PALISADE TRAP 



373 



able width. We shall further find for K a value near 4.8, and for ay/ii^ 

 the value .000016. 



We notice that the initial temperature appears to be relatively higher 

 in the intrusive sheet than in the surface flow of very similar composi- 

 tion. This might mean that the magma itself was hotter or that the 

 country rock was hotter or both. 



We must, however, observe that the curve of grain, especially at the 

 margin, suggests a curve of the type of .49 of plate 57, and that by taking 

 s within the limits of the error of observation, for instance, .000047, Ave 

 could bring it under this type ; but the contact zone would have to be 

 very small, and this would be especially forced in view of the high 

 initial temperature which would be implied. The other explanation of 

 this somewhat greater grain at the margin would be that the effect of 

 stirring and cooling, down to the point of formation of augite before 

 motion has ceased, had produced the observed effect. 



In passing, it will be interesting to compare the grain of the feldspar, 

 for which we have the following table : 



Distance. 



Difference. 



Grain in milli- 

 meters. 



Difference. 



Quotient = s. 









.053 



.00013 



390 





.053 









1110 





.0265 



.000024 



1500 





.0795 









2410 





.0800 



.000033 



3910 





.1595 









3780 





.1245 



.000030 



7690 





.2840 









3900 





.0210 



.0000054 



11590 





.3050 









5810 





.1100 



.000019 



17400 





.4150 









5810 





.1420 



.000024 



23210 





.5570 









11280 





.1230 



.000011 



34490 





.6800 







The average gradient s from 1,500 to 17,400 millimeters is .000021, and 

 it is obvious that the gradient remains equally steep up to 23,210 and per- 

 haps beyond. This is farther than in the case of the augite and is an indi- 

 cation, by equations (18) and (19) of page 397, that the period of formation 

 of the feldspar is earlier than that of the augite, its temperature of con- 

 solidation being nearer the initial temperature. This is true, as Andrew 

 and Osann have remarked on page 4. From this we should compute an 

 average gradient s of .000022, and if we suppose s to represent A and 

 compute B we find half the values of B will be possible negative values. 



