Vol. 70.] COMPOSITION OF ROCKALLITE. 297 



phosphorus pentoxide obtained by me would almost justify the 

 application of the term ' trace " to them 



I. II III IV. 



SiO., 73-60 69-80 68*70 68*29 



Al.,6., 4-70 5-10 6-85 8'71 



Fe*6, 13-10 13-23 9*93 11-60 



Feb' n. d. 0-78 1-14 trace 



MgO 0-11 0-11 0*26 0-40 



CaO 0-37 0-72 1'34 0*51 



Xa.,0 6-96 8-04 7-01 10-63 



K 2 trace 0-22 1-58 0-67 



H 2 0-f- n. d. 0-46 0'50 n. d. 



H 2 0- 0-31 



CO., ... none 



Ti6., 0-34 0-26 



ZrO" 1-17 3-71 "incl. in SiO,.~ 



P 2 O o : 0-07 



S0 3 none 



CI ... none 



S ... none 



Ce.,0, 0-37 



MnO 0-93 0-12 trace 



XiO 0-06 n.d. 



BaO none 



Totals 99-83 100-84 101-28 100-81 



I. Rockallite : (II) III. (2) 3. 1. 5. Eockall. C. J. S. Makins. analyst. J. W. 

 Judd, Trans. R. Irish Acad. vol. xxxi (1897) p. 54. 

 II. Eockallite, "in. 3. 1. 5. Eockall. H. S. Washington, analyst. 

 ITI. JEgirite-granite. IT'. 3''. 1. 4. Ampasibitika. Madagascar. Pisani, analyst. 

 A. Lacroix, ' Materiaux pour la Mineralogie de Madagascar ' pt. 2, 

 1903. p. 235 (Xouv. Arch. Mus. Hist. Nat. Paris, ser. 4, vol. v). 

 IV. jEgirite -granite. (11)111. 4. 1. "5. Capo alia Cucnlla (Corsica). J .Deprat, 

 analyst. Bull. Serv. Carte G-eol. France, vol. xvii (1906) No. 114. p. 46. 



Analysis II is of especial interest, because of some of the minor 

 constituents. The absence of carbon dioxide, sulphur trioxide, and 

 sulphur was to be expected, and they were only looked for in the 

 regular analytical routine. The same applies to the absence of 

 barium, as this element favours potassic rather than sodic rocks. 1 

 The amounts of titanium oxide and manganese monoxide are 

 decidedly smaller than was expected, and the presence of what 

 amounts to scarcely more than a ' trace " of phosphorus pentoxide is 

 in accord with the rarity of apatite in the sections examined by me; 



The two minor constituents that are of most interest are 

 zirconia and ceria. both of which were found to be present in 

 unexpectedly large amounts. In his letter referred to above, 

 Prof. Judd suggests that the former may be due to zircons 

 present in submicroscopic crystals, as no zircons were visible in 

 his sections. In mv two sections some Very minute zircons were 



1 Cf. H. S. Washington, Carnegie Inst. Publ. No. 57 (1906) p. 188 : and 

 Trans. Am. Inst. Min. Eng. vol. xxxix (1909) p. 754. 



