THE MINERALOGRAPHY OF THE FELDSPARS 251 
the specimen more accurately than the intersection of the isogonic 
line and the side of the triangle. The latter procedure is the one 
commonly followed although the petrologist does not express the 
method in these terms. 
An important fact, which needs to be emphasized, is that in 
using the conventional extinction curves of the textbooks to 
determine the percentage of the soda component in a plagioclase, 
it is not possible to determine the percentage of the lime and 
potash members with anywhere near the same accuracy. The 
reason for this fact is that the isogonic lines are nearly parallel to 
the potash-soda side of the triangle. It follows then that the 
most satisfactory means of determining the true composition. the 
percentage of each of the three components, is to consider that 
the value of the extinction angles gives the percentage of the soda 
component only. Knowing this fact it can be ascertained what 
particular subdivision of the series is being examined. By con- 
sulting the compositional line or by reference to the average per- 
centage of the potash component in the natural plagioclase, the 
approximate amount of the third component is found. The 
amount of the line component is the remainder. This can be illus- 
trated: Suppose the extinction angles of a specimen were found to 
be 23° on the (oro) face and 10° on the (oor) face. The theoretical 
composition of this specimen would be found from the plot shown 
in Figure 3 to be Ab,,An¢., but by the method here suggested the 
value of the Ab alone is correct. Now it has been found that most 
labradorites contain an average of 6 per cent of the potash com- 
ponent. Thus the amount of the lime member is 60 minus 6, 
or 54 per cent. Therefore the composition of the specimen is 
K-feldspar, 6; Na-feldspar, 40; Ca-feldspar, 54. 
The fact that the isogonic lines are nearly parallel to the 
potash-soda side of the triangle is the reason why it is easier to 
determine the composition of the plagioclase feldspars than that 
of the potash-soda series; for the isogonic lines intersect that side 
of the diagram more frequently. To state it in another way, there 
is a greater change in the value of the extinction angles per unit 
change in composition. As most petrographic determinations of 
extinction angles are only approximate it follows that an error in 
