460 A. A. JULIEK — AMPHIBOLE SCHISTS OF MANHATTAN ISLAND 



tion at hand, and so far as concerns the hornblende schist of Manhattan 

 island, no fluorine could be detected by Doctor Jouet in his sample. 



IDENTIFICATION OF THE HORNBLENDE 



Comparison of methods of graphic analysis. — The peculiarities of the 

 chemical analysis of the hornblende schist (IV) sugojested a further effort 

 toward deduction of the chemical composition of the hornblende itself. 

 On account of the numerous minerals of the rock with closely approxi- 

 mate specific gravities, a process founded on dissection of the thin- 

 section, akin to the " mechanical process " of Delesse,* seemed best fitted 

 for the object in view. 



For the various processes founded on measurement, drawing, or pho- 

 tography of mineral plates in a thin-section, I would suggest the general 

 name o( graphic. Such a process, wdth the ensuing calculation, may be 

 based upon estimation of various dimensions. If the average diameter 

 of the grains of a mineral, as seen in a thin-section, be represented by 

 d, or the actual average area of its plates by a, the following possible 

 conditions may enter into such calculation : 



First, d. On relation to this, the " geometric rock analysis " of A. 

 Rosiwal t is founded — the application of an ocular micrometer to suc- 

 cessive measurement of length of cross-sections of plates, along a series 

 of lines over a section of the rock, of such thinness that all the grains are 

 cut by the two parallel faces. With a rock of uniform grain, especially 

 one whose grains approach cubical form, coincident results may be 

 expected from such a process ; but the relation of the edges of a series 

 of cubes can not, of course, be the same as that of the areas of their faces 

 or as that of their volumes. The concordance of results presented in 

 Rosiw^al's determinations seems to me, therefore, to have been largely 

 due to their mutual comparison rather than to comparison with some 

 established standard, such as corresponding results obtained on the 

 same rocks by some other exact process — for example, by means of 

 liquids of high density. So far as the relation in question may yet be 

 found sufficiently close for practical purposes, the process as described 

 may be well applied, particularly to massive rocks of uniform grain, both 

 as to cubical form and approximately equal dimensions. 



Second, d^. The volume of the mineral grain is beyond question the 

 factor called for in all cases toward exact calculation. Where the average 

 forms are nearly cubical and great inequality exists in the third dimen- 

 sion, thickness of the plates, of various minerals, d ma}^ well be deter- 



* Compt. rend., vol. xxv, 1847, p. 544 ; Ann. d. Mines (4), vol. xiii, 1848, p. 379 ; Proced6 m6canique 

 pour determiner la composition des roches, Paris, 1862. 

 f Ueber geometrische Gesteinsanalysen, Verh. Wien. geol. Reichs., vol. xxxii, 1898, pp. 143-1''5. 



