72 DK. H. S. WASHIXGTOX OX THE TITAXIFEROTJS [Feb. I907, 



of what is apparently leucite, though the only analyses yet made of 

 these types render the identity of this mineral somewhat uncertain. 

 Hornblende is never seen in the thin sections of any of these rocks, 

 but phenocrysts of biotite occur in one rare type at Monte Ferru, 

 as mentioned above. 



As might be expected from the chemical composition, segirite is 

 not present in these salfemanes, nor do the augites proper appear 

 to carry much of the segirite-molecule. It may be mentioned here 

 that, in the more salic rocks, in which its presence might be looked 

 for, segirite is not a common mineral, although the occurrence of 

 the soda-hornblende, senigmatite, in some of the Pantellerian lavas 

 is well known; but it appears to be far less abundant than is 

 commonly supposed. 



Concerning the textures little need be said. They are those 

 usual in such femic eruptive rocks, but the generally very slight 

 megaphyric development may be noted, as well as the almost 

 complete absence of true ophitic textures. Many of the basalts, 

 especially the more compact ones, are noncrystalline, while others 

 are more or less vitreous, the vesicular forms consisting largely of 

 glass, which is either dark in colour, or dusty and opaque with very 

 numerous, minute, black microlites. 



III. Chemical Cojipositiox. 



The chemical composition of these basalts is shown in the accom- 

 panying table (I, pp. 74-75), which includes only a portion of those 

 analyses that have been made by me. Those here presented are 

 regarded as representative of the majority of the salfemanes (basalts) 

 of the several districts, and most of the other available examples 

 much resemble them in their essential features. 



The methods of analysis were those commonly employed in the 

 United States. 1 The alkalies were determined by the Lawrence- 

 Smith method, and titanium-dioxide by the colorimetric method of 

 Weller. The latter, it may be remarked, is far more accurate, as 

 well as more expeditious, than the gravimetric method so often 

 employed ; and, as the figures for titanium-dioxide are in all cases 

 the means of three closely-agreeing determinations, they may be 

 relied upon. Combined water (H 2 + ) was determined directly by 

 Penfield's method, since the loss on ignition gives rise to very 

 misleading figures, owing to the oxidation of the large amounts Gf 

 ferrous oxide present in these rocks. Some of the analyses here 

 presented are not complete, although sufficiently so for the object 

 immediately in view, and are to be amplified in the future by 

 determinations of some minor constituents, such as zirconia, nickel- 

 monoxide, manganous oxide, cuprous oxide, and baryta. In their 

 present form, therefore, they are to be regarded as provisional, 

 although the corrections involved by such determinations in the 

 figures given here will be presumably of small magnitude. It may 



1 W. F. Hillebrand, * Some Principles & Methods of Eock- Analysis ' Bull. 

 U.S. Geol. Surv. Ko. 176 (1900) ; and H. S. Washington, ' Manual of the Chemical 

 Analysis of Rocks ' New York & London, 1904. 



