April 14, 1905.] 



SCIENCE. 



589 



faunas which inhabit them. The fresh-water 

 shells of one drainage system may be dis- 

 tinctly different from the shells found in the 

 neighboring drainage systems on either side — 

 so much so, in fact, that the student of these 

 forms can often tell at a glance from what 

 locality a given "museum specimen has come. 

 Now it is evident that if river-capture takes 

 place, the shells of the captured stream will 

 mingle with those of the capturing stream. 

 Such a commingling of faunas becomes an 

 evidence of drainage modifications, and it is 

 with this type of evidence that the present 

 paper is concerned. It is important to know 

 whether this evidence belongs to the first of 

 the two classes above outlined, being a positive 

 proof of river-capture; or whether it belongs 

 to the second class, and is, therefore, only of 

 suggestive value. 



The line of evidence in question is not new. 

 It has already been advanced in support of a 

 great example of supposed river-capture. The 

 Tennessee River after flowing southward 

 through eastern Tennessee to a point near the 

 present city of Chattanooga, leaves its broad, 

 open valley and turns abruptly westward 

 through a deep, narrow gorge in Walden 

 Plateau. Dr. C. W. Hayes and Mr. M. E. 

 Campbell, in a paper on ' The Geomorphology 

 of the Southern Appalachians,' published in 

 1894, advocated the theory that the Tennessee 

 formerly continued on southward via the 

 Coosa and Alabama rivers into the Gulf of 

 Mexico, but near the close of the Tertiary 

 period was captured at the point near Chatta- 

 nooga by a branch of a stream farther west. 

 Six years later Mr. Chas. T. Simpson, study- 

 ing the fresh-water shells of this region, came 

 to the same conclusion; and in a paper on 

 ' The Evidence of the Unionidse Regarding 

 the Former Courses of the Tennessee and 

 Other Rivers,' published "in Science . in 1900, 

 independently supported the theory of capture 

 on the basis of the biological evidence alone. 



This latter line of evidence was urged as 

 absolute proof of the supposed river-capture, 

 it being held that the shells must have direct 

 water communication in order to pass from 

 one stream to another. It is true that else- 

 where the author has advocated other means 



of dispersal than river-capture, but these 

 means were not considered in connection with 

 the Tennessee problem. In this case the 

 whole strength of the argument lay in the 

 statement, ' these forms can not travel over- 

 land from river to river, but must have water 

 communication in order to pass from one 

 stream to another.' And since shells pe- 

 culiarly like those of the Tennessee drainage 

 were found mingled with the usual forms of 

 the Coosa and Alabama rivers the conclusion 

 was reached that the Tennessee must have 

 been diverted from a former southward course 

 by capture near Chattanooga. The paper has 

 been widely accepted and quoted as an ex- 

 ample of the definite proof of river-capture, 

 and some who could not accept as conclusive 

 the physiographic evidence presented by 

 Messrs. Hayes and Campbell in support of 

 the theory of capture were impelled to regard 

 the biological evidence a final proof that the 

 capture had taken place. 



In connection with a study of the Tennessee 

 problem I have been especially interested in 

 the evidence furnished by the distribution of 

 the fresh-water faunas, and have become con- 

 vinced that the evidence should be included 

 in the second of the two classes above out- 

 lined, being produced by river-capture in some 

 cases, but also being produced by other agen- 

 cies as well, and, therefore, not being con- 

 clusive in favor of capture. The reasons for 

 this will best appear if we consider some con- 

 crete case, as that of the Tennessee, in which 

 this line of evidence has been especially em- 

 ployed. 



The facts brought forth by a study of the 

 Unionidffi are as follows: Pleurohema, a genus 

 of Unio, has its metropolis in the Tennessee 

 River. It is not found throughout the other 

 portions of the Mississippi basin. But it is 

 found abundantly in the Coosa and Alabama 

 rivers. Also certain other forms of Unio 

 common to the Mississippi-Tennessee basin 

 are found in the Coosa-Alabama basin. From 

 these facts it was concluded that at some time 

 the upper Tennessee River must have flowed 

 southward into the Coosa-Alabama River. On 

 the basis of this line of argument we must 

 necessarily assume that the fresh-water mus- 



