LOCAL EFFECTS OF SAN FRANCISCO EARTHQUAKE 313 



that the water in a large wooden tank had splashed against a roof 

 placed over it with sufficient force to drive shingles from the north- 

 west side. In some places large water-tanks holding 3,000 or 4,000 

 gallons were almost emptied by the splashing. 



The wave-motion was responsible for most of the damage done 

 outside of the narrow belt along the line of fracture. Several people, 

 who were out of doors at the time of the earthquake and several 

 miles from the fault-line, state that the ground appeared to move 

 like the waves of the sea. While these statements cannot be used 

 as conclusive evidence, there are many facts that indicate a true 

 wave-motion, having distinct crests. Water in reservoirs and tanks, 

 standing at a distance of several miles from the fault-line, splashed 

 out on the northeast and southwest sides. At King's Mountain 

 House, a little over 2 miles southwest of the fracture, there were a 

 number of milk-pans setting on shelves. All of the cream went 

 out on the southwest side of the pans, and afterward the milk splashed 

 back and forth, spilling out on both the southwest and northeast 

 sides. At a barn 3 miles northeast cf Wocdside, heavy carriages 

 standing with their wheels parallel to the fault-line were moved side- 

 ways a distance of 6 inches, but did net roll forward on their wheels. 



At Stanford University, 4^ miles northeast cf the fault-line, the 

 sandstone buildings afford evidence of the wave-motion. Walls 

 running northwest and southeast, when free to fall, fell by toppling 

 over, and the stones lie on the ground in nearly the same relative 

 positions that they occupied while standing. Walls running more 

 nearly parallel to the direction of wave-motion were crushed, and 

 the stones fell in irregular piles, while the walls that are still standing 

 show 45° shearing cracks. Perhaps the best evidence of a true 

 wave-motion is to be found in the arches. When the crest of a 

 wave struck an arch running northeast and southwest, the arch 

 was pulled apart, allowing the keystones to drop a short distance. 

 There are forty-six arches running approximately northeast and 

 southwest in which the keystones dropped, while only twelve arches 

 running northwest and southeast had their keystones lowered, and 

 some of the latter may be accounted for by the faUing of neigh- 

 boring walls. It might be well to state that there were more arches 

 running northwest and southeast than at right angles to that direc- 



