METHODS OF CORRELATION 461 



arrange themselves in a narrow northwest-southeast belt — probably an 

 erosion valley — (some 50 miles long and 4 or 5 miles wide) that passes 

 between the localities of the two cuts in which the true Mendota type of 

 rock is absent. It was further agreed that the beds immediately under- 

 lying these Mendota outcrops vary decidedly in age from place to place; 

 hence that the contact is unconformable, and that differential movement, 

 emergence, and locally varying amounts of surface erosion had occurred 

 before the deposition of the Mendota began. Finally, it was decided that 

 the lower part of the sandstone in both of the railroad cuts represents 

 partially eroded remnants of typical Jordan sandstone on which Madison 

 sandstone was laid down without the intervention of the Mendota; and 

 we found between them a slightly uneven plane of contact, the distinct 

 nature of which, as compared with the ordinary bedding planes, is marked 

 by the unmistakable criteria of a reworked sand deposit. Most convinc- 

 ing of these criteria is the relative coarseness of the quartz grains that 

 make up the first few inches or more of the overlying sandstone — a con- 

 dition resulting from washing and sifting by wave action to which the 

 previously exposed and weathered surface of the Jordan was subjected at 

 the time. of the early Ozarkian marine transgression. In confirmation of 

 our conclusion, it remains to be said that the characteristic fauna found 

 elsewhere in this vicinity only in the Madison sandstone above the Men- 

 dota dolomite is confined in these cuts to beds above the break. 



CONTINENTAL SEAS USUALLY SMALL AND FREQUENTLY WITHDRAWN 



This brings us to one of the old conceptions that has done more to 

 retard progress in correlation than any other. I refer, namely, to the 

 long prevailing view of continuity of submergence of epicontinental basins 

 and the consequent wide extent of contemporaneous marine deposition in 

 them. It is this conception that is chiefly responsible for the common 

 failure of geologists to recognize stratigraphic hiatuses due to emergence. 

 The physical and faunal criteria which indicate such breaks in the process 

 of marine sedimentation are, as a rule, readily detectable after the ob- 

 server has once and for all accepted the principle of small, frequently 

 shifting continental seas; and the true significance of the phenomena is 

 always recognized when they have been discovered and explained and 

 shown in the field by another. 



In accord with this old conception, simulating rocks and faunas that 

 seemed to occupy the same stratigraphic position were unequivocally cor- 

 related and commonly referred to under the name of the formation with 

 which they were identified. This was the practice whether the correlated 

 bed was much thicker or attained but a fraction of the thickness of the 



