104 CONTRIBUTIONS TO PAL^ONTOLOGY 



covery of a fair sample, it would indeed be rare to find absolute 100 per cent identity 

 between two or more sets of fossil assemblages. Moreover, sUght differences in 

 environmental conditions, which may follow each other in short spaces of time or in 

 short distances, may be expected to have produced shght or even marked differences 

 in various assembUiges of organisms. FinaUy, it must be recognized that similar 

 assemblages may be homotaxial rather than synchronous. For most problems of 

 correlation, however, this quahfication is not necessarj'^, since the span of time 

 needed for the migration of newly evolved organisms from one region to another is 

 negUgible as compared with the time required for the deposition of the average 

 geological formation. If barriers to migration are known to have existed, however, 

 homotaxis must be considered. 



2. The most useful index fossils are those which are abundant, easily identified, 

 and of wide geographic range and short stratigraphic range. Although generally 

 understood and usually reUable in appUcation, this principle needs some quaUfica- 

 tion. Each investigator, for example, is entitled to an opinion as to the exact mean- 

 ing of such terms as "abundant," "easily," "wide," and "short." Moreover, fos- 

 sils that seem reUable today may become unreUable tomorrow as a result of new 

 discoveries extending their stratigraphic range. Furthermore, a short stratigraphic 

 range of a species or genus cannot be estabUshed unless its position in time can be 

 determined at several stations by clear stratigraphic relations and by associations 

 with other organisms. It is also considered more satisfactory for correlation to use 

 indcx associations of as many species as possible, rather than merely a single or a 

 few index species. Fiiiany, in cases where index fossils present conflicting evidence, 

 careful consideration must be given to such possibiUties as mixed coUections, in- 

 correct idcntifications, and inadequate deUmitation of zones. 



3. The earliest appearance of new forms in abundance is a valuable criterion for 

 correlation. 



4. The latest appearance of old forms is sometimes an important guide, par- 

 ticularly when they are accompanied by other organisms of restricted range. The 

 earUest appearance of new forms is, I beUeve, a safer criterion than the latest ap- 

 pearance of old forms, since the lattcr may be surviving reUcts of long-ranging 

 organisms, wliereas the incoming of abuudant new forms to a region is generaUy a 

 result of invasion or major environmental change, eithcr of which would result 

 in widespread biotic changcs which would be useful in exact correlation within that 

 region. 



5. The dominance of certain genera or spccies is a reUable adjunct to other 

 palaeontological methods of correlation. Dominancc is here meant to include both 

 abundance of individuals and widespread occurrence at many locaUties. It is often 

 found that dominance of a form or group of forms may be of time significance even 

 though the form or forms occur in more Umited numbers in both older and younger 

 rocks. In cases where a single form or several forms are dominant, a "zone" may 

 be defined and named for the dominant form or forms. 



6. Similarities of stratigraj^liic succession, though often unreUable criteria, may 

 contribute materiaUy to conclusions reached by palajontological methods of correla- 



