592 



SCIENCE. 



[X. S. Vol. XXI. No. 537. 



which seems to indicate that the Tennessee 

 has persisted in its present course for a long 

 period of time.* 



If the distribution of fresh-water faunas 

 can not be regarded as a proof of river-capture, 

 it is pertinent to ask, of what value is this 

 class of evidence? 



It seems to me that this will largely depend 

 upon local conditions which must be taken 

 into account in each individual problem. If 

 there is a marked similarity or identity of 

 forms in the captured stream and the stream 

 representing the supposed former, lower 

 course, and no such relation is found in any 

 other two streams of the region, the evidence 

 would be very suggestive. If the special forms 

 thus distributed are so constructed anatomic- 

 ally as to be poorly adapted to dispersion by 

 birds, insects, etc., the evidence would become 

 much stronger. But if the divide between 

 the two basins is low and indistinct and 

 occurs in a broad, open valley along which 

 aquatic birds are known to migrate habitu- 

 ally, and the shells in question are adapted to 

 the various means of dispersal, then the oppor- 

 tunities for transference of forms between the 

 two basins would be so excellent that the 

 faunal evidence would be worthless as a proof 

 of river capture. On the other hand, where 

 no commingling of forms occurs, it might 

 appear that no recent capture could have 

 taken place and the evidence thus become of 

 negative value. But even here we must take 

 into consideration the restricted distribution 

 of some forms along the same stream, due to 

 the character of shores and stream bed, the 

 intervention of falls or rapids and other fea- 

 tures. Even where capture has taken place, 

 the forms may not be transferred to the lower, 

 new course of the stream, since they may not 

 be found in the lower courses of streams long 

 established in their present relations. 



In conclusion, it is believed that the dis- 

 persion of fresh-water faunas is effected by 

 so many different agencies, and the fea- 

 tures of distribution are dependent on so 

 many different factors, that such distribution 



* The full discussion of the Tennessee problem 

 will appear in a forthcoming issue of the Journal 

 of Geology. 



can have but very limited value as an evidence 

 of drainage modifications. In the cases where 

 this evidence has already been offered as a 

 proof of river-capture, it is believed that the 

 conditions are such as to render its use in- 

 valid. It seems necessary to subject such 

 evidence to unusually critical examination 

 before offering it in support of any theory of 

 drainage modifications, or accepting it as 

 proof of the correctness of any such theory. 



Douglas Wilson Johnson. 

 Department of Geology, 

 Massachusetts Institute of Technology. 



CURRENT NOTES ON METEOROLOGY. 



LONG-R.ANGE WEATHER FORECASTS. 



The Weather Bureau has wisely published 

 a Bulletin (No. 35) on the subject of ' Long- 

 Range Weather Forecasts,' prepared by Pro- 

 fessor E. B. Garriott, in order to counteract, 

 so far as is possible, the misleading predictions 

 for a month or a season in advance which are 

 constantly finding their way into our news- 

 papers. Indeed, such spurious long-range 

 predictions are actually sold to the papers and 

 to the public, and are most injurious in their 

 effects. Long-range predictions are of vari- 

 ous kinds, ranging from those based upon sup- 

 posed planetary influences to such well-known 

 statements, found in farmers' almanacs, as 

 ' About — this — time — expect — showers,' these 

 five words being so printed that they apply to 

 a week or ten days of time. There are also 

 other classes, based upon a careful study of 

 sunspot periods, lunar periods, etc., some of 

 which, as in the case of the recent investiga- 

 tions of Sir Norman Lockyer and Dr. W. J. 

 S. Lockyer, seem to promise something in the 

 way of more definite results. As to lunar in- 

 fluences, although much time has been spent 

 on this matter, and faint lunar tides in the 

 atmosphere have been made out near the 

 equator, in the present state of our knowledge, 

 as Angot put it two or three years ago, * it 

 can not be affirmed that the moon does exert 

 any influence upon the weather, but at the 

 same time it should not be denied that this 

 influence may possibly exist.' As to seasonal 

 predictions based upon the behavior and con- 

 dition of animals, it is clear that the physical 



I I 



