588 



SCIENCE. 



[N. S. Vol. XXI. No. 537. 



Colo., in the last four months, I find a num- 

 ber of specimens of Ilyocypris hradyi Sars. 

 A collection made from a small stream in 

 central Illinois in August last, consists en- 

 tirely of a species of Ilyocypris, allied to I. 

 iners Kaufmann, which appears to be unde- 

 scribed. 



The genus Ilyocypris Brady and Norman is 

 widely distributed in Europe but has not hith- 

 erto been found in America. Including the 

 two forms above mentioned, the family Cypri- 

 didse is represented in North America by 44 

 recognizable species comprised in 12 genera. 

 Of these, 2 genera, comprising 3 species, are 

 exclusively American ; the remaining 10 genera 

 are represented by 13 species common to 

 Europe and North America, and by 28 species 

 which have been found only in North America. 



A full description, with drawings, of the 

 Ilyocypris from Illinois is in preparation. 



Arthur E. Beardsley. 



State Normal School, Greeley, Colorado, 

 March 17, 1905. 



SPECIAL ARTICLES. 



THE DISTRIBUTION OF FRESH-WATER FAUNAS AS 

 AN EVIDENCE OF DRAINAGE MODIFICATIONS.* 



As the result of careful studies of stream 

 development, it has been well established by 

 a number of investigators that very important 

 changes in the arrangement of drainage lines 

 are often produced by the capture of a portion 

 of the waters of one stream by a tributary of 

 some neighboring stream. It is but seldom 

 that the actual process of immediate capture 

 is witnessed. We most frequently see the 

 evidence of conditions which we believe will 

 ultimately lead to capture, or results which 

 we believe have been produced by capture 

 sometime in the past. 



Whenever one stream succeeds in capturing 

 a portion of the drainage system of one of its 

 neighbors, there are certain results which 

 must necessarily follow, just as there must 

 have been certain conditions present to make 

 the capture possible. By a study of the re- 

 sults produced it is often possible to learn 

 what were the former relations of streams in 



• Paper read before the Philadelphia Meeting 

 of the Association of American Geographers. 



a given region, and so prove the fact of cap- 

 ture, and even the approximate time of its 

 occurrence. The evidences of drainage modi- 

 fications, therefore, are of prime interest to 

 the student of geography. 



It is not my purpose to review the several 

 results which are produced when one stream 

 captures another, but rather to direct atten- 

 tion to one of the results produced, and to 

 consider its value as an evidence that capture 

 has occurred. At the outset it is necessary 

 to divide the features produced by river cap- 

 ture into two distinct classes: (1) those fea- 

 tures which are produced by river-capture and 

 which can be produced by nothing else; (2) 

 those features which are produced by river- 

 capture, but which may also be produced by 

 some other agency. Features belonging to the 

 first class are of themselves definite proofs 

 that river-capture has taken place. As an 

 example of this type of evidence we may note 

 the occurrence, along the former course of the 

 stream which has suffered capture, of river- 

 brought gravels which are so distributed that 

 they could have reached their present position 

 only through the agency of the captured 

 stream. Features of the second class, how- 

 ever, when taken alone can not be regarded 

 as proofs of river-capture, since, according to 

 the basis of classification, they may also be 

 produced by other agencies. Considered by 

 themselves they are only of suggestive value; 

 other lines of evidence must be appealed to 

 before the river-capture, of which they may 

 be the direct resirlt, can be proved to have 

 taken place. As an example of this type of 

 evidence we may cite the continuation of a 

 broad, open valley along the former course of 

 a large, mature stream which has been diverted 

 by capture. A similar valley may also be pro- 

 duced by a relatively insignificant stream, 

 provided it is working on a band of soft, easily 

 soluble rock, as has been the case along certain 

 headwater branches of the James and Roa- 

 noke rivers. The existence of such a valley 

 alone is, therefore, not conclusive evidence of 

 capture, however strongly it may seem to sug- 

 gest it. 



It is well known that different streams are 

 often marked by certain peculiarities of the 



