DESCRIPTION 



305 



minutes 3 search. Since my attention was specially directed to the ques- 

 tion of their origin I have examined several hundred. A locality of ex- 

 ceptional abundance is represented in plate 39. 



In length they measure from a few inches to more than 6 feet, measure- 

 ment being made in a straight line from horn to horn of the crescent. 

 Within the range of my observation chatter-marks are comparatively 

 small, the largest observed being less than a foot in length. Solitary 

 gouges are often seen, but in the majority of cases they occur in sets of 

 from two to six or seven. Ordinarily the members of the same set have 

 about the same size, but in a few cases a progressive increase was ob- 

 served, the individual most advanced in the direction of ice motion being 

 largest. It seems legitimate to infer from this arrangement in sets that 

 the cause of the gouge, whatever it may be, moved forward with the ice. 

 As already mentioned, the convexity of the crescent is turned forward; 

 but to this rule there are occasional exceptions. Two or three indi- 



Figure 1.- -Profile of Part of Glacier Bed. Figure 2. 



The ar.' ow shows the direction of ice 

 movement. Crescentic gouges occur on as- 

 cending s opes, from a to 6. 



-Diagrammatic Sketch of Cres- 

 centic Gouge. 



The arrow shows the direction of ice move- 

 ment. Compare figure 3. 



vidua!', were seen with the concavity turned forward, and a few also with 

 the longer axis in the direction of ice motion (see plate 38, figure 1). 

 The gouges were seen only on the upstream sides of projecting bosses 

 (figure 1). They are not restricted to the bottoms of glacial troughs, 

 but occur also on the walls, and in that case are on the upstream faces of 

 salients. 



Conoid Fracture 



The cross-profile of the crescentic gouge (figure 3) exhibits an angular 

 notch bounded by two unequal slopes. The slope from the upstream edge 

 is gentle, that from the downstream edge approximately vertical. This 

 character is exhibited in all parts of the crescent (figure 2). The gentler 

 slope radiates from an axis somewhere within the curve of the crescent and 

 is essentially a portion of a conoid surface. It is one wall of a fracture, 

 or crack, which does not end at the bottom of the gouge, but continues 

 on into the rock to an undetermined distance. The fact that the vertical 

 fracture terminates against the oblique fracture shows that the oblique 



