THIRD HYPOTHI^:SIS Ot" DH^RlVATION 



459 



XXIII. 



XXIV. 



XXV 



XXVI. 



XXVII 



XXVIII. 



XXIX 



XXX 



XXXI 

 XXXIl 



SiOo 



46.68 



49.19 



44.49 



54.13 



52.39 



46.31 



48.2 



47.3 



50.0 

 43.77 



TiOa. 



AloOo 



Trace 



17.12 



18.71 



16.37 



14.53 



16.13 



11.14 



17.9 



16.9 



13.4 

 17.07 



FeoOo 



2.18 

 5.03 

 5.07 

 1.50 

 1.64 



5.3 

 1.7 



4.3 

 4.17 



FeO. 



7.61 

 4.04 

 5.50 

 5.25 

 1.44 

 21.69 

 5.0 

 5.6 



7.3 



7.14 



MnO. 



P2O5 



CaO. 



13.46 

 5.92 

 7.94 

 4.91 

 8.76 

 9.68 

 10.9 

 13.3 



8.1 

 9.32 



MgO. 



10.34 

 7.98 

 7.50 



10.93 

 4.70 

 Trace 

 8.1 



11.3 



11.0 

 6.22 



KoO. 



Trace 

 .77 

 .56 

 .32 

 1.42 



2.7 

 .4 



1.6 



.81 



Na^U, 



1.75 

 1.44 

 2.59 

 3.53 

 2.59 

 6.91 

 .5 

 4.3 



2.6 

 3.15 



HoO. 



5.05 

 4.99 

 4.21 

 .17 

 4.44 

 1.5 



1.7 



5.63 



CO,. 



5.38 



In comparing the analysis of the hornblende schist of Manhattan (IV) 

 with the foregoing, the following differences are apparent : 



First, its low content of silica, agreeing only with the figures in XXV 

 and XXXII, and, secondly, its excess of iron oxides, only equaled in 

 XXVIII. 



A comparison of the analysis of diabase greenstone, XXII (which 

 closely approached that of our schist, IV), with that of the derivative 

 hornblende schist, XXV, shows in the latter a notable loss in iron oxides 

 and (on deduction of carbonates) in earthy bases. From the investiga- 

 tion of Rosenbusch (place cited), it has been too hastily inferred that the 

 alteration of a massive rock into a schist has been generally attended 

 with little change in chemical composition ; but conditions of altera- 

 tion have varied widely in different regions, and the Wisconsin rock has 

 been probably subjected to a process attended with partial leaching out 

 of silica and bases. To the varying amount of such leaching which has 

 generally prevailed the variation of the analyses in the foregoing table 

 may be reasonably attributed and their excess in silica and iron oxides 

 over those in our schist. 



The detection of fluorine in many samples of hornblende suggested 

 that its presence or quantity, as the result of the action of mineralizers 

 might serve as an indication of trappean origin. A review of over 200 

 analyses of the mineral, reported by Dana, Roth, Rammelsberg, etcetera, 

 yielded the following deductions : In 133 samples of amphiboles con- 

 taining little or no alumina, fluorine was found in 18, from a trace to 

 1.16 per cent, especially in actinolites and tremolites. In 70 analyses 

 of aluminous amphiboles, fluorine has been reported in 16, from 0.21 to 

 2.86 per cent. No definite conclusion could be drawn from the informa- 



