SERPENTINE. 277 



Still less sharp is the distinction Ijctvveen tlie serpentine and the other altered 

 rocks, though, so far as practicable, the areas covered by each are indicated 

 on the map by different colors. There is nevertheless abundant evidence 

 that serpentinization was a distinct process at Knoxville, from the formation 

 of granular pseudodiabast*, pseudodiorite, and glaucophane schist, all of 

 which were formed under similar conditions and at the same time. The ser- 

 pentine was formed in part from these granular rocks, but in part also from 

 sandstone, and the microscopical evidence of this fact has been full}* stated 

 in Chapter III. This ^^■ould be inferred from the occurrence in the field, but 

 the difficulty of distinguishing- partially altered sandstones from sandstones 

 which have been convei'ted into fine-grained, holocrystalline rocks is such 

 that it would be impossible to make absolutely sure that a preliminary 

 recrystallization did not invariably precede serpentinization. In a great 

 number of occurrences at Knoxville serpentinization has evidently begun 

 along cracks in rock retaining macroscopically the appearance of somewhat 

 altered sandstone. When some progress has been made in the conversion, 

 the structure may be illustrated by the following diagram jDrepared from 

 sketches. Here the serpentine is repi'esented by dark bands of nearly even 

 width, but the corners of the intervening blocks are rounded, and it is evi- 

 dent that, were the serpentine removed, the remaining masses would no 

 longer fit together as they originally did (Fig. 8). When the process has 

 been carried further rounded balls are formed and 

 in some cases these have weathered out and strew 

 the ground like water- worn pebbles. The i)rocess is 

 mechanically strictly analogous to the formation of 

 balls of basalt in the Sulphur Bank, of which mention 

 was made in the last chapter, and a theory of the 

 {process has been given on page G8. 



The fibers of serpentine usually follow the ,.lV.r,'^^::ZT'7Z 

 direction of the veins separating the more or less "'«=i^' ■"«>« "">««■ 

 rounded masses which they include, but in a few cases stand vertically to 

 the surfaces of the unserpentinized nuclei, and the lines of division then 

 clearly mark the exact position of the original crack, as shown in Fig. 9. 

 One very fine case was found in which a subangular mass when broken 



