32 



SCIENCE. 



[X. S. Vol. XXI. Xo. .V2:3. 



Luray, Va. The hill referred to has extensive 

 caverns beneath it, and, as appears evident, 

 has been left in relief owing to the more rapid 

 denudation of the surrounding country; the 

 reason being that rain falling on the area 

 where the rock is cavernous percolated down- 

 ward and was prevented from forming surface 

 streams and in consequence lost its ability to 

 mechanically erode, while the surrounding 

 country where the existence of surface streams 

 was possible was degraded more rapidly. 



The influence of subterranean drainage, as 

 must be well known although seldom men- 

 tioned, is frequently indicated by minor eleva- 

 tions, especially in limestone regions where 

 joints and other openings permit of the ready 

 descent of surface water. Similar conditions 

 on a larger scale, as just stated, may reason- 

 ably be held accountable for the origin of the 

 hill above the caverns at Luray, and seem- 

 ingly furnish the basis for an hypothesis which 

 meets the conditions present at Mackinac 

 Island and Gibraltar. If this hypothesis is 

 sustained by future tests, it not only furnishes 

 an explanation of the origin of the elevations 

 just mentioned, but embodies a principle 

 which is widely applicable. For example, it 

 is frequently stated in modern text-books of 

 physical geography, that residual hills stand- 

 ing on plains of subaerial denudation or 

 ' monadnocks,' owe their prominence to the 

 greater resistance of the rocks of which they 

 are composed,, mainly because of their hard- 

 ness, in comparison with the rocks about 

 them ; or have been spared on account of their 

 geograi)liical position, that is, they occur at 

 localities where streams originated and flowed 

 away in various directions, and in conse- 

 quence were left in relief after the country 

 about them had been conspicuously degraded. 

 To these explanations of the origin of monad- 

 nocks a third may now be added, namely: If 

 the rocks of a given area arc more open and 

 porous, or traversed by fissures or caverns to 

 a conspicuously greater degree than the rocks 

 beneath the surrounding region — the general 

 elevation being sufficient to favor subterranean 

 drainage — tlioy may be left in relief because 

 the water reaching them will be conducted 

 away by means of underground channels and 



thus in a great measure and in general almost 

 entirely deprived of its power to mechanically 

 erode, while adjacent areas are not favored in 

 this manner. 



A consideration of all the known facts re- 

 lating to the rocky heights forming Mackinac 

 Island and Gibraltar indicates that at each 

 of these localities a residual of the nature of 

 a monadnock has been left as the region about 

 it was lowered by erosion; the controlling 

 condition being that the rocks left in relief 

 are fissured and cavernous, thus facilitating 

 subterranean drainage, while the country 

 about them was denuded at a more rapid rate 

 through the agency of surface streams. 



Israel C. Russell. 



University of Michigax. 



a notable advance in the theory of 

 correlation. 

 To Professor Karl Pearson the new science 

 of biometry is indebted not only for its name, 

 but also for those refinements and extensions 

 of the methods of statistical analysis without 

 which it would be far from occupying the 

 position which it holds to-day. In the re- 

 markable series of memoirs which have ap- 

 peared under the general title ' Mathematical 

 Contributions to the Theory of Evolution,' 

 Pearson and his assistants have laid a founda- 

 tion on which a superstructure of great im- 

 port to biology can, and will be, reared. The 

 most recent of the memoirs in this series* 

 brings forth a very interesting extension of 

 the theory of correlation which at once greatly 

 widens the range of problems and material 

 which can be effectively handled by hiometric 

 methods. 



In the development of the method of de- 

 termining the degree of correlation between 

 characters not admitting of quantitative meas- 

 urement,f it was thought necessary in form- 

 ing the correlation table to arrange the classes 



* ' ^latlipmatical Contributions to the Theory of 

 Evolution,' Xlll. ' On the Theory of Contingency 

 and its Rehition to Association and Xornial Cor- 

 relation.' Drapers' Company Research Memoirs, 

 Biometrie Series, I., pp. 1-3.5, 2 pi., 1004. 



■f Phil. Trans., Vol. 19.5 A, pp. 1-47, and pp. 

 79-l;-0. 



