540 



THE AMERICAN NATURALIST 



[Vol. XLII 



the lack of anatomical details because lie could not dissect 

 the one precious specimen ! 



The best way to avoid trouble in taxonomy is to begin 

 with making the original description as complete as pos- 

 sible. It is then a simple matter to condense for practical 

 diagnostic purposes. The biometric determination of 

 the variability of the different characters allows of their 

 arrangement in the order of increasing variability and 

 will perhaps demonstrate which are the "best" char- 

 acters. It is too soon to state exactly what percentage of 

 divergence justifies the creation of a variety, or of a 

 species, and it is not probable that any universally ap- 

 plicable measure will be established. 



Biometry gives us data of value bearing upon one or 

 more of the factors of evolution and records them in the 

 best form for use. We know that under varying en- 

 vironmental conditions different varietal, or specific 

 forms respond. The new forms may be only temporary 

 and followed by still other new forms, under other 

 changed conditions, or with reversion, when the old condi- 

 tions are restored. Characters, whether expressed in 

 the terms of biometry, or not, are not permanently fixed 

 by the publication of a specific description. It is to be 

 hoped that the spirit of Darwin is with us yet and that 

 we realize that species are in a state of evolution, either 

 continuously or discontinuously, slowly or rapidly. If 

 we are to follow species in their evolution we must have 

 exact and comprehensive statements of their characters 

 from time to time as it is possible, and, if in the terms of 

 biometry, these statements are always in harmony and 



If then, while performing the necessary work of tax- 

 onomy, we can make our descriptions more complete and 

 nearer the truth ; if occasionally we may be relieved of 

 the necessity of creating a new species ; if our work may 

 contribute to the advancement of the philosophy of biol- 

 ogy, should we not welcome biometry as a method which 

 can well serve in one and all of these things? 



