208 Geographical Collections. 



considerable height. The cuhninating points of lateral chains may surpass in 

 height the elevation of the summits of the principal chain, while the mean height 

 of the crest of the latter is greater than that of the former, as the height of the 

 culminating points in one chain of mountains may exceed that of another, while 

 the mean height of the crest may be greater in the latter, and this is the ground 

 work of a difference between the Alps of Switzerland and the Pyrenean Moun- 

 tains. 



Superiority of height of crest, as well as pre-eminence of smnmit, may also 

 exist in parallel or in transverse chains and not in the principal ; and, as a gene- 

 ral fact hitherto observed, the culminating points of countries of mountains are 

 seldom in the centre of the chain but at the extremities ; whilst, when the highest 

 summits occur towards the centre, they almost invariably exist in small trans- 

 verse branches, and sometimes between two parallel chains. 



It must be observed here, that transverse chains may be parallel to one 

 another while at right angles to the line of the crest : they are then parallel trans- 

 verse chains, but not lateral ones, and the structure of the ridges most generally 

 differs from that of the lateral chains. The same may be observed of the latter 

 when opposed to the principal crest ; but when the last is wanting, and the crest 

 exists in a series of parallel ranges, the structure will be found similar, or at least 

 pretty nearly of the same age. 



The culminating points or the maximum of the line of the crests of the prin- 

 cipal chains of mountains in Europe, in America, and in Asia, are, according to 

 De Humboldt, as the numbers 10, 14, 18 ; that is to say, they foUow pretty 

 nearly a progression by differences whose relation is one half. But in the seven 

 chains of the Alps, the Andes, the Himmaleh, the Caucasus, the Alleghanies, 

 and Venezuela, the relation between the mean height of the crest and the culmi- 

 nating points is as 1 to l^^j, or as 1 to 2. 



Mr. Ramond had already remarked that the crest of the Pyrenees is only a 

 little lower than the mean height of the Alps, while that which characterizes the 

 last chain is the great relative elevation of its summits ; that is to say, the rela- 

 tion of these summits to the mean height of the line of the crest. From De Hum- 

 boldt's calculations, founded upon the mean height of the passes or ports and that 

 of the culminating points, the mean height of the line of crest is equal in the An- 

 des to the culminating points of the Pyrenees, and in the Himmaleh to the cul- 

 minating points of the Alps. From considerations founded on the same data, the 

 relation of the mean height of minimum of crest to the culminating point would 

 be, in the Pyrenees, as 1 : 184 ; in the Alps, as 1 : 2 ; in the Andes, as 1 : 1,8 ; 

 in the Venezuelas, as 1 : 1,8 ; in the Caucasus, as 1 : 2 ; in the Alleghanies, as 

 1 : 1,8 ; in the Himmaleh, as 1 : 1,8. 



Considerations founded on a physical conception of the height of chains, which 

 will improve as the facts connected with their physical laws will be more perfectly 

 developed, are of the highest interest to the oryctography of the earth. The gra- 

 phic sections used by De Humboldt and others, and which are founded on simple 

 barometric or geodesical operatious, have been of much utility in obtaining con- 

 clusions of this nature. M. Marquis Laplace was led to consider that the sur- 

 face of the earth when liquid would be pretty nearly in a state of equilibrium ; 

 and remarks that, from the harmony which experiments on the pendulum offer 

 with the results given by the mensuration of terrestial degrees, and with the 

 lunar inequalities, it would follow that the mean depths of the sea might be of 

 the same nature as the height of continents and islands. It must be evident 

 ' that th^ mean height has little connection with the culminating points of moun- 

 tain chains, while the mean height of the crests forms an indispensable acces- 

 sory to the evaluation ; and, as de Humboldt has remarked at greater lengtli, 

 the chains and mountains which attract the curiosity of the vulgar have much less 

 importance in such considerations than the vast plateaus, and undulating plains, 

 and alternating slopes, which influence, by their extent and their mass, the portion of 

 the mean surface ; that is to say, upon the height of a plain so placed that the 



