FRAGMENTATION — BARRELL 285 



substances.^ Barrell, both in bis cli-scussiou and in liis diii.;;i;un (fig. 1), pustu- 

 lates that granite near the surface grades downward into diorite, with a corre- 

 sponding decrease in the radioactive materials. Even so, it is probable, from 

 the best data available, that the upi)er limit of the zone with basaltic com- 

 position lies higher than is indicated in Figure 1. Possibly the lower limit 

 of isostatic compensation lies wholly within this basic zone, and not in a dioritic 

 zone as indicated by the figure. If this conception is adopted, the black of 

 the figure should be extended upward into the lower part of the lithosphere, 

 and conduits leading to the higher intrusive bodies should be pictured as trans- 

 gressing a zone of solid basic material before entering the more acidic portion 

 of the crust. Under such a conception the mechanism proposed by Barrell would 

 encounter ditficulties, as tliere can be no effective change of volume in the 

 lithosphere so long as the material at its base has essentially the same density 

 as the ascending magma. 



If the features of Figure 1 are taken literally, the form of the Intrusive bodies 

 may be open to question. In his later years Barrell was impiessed with the 

 lopolithic form of certain large basic intrusions ; and, judging from the figure, 

 he considered this form as probably predominant. Clearly such a conception 

 must assume a horizontal gneissoid or schistose structure in the granitic shell. 

 However, this question is of only incidental interest so far as the present paper 

 is concerned. In the text Barrell suggests that the intrusions in the upper 

 part of the shell have the form of " lopoliths and batholiths." Figure 1 is to 

 be considered merely as a diagrammatic representation of the mechanism 

 assumed. The essential point is concerned with the volume of basic intrusions 

 and not their form. As Barrell points out, the volume necessary for the task 

 involved in the problem is formidable, and the mechanical ditficulties of in- 

 trusion on the scale assumed are increased by the present-day conception of a 

 tliin granitic shell. It is a question whether he might not now want to assign 

 a larger rule to extrusion as a cause of crustal foundering. Arguments for such 

 a conception might be based t)u the great thickness of Keweenawan lavas, men- 

 tioned in the paper, as well as on phenomena of the lunar maria. 



The merits and weaknesses of Barren's hypothesis, as seen from our present 

 viewpoint, may be discussed at much greater length ; but it is not the purpose 

 of this preface to iireseut a full critical analysis. 



ON CONTINENTAL FRAGMENTATION 



INTRODUCTORY 



The outer crust of the earth, granitic in its upper part and some- 

 what more basic at depth, is held to have a thickness of from 50 to 

 75 miles. It is ver}^ strong and is marked by broad variations in 

 density amounting to as much as 5 per cent, and by more local 

 variations up to 10 per cent, these differences corresponding to the 

 broader relief of the earth's surface. Below this lies a thick, hot, 

 basic, rigid yet weak shell, the asthcnosphere, or sphere of weakness. 

 The problem of the origin of the ocean basins and the continental 

 ])latforms consequently resolves itself into one of the origin of the 



» Jeffreys made an estimate based on the assumption that the radioactive content faRa 



off exponentially witli depth. 



