G. K. Gilb„i—nn^n o/Jninhd Strwture. 51 



cooling of lava beds. The system of cracks formed on the 

 surface of a congealing lava stream are carried downward as 

 solidification and contraction progress, and cause the rock to 

 be divided into a system of polygonal, prismatic blocks. 



Cracks of this type are included by some writers under the 

 head of joints, but it will be convenient here to call them dis- 

 tinctively shrinkage cracks and follow Professor Dana in giving 

 to the word joints a more restricted meaning. The joints 

 11 hi h ncdir so general 1\ in i ml u rati d rocks an <•! ra tcri i l 

 primarily by parallelism. By means of parallelism they are 

 grouped in systems, and most rock-bodies are traversed by- 

 two or more of these systems. Their trad igs at t ie >i,i 

 constitute a lattice of straight lines or of lines nearly straight. 

 The lines of two systems cross each other without interference. 

 From each point of intersection lines go in four directions, and 

 the alternate lines are prolongations of each other. Excep- 

 tional points can be found in which three lines meet, but the 

 meeting always makes a letter T and never a letter Y — that is 

 to say, two of the three meeting lines always agree in direc- 

 tion so as to constitute a continuous lire, amutist which the 

 third terminates. Usually the hammer will reveal an inchoate 

 joint in th< the third line. In each of these 



respects joints differ from shrinkage cracks. 



Joints of the same system are parallel ; shrinkage cracks arc 

 not parallel. Joints of different systems cross each other: 

 shrinkage cracks do not cross each other. lit jointed .structure 

 the joint is the leading feature, the block is incidental, and the 

 wide-spread evidence of system everywhere observed shows 

 that the causative i'orc< eit ier is dill used or is extraneous. In 

 shrinkage-crack structure the causative force is localized in the 

 shrinking block, and the crack is incidental. 



Yet other points of difference could be enumerated, but 

 enough have been adduced to show that the two series of phe- 

 nomena are not closely parallel. Perhaps the shrinkage 

 hypothesis for joints should not )><• -<-t aside as absolutely un- 

 tenable, but it certainly cannot be adopted until division planes 

 demonstrably due to shrinkage are in some instance shown to 

 have the peculiar characters of joints. 



If we turn, now, to the relation between jointed structure 

 and slaty cleavage, we find, to sav the least.' a close analogy. 

 They agree in the parallelism of division planes. Just as one 

 system of joints intersects another, so a system of joints is 

 found to intersect a system of cleavage planes; and it is prob- 

 able, though perhaps not certainly known, that two or more 

 systems of slaty cleavage may exist in the same rock mass. 

 The most striking difference "between the two structures may 

 possibly be one of degree merely: — in the case of joints the 



