72 MEMOIRS NATIONAL ACADEMY OF SCIENCES. [voL.xn. 



But on the other hand the habits of the animal help to resist this persistent tendency to 

 wash the shells downstream. Besides the muscular power which enables the snails to hold 

 fast to the substratum, it probably has a positive rheotropic response which leads it into rapid 

 water and consequently at times upstream. That there is a tendency to migrate upstream 

 is also suggested by the fact that what seems to be the most ancient forms, the smooth shells, 

 are maiuly confined to the headwaters. The presence of these ancient kinds of shells in the 

 headwaters is also due in part to the slow migration of the habitat itself. It is a weU-known 

 physiographic principle that as drainage lines are perfected or developed the obstacles in their 

 paths (falls, rapids, shoals, or riffles) are removed and the areas of rapid water tend to migrate 

 upstream toward the divides (Adams, '01) and take their animals with them. The river pro- 

 files of the upper Tennessee system (Kingman, '00c; Kingman and others, '00, '00a) show clearly 

 that the course of these streams forms a succession of pools, progressively elevated one above 

 another toward the divides. These pools overflow at their rim and as their waters fall to the 

 lower basin the rapid water habitat is formed. In general these obstacles, the rims between 

 the basins, are removed first in the lower parts of the streams and progressively upstream 

 quiet waters advance and the rapids recede. This migration of the habitat is only a part of 

 the general problem of changes in drainage to which reference "nnll be made later. 



4. LONGITUDINAL DISTRIBUTION IN STREAMS. 



lo is a river snail. It does not live in creeks or brooks, and for this reason it does not 

 occupy the extreme headwaters of streams. However, other members of this family of shells, 

 as some Pleurocera do inhabit such situations exclusively and shun the very conditions in 

 which lo flourishes. After a stream reaches a certain degree of topographic development and 

 in other favorable conditions, as at Dryden on the Powell, Artrip on the Clinch, Saltville and 

 Fishdam on the Holston, and Conkling on the Nolichucky, lo makes its appearance and per- 

 sists in the downstream reaches of this drainage to the Muscle Shoals in northern Alabama. 

 In no case does the kind of lo found farthest upstream extend downstream to the other limit. 

 In the Powell, Clinch, and Holston a relatively smooth shell occurs in the headwaters for a 

 variable distance, only to be replaced sooner or later downstream by a very spinose kind. 

 Even in the Nolichucky, which has a spinose shell in the headwaters, the shell, while it remains 

 spinose, becomes more so downstream. There is then in aU the streams a replacement in the 

 kinds of sheUs and in the degrees of spinosity progressively downstream. Somewhat similar 

 conditions are foimd in some elongated lakes. In Lake Tanganyika of central Africa, Moore 

 ('03, pp. 149-150, 261) fomid and figures a Viviparid sheU, Neothauma tanganyicense, which 

 varies in its habitats in the lake. A strongly ribbed or keeled shell lives "exclusively at the 

 south end of the lake, swarming in the broad and more or less sheltered reaches into which the 

 southern end of Tanganyika expands. In the narrow, surf-swept, and turbulent portion of 

 the lake which stretches between the north of Cameron Bay and Tembwi, Neothauma is only 

 found in the little bays and sheltered places occurring along both shores, and here the character 

 of the form changes, the double-keeled sheU of the former variety being replaced by the elon- 

 gated type." * * * [A sheU with larger body whorl and aperture which is smooth and 

 lacks the keel.] "Northward the lake terminates again in more or less sheltered expanses 

 like the Gulf of Ubuari, the deep bays near Ujiji, and the extreme northern extremity of Tan- 

 ganyika. In these the form of the genus again changes, the two more southern varieties being 

 generally replaced by the curious rounded form." This is a less elevated shell similar to the 

 one from the central part of the lake. It should be noted that the shells living in the quiet 

 waters at each end of the lake are quite different, but the one from the exposed shores is much 

 like the one from quiet waters at the north end of the lake. This appears to be a replacement 

 of races comparable to the downstream change in the Tennessee system. 



This law of replacement is not confined to this family or genus of shells, but has long ago 

 been observed in many other kinds of animals. Thus MoUendorff (1873, pp. 59-60) has shown 

 that in the Bosnia River, a tributary of the Upper Danube, in Bosnia, the snail Melania holandri 

 Fer. (Family Melaniidse) varies greatly in passmg from the headwaters downstream. In the 



