Contributions to the Mineralogy of New Zealand—Part I.

By C. Osborne Hutton, Petrologist, Geological Survey, and F. T. Seelye, Chemist, Dominion Laboratory, Wellington.

* All refractive index measurements for this investigation were made in sodium light.

Corundophilite.

A ferriferous corundophilite was separated by electromagnetic and centrifuge methods, from a dravite-chlorite rock; these chlorite rocks are (Bell, 1907) altered ultrabasics, and they have been injected into the Aorere Series of argillites, greywackes, quartzites and schists. An analysis of this chlorite is closely comparable to that of one from Western Australia.

The optical properties of the north-west Nelson corundophilite have been determined as follows:—

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Formula: (OH)_{7 7 04} (Mg, Fe'', Ti, Al)_{6 07 07} [(SiAl)₄ O₁₀].

The refractive indices for this chlorite fall within the limits defined for prochlorite by Winchell (1936, p. 648)—viz., Nm = 1.59–1 62, Ng-Np = 0.004–0.010; in the broader sense, however, the mineral lies in the field of ferriferous corundophilites (Winchell, 1933, p. 284).

When the analysis is calculated on the basis of 18 (O.OH) atoms to the unit cell (Pauling, 1930) good agreement is obtained with the structural formula (OH)₈ Y₆ [(SiAl)₄ O₁₀], derived by X-ray analysis (Table V), the only noteworthy variation being the low figure for the (OH) group. This point was observed with two prochlorites from chlorite zone schists in Western Otago (Hutton, 1940, pp. 18–19) and one from similar schists in South Devon (Tilley, 1938, p. 503). It is difficult to account for this consistent low value, especially in view of the absence of noteworthy oxidation of the ferrous iron and the lack of fluorine. Some of the hydroxyl groups are linked to Mg, Al, and Fe, as in micas, and the remainder are grouped in a brucite layerlattice. Thus it would appear as if the hydroxyl group could be completely displaced from chlorites by moderate heating, and that there should be no need to resort to severe treatment with the use of the oxycoal gas blowpipe flame such as is required in the case of the amphibole and epidote groups (Smethurst, 1935, pp. 173–179; Hutton, 1940, p. 12).

Dravite.

Dravite occurs in large, simple, six-sided prisms, sometimes up to 6 inches in length, embedded in a matrix of fine, flaky corun-dophilite; in some specimens the tourmaline has crystallised in beautiful radiating forms. A large fragment of dravite was readily separated from its host, and a pure fraction obtained by centrifuge methods. The composition (Table VI, anal. A.) of this tourmaline is comparable with that of dravite from a marble in Connecticut, U.S.A. (anal. B), and with a similar type occurring in migmatized serpentinites and amphibolites in Sweden (anal. C.); the paragenesis of the latter example is somewhat similar to that of the Nelson tourmaline.

In an investigation concerning the composition of tourmaline in relation to their host rocks, Agrell (1941) concludes that with only

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Formula: X_{2 00} Y_{27 17} B_{8 84} Si₁₈ H_{10 57} O₊₊₊ or, fully expanded, we get: [Na_{1 85} Ca_{0 84}] [Mg_{0 80}, Fe''_{1 10}, Fe'''_{0 78} Ti_{0 20}, Al_{17 72}] B_{8 81} Si₁₈ H_10.57 O _{[ unclear: ]} .

The formula thus derived compares reasonably well with those proposed by Kunitz or Machatschki; however, it should be observed that an exceedingly small amount of aluminium occurs in the 4-fold co-ordination and further, there is a slight deficiency in boron and the X group.

In connection with the analytical work, it should be stressed that the figures for FeO and Fe₂O₃ given in Table VI be regarded as only approximately correct as the mineral is attacked extremely slowly by mixtures of HF and dilute H₂SO₄. In determining the FeO repeated attacks with the mixed acids were necessary with intermediate further grinding of the residue until the whole of the sample was decomposed; this treatment necessarily multiplies any error of experiment, and tends to decrease the FeO: Fe₂O₃ ratio. Although Rowledge's method for FeO (Groves, 1937, p. 85) by fusion with sodium meta fluo-borate gives a good decomposition and a somewhat higher result for FeO it was not thought advisable to accept this figure as there is some evidence that the flux itself will reduce at least some ferric compounds to the ferrous state and so produce too high a result for FeO.

The following physical constants have been determined:—

Beryl.

Beryl is a rare constituent of granite and granite pegmatites in New Zealand, occurring at Dusky Sound, Paterson's Inlet, and Charleston, where it can be considered no more than a mineralogical curiosity. At Charleston a few large, coarse, columnar crystals, up

to 5 inches in diameter, have been found, the analysis and optical properties of which are presented in Table VIII. The crystal analysed (P. 6893) is very pale green to nearly colourless, and is considerably stained with iron. Prism faces are well developed, but terminal faces, except basal plane at one end only, are undeveloped.

When alkalis are plotted against α and γ on Winchell's diagram (1933, p. 213) showing the effect of alkalis on refractive indices in beryl, the points for Charleston beryl, though very close to Winchell's curves, lie slightly above them.

The structure of beryl has been analysed by Bragg and West (1926), who find that the unit cell contains two molecules of Be₃ Al₂ Si₆ O₁₈. On the basis of 18 oxygen atoms the beryl analysis has been recalculated (Table IX); magnesium and iron have been grouped with aluminium, the total amount of the later being in 6-fold coordination, whereas sodium has been grouped with beryllium (Ford, 1932, p. 580). It would appear to the authors that the extent of this latter substitution would need to be exceedingly small when the great difference between the radii of the sodium and beryllium ions is considered.

Formula: (Be. Na)_{2 [ unclear: ] 1} (Mg, Fe'', Fe''', Al)_{2 1} Si₆ O₁₈.