340 
nite and the syenite is probable, and it appears to be indicated 
also by the presence of granite enclosures in the augite-syenite 
of Iviangusat (see p. 53). 
II. The transition zone between the granite and the ag- 
paites consists of three layers: quartz-syenite, pulaskite, and 
foyaite, the total thickness of which only amounts to 10—50 
meters. Only the central one of these layers, the pulaskite, has 
been found constantly separating the granite from the agpaites, 
while the two other layers are wanting at some places, and are 
always somewhat variable in their mineralogical constitution. 
The quartz-syenite has not been chemically analyzed. The 
pulaskite (analysis 3) is chemically very closely related to the 
syenites (nordmarkite and augite-syenite) which occur in the un- 
stratified part of the batholite. The foyaite (analysis 4) has about the 
same silica percentage as the syenites, but differs in its alkalies 
which are a little higher, in alumina which is about 5 per cent 
higher, and in iron oxides and lime which are lower. The 
presence of this rock as a subordinate sheet within the Ilimau- 
sak batholite is remarkable, for the rock is chemically of a type 
which is very different from all the other rocks of this ba- 
tholite, but shows a certain analogy with the foyaite of the 
Igaliko batholite (compare Е, and F,, Fig. 32) and is very like 
the foyaites of many localities outside Greenland. 
The rocks of the transition zone cannot apparently be re- 
garded as having been brought in by successive intrusions along 
the boundary plane which separates the granite from its sub- 
stratum. Here, as elsewhere, the hypothesis of successive in- 
trusions fails to explain the differences between contiguous 
sheets, and only accounts for contact phenomena which ap- 
pear at the boundaries between the sheets. But at Ilimausak 
such phenomena are entirely wanting; the sheets are uniformly 
coarse-grained throughout; and there is everywhere a gradual 
transition without any trace of a contact plane between them. 
These facts seem to point to the conclusion that the junctions 
