420 N. L. BOWEN 



If we return to a consideration of the fundamental action 

 involved, namely the formation of a bridge as a result of the refusal 

 of a certain layer to bend in conformance with the bending of the 

 layer immediately beneath it, we find that this could take place 

 only when the crystal mesh is of such a nature as to have fair 

 compressive strength but very little tensile strength. If the 

 tensile strength were even moderate the upper layer would be pulled 

 into conformable shape by the lower layer, for the tension acts over 

 a relatively large area. The crystals of the mesh must therefore be 

 in contact but must be grown together only to a very limited degree. 

 There is a further reason why the proportion of crystals to liquid 

 must be limited. The bridging action can take place only when 

 the crystal mesh is so open that pressure on the liquid is transmitted 

 to all parts of it with great freedom and with little frictional resist- 

 ance, for the bridge has no real strength to support a significant 

 superincumbent load, and liquid must be free to flow into the space 

 immediately below it if the potential bridge is to become a reality. 

 On the basis of some rough experiments it is estimated that the 

 action could take place when the mass was 50 per cent crystalhne 

 but probably not when the proportion of crystals was significantly 

 less than that value. On the other hand, it probably could not 

 take place when the proportion of crystals was as great as 65 per 

 cent. The contrast between bands is therefore of a different order 

 of magnitude from that shown in the gabbro-granophyre associa- 

 tion. There is likewise a rather limited period for any particular 

 portion of a mass during which this action can occur. It is believed 

 that during this period and as a result of the action described the 

 banding of the Duluth gabbro was produced. In this case it is 

 possible to show the relation between the direction of banding and 

 the forces that produced it. For many igneous masses having 

 primary banding this relation would perhaps not be so plain, yet 

 it would seem to be safe to state it as a general proposition that 

 primary banding is produced by the action of warping stresses^ 

 torsional stresses, and in some cases shearing stresses, acting at the 

 appropriate stage of crystallization. 



Origin of the red rock. — Returning now to the Duluth lopolith, 

 we find it at a stage when crystallization is fairly well advanced. 



