1916] RURAL ENGIJiTEERING. 581 



In the sand bed tests it was found that there was arching of the sand under 

 pressure. " It was found that the lateral thrust rose to a maxunum under a 

 shallow depth and remained practically constant. With a slightly yielding bot- 

 tom the sand arched to relieve the bottom of load which was carried by fric- 

 tional resistance of sand against the box sides. In the main tests, it was found 

 that when the platen was wider than the pipe the pressure was largely supported 

 by the column of sand over the pipe and less by the more yielding sand beyond 

 the pipe walls. Then the load imposed on the pipe was greater than as if con- 

 sidered uniformly distributed by the platen. The results exhibited the con- 

 veyance and distribution of sand pressures, but were not considered sufficient 

 to warrant developing a formula for calculating culvert pressures. For in- 

 stance, it was seen that the top pressure on pipe decreases with a given load 

 placed on increasing depths of sand, but the proportional relations were not 

 disclosed. The lateral pressures were low and no definite ratio was estab- 

 lished. The lateral pressure increased rnpidly from the top down to a point de- 

 termined by the angle of repose of the sand and then decreased. . . . 



" It is concluded that under the heaviest load that can now be applied to railway 

 ties by any existing locomotive or car, a 24-in. 14-gage pipe with ^-in. depth 

 of corrugation, and under 24-in. cover of dry sand, can not be deflected beyond 

 its elastic properties of complete shape recovery. The 36- and 48-in. culverts 

 are held to be uucrushable under a cover equal to their own diameter. The weight 

 of cover in a wide loose fill or bank is not so well supported as in a ditch, 

 and it was concluded that conditions of unusual severity might be induced which 

 should be provided for by heavier metal." 



Experiments on the distribution of vertical pressure in earth, R. B. Fehk 

 {Ann. Rpt. Penn. State Col. 1914, pp. 111-12S, pis. 13).— The results of tests 

 with dry sand, silty yellow clay, gravelly silt loam, and brown dry river sand, 

 and on a mixture of these, to determine the distribution of pressure due to a 

 concentrated load through various depths of the soils are graphically reported, 

 together with a description of methods and apparatus used. 



" The following depths of sand were tested : 3, 6, 12, 18, 24, 36, 48, and 59 in. 

 For each depth the eccentricity of the load was varied from to 42 in. both 

 right and left. The loads were applied in increments of 1,000 lbs. as indicated 

 by the gage up to the point where the loading strip sank into the sand as fast 

 as the load was applied. As determined from the calibration curve these loads 

 varied from about 600 lbs. to between 2,200 and 6,000 lbs., according to the 

 depth of the sand." 



It was found that " there was a very marked change in the percentage of 

 transmission when the eccentricity of the load was equal to the width of the 

 weighing strip and at this point the maximum value was 13 per cent. For 

 greater eccentricities this value was never exceeded. . . . The size of the 

 restricting box does not make any appreciable difference in the results. The 

 maximum bearing power of the sand decreased with the depth of sand. . . . 

 There was a distinct tendency for the percentage of transmission to increase as 

 greater loads were applied." 



With the clay loam sand mixture " the method of procedure was exactly the 

 same as in the sand tests except that the eccentricity of the load was varied 

 from 24 in. right to 24 in. left, and back across the soil to the starting point. . . . 

 The depths tested were as follows : 6, 9, 12, 18, 24, and 36 in. . . . 



" In practically all cases there was quite a regular increase in the percentages 

 of transmission as the load varied from 600 to 10,000 lbs. per square foot, the 

 maximum loads causing an average increase of 36 per cent in the transmission 

 as produced by the minimum loads. . . . The tests in loam were run in exactly 



