Isolation of Minerals.

The crystalline rocks examined comprised: (1) those in which certain minerals were distributed in large crystals or arranged in folia, so that clean samples could be picked out by hand; and (2) those from which the various minerals had to be separated by means of a liquid of high specific gravity.

The first class included many samples from the gneiss, syenite and granite of the Manipori formation, N. Z., from which the separation of mica and hornblende was specially easy. In various parts of this formation mica-veins are abundant. It was sometimes possible to pick in a few minutes 2 or 3 pounds of clean mica in plates up to 2 or 3 inches square. The syenite and syenitic gneiss often contain bands of almost pure hornblende, quite an inch thick, permitting a similarly easy isolation. Finally, magnetite can, of course, be easily separated from any rock in which it occurs.

The treatment of the second class, comprising many of the rocks examined, may require some explanation.

The separation of minerals from rocks by means of heavy solutions is extremely tedious, and it is therefore desirable to reduce the quantity of the rock to be operated upon if this can be done without material loss of the mineral to be isolated. As all crystalline rocks contain a large percentage of material below 3 in sp. gr., which can be effectively removed by panning, each sample was concentrated first by panning to get rid of quartz, feldspar and other specifically light materials. This procedure would, of course, be inadmissable if the object had been to determine accurately the gold and silver per ton of the rock examined; but the purpose in each case being simply to get as large a quantity as possible of a given mineral, and to get that sample approximately pure, this preliminary crude concentration seemed to be unobjectionable.

As an illustration, the following record of the process, as applied in one instance (Sample 7, Table XVIII.), is here given. With suitable modifications, this method was followed in all similar cases:

Sample of Syenite.

1.	The sample was roughly broken and examined with a good lens, but no sulphides were discovered. A chemical analysis of 10 grammes of the sample also showed that sulphides were absent.page 30
2.	A portion of 500 grammes, pulverized so as to pass a No. 30 sieve, was reduced by careful panning to 217 grammes, of which 58 grammes was removed by a weak magnet. The portion thus removed proved, under the microscope, not to be pure magnetite, but to contain a good proportion of hornblende and feldspar (chiefly orthoclase) grains, probably drawn to the magnet by small adhering particles of magnetite.
3.	The 159 grammes not attracted by the magnet consisted chiefly of hornblende and feldspar, with a little mica (biotite) and quartz. This powder was introduced into an apparatus modeled on Thoulet's, but made much larger, to save time. The heavy liquid used was Sonstadt's: mercuric iodide, dissolved in excess of potassic iodide, and having 3.175 sp. gr. On exhausting the air from the apparatus, 77.5 grammes of practically pure hornblende fell to the bottom of the liquid. A dilution of this liquid to 2.95 sp. gr. was followed by the precipitation of 35.5 grammes more—also nearly pure hornblende. The total of 113 grammes, reckoned as hornblende, was finely powdered and assayed with pure litharge. No trace of gold or silver was obtained.
4.	The 58 grammes removed by the magnet was similarly powdered and assayed, with the same result.

Sonstadt's solution is cheap and easily prepared, but has the great disadvantage of being extremely corrosive. Even with the greatest care, it was found almost impossible for the operator to avoid burning his fingers. Hence, in a number of instances Klein's solution of boro-tungstate of cadmium was employed. This has another advantage over Sonstadt's solution, namely, that when a partial separation of hornblende from augite is desired, the greater part of the hornblende can be kept from sinking in the liquid by using the latter in the concentrated state.^*

The results of the analysis of minerals separated from 47 samples of rock are given in Table XVIII. On this table the following additional notes are presented:

1. Sample 17 represents a rock of peculiar occurrence, occupying a narrow strip in an area of Tertiary trachyte at Porto-bello, on the east coast of the Otago peninsula. Samples taken short distances apart vary much in character; but, on the whole the rock may be classed as a diorite. Gold was said to have

^* It was found impossible to separate augite from hornblende completely by the use of heavy liquids; but since no gold or silver was found in either, this fact did not affect the practical results of this investigation. The same remark applies more or less to all the minerals thus isolated for analysis. The writer knows of no method which will perfectly isolate considerable quantities of one mineral from other minerals not widely removed in specific gravity. In Table XVIII therefore, the terms "hornblende," "augite," etc., simply designate the greatest part of the mineral samples to which they are applied.

page 31 been found in the rock, up to half an ounce to the ton. If this were the case, such an occurrence of gold would have been, so far as the writer knows, unique.

Of the three samples chosen, one (No. 17) contained pyrite, and this pyrite was—but the rock itself was not—auriferous. The occurrence of auriferous pyrite in such an area (of recent volcanic rocks) is most unusual, and Prof. Ulrich, Director of the Otago School of Mines,^* who has examined the locality and studied rock-sections from it, thinks that this dioritic rock probably underlies the basic volcanics which form the rest of the peninsula. In that case, it would be allied rather to the dikes of the Upper Silurian in Victoria than to the newer volcanic rocks of its immediate neighborhood.

2. Samples 29 to 34 are from the Thames district. It was found difficult to get samples of the Thames andesites in which the analysis of 5 grammes would show no trace of sulphides—this being, as already explained, a requisite condition for the particular investigation in hand. A large number of samples had to be rejected on this account; but in Nos. 29 to 34 no trace of sulphides was found. Tables XVI. and XVII, on the other hand, give numerous samples of the Thames andegite country-rock, containing pyrite and other sulphides, carrying in many cases notable quantities of gold and silver.

3. Nine samples in Table XVIII.—namely, Nos. 6, 11, 13, 17, 26, 36, 38, 46 and 47—contained sulphides. The results of further examination were as follows:

No. 6. Sample of 500 grammes gave 53 grammes of sulphides, mostly pyrrhotite and chalcopyrite, which contained 0.0053 grain of silver.

No. 11. Sample of 500 grammes gave 18.56 grammes, chiefly pyrrhotite. No gold or silver.

No. 13. Sample of 1000 grammes gave 12.18 grammes of pyrrliotite and arsenopyrite. No gold or silver.

No. 17. Sample of 1000 grammes gave 4.934 grammes of pyrite, containing 0.037 grain of gold and 0.0041 grain of silver.

No. 26. Sample of 1000 grammes gave 28.92 of arsenopyrite and pyrrliotite, with a little galena, containing 0.0032 grain of gold and 0.0019 grain of silver.

No. 36. Sample of 1000 grammes gave 16.48 grammes chal-copyrite, with a small quantity of bournonite, carrying 0.407 grain of silver and no gold.

No. 38. Sample of 1000 grammes gave 16.03 grammes of pyrite. No gold or silver.

No. 46. Sample of 1000 grammes gave 8.42 grammes of pyrite. No gold or silver.

No. 47. Sample of 500 grammes of pyrite and sphalerite, containing 0.0016 grain of gold and 0.0037 grain of silver.

Conclusions.—The results summarized in Table XVIII. were greatly surprising to the writer. In view of the usually tedious character of the operation of isolating the various constituents of a rock, he would not have examined so large a number of samples had he not expected, at each new analysis, that he might succeed in discovering gold in some mineral other than a sulphide.

It is, perhaps, comparatively easy to conceive why, in a stratified area, gold may occur only in connection with sulphides; but that in such a rock as gneiss, granite, syenite, or diorite, it should form no part of the crystalline constituents, but, on the contrary, should occur only in the sulphides found in these rocks, seems more remarkable and significant.