MUSEUM OF COMPARATIVE ZOOLOGY. 299 



Percival noted that the amygdaloids are generally found in the lateral 

 portions of the trap ranges, and occasionally make a large part of the 

 lateral ridges (315); Hitchcock wrote that amygdaloids occupy "the 

 easterly [i. e. posterior or upper] part of the ridges wherever I have 

 examined them " (e, 645) ; and Rogers, that amygdaloids are common 

 "near the borders of certain of the larger dikes" {g, 671). 



The occurrence and position of the amygdaloid has been variously 

 explained. Jackson and Alger (6, 265) thought this texture resulted 

 from the combination of the trap with sandstone and shale, liogers 

 stated that the amygdaloid occurred "in immediate contact with the 

 altered red shale, by the reaction of the trap upon which this amygda- 

 loidal character has been acquired" {g, 761, 763). Cook says tliat if 

 the cooling of the traps had been " rapid and not under much pressure, 

 they would be more or less cellular" (6, 215). Professor Dana con- 

 siders all the original trap to have been equally anhydrous at its deep 

 source, and to become vesicular by the expansion of steam formed 

 wherever water was met in the process of eruption [e, 107). E. S. 

 Dana and Hawes accept this cause. 



Several of these explanations are undoubtedly true and possible, but 

 they do not show why the overflows should always be amygdaloidal on 

 the back, and why the intrusions should so generally be compact. It 

 seems most probable that the vesicular texture was produced in these 

 old traps as it, is in modern lavas ; not so much by meeting water during 

 their eruption, as on account of a decrease of pressure which allowed the 

 occluded gases and vapors to separate from the surface of the overflowing 

 molten mass (Dawson, d, 63; e, 87). The difference in the composition 

 of the eastern and western traps found by E. S. Dana and Hawes in 

 lower Connecticut has already been referred to as resulting naturally 

 from the better chance the eastern traps have had for alteration. 



The area marked by some of the larger ranges is very considerable. 

 The Mount Tom — Hanging Hills range has a front sixty five miles long, 

 and, judging by the curves at either end, its breadth must be six miles, 

 giving an area of nearly four hundred square miles. If the several loops 

 down as far as Saltonstall Lake all belong to the same sheet, broken by 

 faults, as is suggested below, then the length and breadth would be 

 much increased, and the area might be over seven hundred square miles. 

 The Newark Mountains in New Jersey would in the same way have an 

 area of above three hundred square miles. 



But it is not by any means proven that these sheets are the product 

 of single eruptions. The heavy trap which forms Mount Tom (B) de- 



