J. D. Dana — Subdivisions in Archaean History. 459 



atmospheres it is still an enormous amount beyond what ordinary- 

 life, even aquatic life, will endure. Reducing the estimated mean 

 thickness for the limestone layer over the globe from 1000 to 500 

 feet would make the amount nearly one half less.* 



The making of carbonates early began the work of storing 

 carbonic acid and purifying the atmosphere; and the introduction 

 of life increased the amount thus stored, and added to it through 

 the carbonaceous materials from living tissues contributed to the 

 earthy deposits. But with all the reductions that can be explained, 

 the excess is still very large. It has been pi"oved by experiment 

 that an excess also of oxygen diminishes the deleterious influence 

 of carbonic acid on plants; and that if the amount of this gas is 

 made equal to that of the oxygen in the present atmosphere, 

 plants will still thrive. How far this principle worked in early 

 time cannot be known. 



2. Subdivisions based on Stratification. 



The stratification in an Archaean region affords the only 

 safe and right basis for subdivisions. This method has been 

 used in the separation of the Huronian from the older Archaean ; 

 and recent]) 7 , with good success, by Irving and Tan Hise in 

 the study of the Penokee-Marquette region, or the Huronian 

 belt of "Wisconsin and Michigan. The intimate relation of 

 the beds in the series has been worked out and their uncon- 

 formability with the lower rocks thus ascertained, besides 

 the stratification and constitution of the iron-ore series within 

 the belt. This is the first step toward that complete study 

 which should be carried on throughout all Archaean areas, 

 however "complex." The distribution of the rocks and their 

 apparent or real stratigraphic . succession, whether massive or 

 schistose, the positions of the planes of foliation or bedding, 

 the unconformities in superposition, and those of mere fault- 

 ing, and all structural conditions, should be thoroughly inves- 

 tigated. Correlation by likeness of rocks has its value within 

 limited areas, but only after much questioning, f The work is 



* A right estimate is very desirable. If made for North America, it could not 

 be far out of the way to assume it to be a mean for like areas of the other con- 

 tinents as regards the limestone. But with the best possible result for the con- 

 tinents, the oceanic area, three times that of the continents, and out of the reach 

 of investigation as to depths of bottom deposits, remains a large source of doubt. 



f As a preliminary in the study of any such region, thousands of clips and 

 strikes of planes of foliation or bedding should betaken (in imitation of Percival's 

 work before 1842, mentioned in the note on page 440 of the last volume of this 

 Journal), and all should be plotted on maps of large scale by means of symbols 

 with affixed numbers recording the dips and strikes, for full comparison in the 

 final elaboration. Even the Penokee-Marquette region needs further investigation 

 with a clinometer-compass in hand. 



Before commencing the study of any crystalline rocks, models of flexures 

 should have been studied until the fact is fully appreciated that a flexure having 

 an inclined axis — the commonest kind — ranges through 180°, or nearly, in its 

 dips and strikes, and until the characters of the bedding in different transverse 



