142 
actually erystallized later than the main mass of the other con- 
stituents of the rock. 
Chemical composition. — A chemical analysis of specimens 
of this rock, taken at Tupersuatsiak, has been made by Dr. C. 
Winteer, with the results given in №.5 of the above table. If 
this analysis is compared with the analyses of other nepheline- 
syenites it will be found to differ considerably from most of 
them. Among the rocks of Lujavr-Urt, Kola Peninsula, how- 
ever, a closely related type occurs, namely the tavite of Ramsay 
(see above, analysis A). The analysis of the tavite is incom- 
plete, chlorine not having been determined, but as a variety of 
sodalite (hackmanite) is a main constituent of the rock it is 
probable that a more complete analysis would show a still greater 
agreement with the analysis of the Greenlandic rock. On the 
other hand it deserves notice that the tavite as to structure is 
much more similar to the Greenlandic naujaite than to the 
sodalite-foyaite. 
From the chemical analysis the mineral composition of 
the sodalite-foyaite of Tupersuatsiak is found to be approxi- 
mately: 
Per cent. 
ПОНИ ооо ee 16 
AUDITE SA ORAN ARE О RES 20 
ЗО 22 
nephelines es cer re ee 5 
Опа еее ee ome ee 4 
ОВ. 12 
НА о ро воо вооон авовев а 6 
за ее 7 
zeolites (analcime and natrolite).... 6 
fluorite. маке ее eee TE PE 2 
100 
In calculating these figures it has been assumed that the 
eudialyte, the ægirine, and the arfvedsonite are of the same 
composition as those of the pegmatites as analysed by J. Lo- 
RENZEN. Likewise the analysis of nepheline by Crarke and that 
