339 
(2) Soda predominates strongly over potash. 
(3) Magnesia is almost entirely absent. An exception is 
found in the magnetite-pyroxenite of Narsak (analysis 19), but 
this rock only occurs within a limited area of quite insignifi- 
cant dimensions. 
(4) Lime is relatively low. 
(5) When the basic rocks (essexites and trachydolerites) are 
left out of consideration alumina in relation to alkalies is low. 
While these characters illustrate the general consanguinity 
of the Ilimausak and Igaliko rocks, a close relationship asserts 
itself within some of the main divisions of the complexes. As 
only a few analyses of the effusive and hypabyssal rocks have 
been made, these rocks are omitted from the following dis- 
cussion. 
Igaliko batholite. — The main rock is a nepheline-syenite 
(analysis 25) which agrees with the habit of the ordinary Foya 
type. It is also chemically related to the Foya type, but shows 
a marked tendency towards the agpaites mentioned below. As- 
sociated with the main rock is an augite-syenite (analysis 26), 
closely related to that of the Ilimausak batholite. This rock is 
the only type which is common to the two batholites. 
Unstratified part of the Ilimausak batholite. — The ana- 
lyses of the principal rock types are graphically represented in 
Fig. 311. The figure shows that the compositions of the syenites 
(nordmarkite and augite-syenite) are generally intermediate be- 
tween the compositions of the granite” and the essexite. This, 
1 In the Figs. 31 and 32 the horizontal rows represent the analyses cal- 
culated at 100 per cent., omitting Н.О, Е, and Nb,0,. TiO, and ZrO, 
have been added, and the sum is given as TiO, if TiO, be predomi- 
nant, and аз ZrO, if this be predominant. The percentage weights are 
given as abscisse. 
In Fig. 31 the first column ought to give the composition of the arf- 
vedsonite-granite of Narsak, described (p. 115). As, however, no chemical 
analysis has been made of this rock, the corresponding rock from North 
Siorarsuit has been taken to represent the Narsak granite, with which it 
agrees very closely. 
12 
