September 2, 1898.] 



SGIENGE. 



295 



IV. The unstable flank of a mountain 

 chain is most unstable in its steepest parts. 



V. The steeper sides of valleys are, like- 

 wise, the more unstable. 



VI. When two mountain chains cross, 

 making an angle of less than 90°, the area in- 

 side the exterior angle is the most unstable. 



a 

 IX 



(The region a will have more shocks than 

 the region b.) 



VII. When a mountain chain (a) has a 

 buttress (6) the flank opposite b is the more 

 unstable (c). 



VIII. Mountain masses are more unsta- 

 ble on their flanks than within the mass. 



IX. Abrupt changes of slope are espe- 

 cially favorable to instability. 



X. The highest parts of valleys are fre- 

 quently more stable than those at the aver- 

 age level, and the lowest parts are gener- 

 ally more stable than those of average level. 



XI. Narrow mountainous peninsulas are 

 unstable. 



XII. An isthmus in a sunken region is 

 unstable. 



XIII. Narrow straits are often the cen- 

 ters of earthquakes. 



XIV. Eegions of great earthquake fre- 

 quency usually do not coincide with regions 

 of many volcanoes ; or, earthquakes and 

 volcanic phenomena are, in general, inde- 

 pendent of each other. 



Several of these laws are well known ; 

 some of them would be announced by an 

 expert even before seeing the data ; but, on 

 the other hand, some of them are genuine 

 surprises. In their collected shape they 

 constitute an important contribution to the 

 subject. Law XIV. is not proved in the 

 pamphlets cited by title, but the first half 

 of it is well known to be true in very 

 many regions of the globe, and the last 

 half follows as a statistical consequence. 



It is a very ancient opinion that earth- 

 quakes are decidedly more frequent at some 

 seasons of the year than at others. Aris- 

 totle, for example, declared that the autumn 

 and spring were seasons of frequent shocks, 

 while summer and winter were seasons of 

 few shocks. Perrey, Mallet and others 

 have announced similar laws. From a dis- 

 cussion of 63,555 shocks in 309 regions of 

 the globe M. de Montessus shows that, 

 taking the whole earth together, shocks are 

 equally probable at any season. This gen- 

 eral law may not be true for certain special 

 localities, but it is true for the whole earth. 



In order to study earthquake statistics to 

 advantage and to compare one region with 

 another it is desirable to have some uniform 

 method of expressing earthquake frequency 

 numerically — of deducing the coefficient of 

 earthquake frequency, as a mathematician 

 might express it. M. Montessus forms such 

 numbers in the following way : The region 

 to be studied is divided into smaller areas. 

 Each of these areas is chosen so as to be 

 fairly homogeneous in physical charac- 

 teristics, geographically, geologically, etc. 

 The areas are now divided into as many 

 small squares as there are earthquakes per 

 year. The greater the frequency the greater 

 the number of squares, of course. The 

 side of one of these small squares is chosen 

 as the coefi&cient of frequency,* and the 

 greater the number of shocks per year the 

 smaller is this coefiicient. There is some- 

 thing arbitrary in this process ; but, at the 

 same time, it leads to results of importance 

 because, after all, it is only the relative 

 earthquake frequency that is sought, not 

 the absolute. Of two regions, which is 

 most shaken and in what ratio ? is the 

 question to be solved. 



In eastern Java, for example, S (the seis- 

 mic number) is equal to 56 kilometres. That 

 is, there is, on the average, one earthquake 

 per year in each square of 56 kilometres on 



*M. Montessus calls it the 'sismioit^.' 



