ROCK SPECIMENS FROM CENTRAL AND WESTERN AUSTRALIA. 311 



Central Australia. The nearest allies known to the writer 

 are to be found in the dyke at Norseman already referred 

 to. A study of eight specimens from different parts of the 

 dyke, kindly presented to the writer by Mr. A. Gibb Mait- 

 land, Government Geologist of Western Australia, shows a 

 series of rocks ranging from hypersthene through olivine 

 norite to quartz norite, with enstatite-augite and pyroxene- 

 perthite in many of the specimens. The dyke runs east 

 and west through a series of amphibolites and schists that 

 lie in N.E-S.W. belts. It is distinctly the youngest of the 

 solid rocks of the field, later even than the granites and 

 accompanying quartz-porphyries, which run in the plane 

 of the foliation. In these respects, and in its petrological 

 characters (though not its coarseness) it is typical of a large 

 series of dykes traversing the rocks of the different gold- 

 fields. These "later dykes" as the writer proposes to 

 term them, are in many fields distinct from the gold bearing 

 lodes, which they intersect and fault. The areas occupied 

 by them in Central Australia may therefore be presumed 

 to be non-auriferous. It does not necessarily follow, how- 

 ever, that the whole area is non-auriferous. Owing to 

 differential erosion, these later dykes frequently stand up 

 above the softer auriferous rocks, and afford the best 

 opportunities of collecting specimens. So it may be that 

 in central Australia the later dykes are intrusive through 

 auriferous rocks also. The evidence on this point is quite 

 non-conclusive. 



The presence of enstatite-augite and quartz-dolerites in 

 Central and Western Australia leads to theoretical con- 

 siderations of more than local interest. Though those 

 described are fairly deep-seated and approaching gabbros 

 in crystallisation, there are petrologically similar rocks of 

 finer grain among the later dykes of Western Australia 

 known to the writer which clearly show the close connec- 



