ON GEOLOGIC TIME-DIVISIONS 339 



(Questions I and 2.) Though time is universal, faunas and 

 histories are more or less local. A refined time scale cannot be 

 used to advantage in the correlation of formations widely sepa- 

 rated. Therefore, only the major orders of a time classification 

 should be treated as universal, and the minor should be recog- 

 nized as local. I suggest that the line of discrimination be 

 arbitrarily drawn between divisions of the second and third 

 ranks, periods and epochs. 



Pursuant to this suggestion, I propose the following auxiliary 

 criterion for periods (not replacing but supplementing other 

 criteria) : Periods should have such magnitude that their applica- 

 tion to the correlation of formations anywhere in the northern 

 hemisphere will yield areas of certainty which are large as com- 

 pared to the unavoidable zones of doubt. 



This criterion is used in the selection of the subjoined scheme 

 of periods, but is subordinated to other considerations in the 

 admission of Pleistocene and Algonkian. Jurassic and Triassic 

 are given separate place despite their broad zones of doubt 

 when applied to American terranes, because the breadth of 

 those zones is due to dearth of the most important data for 

 correlation, marine fossils. 



Periods. 



12. Pleistocene (or Quaternary). 6. Devonian. 



II. Tertiary. 5. Silurian. 



10. Cretaceous. 4. Ordovician. 



9. Jurassic (or Jura). 3. Cambrian. 



8. Triassic (or Trias). 2. Algonkian. 



7. Carboniferous. I. Archaean. 



(3.) Four time-nouns have been used in this rank by various 

 authors : Era, age, eon, and time. Time cannot be spared from 

 its general sense. Of the others eo7i alone has a good connota- 

 tion for this place ; its untechnical meaning always includes long 

 duration. 



Group is not well placed in this rank. Prevalent American 

 usage, which puts it next above the ux\\\. formation, is in harmony 



