374 Transactions. 



the denuded plug that fills the pipe through which magmatic matter rose to 

 the crater. The variety of plutonic rocks found here, however, was quite 

 small, a fact that can be explained by the smaller size of the Taiarabu volcano, 

 and therefore by the less extent to which denudation has laid bare the 

 deeper portions of the plug. There was, unfortunately, no opportunity of 

 making geological observations in the valley of this river. 



It will at once be seen that there is in Tahiti a splendid opportunity 

 of studying the differentiation that has taken place in the magma that 

 filled the pipe of a volcano of large dimensions at a depth of 2,000 metres 

 below the actual crater, and of the correlation of these differentiates with 

 the lavas that have issued from the crater from time to time. It happens, 

 however, that a lack of time and means have prevented anything more than 

 a mere locating of a few of the plutonic types, and but little has yet been 

 done in finding the extent of the different lavas. The present paper 

 merely calls attention to the excellence of this field for study. 



It may, however, be noticed that the rocks of Tahiti appear to be 

 merely typical of those of most of the volcanic islands of the Eastern 

 Pacific. Thus, nothing but basic and alkaline rocks have up to the present 

 been recorded from the following islands : Samoa, Rarotonga, Aitutaki, 

 Mangaia, Raiatea, Huaheine, Mangareva, Pitcairn, Rurutu, and the Sand- 

 wich Islands. 



The association of the rock types in Tahiti appears to have a close 

 relationship to the series of alkaline and basic rock types at the Otago 

 Peninsula, New Zealand. Here, as previously pointed out, there is a large 

 area of white trachyte c< imposed of almost pure feldspar, a large develop- 

 ment of phonolites, and a great many basalts, as well as connecting types 

 such as trachydolerites.* Some of the trachydolerites contain large crystals 

 of brown hornblende which is largely resorbed, and is associated with the 

 violet augite in exactly the same manner as in the theralites and essexitcs 

 of Tahiti, and in similar rocks of other localities, such as Massachusetts. 

 The composition of these trachydolerites is intermediate between the 

 composition of the phonolites and that of the basalts. 



At the north head of Otago Harbour a trachydolerite lava nearly always 

 lies beneath a phonolite, which is in turn followed by basalts. The author 

 has suggested that this association can be explained by the assumption that 

 the lavas were supplied from a reservoir at an intermediate level, to which 

 theralitic matter party crystallized was moved from time to time. Within 

 this reservoir resorption of the crystals that had already formed might have 

 taken place if the mineralizers of the magma escaped. Differentiation would 

 then take place in this magma, with the result that phonolitic and basaltic 

 matter might be separated and emitted at different eruptions. 



Inspection of the old solidified plug of Tahiti shows that much differenti- 

 ation of a magma can take place at depths of no more than 2,000 metres 

 — -in fact, that it may be so complete as to produce a syenite composed 

 almost solely of feldspar as well as a non-feldspathic peridotite. 



This observation may therefore explain the frequent association of 

 alkaline and basic rock types in the Pacific region, to which attention has 

 previously been called, and for which a special explanation has been re- 

 quired. ' Whether the different species of rocks in Tahiti have resulted 



* Marshall, P., "The Geology of Dunedin," Q.J.G.S., vol. (52, 1906, p. 388. 



