Vol. 70.] LAVAS AT NOltTlf ItEAD, OTAGO harbour. .'390 



types is the old theory of Bunsen, with its further extension by 

 Durocher. This theory, so far as its general application is con- 

 cerned, lias at the present time a historical interest only. It is, 

 however, worth while to consider the possibility of its application 

 to a limited series of rocks such as the one now being considered. 

 The two extreme types of rock found in this scries would have to 

 he taken as representing the composition of the two original primary 

 magmas, and all the other types (phonolite. kaiwekite, trachydolerite, 

 and basalts) would have to be regarded as formed from these two 

 magmas mixed together in different proportions. 



The two extreme types are the trachyte Xo. 1 and the basalt 

 No. 0. It is, however, impossible to regard the phonolite No. 14, 

 the most alkaline rock of the whole series, as formed from the 

 mixture of these two primaries in any proportions whatsoever. 



(3) The nature of the rock-sequence shows that the different 

 members could not all owe their origin to the processes of differ- 

 entiation that might take place in a single magma. Although 

 temperature-differences and relative miscibility of various com- 

 ponents of the original magma would almost certainly play their 

 part in promoting such differentiation, yet it is not perhaps too 

 much to say that, in the' main, the separation of the magma into 

 portions of different chemical composition would depend upon the 

 action of the force of gravity. If this be admitted, it follows that 

 the various lavas were emitted roughly in the order of their specific 

 gravities. This, however, is far from being the case, for the 

 phonolites are of much lower specific gravity than the basalts, yet 

 they occur higher in the series than the basalts, and the very last 

 lava of all is a phonolite. 



What is true of the rocks as a whole is also true when applied to 

 a consideration of the occurrence of different minerals in the rocks 

 of the series. Thus barkevikitc, a mineral with a high specific 

 gravity, does not occur in any of the lower lavas, but only in those 

 that are relatively high in the series, and in those lavas in which it 

 does occur it is associated with anorthoclasc, a mineral of relatively 

 low specific gravity. While the occurrence of the rocks in the 

 order of their specific gravity would, in the main at least, be due to 

 the action of molecular differentiation, the occurrence of minerals 

 in their order of specific gravity would be the result of fractional 

 crystallization. Since neither the rocks nor the minerals occur in 

 this series in an order which on the whole has any relation to the 

 order of their specific gravities, it follows that the order of occur- 

 rence of both the minerals and the rocks i>'oes far to show that the 

 series of lavas has not originated from the differentiation that 

 might have taken place within a single magmatic ma--. 



(4) The chemical relationship of the kaiwekite No. 13 and the 

 trachydolerite Xo. 21 to the rocks above and below them shows at 

 once how remarkably intermediate a chemical composition they 

 possess. In regard to several of the constituents, notably lime, 



