Vol. 4] Thelen. — Thermal Conductivities of Certain Schists. 219 



The end section showed well developed prismatic cleavage in 

 nearly every specimen. Here none of the minerals were cut 

 even approximately parallel to c. The acute prismatic angle 

 varied but little on either side of 56°. The longer diagonals of 

 the cleavage parallelograms of the sections were roughly parallel 

 to b. The uniformity of the orientation was not as marked as 

 that of the c axes in the bottom or side sections of the rock, but 

 it was strongly developed. 



The orientation of those basal sections which were devoid of 

 the traces of the prismatic cleavage was readily determined by 

 the pleochroic colors. In fact, the whole scheme of orientation 

 was first worked out independently by means of the pleo- 

 chroism, then by means of the cleavage. 



The arrangement of the other minerals follows no law. If 

 the hornblende were eliminated from the rock, there would be 

 no reason why the wax figures should vary from the circle. 

 And the smoothness of the circle would be marked because, in 

 general, the crystals are so small that their heterogeneously 

 arranged variations in conductivity could not seriously affect 

 the outline of the circle. 



The relative thermal conductivities of the rock are very 

 satisfactorily explained by taking account of the relative con- 

 ductivities of the hornblende. 



These are 1.00 : .71- : .80-. 



The first number is for an axis which very nearly coincides 

 with c; the third is for an axis parallel to b, and the second 

 is perpendicular to these two. 



If we assume one-third of the rock to be hornblende, or im- 

 agine the whole rock to be one large crystal of a hornblende in 

 which the differences in thermal conductivities have been 

 diminished by two-thirds, we would have a ratio of 



1.00: 1.00— (1.00— .7l 2 ) /3: 1.00— (1.00— .80 2 ) /3. 

 This yields a crude approximation to the values we should find. 

 The ratio reduces to 1.00 : .914- : .938-. The rock shows a 

 ratio of 1.00 : .885- : .942 2 for directions which nearly coin- 

 cide with those of the hypothetical hornblende crystal. The 

 coincidences of the second members is not remarkable, but that 



