504 Transactions. 



successful application of magnetic concentration, it appeared desirable to 

 obtain some data as to the extent to which the percentage iron-content 

 could be increased and objectionable impurities eliminated by such means. 



Eepresentative samples were obtained by Mr. W. C4ibson, B.E., Assistant 

 Geologist, from Patea, covering an area estimated to contain at least 

 5,000,000 tons of sand, and also from the neighbourhood of New Plymouth. 

 These were analysed for iron, titanic oxide, phosphorus, and vanadium. 



Iron. — This was estimated by reduction of 0-5 gram of the finely ground 

 sample in a current of hydrogen gas at a red heat, follow^ed by solution 

 in semi-dilute sulphuric acid, and titration with deci-normal potassium 

 permanganate.* The titrations were made at a temperature of 70° C, 

 and corrected for vanadium by deducting 0-4 c.c. for every 0-001 gram of 

 vanadium present. 



Titanium. — Titanium was estimated colorimetrically by adding hydrogen 

 peroxide to the solution after the permanganate titration, and comparing 

 with a standard solution (\V"eller"s method). 



Phosphorus. — The determination of phosphorus in ironsand is rather 

 difficult. In the presence of titanium there is a tendency, when eliminating 

 silica, for a highly insoluble titano-phosphate to be formed. An excess of iron 

 in solution would appear to retard the precipitation of traces of phosphoric 

 acid. On the other hand, vanadium, when present, is partly carried down 

 with the phosphorus, and contaminates, though only slightly, the final 

 precipitate. 



The method finally adopted was to fuse 1 gram of ironsand with 8 grams 

 sodium carbonate, and extract with water. The insoluble residue was re- 

 fused, and again extracted. A complete separation was thus obtained 

 between the titanium and iron, which remained in the insoluble portion, 

 .and phosphorus and vanadium, which were entirely soluble. 



The water-extract was acidified with nitric acid, evaporated to dryness, 

 taken up with hot dilute nitric acid, filtered to remove silica, and again 

 evaporated, this time to small bulk. The phosphorus was precipitated 

 by the method of Stunkel, Wetzke, and Wagner (as described in Crooke's 

 " Select Methods in Chemical Analysis," 3rd ed., p. 509), and weighed as 

 magnesium pyrophosphate. Any vanadium in the precipitate was deter- 

 mined colorimetrically, and deducted. 



Blank determinations with known amounts of phosphorus, iron, and 

 titanium gave excellent results. 



Vanadium. — This was estimated colorimetrically with hydrogen peroxide 

 in the presence of nitric acid. One gram of the sample was fused with 

 •6 grams of sodium potassium carbonate, and extracted with water. The 

 extract was acidified very slightly with nitric acid, and 1 c.c. of hydrogen 

 peroxide (10 per cent, by volume) added. The brown tint developed was 

 compared with that of a similar solution containing a known amount of 

 vanadium. 



Magnetic Separation. — The samples were also separated by the action of 

 first a weak and then a strong magnet into three portions : No. 1, strongly 

 magnetic, separated by a very weak magnet ; No. 2, feebly magnetic, by 

 the use of a triple horse-shoe magnet, after removal of portion 1 ; No. 3. 

 non - magnetic residue. Portions 1 and 2 were then analysed for iron, 

 titanium, phosphorus, and vanadium, and the results compared with the 

 analyses of the original sands, as shown in Table I. 



* Trans. N.Z. Inst., vol. 41 (1908), pp. 49-51. 



