HOLMES.] THE GEAND CANON. 39 



siderable irregularity, the columns being smaller and mucli warped and 

 indented. To the right of the spur which T descended there was a deep, 

 narrow recess or gulch (/, Plate XVII), and beyond that an irregular 

 spur, in the face of which I could see a mass of dark rock {d, Plate XIX) 

 projecting from beneath the ledge of rhyolite that appeared to be con- 

 glomerate. Determining to examine it more closely, I descended into 

 the deep gorge which intervened, and was not a little surprised to find 

 myself in an old volcanic pipe or crater, three sides of which were still 

 preserved, the other having broken down. Portions of the loose materi- 

 als that had" once filled the space are stiU clinging to the compact walls. 

 They are scoriaceous materials and brecciated masses that seem to 

 grade imperceptibly into the surrounding rhyolites. Judging by the 

 remnants, the contents of the i)ipe have been of a most heterogeneous 

 character. The light, often beautifully-colored scoriaceous breccias, are 

 filled with pockets, which contain the greatest variety of rhyolitic frag- 

 ments, pitchstones, sphaerulitic obsidians, and volcanic bombs. Some of 

 the latter are over a foot in diameter. At the back part of the pipe, 

 and rising 200 feet above the slide of debris which partially fills it, is 

 the wall of trachyte that forms the brink of the cahoii. Crossing to the 

 spur beyond, I found that the dark mass of rock which I had taken 

 for a conglomerate was a heavy bed of black, finely-jointed rock that 

 has the appearance of a basalt, although a specimen from the middle of 

 the mass is pronounced an augite andesite by Captain Dutton (No. 27, 

 catalogue). This bed of rock, which I shall continue to call basalt on 

 account of its perfect resemblance to other unidentified rocks that I 

 have found it convenient to class as basalts, is nearly 200' feet thick, and 

 forms a steep ledge, which extends for a considerable distance along the 

 middle part of the canon wall. At the base there is a columnar band 

 some twelve feet thick, which rests upon a great series of Tertiary 

 strata, sandstones, and conglomerates. These strata are horizontal, 

 and where in conjunction with the basalt are much metamorphosed. 

 They extend down to the river bed, some 400 feet below. In following 

 along the base of the sheet of basalt I came upon the section of an old 

 river bed (e, Plate XIX) which had been filletl up by the basalt. The 

 stream had flowed in a shallow bed eroded from the Tertiary conglomer- 

 ates. The well-rounded pebbles, extend over a width of 150 feet, and 

 are of a great variety of rocks, the granitic, however, prevailing. That 

 part of the basalt which lies in the old river channel has, on the under 

 surface, a layer of j)umice about a foot in thickness. 



Something of the past history of this valley can be learned from the 

 facts here observed, A stream — whether the Yellowstone or not, no one 

 can say with certainty — had cut down through the Tertiary strata to a 

 depth of 600 feet below the top of the present valley, and about 400 

 above the bed of the present river. The lava has flowed into it, filling 

 the vallej' to the depth of 200 feet or more. Whether this occurred 

 before or after the flow of the rhyolite I am unable to say from any 

 facts observed, although on general principles we would probably 

 be safe in concluding that it is more recent. The debris at the ends 

 and on the top of the mass of basal fc make ir diflicult to say whether it 

 passes beneath the rhyolite of the main wall of the canon or is simply 

 set against it. Tlie fact that it does not continue across the crater to 

 the right, as seen in the drawing, proves nothing, as the rhyolite, which 

 extends down the opposite spur to a point considerably below the base 

 of the basalt, may have been deposited either before or after it. I had 

 not time to go on with this interesting investigation or I should, with- 

 out a doubt, have been able to solve the whole problem of the relative 



