THE SMITHSONIAN INSTITUTION. 273 



great measure, we must try to fill up the lacuncb oi' our knowledge by 

 generalizations and ideal connexions. It is natural that, in so doing, 

 we should be exposed to error ; but we shall keep our mistakes within 

 the narrowest possible hmits, if we proceed by the M'ay of simple induc- 

 tions, and refuse to submit to premature theories. No doubt the pro- 

 pensity of the human mind to bring isolated facts into an ideal con- 

 nexion originates in our highest intellectual facult}^ b}^ which alone we 

 are able to discover the general laws which govern the endless variety 

 of cases. But there is scarce^ one science which has not been led 

 astray from time to time by this same propensity, and no science, per- 

 haps, more so than geology, of which orography, or the knowledge of 

 the external form of the dry surface of our globe, may, in some respects, 

 be said to be a chapter, while physical geography in general is its 

 descriptive department. 



Among the many mistaken notions still prevalent on that subject, 

 is the opinion that the principal S3"stems of water-courses or the great 

 river basins and continental depressions must be divided by mountain 

 chains. In America this is not more true than in any other part of 

 the world. But great and important as is the number of well known 

 facts which prove that the less striking differences of level followed 

 by the w^ater-courses of a country may be independent of the system 

 of real mountain chains, both being very often the results of two 

 entirely different series of causes, still these facts are regarded as mere 

 exceptions to a general rule, and, wherever positive observations are 

 wanting, geographers continue to fill up the blanks in our maps accord- 

 ing to that supposition. Thus, to separate the Pacific from the Atlantic 

 slope, and especially from that towards the ]\|exican Gulf, the Rocky 

 Mountains have been brought into an imaginary connexion with the 

 Sierra Madre of Mexico, and this latter chain has been forced on our 

 maps to take a direction which it does not take in reahty. I have 

 often heard the name of the former unhesitatingly extended to the 

 latter by Americans living in northern Mexico, though there is an 

 interval of several hundred miles in longitude and. latitude between 

 their nearest points. A generalization even of a bolder character is 

 sometimes made, when the Sierra Madre and the Rocky Mountains 

 together are said to be the continuation of the "Cordilleras" of South 

 America. But the system oi the Andes does not continue through the 

 Isthmus of Darien ; and the hills of the Isthmus of Panama have little 

 to do with them. These hills, again, are not connected with the moun- 

 tains and table lands of upper Mosquitia, of Honduras, and Guatemala, 

 ' nor with the volcanic cones which rise in isolated beauty from the 

 plains of Nicaragua and San Salvador. 



It may be observed that these interruptions of continuity are not 

 important enough to affect a general view of the subject, and it may 

 be conceded that this is true. Certainly we may speak with all pro- 

 priety of the mountains and table lands of the western side of the new 

 world as of one great system following the course of its western coast 

 from Terra del Fuego to the northern Polar ocean, and separated by 

 wide tracts of flat and, comparatively speaking, low country, from 

 the groups and chains which occupy certain sections of the eastern side 

 of both the northern and southern continents. But this is only repeating 

 Mis. Doc. 24 18 



