

400 IMiOF. 1'. fctAKSHAlX OS THE SEQUENCE OF [Dec. I9M, 



magnesia, soda, potash, and, to a less extent, silica, it is seen at 

 once that the proportions are almost the exact mean of those present 

 in the associated rocks. This is clearly shown in the curves which 

 represent the molecular proportions of the substances and in those 

 which represent the molecular proportions of these substances as 

 compared with silica (Pis. LIII & LIV). It can hardly be doubted 

 that this indicates a genetic relationship. It must, however, be 

 insisted that although the kaiwekite, for example, lies between a 

 basalt and a phonolite, it must not be regarded as especially related 

 to the basalt lying beneath it, but to the basalts that rest on the 

 succeeding phonolite. The general similarity of the basalts of the 

 series is, however, so great as not to invalidate the value of the 

 comparison made. In point of fact, the probability of the genetic 

 relationship is still further emphasized when in tins case com- 

 parison is made with the basalts that rest upon the phonolite 

 (No. 14). The lowest of these is distinctly less basic than the 

 normal basalt of the series, and it can therefore be reasonably 

 claimed as a differentiation -product of the kaiwekite-magma, 

 connecting in a definite manner the phonolite and the basalt- 

 differentiates. We are thus clearly led to the opinion that the 

 kaiwekite represents the composition of the original magma in 

 the volcanic reservoir, and that the phonolite and succeeding basalts 

 are chemical differentiates from it. 



The same relationship can be seen in connexion with trachy- 

 dolerites Xos. 24 & 21 : although in the former case any basalt 

 that may have rested on the phonolite has been removed by denu- 

 dation, and in the latter case there is no phonolite. This, however, 

 is not surprising, for, as has been previously pointed out. all the 

 lavas that issue from the crater of a volcano are not represented 

 on a single slope of the mountain. It is indeed remarkable that 

 a succession of lavas that can be regarded as a complete series, 

 such as that above the kaiwekite, should be found on one slope. 

 The occurrence of the trachyte at the base of the whole series is 

 not explained by this suggestion. The rock on which it rests is not 

 shown, but, as will afterwards be pointed out, there is, in the great 

 magmatic reservoir of Tahiti, now exposed by the denudation of 

 b'000 feet, at the least, of volcanic material, a mass of a syenite 

 composed solely of felspar, and this has probably been derived 

 from the differentiation of an essexite magma. 



The foregoing statements support the belief that the lavas of 

 this series have issued from a reservoir at an intermediate 

 level, and that within this reservoir the conditions allowed of 

 differentiation. On four separate occasions this reservoir is 

 supposed to have been supplied from some deeper source with 

 an essexitic magma in which a certain amount of crystallization 

 had taken place. There is even some evidence that a certain 

 amount of differentiation had already occurred in this deeper 

 source, as the markedly porphyrinic types of rock — trachyte 

 X<>. 1, kaiwekite Xo. 13, trachydolerite No. 20, and trachydolerite 

 Xo. 121, show a progressive decrease in silica and corresponding 



