interme- 

 diates are lacking, or largely lacking, it becomes a comparatively 

 easy matter to fix the lines of demarcation, and the system appears 

 extremely natural simply through the absence of the immense mass 

 of intermediate forms that at one time existed. But when these 

 numerous intermediates and intergrades are extensively present, any 

 attempt to apply an arbitrary system of classification to the group 

 can not but result in disaster. A system can be thoroughly natural 

 only in so far as it indicates natural types of families, subfamilies, 

 tribes, and genera, and groups the intermediates and intergrades 

 around them. Properly conceived and executed, such a system is 

 the only natural one, since it must accord with the facts as known. 

 At the same time the fact must not be lost sight of that taxonomy 

 is at best merely a means to an end, and does not exist in nature. It 

 is artificial in its original conception, because it is practically in- 

 tended to ignore numerous steps in the development of life — steps 

 that have been lost during the evolution of forms now existing, and 



