THE RIVER SYSTEM OF CONNECTICUT 483 



lines, if it be assumed with Daubree that the gaping fissure 

 planes have directed the streams in their courses. The pre- 

 sumptive evidence is, the writer believes, in favor of the former 

 development of the Newark rocks over a much larger territory 

 than that which they now occupy and probably over the entire 

 state of Connecticut. 



It is an observation of much interest that the minor twig-like 

 branches of the streams, which in the deeply eroded mass of 

 crystallines must have been adjusted after the capping of sand- 

 stones had been removed, show an equally strong tendency with 

 the master streams to follow the special directions indicated by 

 the system as a whole. ^ 



The study of the fault system of the Pomperaug Basin offers, 

 however, another rational and natural explanation of the network 

 of streams, provided the assumption is made that the drainage 

 is adjusted to that formed in the geographic cycle which suc- 

 ceeded the deformation of the area. The system of parallel 

 faults has divided the area into vertical triangular, rhombic, or 

 rhomboidal prisms, which stand at different relative altitudes. 

 These prisms are found to be grouped into composite blocks of 

 increasingly higher orders, the peculiar property of each of 

 which is that the average altitude of its component prisms 

 approximates (however roughly) to a fixed value — the com- 

 posite blocks have an average level surface, although alternate 

 prisms or alternate subordinate blocks for short distances pro- 

 ject above or stand below the general level. The initial surface 

 formed by these prisms would be marked by canal-like structure 

 trenches {Grabeti) which follow the directions of fault planes 

 and which have stronger directive power, as regards streams, at 

 the junction of the trenches and at the crossings with the similar 

 trenches of other series. 



Although the numerous generally curving fault planes dis- 

 covered by Davis in his extensive studies of the Newark of the 

 Connecticut valley have here been omitted from consideration, 

 for the reason that no close relationship to the orientation of 



'Twenty-first Ann. Rept. U. S. Geol. Surv., Pt. Ill, p. 145, Fig. 52. 



