310 J. A. THOMSON. 



According to the norm of the American authors there is 

 17*86 diopside and 20*30 hypersthene. In the mode, how- 

 ever, there is not so much hypersthene, and most of the 

 metasilicate is contained in an augite poor in lime. It is 

 difficult to account for the presence of quartz in the norm 

 in a rock from which olivine has crystallised. 



The analysis is fairly closely paralleled by that of the 

 "norite" of Norseman, which is a very similar rock as is 

 pointed out below. It contains, however, a higher propor- 

 tion of hypersthene. The diabase from British Guiana, of 

 which I have not been able to compare the mineral com- 

 position, also agrees sufficiently well to place it in the same 

 class, but with a higher percentage of soda there is a distinct 

 drop in the amount of lime, reflected by a smaller drop in 

 the alumina. The Hunnediabase from Launceston shows 

 the same differences and also lower iron and magnesia per- 

 centages, due probably to the absence of olivine. In spite 

 of these minor differences, the four analyses obviously show 

 a close agreement. 



The nomenclature of this group of rocks is a matter of 

 difficulty. Overlooking the presence of hypersthene, those 

 rocks with enstatite-augite and without quartz come under 

 Rosenbusch's group of Hunnediabase; those with quartz, 

 under his group of Kongadiabase, while No. 7 may be 

 described as an ophitic norite or a hypersthene diabase. 

 The writer prefers, however, to call them all dolerites, 

 indicating their special characters by the prefixing of the 

 names of those minerals not common to all dolerites, e.g. 

 hypersthene enstatite-augite olivine dolerite. This method 

 though admittedly clumsy, gives due weight to each element 

 of composition. 



The affixing of correct names to the rocks is, however, 

 only a matter of secondary importance. What is more 

 important is the recognition of this group of rocks in 



