Parallel and Intersecting Joints. — Crosby. :{7."> 
ished, since no pains were taken to obtain strips of uniform 
width and each must have possessed a weakest line, to find 
that the breaking was not localized, but quite uniformly dis- 
tributed, the glass being shivered from end t<> end, in some 
eases into hundreds of rectangular fragments. I am satis- 
lied, upon reflection, that the only rational explanation of 
this phenomenon is that the glass does actually break first at 
its weakest point and so suddenly, being a highly clastic sub- 
stance, that the resulting shock or jar, spreading swiftly 
toward either end of the sheet, precipitates the breaking 
through the entire length. Applying this explanation to the 
earth's crust, it is seen that earthquake vibrations arising in- 
dependently of the torsional strains with which they co-oper- 
ate are not strictly necessary to the satisfactory working of 
tlie composite theory here proposed. 
Plain indications are not wanting that the joint-structure 
of the rocks is developed, as a rule, very early in their his- 
tory. That is sufficiently obvious for the contraction joints, 
including the sun-cracks in clay and the columnar jointing 
of eruptive rocks, and for the parallel and intersecting joints 
of the Quaternary and Tertiary clays and marls. In the 
older stratified rocks, the dikes and veins, where occurring, 
usually conform in direction with the joints of this class. 
This is conspicuously the case in the Boston basin, the 
numerous dikes traversing the beds of conglomerate and 
slate showing a close correspondence in trend and hade with 
the joint-planes. The dikes must be later than the joints; 
and yet we have reason to believe that the dikes themselves 
are, geologically speaking, nearly as old as the strata which 
they intersect. As soon as the sediments are sufficiently 
hardened to make joint-cracks a possibility, they are broken 
or jointed by the first torsional or vibratory strains of ade- 
quate intensity to which they are exposed; and ever after- 
ward similar strains, unless of unusual intensity, are relieved 
by slipping along the existing fractures, which are thereby, 
in many cases, we may suppose, kept open and prevented 
from healing by slow cementation, but are converted into 
minor faults with the attendant slickenside phenomena. 
