300 



DR. H. S. WASHINGTON ON THE 



[June 1 9 14, 



petrographic provinces, and two of these furnish analyses of aegirite- 

 granites which resemble rockallite quite closely. That which 

 resembles it most is the aegirite-rich portion of the alkali-granitic 

 mass of Ampasibitika in Madagascar, the analysis of which is given 

 under III in the Table on p. 297. This has been described by 

 Prof. Lacroix, who calls attention to its close analogy with 

 rockallite. Another is an segirite-granite from Corsica (Analy- 

 sis IV), in which island Dr. Deprat has discovered a small province 

 of highly sodic rocks. The symbols of these two granites are 

 respectively II". 3". 1. 4 and (II) III. 4. 1. "5, which express very 

 clearly their relations to the two fragments of rockallite. 



The Pyroxene. 



The most interesting mineral in rockallite is the pyroxene, 

 some optical data of which, as determined by Wright, have been 

 given on a preceding page. The chemical composition of the two 

 pyroxenes as a whole is very readily calculated from the analytical 

 data, and the results obtained by Judd and myself are given below, 

 together with some recent analyses of other segirites, and one 

 acmite. The earlier analyses may be consulted in the mineralogical 

 text-books of Dana and Hintze. 



V. VI. VII. VIII. IX. X. 



SiO, 51-97 50-63 50-45 51-73 51-08 51-35 



TiO~ 0-77 0-90 1 0-64 0-66 



ZrO~ 2-67 0-08 1 



A1 2 3 none none 1'78 2 1*91 0-80 1*59 



Ce 2 3 0-84 



Fe.,0., 33-48 29-92 23-42 31-86 29'30 32*11 



FeO 1-64 5-26 0-87 2-29 2*59 



MnO 2-41 0-28 0-10 0-60 111 0-37 



MO 0-20 



MgO 0-37 0-25 1-48 0-14 trace none 



CaO 0-52 1-41 5'92 0-87 2-54 trace 



Na 2 10-97 11-59 9-84 11-43 11-50 11-39 



K 2 none none 0-24 0*40 ... trace 



H 2 none none 0'55 0-20 1-04 



Totals 99-92 100-00 100-02 100-65 100-32 • 99-40 



1 Determined by H. S. W. 2 Corrected for TiO, and ZrO. 



V. ^Egirite-acmite, Rockall. Calculated by Judd. 

 VI. iEgirite-acmite, Rockall. Calculated by Washington. 

 VII. JSgirite, Magnet Cove (Arkansas). G. Steiger. analyst. Am. Journ. 

 Sci. ser. 4, vol. xiii (1902) p. 36. 

 VIII. .ZEgirite, Quincy (Massachusetts). C. H. Warren, analyst. Am. 

 Journ. Sci. ser. 4, vol. xxxi (1911) p. 553. 

 IX. JEgirite, Brevik (Norway). S. Hillebrand, analyst. Tscherm. Min. 

 Petr. Mitth. vol. xxxii (1913) p. 249. 

 X. Acmite, Brevik (Norway). C. Dcelter. analyst. Tscherm. Min. Petr. 

 Mitth. vol. i (1878) p. 381. 



It will be seen that the two calculated compositions of the 

 Roekall a?girite agree fairly well, apart from the presence of minor 

 constituents and the probably too high manganese oxide of No. V. 



