Plan view and profile sketches appear in figure 24, along with pertinent seismic 

 profile and test pit information. 



Following are the seismic interpretations for site 917-3: 



Traverse 1: 21.5 feet of loose material (1,150 f.p.s.) overlying 

 at least 40 feet of fractured rock (3,000 f.p.s.). 



Traverse 2: 4.8 feet of very loose material (830 f.p.s.) overlying 

 at least 25 feet of fractured rock (2,500 f.p.s.). 



Traverse 3: 3.1 feet of very loose material (770 f.p.s.) overlying 

 at least 20 feet of loose material (1,100 f.p.s.). 



A berm, although apparently not of soil cement, is located along the outside edge 

 of the road surface, as at site 917-1. However, the road has negligible grade through 

 this section, and the berm apparently acts as a small dam. This causes water to collect 

 on the road. In fact, there may even be a slight dip in the longitudinal axis of the 

 road, as evidenced by the fact that water had been flowing along the cracks and into 

 the failed section from both directions. 



Apparently the incipient failure surfaces, as is the scar left by the narrow 

 failure, are at relatively shallow depths below the fill slope. Ignoring for the moment 

 any water infiltrating from the road surface, the ground water conditions represented 

 by figure 11 would seem to have minimal influence in the outer portion of the fill 

 where failure occurred and is impending. To evaluate the accuracy of the laboratory 

 estimates for (j> and c, a series of analyses were conducted assuming circular failure 

 arcs, with an assumed saturated unit weight of 125 p.c.f., but without excess boundary 

 neutral stresses along the failure arc. (From table 2, test pits 917-3-1 and 917-3-3, 

 an average dry unit weight of 100 p.c.f. and e = 0.66 provides for a saturated unit 

 weight of 125 p.c.f.). The analyses were made under the condition that all failure arcs 

 passed through the toe of the fill slope, where the fill slope was 70 feet in length 

 and at a slope of 38.6° (profile A-A) . 



The results of these analyses are shown in figure 25, wherein R and a are again 

 defined as in figure 17. 



Under the assumption that $ = 35° and c = 0, several trial failure arcs were 

 found with a safety factor less than 1.0. As discussed in connection with site 917-1, 

 it is possible that this material retains some cohesive strength, even under the condi- 

 tion of saturation with seepage. Noting that the lowest safety factors were obtained 

 when a = 38.6°, a series of analyses were conducted at this steepest chord angle when 

 <J) = 35° and c = 1 p.s.i. As seen in figure 25, the safety factors were significantly 

 increased. However, these increases were more pronounced at a greater R value. Since 

 the appearance of the failure indicates a failure arc of low curvature, i.e., of long 

 radius, the safety factor vs. radius curve for <f> = 35° and c = 1 p.s.i., in figure 25, 

 does not seem to comply with the actual failure surface nearly as well as do the curves 

 for <f> = 35° and c = 0. 



Finally, a series of analyses were conducted with $ increased to 38° and c = 0. 

 The results of these analyses are also shown in figure 25, and seem to be most compati- 

 ble with the observed conditions at the site. 



The low depth-to- length ratio of the failure at this site suggests that the infi- 

 nite slope stability theory might be applied. Assuming a depth no greater than one- 

 twentieth the length, or about 3.5 feet, developed cohesion values of only 0.18 and 

 0.03 p.s.i. are necessary for stability in the absence of seepage forces, for <J> = 35° 



25 



