328 Journal of Agricultural Research voI.x.no. 7 



The last column of Table X gives the calculated unfilled voids in each 

 type of test piece. They represent the percentage difference between 

 the actual specific gravity of the test pieces found by dividing their 

 weight by their volume, and their ideal specific gravity calculated from 

 the specific gravities and the percentage by weight of their constituents. 



When it was found that no physical tests could be correlated directly 

 with the behavior of the rocks in mixtures, it was suggested that the 

 surface character and cleavage of the particles might be responsible for 

 the differences. An examination with a hand glass seemed to show 

 some interesting differences in surface character, and in order to permit 

 a more careful investigation enlarged photographs of a few lo-mesh 

 particles of each rock were made. Reproductions are shown in the 

 accompanying plates. Plate 49 shows the first six rocks in Table VIII. 

 These are of low average toughness and all fail to show any toughness 

 with the light tar, except the biotite gneiss (9095) which gives a 

 toughness of i at seven days. These rocks are large grained and made 

 up of individual glassy particles, and therefore may be characterized 

 as possessing smooth surfaces. This is most evident in the case of the 

 sand, which had the lowest average toughness of any of the materials 

 shown in Plate 49. In Plate 50 are shown fragments of the rocks which 

 gave the relatively higher toughness values, except the sandstone (8136). 

 It will be noted that the surface condition of these particles is quite 

 different from those in Plate 49 on account of the fine-grained, relatively 

 rough surface. The photograph of the feldspathic quartzite (9321) fails 

 to show this character as clearly as desirable, but it can be stated that the 

 fragments have a decidedly dull or frosted appearance, probably due 

 to the feldspar content, as compared with the glassy quartzite (10108) 

 in Plate 49. A comparison of the photographs with the toughness 

 results would therefore seem to offer one explanation for the difference 

 in the behavior of the various rocks, and these differences appear, to some 

 extent at least, to be due to the fact that polished, glassy faces of the 

 rocks in Plate 49 fail to hold the bitumen as well as the rougher surfaces 

 of the rocks shown in Plate 50. An exception is noted in the case of the 

 sandstone (8136) shown in Plate 50. The particles of this rock present 

 a decidedly rough, dull surface, but the rock in combination with all 

 bituminous materials showed the lowest average toughness of any tested, 

 owing to the fact that the individual particles of the specimen were 

 bound together loosely. In fact, the rock could be crumbled in the 

 hand without very great difficulty. As noted above, a bituminous con- 

 crete section in which this rock was used failed rapidly and this was 

 evidently due to the failure of the individual rock fragments rather than 

 to the failure of the adhesion of the bitumen. 



