Tests 121 



groove end, but is likely to vary a quarter inch or more at the tongue 

 end. The reason for this eccentricity is that the pipe is made v^rjth 

 the groove end on the ground, and that the forms become displaced 

 somewhat by reason of unequal tamping as the forms are gradually 

 filled. The machine-made pipe is accurate as to length and uniform 

 in weight. 



Usually the breaks occurred suddenly and a single crack opened 

 about one-sixteenth inch or less, extending vertically from bottom 

 to top. In one case the crack opened slowly, beginning at the bot- 

 tom, and in several cases the crack appeared to start definitely at 

 the bottom. A few of the cracks extended upward spirally. Inas- 

 much as longitudinal breaks in pipe lines produce cracks in both 

 top and bottom, it might be expected that there would be two cracks 

 in the broken test pieces. There was no indication, however, of a 

 second crack in the pipe wall with two exceptions. The cracks in 

 the machine-made pipe opened much wider than those in the hand- 

 made pipe. 



In the author's tests there were six or eight cases in which the 

 pipe broke at very low pressure. Three of the machine-made pipe 

 broke at less than ten pounds per square inch. Several possible 

 explanations of these tests should be considered. They are : 



1. Injury to the pipe in hauling or unloading. 



2. Water hammer or pulsations due to the proximity of the 

 testing machine to the city pumps. 



3. Eccentric loading or combination of central load at top with 

 two side supports at bottom. The location of the breaks might in- 

 dicate this. 



4. Reduction of strength due to saturation. 



5. Expansion of the interior part of the pipe wall, perhaps one- 

 third or one-fourth of the thickness, due to partial penetration of 

 the water. The pipe were dry when tested. 



It is believed that the last named was the main cause of the early 

 failures. Expansion of the inner shell would produce tensile stresses 

 in the outer portion which would increase with the penetration of 

 water. This tension, added to that due to internal pressure, might 

 crack the outside portion, and the remaining wetted portion would 

 be too thin to resist even the low internal pressure. 



More recently the author, while visiting another pipe yard, 

 watched the testing of two specimens. Both withstood 80 pounds 

 pressure. He then asked to make a test in the same manner as 

 practiced at Tucson, and picked a pipe which had been lying in the 

 sun for two months. It broke at 15 pounds. The first tests were 



