88 Washington — Pyroxenite and Hornblendite in Brazil. 



Since high alumina in one is connected with low magnesia in 

 the other, the difference being almost exactly the same, while 

 the other components of the rock are almost identical in both* 

 one is inclined to suspect the common error in the determina- 

 tion of these two constituents in the diorite. If this were the 

 case, the close resemblance between the two would have been 

 still greater. 



The two authors determined the modes at several points reg- 

 ularly distributed between the two extremes, and have plotted 

 them on a diagram. This shows very clearly the inverse 

 behavior of the hornblende and feldspar, while the minor con- 

 stituents, as quartz and biotite, remain comparatively constant. 

 The significance of this is not commented on by the authors, 

 but it is clearly evident, both from their diagram and from the 

 norms, that much of the actual feldspar of the diorite has lost 

 its modal existence, and become incorporated in the complex 

 hornblende as one approaches the border. 



In 1901, Lacroix* described some pyroxenites from the Pyr- 

 enees, an analysis of one of which is given in Table III. 

 These pyroxenites, called by Lacroix " ariegites," which con- 

 tain varying amounts of hornblende or garnet with magnetite, 

 and only traces of olivine, are distinctly lower in silica and 

 iron oxides, and much higher in alumina and lime than that 

 from Bahia. Lacroix calls attention to their chemical affinities 

 with the gabbros, and on this account he undertook some 

 experiments in recrystallizing them from a fused condition. 

 On reheating (at a temperature not stated) the glass obtained 

 by fusing the rocks in a platinum crucible, he obtained masses 

 made of " microlites of augite imbedded in larger crystals of 

 bytownite." Olivine was also formed readily in the melt. On 

 referring to the norms of these rocks,t it will be seen that they 

 correspond well with the modes of these artificially recrystal- 

 lized melts. We may, therefore, consider that, even for rocks 

 of such complicated abnormative modes, the norm as calculated 

 in the quantitative system is justified as a basis of comparison. 

 Lacroix himself points out (p. 360) the theoretical interest of 

 his experiments as demonstrating that the differing modes are 

 due solely to differing conditions during consolidation (as well 

 as, almost undoubtedly, to the presence or absence of gases or 

 water vapor), and remarks that had his ariegites solidified as 

 volcanic lava flows, they would have formed pyroxene-rich 

 basalts. 



Mr. G. A. Rankin, of the Geophysical Laboratory, to whom 

 I desire to express my thanks, very kindly undertook similar 



* A. Lacroix, C. R. xxxii, p. 358, 1901. 



+ H. S. Washington, U. S. G. S. Prof. Paper 14, 1903, pp. 335, 337, 339, 

 347. 



