78 



are the terraces coutinuoiis for the whole distance indicatetl on the pro- 

 files, but consist of remnants, part on one side, part on the other side of 

 the stream. 



At the ui)per level, there is first a eontiuuous terrace on the west side 

 of the stream for about half a mile; next, in the down-stream direction, 

 for about ] mile on tlie cast side; below this for some distance not more 

 than a trace on either side, succeeded by nearly continuous terraces on 

 both sides for more than i mile; aj^ain a distance of about i mile with 

 either no trace remainini;: or merely sliouhli-r-like projections here and 

 there; and finally at the south end of the canyon-like part of the valley, 

 a terrace i mile in lenirth on the east side, with a remnant only a few 

 rods in length on the west side. 



At the lower level the terraces are likewise discontinuous in places 

 and fonnd only at times on both sides of the stream at the same ix)lnt. 



The width of these terraces is, at the maximum, about (iO feet; aver- 

 age, perhaps 20 to 30 feet ; see Figs. 4 and a. At both levels the terraces 

 are chiefly rock cut, only a small amount of soil, talus, glacial debris, and 

 rock waste being found upon their surfaces, and the rock in place rising 

 in each case on the sidr away from the stream stmietimes as a steep, wall- 

 like slope until, at an ehnatiou above the upper terrace nearly as great 

 as the general level of the country, a considerable amount of glacial drift 

 in found. 



Perhaps the most interesting and significant feature about these ter- 

 races is their gradient as compared with the gradient of tlie stream, and 

 with sea-level as a datum. The terraces of the npiier levi'l vary in height 

 above the stream from ("> to s feet at the jtoint fai'thost upstream where 

 they are clearly marked, lo (M feet abnve the slreiiiii at the farthest down- 

 stream ])oint. a difl'ereiil i;il eh'vation of about Th fi'et, rising higher and 

 higher the farther downstream they are found. This suggested at first 

 that there might be an error in correlating the separati' renuiant.s as parts 

 of the same terrace; but it was ob.served thai in eaeli case where the ter- 

 race was continuous for about A mile, there was this consistent rise in the 

 downstream direction; in one case a rise of about 2") feet in a half-mile's 

 distance; in another c;ise a rise of 11 feet in ;i Utile less tiian i mile. The 

 shoulder-like i»oints and projections and smaller terrace renmants where 

 the terraces are disconi innous, have elevations agreeing closely with the 

 general rise of the gradient in llie dowiist ream direction. 



