42 REPORT OF THE CALIFORNIA EARTHQUAKE COMMISSION, 
square inch. The compressive force due to the vibrations calculated in the example 
we have used (with Young’s modulus for wrought-iron equaling 28,000,000 pounds per 
square inch) is only about one-eighth as great, but at this short distance from the fault- 
plane it is possible that the vibrations may have been greater, and without doubt, the 
pipe itself, on account of the joints, would give way under a much smaller pressure than 
is required by the above formula; we must believe that the pipe yielded like a column 
under compression, and the sudden removal of the resistance when the rupture came 
allowed the elastic forces to throw the pieces 50 or 60 feet to the side. 
The Crystal Spring 44-inch pipe suffered in the same way where it crost the San Bruno 
marsh near South San Francisco and the Guadeloupe and Visitacion marshes a little 
further north. The trestle which carried the pipe over these marshes was built on deeply 
driven piles. The pipe was broken in many places and the pieces flung 4 or 5 feet to 
right and left; the trestle was also demolisht, but the piling and its capping were in 
general uninjured. In some instances, however, the pipe seems to have been raised into 
the air and to have come down with sufficient force to destroy the trestle and crush the 
heavy timbers bolted to the tops of the piles. Altho the vibrations in these marshes 
must have been very violent, it was found after the earthquake that no permanent dis- 
placement had taken place; the piling had not lost its alinement nor its grade. 
It does not seem probable that the lateral vibration was strong enough to break the 
pipe and throw the pieces 4 or 5 feet; the pipe must have been quite flexible enough to 
yield to such vibrations without breaking; nor is it probable that the vertical vibration 
was strong enough to throw the pipe upwards; it is most probable that we have here 
again to do with compressional vibrations, acting upon parts of the pipe as upon columns 
with round ends, for the ends of the short lengths of the pipe, over which the compression 
was strongest, were practically free to turn the small amount required. We suppose the 
vibrations to be communicated to the pipe thru the trestle and to be transmitted along the 
pipe as forced vibrations, with the same period and velocity, and therefore with the same 
wave-length, as in the underlying marsh; but there would undoubtedly be propagated in 
the pipe vibrations having a velocity appropriate to the material of the pipe, and these 
would in places combine with the forced vibrations, to produce unusually large forces 
of compression and tension. 
Rankine’s formula for the yielding of columns with round ends becomes 
16 
4 aL\4 
1+ Z500(7) 
With a 44-inch pipe L/d would be 40 for a length of about 150 feet, and p, the force neces- 
sary for collapse, then becomes 6.6 tons per square inch; which is about 4 times the pres- 
sure calculated in the example we have taken above; but in the marshes the wave-length 
would be greatly reduced, and there seems no difficulty in believing that the compressions 
were in places sufficient to break the pipe regarded as a column with rounded ends 
(especially at the joints), and then to fling the pieces to the side. Where the pipe had a 
slight bend in the vertical plane, the compression would throw it up rather than to the side, 
and in this way its subsequent blow upon the support is made clear. One piece of pipe, 
about 800 feet long, was found lying on the ground by the broken trestle uninjured 
except at its ends; it must have rolled off the trestle after the supporting sides had been 
battered off. 
The San Bruno marsh is about 2 miles from the fault-line and the other two marshes 
about twice as far. The increast intensity of vibration due to the character of the 
foundation far more than made up for the diminisht intensity due to distance, as shown 
by the distribution of isoseismals on map No. 23. 
