magnesia .attain their maxima at difiereiit parts. Some geologists 

 place in separate divisions those rocks which contain nepheline or 

 leucite, and those which go by the name of Mica-trap, or Lampro- 

 phyres. But while the former of these divisions may conveniently 

 be kept separate, the latter appears to be provisional only, and 

 will no doubt, be merged into the main sequence. No member of 

 the Nepheline-leucite division has as yet been described from New 

 Zealand,* but the main sequence is very fully represented. 



Omitting then the Nepheline-leucite Rocks, the Class of Massive 

 Eruptive Rocks is here separated into three Divisions by the 

 nature of the felspars when present, or by their absence and 

 each division is broken up into two or more Series by chemical 

 composition. These series are again divided by macroscopic 

 texture into Groups to which distinguishing names are given. 

 The groups are separated into Species by the ferro-magnesian 

 silicates they contain, and, when necessary, these are again sub- 

 divided by microscopic texture into Varieties. The name of the 

 species is oV>tained by prefixing the name of the ferro-magnesian 

 silicate, or muscovite when ferro-magnesian silicates are absent, to 

 that of the group ; but for brevity only the most important ferro- 

 magnesian silicate is named. Tlie question arises here. Which is 

 to be considered the most important ferro-magnesian silicate when 

 more than one are present ? I have answered this question by 

 selecting the one which is most rarely found in the group, because 

 attention is thus directed to it. For example : nearly all the 

 andesites contain augite, therefore I limit the species Augite 

 Andesite to those andesites which contain no other ferro-magnesian 

 silicate ; and when a second is present it gives the name to the 

 species. In the Trachyte group hornblende is the common ferro- 

 magnesian silicate, and in the Rhyolite group it is biotite. 



In the non crystalline rocks the species depend upon texture or 

 on chemical composition Varietal names are usually not indi- 

 cated but cm easily be formed by adding the texture name. 

 The classification is, theiefore, founded on chemical composition 

 and texture, but a chemical analysis is not generally necessary 

 for namitiy a rock. Usually its clien»ic,il composition cm lie 

 estimated with sufficient accuracy by a knowledge of its muieral 

 composition and it', specific gravity. It is only in doubtful cases 



rocks from Diinedin " (Trans.' N. Z. Inst., 

 fiption has been published of either. 



