OfDliixis : A Mel hi, (I ot' Corn hdioii. — Ki-ijis. 2l>7 
any one set of external features, but ratlier in tlie direct caus- 
es which give rise to the phenomena. The Jinal outcome is 
derived from a comparison of all groujjs of pertinent data in 
the physical histor3% taken ensemhlc 
A little consideration of the practical bearings of correla- 
tive inquiry shows that in a comparison of geological forma- 
tions four as])ects of the case are presented. The [)robleni 
may be looked upon from a local, a provincial, a regional, or 
from a general standpoint. With the various methods of cor- 
relation which have been followed from time to time the uni- 
versal practice has been to attempt to base the broader gen- 
eralizations upon criteria that are, in realitj', applicable only 
to limited areas. Hence, in passing from the more local to 
the more general, dittlculties arise which become more and 
mare formidable in direct proportion to the extension of the 
local scheme. Most of the methods that are applied and that 
are found to answer locally, fail when extended over larger 
districts. The real problem, then, is to find some means of 
solving the difliculties, for the latter or more general. When 
broadly applied most correlation criteria prove inadequate. 
The reasons are evident. As the specific distinctions are ex- 
tended more and more widely they change and all are gradti- 
ally replaced by dilt'erent ones. It is manifest that in no 
case must the critical criteria deal with the intrinsic features 
as such, but with the causes giving rise to them. Moreover, 
the one great factor to be taken into account in every stand- 
ard of comparison which has to do with correlation of strata 
is a definite basis to which the various minor, or local and 
provincial, successions can be referred. The fundamental 
conception is believed to grow out of a consideration of the 
nature of sedimentation itself. Natural phenomena rarely 
result from the action of single, simple laws. Each originates 
in many complex and intricate processes, some of which may 
be primary in (;haracter, but the majority of which are entire- 
ly secondary. For this reason the particular etiects of the 
real causas are usually more or less completely obscured by 
conspicuous 3'^et accidentally associated features. This is 
true in every department of science, as is clearly shown by the 
history and especially by the classifications that have been 
proposed during the ditferent stages of development. 
