96 W. M. DAVtS RELATION OF GEOGRAPHY TO GEOtOGV 



scriptive phrases like the bench of horizontal sandstone overlying the 

 disordered crystalline rocks which are exposed in the steep-walled gorge 

 at the bottom of the Colorado Canyon, or the long ridges of steeply in- 

 clined sandstone strata, which turn in sharp zigzags around the gently 

 pitching anticlines and synclines of the Alleghenies in central Pennsyl- 

 vania. Such descriptions must be doubly geological if rock attitude as 

 well as rock composition are allowed, mentioned only under geology and 

 never under geography ; but they surely deserve a place in geography also, 

 for such descriptions are of great value in describing the visible features 

 of existing landscapes. 



If the mention of past processes of deposition, deformation, or erosion 

 withdraws a statement from geography and transfers it to geology, then 

 such phrases as wind-heaped sand dune, wave-built sand reef, wave-cut 

 cliff, river-cut gorge, glacial cirque, tilted fault block, and uplifted pene- 

 plain are all geological. Indeed many simple terms, like delta, atoll, vol- 

 cano, landslide, moraine, peneplain, in which it is tacitly implied that 

 some process has acted through some portion of past time to produce a 

 certain actually visible form, all become geological terms and must no 

 longer be regarded as geographical. All these terms will be lost to geog- 

 raphy if the explicit or implicit introduction of past process determines a 

 term to be of geological nature ; yet surely they belong in geography, and 

 geography can not let them go. Doubly geological would be such phrases 

 as a dissected volcanic cone; a maturely dissected landslide; an uplifted 

 and slightly dissected delta ; a tilted, dissected, and then glaciated pene- 

 plain; for these phrases suggest the action of past processes in double, 

 triple, or quadruple succession, and thus imply the division of past time 

 into chapters. Surely it is not to be questioned that such a treatment of 

 existing forms has a highly geological flavor; yet the phrases, nevertheless, 

 are to my reading good geographical phrases, because they so concisely 

 and definitely characterize existing land forms. 



If mention of process acting through time be associated with mention 

 of rock composition and attitude, then descriptions thus phrased would 

 be of a still more pronounced geological quality, provided that geological 

 quality is really determined by mere mention of process, composition, and 

 attitude ; witness the following examples : A dissected bench of horizontal 

 sandstones unconformably overlies the ancient peneplain of disordered 

 fundamental crystallines in which the inner gorge of the Colorado Canyon 

 has been eroded, or Snowden is a mass of deformed slates and lavas of 

 subequal resistance, as a whole worn down to subdued form by normal ero- 

 sive agencies, but possessing great cirques recently excavated in its valley 

 heads by local glaciers. Yet both of these descriptions belong very prop- 



