190 KENDALL: WHAT CHARACTERS DISTINGUISH SPECIES? 



ing the fact, that, if we were ignorant of the existence of the 

 connecting forms, the two extremes would be sufficiently dif- 

 ferent to be regarded as distinct taxonomic species. 



Phylogenetically the form at the center of distribution would 

 typify the species. The fact that one extreme or the other may 

 be the center of distribution does not affect the question. Un- 

 fortunately the species of the present-day taxonomist is often 

 already named and may have come from any point in the hy- 

 pothetical area mentioned. It may or may not be one or the 

 other extremes or it may or may not represent the center of 

 distribution. Consequently a subdivision of the species may in like 

 manner represent almost any point more or less remote from the 

 locality represented by the taxonomic species. In fact, accord- 

 ing to his niceties of discrimination or his idiosvncrasies, one 

 might make any number of subspecies, or whatever subdivisions 

 of a taxonomic species it is decided to call these geographical 

 representatives of a single natural species. 



Now, if the two lines of differentiation, previously mentioned 

 as developing at each margin or extreme in more or less remote 

 time, should gradually separate, leaving a gap in the intergrada- 

 tion thus restricting the interbreeding to two independent 

 lines of further development, which proceed without further 

 interruption to the present time, they would result in two re- 

 lated natural species, conforming to the specific taxonomic 

 definition of species of the aforesaid authorities. The degree 

 of difference between these two species would depend, partly 

 at least, upon the remoteness of the period of divergence. 



One of these independent lines, at one period or another, 

 may have repeated either one or both of the previously mentioned 

 situations. The more recent the divergence, the less distinct 

 the differences, until at the present time they are perhaps almost 

 indistinguishable from the first mentioned horizontal or real 

 intergradation. In fact, they may be quite indistinguishable to 

 the systematist having before him only small collections or col- 

 lections from a few localities only. 



Again, suppose that, in some past period of time, more or 

 less remote, a small portion of one of the main divergent lines of 



