478 WILLIAM HERBERT HOBBS 



line which approximates to a diagonal of the map, along the 

 N. ±: 33° E. line which meets the lower margin of the map near 

 where the Housatonic meets it, etc. It will hardly escape notice, 

 also, that the most marked lines in each series are spaced with 

 considerable regularity, and if one space is wider than the 

 others, it is often so much wider as to suggest that the space- 

 interval is a multiple of the space which has been regarded 

 temporarily as the unit. 



The river system of Connecticut. — An examination of the 

 larger area approximately coextensive with that of the state of 

 Connecticut has been made by use of the "two-sheet map of 

 Connecticut," on a scale of one half inch to the mile, prepared 

 by the United States Geological Survey Those lines and char- 

 acters of this map, such as topography and culture, with which 

 we are not now concerned and which would obscure the rela- 

 tionships sought, were eliminated by preparing a careful tracing 

 of all the streams and their minor branches. Upon this map 

 tracing (about six feet long by four wide) approximately recti- 

 linear stretches of river channel, and especially the stretches of 

 neighboring streams which hold approximately to the same line, 

 were sought. If these directions were found to agree closely 

 with any of the fault directions observed in the Pomperaug 

 basin, dotted lines were drawn following those directions and 

 coinciding as closely as possible with the river courses. If such 

 an observed direction was found not to coincide closely with any 

 of the fault directions determined, a direction was sought which 

 would approximate most closely to it, and a similar dotted line 

 with this direction drawn along the course of the river. The 

 term "trough lines" used to designate these lines, need for the 

 present be given no further signification than lines so favored by 

 nature that the waters of the region have been induced to 

 adopt them for their channels over longer or shorter distances. 

 On a map of this scale the trough lines, if rectilinear, should be 

 slightly curved, but inasmuch as the present river courses, 

 because of the many accidents of their history, can only roughly 

 approximate to the direction initially given them, it would bean 



