June 2, 1911] 



SCIENCE 



861 



species from the latter's synonymy. My own 

 publications on the subject, -which have been 

 quite extensive and in which I have proposed 

 many new genera and species, no small part 

 of which may quite possibly have to go in the 

 final synonymy, were produced almost in 

 whole without the above knowledge but with a 

 very considerable appreciation of the necessity 

 for a most careful and minute comparison of 

 external characters. Brauer and von Bergen- 

 stamm's monumental work, performed under 

 the same conditions, and unquestionably the 

 best and most advanced of all, must be classed 

 here too, along with all other taxonomic work 

 in the Muscoidea to 1908. The results in all 

 these cases have been quite unfortunate, con- 

 sidering the amount of time and energy ex- 

 pended. All the material handled by Mr. 

 Coquillett will have to be restudied with great 

 care in the light of dissections of fresh ma- 

 terial from type localities. My own types and 

 those of Brauer and von Bergenstamm will 

 have to be restudied in the same manner. In 

 fact, all accessible muscoid types the world 

 over will have to be restudied in this new 

 light. Here is an amount of work to be done 

 that almost staggers one to contemplate. 



Brauer and von Bergenstamm possessed a 

 most acute appreciation of the necessity for 

 searching out even the most minute external 

 characters in order to arrive at the true rela- 

 tionships of the forms. They probably car- 

 ried the study of the external adult characters 

 about as far as it can be advantageously done 

 without correlation with the reproductive and 

 early-stage characters. I have perhaps gone 

 somewhat farther in my consideration of the 

 external adult characters in the " Taxonomy 

 of the Muscoidean Flies," " but so far as I yet 

 know without any great improvement in the 

 general results. It is thus evident that, for 

 the future, the older order of taxonomic work 

 in these groups must be exchanged for the 

 newer one, which has come into full light but 

 recently, and which demands the exhaustive 

 study not only of the external and largely the 

 internal characters of the adult, but also of the 



»" Smiths. Mise. Colls.," No. 1803, May, 1908. 



characters of the eggs and early stages. It 

 may even greatly profit by a study of general 

 bionomics, especially host relations. 



It is truly a most remarkable state of affairs 

 that finds us at the present day unable to 

 define some of the most common genera of 

 tachinid flies. Nevertheless, such is the fact 

 and necessarily follows from what has here 

 been said. The type species of each genus 

 must be dissected before we may know what 

 species, themselves dissected, can be referred 

 to that genus. The material for such dissec- 

 tions should be fresh, and that for type dissec- 

 tions should be obtained from the type locali- 

 ties so far as possible. I have already done 

 this work for a considerable number of genera, 

 and the results will, I hope, be published 

 within the year accompanied by necessary 

 drawings. But hundreds of genera, many of 

 them represented by names long in common 

 use, remain to be investigated in this manner, 

 and thus we frequently find ourselves at this 

 late day unable to determine material in these 

 groups with any hope of finality. 



In a recent letter to me. Dr. John B. Smith 

 has restated the conditions in the following 

 apt words, which I can not refrain from 

 quoting : 



It is perhaps not surprising that in the Diptera, 

 which are without any doubt physiologically the 

 most highly developed of all orders, the difficulties 

 in classification should be greatest. Their special- 

 ization has extended in so many directions that 

 thq divergencies have become marked by internal 

 modifications rather than external adaptations. 



He precedes these remarks by stating his 

 belief " that it will require a study of the 

 internal organs to get a satisfactory classifi- 

 cation, which may afterwards be helped out 

 by external characters whose importance is 

 not recognized at the present time." This 

 remark is well worthy of consideration. The 

 correlation of the external adult characters 

 with those of the reproductive system and 

 early stages will define the relative taxonomic 

 value of the first in the various groups, and 

 may reveal unsuspected characters among 

 them which will hold good for considerable 

 series of groups. 



