MUSEUM OF COMrARATIVE ZOOLOGY. 201 



is especially the case for the more important upper contact with the 

 sandstone on the back of the trap ; for during the ages in which these 

 rocks liave been eroded, they undoubtedly stood frequently or always at 

 a higher level above the sea than at present ; the soft beds were deeply 

 worn away, and are now buried under the clays and sands of post-glacial 

 weathering or of glacial deposit. Upper contacts are everywhere rare. 

 Lower contacts are not very uncommon in Massachusetts, Connecticut, 

 and New Jersey ; but nearly all contacts are hidden in Pennsylvania 

 and farther south. The discovery and description of the junction of the 

 trap with sandstone above and below it afford excellent field for local 

 obsei'vations in all the Triassic belts. 



The observations of the past summer, as detailed in the preceding 

 pages, give examples of nearly all the points of evidence required to 

 prove an intrusive or an overflow origin of the trap sheets. The locali- 

 ties where the best evidence of overflow sheets was found are : — Con- 

 formable, unbaked sandstone on the trap, or fragments of trap in the 

 overlying sandstone, at Turner's Falls ; on the back of Mount Tom and 

 its posterior ridge ; and at Feltville, on the back of the First Newark 

 Mountain. Tufix deposit with fragments or bombs of trap, under the 

 second posterior ridge. Turner's Falls. A second lava-flow resting on 

 the uneven amygdaloidal surface of an earlier one, at West Springfield. 

 Percival mentions a trap breccia or conglomerate at several points in 

 Connecticut. Dawson represents a large mass of these fragmental trap 

 deposits in his Nova Scotia section. 



Three distinct forms of occurrence must be admitted for these traps : 

 first, the feeders, or supply dikes ; second, the intruded sheets, generally 

 lying evenly between the enclosing layers ; third, the overflow sheets. 

 The second and third forms are generally closely alike in their present 

 topographic features, and can be distinguished only by detailed observa- 

 tion. Even in the best known districts, there is room for much work of 

 this kind. 



Trap Dikes. — Under this heading will be included only those trap 

 masses that have a greater extension across the sandstone layers than 

 parallel to them. They have been observed as follows : — 



Dawson and Harrington note a single occurrence of trap on Prince 

 Edward Island in a dike form ; it is vesicular and scoriaceous in part, 

 as well as dense and columnar (21). 



Bailey and Matthew mention a dike in the Triassic of Grand Manan 

 (221). 



Hitchcock says there are no well-characterized dikes in the Massa- 



