CHANNELED SCABLANDS OF THE COLUMBIA PLATEAU 643 



decayed/ Cementation with calcium carbonate has begun but has 

 not advanced far. 



Most isolated loessial hills in a scabland tract have a deposit 

 of gravel depending from their down-gradient end. Many knobs 

 and buttes of basalt have similarly situated gravel deposits. 



In many cases, the gravel deposits constitute discontinuous 

 terraces on the margins of the scabland tracts, suggesting remnants 

 of former valley iills.^ But the evidence seems conclusive that all 

 gravel deposits of the scablands are bars, built in favorable situations 

 in the great streams which eroded the channels. 



The rounded profiles and ground plans of many gravel deposits 

 in the scablands are in accord with this interpretation. The 

 unfilled canyons and rock basins, in intimate association with the 

 discontinuous gravel deposits, indicate clearly that both are 

 products of the same episode. The only alternative hypothesis 

 is that channeled scabland was formed, then buried in gravel, 

 then in large part re-excavated by streams little short of the magni- 

 tude of those which eroded the scablands. This has no other field 

 evidence to support it and requires a much more complicated his- 

 tory. Furthermore, such deep canyons were cut when the scab- 

 lands were made, and such noteworthy divide crossings were 

 made that a reoccupation of all the scablands by glacial drainage 

 from a second ice sheet would be impossible. And the hypothesis 

 of dissection by the postglacial streams of the scablands is quite 

 inadequate. Lakes and pools still stand in the rock basins on the 

 channel floors, almost as they were left by the glacial flood. 



Gravel deposits in the deeply canyoned scablands occur on the 

 broad upper scabland surfaces, on the roughened slopes of the pre- 

 glacial valleys and down in the canyons. The interpretation of 

 these deposits as bars requires no change in general conditions, 

 as does the alternative hypothesis; it simply requires that gravel 

 be deposited locally as conditions might favor, all through the 



^ Local exceptions, as i mile southwest of Lamont, where ground water has been 

 especially active. 



2 So the terraces in Pine Creek channel were interpreted in the earlier paper by the 

 writer. That view is here abandoned for one much more consistent with all other fea- 

 tures of the scablands. 



