Variation in the Composition of Muscovite and Albite in a Pegmatite Dike Near Yellowknife: Discussion

Abstract

The analytical portions of a recent paper by Kretz (1970) leave certain questions unanswered, and raise some others. (1) In Table 11, what is the signiiicance of the 10-ml aliquot? If the solution is not diluted, why list an aliquot? It is hard to see how sodium and lithium can be determined at the same dilution level for the concentrations involved, unless an unusual sodium line was used. Wavelengths are not listed in the paper. The lithium concentrations for samples 1 and 8, Table V, are so close to the detection limit that use of two significant figures is questionable. (2) Kretz used Angino and Billings' (1967, p. 96) hydrofluoric-sulfuric decomposition although even those authors state that . . It is probably simpler, however, to use an acid other than sulfuric. . .. Kretz did change subsequently to perchloric acid, but no indication is of which results were obtained with which acid. The 0.64% lanthanum in the solutions for calcium and magnesium determination may not be sufficient to overcome the chemical interference of so much sulfate ion, and may account for the small but persistent negative bias in the results for those metals (Table 111). Solutions from hydrofluoric-sulfuric and similar decompositions are not generally considered suitable for aluminum determination (Abbey 1 9 7 0 ~ ) . The only example in the literature is that of Galle (1968), and he followed up his decomposition with a pyrosulfate fusion. (3) Is it really necessary to show calibration curves? Would not a simple statement be sufficient to indicate that all were linear? If curves must be shown, why is there none for lithium? More significantly, why is there no indication of the composition and manner of preparation of the solutions used for calibration? (4) In Table 111, Flanagan ( 1967) is quoted for the given values for G2, GSP 1, and AGV 1. A more comprehensive compilation is available (Flanagan 1969), and also a critical study of the data (Abbey 1970b). Admittedly, the assigned values in the latter study do not differ greatly from those in Flanagan (1967), but it must be remembered that one set of analyses by one laboratory does not make a sample. (5) Is it really necessary to d e h e mean and standard deviation? I