A rock will fail along any plane in which the stress exceeds the strength. Planes of weakness being, by definition, loci of low strength, very little stress is required to cause failure along them. Unless such a plane happens to coincide with the plane of maximum or minimum normal stress, it will experience shearing stress and become a plane of slip. It may be opened or kept closed depending on whether the normal stress on it is compressional or tensional. If the over-all deforming force is tensional, all planes of weakness are likely to be opened except perhaps those parallel to the axis of tension, but the planes closest in attitude to the plane of maximum normal stress tend to open first and widest.

It does not necessarily follow, however, that a rock will fail along its weakest plane, for this plane may occupy such an attitude that the stress on it, whether tensile or shearing, is small or nil. Stress along other directions may reach an intensity sufficient to rupture the massive rock before the plane of weakness is overpowered. Thus the pattern of fractures becomes a compromise between the ideal fracture pattern and the pattern of planes of weakness.

Pre-existing fractures. Where a rock containing old fractures experiences new stress, the orientation of the new stress is likely to be different from that which caused the old fractures. The new stress, in taking advantage of the pre-existing fractures, resolves itself into components normal and parallel to them and may also create new fractures of its own. The problem of interpretation becomes one of separating the effects of the two periods of deformation.

Compromise between old and new fracture planes is probably the explanation of a very common type of vein pattern in which a vein follows, in turn, one after another of a series of parallel shears, crossing from each to the next by a series of connecting links.

Some of the veins of the Coeur d'Alene district may illustrate the re-opening of a fracture system by stresses with changed orientation. In the Burke-Mullan portion of the district there are two sets of frac¬tures: the productive veins, which strike N70W, and faults, which strike N25W. Although the faults offset the veins slightly, it is evident that they were in existence before the veins were mineralized, because in repeated instances an oreshoot is found to extend up to a fault and end there, the displaced part of the vein continuing but with-out ore. It has been suggested that the vein and the fault constitute a complementary shear system, but that by the time mineralization occurred, the forces had assumed such an orientation that they opened the vein near the intersection and kept the fault closed. Still-later (post¬mineral) movement caused just enough displacement on the faults to give them the appearance of younger structures.

Bedding Planes. As most stratified rocks are weaker along their bedding planes than in any other direction, these planes behave in much the same way as pre-existing fractures. The essential difference is that if not folded they all belong to a single set and that, if folded, they are curved surfaces, often of small radius. The result is that part of the surface may lie in such a position as to be favorable to opening while the rest does not. As with parallel pre-existing fractures, veins often assume a step-like shape, following one bedding plane for a distance, then "linking" over to another. Vein systems en echelon are common and saddle reef s occur occasionally. The relations 'of veins to f olds will be discussed in a later section (page 333).

Igneous Contacts. Contacts between igneous rocks and their hosts are likely to be surfaces of weakness, although occasionally they are "frozen" and strong. Relations between contacts and veins are dis¬cussed in the next chapter.

Contrast in Rock-Types. A fracture, and especially a shear-fracture, on passing from one type of rock into another, very commonly changes its direction by a few degrees. In most cases the change is evidently due to contrast in strength-characteristics of the two types of rock. Each type of material has its characteristic shearing angle (p. 295), so that even though the plane of maximum shearing stress may be the same throughout the rock-mass, the angle between that plane and the actual surface of shear-failure differs with the type of rock. Knopf, in noting this effect in the rocks of the Mother Lode (California), suggests the analogy to the refraction of a beam of light and estimates the "index of refraction" for different types of rock.

The refraction or deflection may be caused by a contrasting sedimentary formation or by a mass of igneous rock such as a dike or sill. The position of greatest opening will vary with the direction of de¬flection and with the sense of the movement on the fault.
Anisotropism. Some rocks, such as uniform slates and schists, are homogeneous in the sense that they lack contrasting layers, yet they are weaker in some directions than in others. In such materials, which may be described as anisotropic, a uniform stress produces a non-uniform strain. Although further study of the mode of failure of anisotropic material is needed, everyone knows that failure tends to take advantage of slaty cleavage and schistosity. The surprising thing is that some veins in schist, instead of following cleavage or alternating between di¬rections parallel to it and across it, actually cut the cleavage at small angles without taking advantage of it.